4406-10 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO31 bud§daff - Anomalous Mental Phenomena: Selected Papers Compiled By: The Cognitive Sciences Laboratory 24 June 1991 ME Sckywe AppOwdons Intamdoml Corporadw An Employee-Owned Company 5150 El Carnino Real, Suite B-31, Los Altos, Califomia 94022 (415) 960-5910 Other SAIC 249", Seattle, Tucson Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 I INTRODUCTION In this volume, we present a selected set of papers on, and/or in support of, anomalous mental phenom- ena. No section could possibly be complete; however, we have chosen papers that are representative of their particular sections. The sections, which are separated by blue sheets, are as follows: Ser,tio I Introduction II Meta-analyses of Anomalous Mental Phenomena 8 III Main-stream Publications 7 IV Anomalous-mental-phenomena Journal Publications 6 V Magnetoencephalography 3 VI Physics 6 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 11 META-ANALYSES OF ANOMALOUS MENTAL PHENOMENA As in all behavioral sciences, replication of experiments in anomalous mental phenomena (AMP) is critical before any putative effects ran be verified as part of nature, Because of the complex nature of most behavioral experiments, drawing conclusions from a body of similar experiments has been prob- lematical. Meta-analysis, however, is a relatively new statistical approach that has been specifically de- signed to address the particular difficulties inherent in the behavioral sciences. The papers in this section have been selected because they represent all such analyses of a substantial portion of the published AMP literature to date. Through replication and meta-analysis, the general scientific community will have tools with which to judge the claims of the AMP literature. The number that appears in the upper right-hand corner of the first page for each publication is keyed to the following descriptions: 1. Utts, J., "Successful Replication Versus Statistical Significance," Journal of Parapsychology, Vol. 52, pp. 305-320, (December, 1988). By defining, in statistical terms, the meaning of replication for few-cy effects, Utts, a Professor of Statistics from the University of California at Davis, sets the statistical basis for meta-analysis. 2. Honorton, C., "Error Some Place!" Journal of Communication, pp. 103-116, (Winter, 1975). This paper predates the development of formal meta-analysis, but Honorton provides a critical review of all the ESP card-guessing experiments from 1934 to 1939. The paper includes a description of the claims and counter-claims surrounding the controversy of the day, 3. Honorton, C. and Ferrari, D. C., "'Future telling:'A meta-Analysis of Forced-Choise Precognition Experiments, 1935-1987," Journal of Parapsychology, Vol. 53, pp. 282-308, (December, 1989). Using the full complement of meta-analytical tools, Honorton provides a critical review of all the ESP'experiments during which the target material (i.e., usually ESP cards) is generated after the guess has been recorded. 4. Honorton, C., Berger, R. E., Varvoglis, M. P., Quant, M., Derr, P, Schechter, E., L, and Ferrari, D. C., "Psi Communication in the Ganzfeld," Journal of Parapsychology, Vol. 54, pp. 99-137, (June, 1990). This paper provides a meta-analysis of Ganzfeld experiments (i.e., a form of anomalous cognition). The database is comprised of 11 series for a total of 355 individual trials. 5. Radin, D. I. and Nelson, R. D., "Evidence for Consciousness-Relate d Anomalies in Random Physical Systems," Foundations of Physics, Vol, 19, No. 12, pp. 1499-1514, (December, 1989). Radin and Nelson analyze over 800 experiments that claim evidence for mental human-machine interactions (i.e., anomalous perturbation). After a careful analysis, which includes accounting for experiment flaws, they conclude that there is substantial statistical evidence to support the claim. 6. flonorton, C., Ferrari, D. C., and Bern, D. J., "Extraversion and ESP Performance: Meta-Analysis and a New Confirmation," Proceedings of the Parapsychological Association 33rd Annual ConventioA Chevy Chase, MD, (August, 1990). In an important link to traditional psychological experimentation, this paper provides a meta-analysis for the correlation of ESP performance and a traditional personality variable, extraversion. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 7. Rosenthal, R., "Meta-Analytic Procedures and the Nature of Replication: The Ganzfeld Debate," i Journal of Parapsychology, Vol. 50, pp. 319-336, (December, 1986)!, Rosenthal, a professor of psychology at Harvard University, is one of the early developers of the meta-analysis techniques. In this paper, he comments about the Ganzfeld controversy. 8. Utts, J., "Replication and Meta-Analysis in Parapsychology," Accepted for publication in Statistical Sciences, In this paper, Utts, provides an independent Ad objective overview of' the AMP meta-analyses that follow. Approved For Release 2003104118: CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001,4 Purnal of Parapsychalogy, Vol. 52, Decernber 1988 SUCCESSFUL REPLICATION VERSUS STATISTICAL SIGNIFICANCE By JESSICA U'I-I*S A13STRACT: The aim of this paper is to show that successful replication in para- psychology should not b-c equated with the achievement of statistical significance,,, whether at the .05 or at any other level. The p value from a hypothesis test is closely related to the size of the sample used for the test; so a definition of suc- ces'sful replication based on a specific p value favors studies done with large sam- ples. Many "nonsignificant" studies may simply be ones for which the sample5izC was not large enough to detect the small magnitude effect that was operating. Con- versely, "significant" studies may result froin a small but conceptually insignificant bias, magnified by a very large sample. The paper traces the history of the definition of statistical significance in para- psychology and then outlines the problems with using hypothesis-testing results to define successful replications, especially when applied in a cookb43ok fashion. Fi- nally, suggestions are given for alternative approaches to looking at experimental data. These include calculating statistical power before doing an experiment, using extimatlon inK(end or, or In conjunction with, hypothetilm tenting, wid implellicn6lig some of the ideas from Bayesian statistics. Replication is a major issue in parapsychology. Arguments about whether a given research paradigm has been successful Lend to fo- cus on what the replication rate has been. For example, the recent review of parapsychology by the National Research Council includes statements such as "...of these 188 (RNG] experiments with some claim to scientific status, 58 reported statistically significant results (compared with the 9 or 10 experiments that would be expected by chance)" (Druckman & Swets, 1988, p. 185). In each section, '. the report critically evaluates "significant" experiments and ignores "nonsignificant" experiments. The extent to which nonsignificant experiniciiLs are ignored is exemplified by LhC f011OWing OVerSight., in which the total number of studies is equated with the number of "successful" Studies: "Of the thirteen scientifically repoi-Led experi- ments [of remote viewing], nine are classified as successful in their outcomes by Hansen CL al. ... As it turns out, all but one of the nine scientifically reported studies of rernote viewing suffer from the Raw of sensory cucing" (1). 183, emphasis addcd). Apparently the au- thors decided that the four experiments that did not attain a p value of .05 or less did not even warrant acknowledgment. Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 106 'FhCJ0117-71(11 f)f 1'"?_('P-7@110100 The practice of defining a successful replication as :111 expCI-i- ment that attains a p value of .05 or less is common in parapsychol- ogy, psychology, and some other disciplines that use statistics. How- > ever, like many other conventions in science, it is based on a series of h:sLorical events rather than on rational thought. In this paper, I will trace. som .c of, the history leading to this definition or a "suc- 0 * cessful" experiment, outline some problems with this approach, and * suggest some methods that parapsychologists should consider in ad- CL -n dition to the usual hypothesis-testing regimen. Rao (1984) and Hon- 0 N orton (1984) have discussed similar problems and solutions in the context of psi experiments. U fu CO) M HISTORY 0 It has not always been the case among parapsychologists that an 0 44. experiment was deemed successful if it reached a significance level of p = .05. In 1917, John Edgar Coover, who was the Thomas Wel- 00 ton Stanford Psychical Research Fellow at Stanford University from 1912 to 193-7, published a book with the results from several exper- 1> iments he had conducted up to that time (Coover, 1917/1975). Al- though hypothesis testing as we know it today had not yet been for- malized, he essentially conducted tests on many facets of this data cp and found no evidence for psi that was convincing to him. His con- clusions regarding these results are typified by an example lie gave 0' in which the hit rate for 518 trials was 30.1%, when 25% was ex- 0 pected by chance (exact p value = .00476): CA) We get 0.9938 (p-value = I - 0.9938 = 0.0062) for the probability t -hat 0 chance deviations will not exceed this limit (of 30A percent).... Since 0 this value, then, lies within the field of chance deviation, although the 0 W probability of its occurrence by chance is fairly low, it cannot be ac- 0 0 cepted as a decisive indication of some cause beyond chance which op- 0 _L crated in favor of success in guessing. (p. 82) 4.. He then rev6aled what level of evidence would convince him that nonchance factors were operating: "...if we meet the requirement of a degree of accuracy usual in scientific work by making P = 0.9999779, when absolute certainty is P = 1; then [there is] saLisfac- Lory evidCnCe for sonic cause in addition to chance" (1). 83). In other words, he was defining significance with a p value of 2.21 X 10-.". Coover was not alone in requiring that results conform Lo arbi- trarily stringent significance levels. In 1940, when Rhine et al. pub- Rel5liceltion 71s. Significance 307 lished Extra-Senso7y Perception After Sixty Years, they included the fol- lowing definitions in the glossary: P-value = probability of success in each trial > SIGNIFICANCE: When the probability that chance factors alone p M r% chiccd a given deviaLion is Sufficiently small to provide relative certaint B that chance is not a reasonable expectation, the deviation is sign@6eantI3-< above or below the chance level. Among ESP results, this is arbitrarilm XL taken to mean a deviation in the expected direction such that the criticatn 0 ratio is 2.5 times the standard deviation (or four times the probable er-1 ror) or greater. (p. 423-424) Thus, significance was defined by z _- 2.5, or p -_ .0062. M fu Seventeen years later, in their book Parapsychology: Fronder Sci-CM0 ence of the Mz'nd, Rhine and Pratt (1957) suggested that .0 1 was the" 0 appropriate threshold: 0 In order for such judgments to have the necessary objectivity, a crz*ter;'071@: of sigw@-rcance is established by practice and gencral agreement among"M the research workers in a particular field.... Most workers in parapsy- - chology accept a probability of .01 as the criterion of significance. (p.Q 186) > Finally, the Journal of Parapsychology has included a definition A s,*gnqkcance in its glossary for many years, but the appropriate pw value has fluctuated back and forth between .01 and .02, finally set-T ding at .02 in 1968. The Following ;ire excerpts from those glossar-0 -4 ies: 00 W December 1949: "A numerical result is significant when it equals;0 or surpasses some cri of degree of chance improbabil-S) Ity.- Co in--m-on criteria are: a pro~b-,Iyili-t--y-v-,.~-11-u---e--o5-f ul or less-. C) 0 March 1950 to June 1957: "The criterion commonly used in this 0 W journal is a probability value of .02 or less." 0 0 0 September 1957: "The criterion commonly used in this Journa) _L is P = .01." 4.. December 1957 to December 1967: "The criterion commonly used in parapsychology today is a probability value of .01 or less." March 1968 to December 1986: "The Criterion Commonly used in parapsychology today is a probability value of.02 (odds of 50 to I against chance) or less.... Odds of 20 to 1 (probability of .05) are regarded as strongly suggestive." > -a 13 I 0 < M CL 0 U) M 1%) 0 0 44. 00 > @13 0 W a) 6 0 J 00 W X 0 0 0 0 0 0 0 0 0 m- 41@h. 108 771t.-Journal il Parapsychology March 1987: The term si@,niftcemcr no longer appears it, tile glos- s a ry. By the mid-1980's, despite the value of .02 given in the journal of Parapsycholog, significance seemed to have been determined'to correspond to a p value of .05. For example, in their bibliography of remote-viewing research, Hansen, Schlitz, and Tart (1984) claim: 11 We have found that more than half (fifteen out of twenty-eight) of the published formal experiments have been successful, where only one in twenty would be expected by chance." As mentioned in my . introduction, .05 was the value used by the National Research Coun- cil in their recent evaluation of parapsychology. Both Hyman (1985) and Honorton (1985) used .05 as the criterion for a successful ganz- Feld study. In discussing the Schmidt REG experiments, Palmer (1985) implicitly used .05 as the cut-off for significance by observ- ing: "Based on Z-tests ... 25 of the 33 (76%) were significant at the .05 level, two-tailed. In two of the seven nonmsignificant studies ...... (p. 102). This definition of significance is obviously not unique to para- psychology. A popular introductory textbook in psychology states that: atistical inference procedure that ives them an Psychologists used a st estimate of the probability that an observed difference could have oc- curred by chance. This computation is based on the size of the differ- ence and the spread of the scores. By common agreement, they accept a difference as "real" when the probability that it might be due to chance is less than 5 in 100 (indicated by the notation p < .05). A sig- nificant difference is one that meets this criterion.... With a statistically significant difference, a researcher can draw a conclusion about the be- havior that.was under investigation. (Zimbardo, 1988, p. 54) Given the weight that has been atLached to .05 as tile criterion for significance, one would think that it resulted from careful con- sideration of the issue by statisticians and psychologists. Unfortu- nately, such is not the case. Its roots apparently lie in the following passage published in 1926 by one of the foundcrs of modern StaLis- tics, Sir Ronald A. Fisher: It is convenient to draw the line at about the level at which we can say: "Either there is something in the treatment, or a coincidence has oc- curred such as does not occur more than once in tiventy trials." ... If one in twenty does not seem high enough odds, we may, iC wc prelcr it, draw tile line at one in fifty (the 2 per cent point), 01' one ill ;I hundred (thc I pC1. cent point). Personally, the writcr prcrcrs to %Ct a 1%!r1)1ir111io)1 Ia. sign@icance .1109 low suilidard of'significance at the 5 per cent. point, md ignore entirely ;111 I-eSUILS which fail to reach that level. A scienLific fact should be re- arded as experimentally established only if a properly designed exper- inient rarelyfails to give this level or signiricance. (Fisher, 1926, p. 504; also quoted in Savage, 1976, p. 471) Thus began the belief that an experiment is successful only if the null hypothesis can be rejected using ot = 0.05. As an immediate consequence of this belief, Fisher and his followers created tables of F statistics that included values only for tail areas of .05 and .01. Since researchers did not have access to computer algorithms to de- Lerinine intermediate P. values, success carne to be measured i' terms n of these two values alone. PROBLEMS wi-ri-i HYPOTHEsis TESTING MiSCOnceptions about p Values Most modern research reports include p values instead of simply discussing whether an experimental result is significant at a' pre@ specified level. Although this is somewhat better than the ol d method of "one star or two" (corresponding to a significant resul t at .05 'or .01, respectively), it is still a misleading way to examine experimental results. The problem is that many -researchers interpret p values as being related to the probability that the null hypothesis is true. Even some so- phisticated researchers tend to think that an extremely small p value Must Correspond to a very large effect in the population and that a large P value (say > .10) means that there is no effect. In other words, the size of the p value is i71correctly interpreted as the size of the effect. It should be interpreted as the probability of observing results as extreme or more so than those observed, if there is no effect. To see how arbitrary it is to base a decision about the truth or falsity of a statement on a p value, consider a binomial study based on a sample of size n which results in z = 0.30, p value = .38, one- tailed. One would probably abandon the hypothesis under study and decide not to pursue the given line of research. Now suppose that the study had been run with a sample of size 100n instead and I-CSUILed ill the exact same proportion of hiLs. Then we would find z = 3,00, P value = .0013. These results would be regarded as highly significant! > M M -1 0 < M CL -n 0 M 2) U) M 1%) 44. 00 - - > CD a) 6 0 -4 00 W X 0 0 0 _L 0 0 0 0 0 0 0 4.. Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 310 TheJournal qf'Parapsychology As another example, consider a chi square test for randomness based on a sequence of n numbers, each of which can take the val- ues 1, 2, ... 10. Suppose that the test results in a chi-square value of 11.0, df = 9, P value = 0.28. Now suppose the sequence was three timcs as long but the proportions of each digit remained the same. Then each term in the numerator of the chi-square statistic would be multiplied by 32, whereas each term in the denominator would only be multiplied by 3. The degrees of freedom would noL change, but the new result would be x' = 33.0, df = 9, p value = .00013. In the first case, the conclusion would be that the sequence was suf- ficiently random, yet a sequence three times as long with the same pattern would be seen to deviate considerably from randomness! This problem was recognized more than 50 years ago by Berk- son (1938): We may assume that it is practically certain that any series of real ob- servations does not actually follow a normal curve with ab@olule a-Cactilude in all respects, and no matLer ]low small Lhe discrepancy bcLWCCII LhC normal curve and the true curve of observations, the chi-square P will be small if the sample has a sufficiently large number of observations in it. If this be so, then we have something here that is apt to trouble th( conscience. of a reflective statistician using the chi-square test. For I sup. pose-it would be agreed by statisticians that a large sample is alway., better than a small sample. If, then, we know in advance the P that will result from an application of a chi-square test to a large sample, there would seem to be no use in doing it on a smaller one, but since the result of the former test is known, it is no test at all. (pp. 526-527, --emphasis in-originaL@- Rel5lication vs. SignVicance 311 respondents demonstrated that they were far more likely to believe results based on the large sample when the p values were the same. (For a discussion of this example and some other problems with hy- pothesis testing in psychology, see Bakan, 1967.) One consequence of this misunderstanding is that researchers misinterpret what constitutes a "successful replication" of an exper- iment. Tversky and Kahneman (1982) asked 84 members of the American Psychological Association or the Mathematical Psychology Group the following question: Suppose you have run an experiment on 20 subjects, and have obtained a significant result which confirms your theory (z = 2.23, p < .05, two- tailed). You now have cause to run an additional group of 10 subjects. What do you think the probability is that the results will be significant, by a one-tailed test, separately for this group? (p. 23) The median answer given was .85. Only 9 of the 84 respondents gave an answer between .40 and .60. Assuming that the valub ob- talned in the first test was close to tile true population value, the probability of achieving a p value --< .05 on the second test is actually only about .47. This is because the sample size in the second study is so.small. The effect would have to be quite large in order to be detected with such a small sample. Inthe same survey, Tversky and Kahneman also asked: An investigator has reported a result that you consider implausible.'He ran 15 subjects, and reported.a significant value, t = 2.46. Another in- vestigator has attempted to duplicate his procedure, and he obtained a nonsignificant value of t with the same number of subjects. The direc- tion was the same in both sets of data. You are reviewing the literature. _-Whatis--the hi-g--h-c-s-t-v-a-l--u-e--o-f-t-in- the second set of __ data that you w ould c1cscribe as a Failure to replicate? (p. 28) Replication The majority of respondents considered t = 1.70 as a failure to Very often researchers simply do not understand the connection replicate. But if the results from both studies are combined, then between the p walue and the size of the sample. For example, Ro- (assuming equal variances) the result is I = 2.94, df = 29, @ value senthal and GaiLO (1963) asked nine faCUlLy members and ten grad- = .003. The paradox is that tile new study decreases faith in the orig- uate students in a university psychology department to rate their inal result if' viewed separately but increases it when combined with degree of belief or confidence in results of hypothetical studies with the original clatal This misunderstanding about replication is quite prevalent in the various p values and with sample sizes of 10 and 100. Given the psi literature, as demonstrated by the emphasis on successful repli- same p value, one should have more confidence in a study with a cation, where success is defined in terms of a specific.p value, re- smaller sample because it would take a larger underlying effect to gardless of sample size. As an example of how unnecessarily dis- obtain the small p value for a small wipple. rt Lely, pprovevcpp(or'Wi'e)eaget~V63/04/18 CIMREROPUO@@ dW3ft04r_*rs, I have shown elsewhere (Utts, > 0 < M CL -n 0 0 (D 44- _L co 0 5; @0 0 w a) 6 -0 4 co CD co 0 W 0 The Journril ty* J'arap.%Ychology '112 1986) Lliat. if' the Lille hit rate in it binomial Study (such -,is a ganzield experiment) is actually 33%, and 25% is expected by chance, then a study based on a sample of size 26 should be expected to be "suc- cessful" (p -- .05) only about one fifth of the time. Even it study based on a sample of size 100 should be "successful" on!- about half I of the tinle. It is no wonder L11aL there ilre so many "unsuccessrul" attempts at replication in psi. As another example of the paradoxical nature of this definition of replication, consider the "unsuccessful" direct-hit ganzFeld studies ' d by the meta-analyses of Hyman (1985) and Honorton covere (1985). Using those studies with p(hit) = .25, there were 13 out of 24'chat were nonsignificant, o: = 0.05, one-Lalled. (See HoriorLon, p. 84, Table Al.) But when these 13 "Tailures" are combined, the re- sult is 106 hits out of 367 trials, z = 1.66, p = .0485! Problei= with Point Null Hypotheses A point null hypothesis is one that speci6es a parti&ilar value ("point") as the one being tested. Most hypothesis testing is done with point null hypotheses. The problem with this approach is that any given hypothesis is bound to be false, even if just by a minuscule amount. For example, in a coin-tossing experiment, the null hy- pothesis is that the coin is fair, that is to say, H,,: P = .5000000. This is never precisely true in nature. All coins and coin-tossers in- troduce a slight bias into-the experiment. This slight bias can pro- duce a very small p value if the sample size is large enough. If, for example, the true probability of heads is .5001, and the observed proportion of heads ffills right ;It this V,111le, then the null hypothesis 107 will be rejected at .05 if the sample size is at least 6.7 x As long its there is any bias al all, the p vflue can be madc arbiLrarily sniall by taking a large enough sample. In practice, this problem was@ rarely serious before it became pos- sible to collect large amounts of data rapidly using computers. Stat- isticians have often used ESP as an example of one of the few cases where it really is possible to specify an exact value for the null hy- pothesis. But ev4en this view is changing, as shown by this comment from a recent issue of a popular statistics journal: It is rare, and perhaps impossible, to have a null hypothesis that can be. exactly modeled as 0 = 0.. One might feel that hypotheses such as H.: A subject has no ESP, or H.: Talking to plants has no effect on their growth, are representable as exact (wid believable) point 111111S, bill, evell here, Re.-plicaliot? vs. signijiCaller 313 ininor biases in the experiniellLS Will LISLI-ally preVc:11t 42.-ulct represellLa- tions as points. (Berger & Delampady, 1987, p. 320) In summary, hypothesis testing as it is currently formulated Lends to be a misleading approach to examining data. Small samples tend to lead to "nonsignificant" studies, whereas large samples c;in> lead to extremely small p values, even if the null hypothesis only,150 slightly wrong. Many researchers do not understand t he meaning ofo, < a p value and do not understand how closely replication issues areCD tied to sample size. Arguments about replication should not beL -n based on p values alone. 0 N ip SOLUTIONS M Power Calculations M bi 0 0 If a hypothesis test is to be done at all, a researcher should ap, least determine in advance whether it is likely to be successful. Thes@. statistical power of a test is the probability that the null hypothesi5g. will be rejected. It obviously depends on what the true underlyingm state of nature is. Because this 'information cannot be known (o;Q there would be no point in doing the experiment), it is a good'ide4- to loo -k at power for a variety of possibilities before conducting theo experiment. The results will tell you whether you are likely to b able to reject the null hypothesis, using the sample size you havP, 0 planned, for specific values of the magnitude of the effect. 0 -4 Statistical power is a function of the sample size; the true under-co W lying magnitude of the effect, the level ofsignificance for which L11% CXJ)Cl-iJ11CnL would be considered it success, and the method og analysis used. It does not depend on the data. As an example, suppose you are planning to conduct a test oE3 the hypothesis H,,: P = .25 using a series of 10 independent trialso Power calculations would proceed as follows: 0 1. Find the cutoff point for the number of hits that would leacP_, to rejection of H.. In this case, the p value for 5 hits is .08, and fol@. 6 hits it is .02, so 6 hits would probably be required to reject the null hypothesis. 2. Power for a specific alternative is the probability that the null hypotlicsis would be rejected if that alternat] ve value is true. In this case, power = P(6 or more lilts). This can be computed directly, using the binomial formula, for any specified hit rate. Here are Some exallipics: 114 ThcJou7nal (;f Parapsycholc)gy choices of potential targets as compared to the response he or she Hit rate Power P(6 or more had produced. By chance, the average rank should be 4.5. If 0.30 .047 psychic functioning had reduced the average rank to 4.0, the p > 0.33 .073 value would have been .298, not significant. Even if the average > M 0.40 .166 rank had been reduced to 3.5, the study would still not have beenm M 1 0.50 .377 M significant, p value = .126. The average rank would have to be 3.0 1 0 < 0 before this study would achieve a significant result. A parapsychol- < M CL Notice that even if the true hit rate is 50% instead of Lhe chance level M ogist experienced in remote viewing should be able to determine in CL -n of 25%, the chances of a "successful" replication are poor, that is, only advance whether such a study would be likely to be successful with -n 0 37.7%. In most psi applications, 30% or 33% is probably a more re- 0 such a small sample. alistic approximation to the true hit rate, so there would be a very The lesson here is that a "nonsignificant" study may be nothing U small chance of having this experiment succeed with only 10 trials. more than a study with low power. Before investing time and U ID 0 As a second example, suppose you are planning to run the same M ID money in a new study, it should be determined whether it is likely 0 M 1%) experiment with 100 trials and are planning to use the normal ap- to succeed if psychic functioning is operating at a given level. 1%) proximation instead of an exact test. Further, suppose you will re- 0 0 0 c-4. ject the null hypothesis if z -_ 1.645, where z is the usual critical Estimation. 0 44- ratio, corrected for continuity: z = (number of hits - 0.5 - 0 44- An approach that avoids many of the problems with hypothesis 0" 25) / V(100 x .25 x .75) = .23(number of hits - 25.5). Using 0 testing is to construct a "confidence interval" or an "interval Cs:Li- 00 simple algebra, note that z _- 1.645 when the number of hits inate" for the magnitude of an effect. This is done by compuLing an 32.65. Thus, the null hypothesis will be rejected if there are 33 or i> hi P 3 0 interval of values that almost certainly covers the true population 5>- more ts, so power = or more hits). Computing this for the (3 same hypothetical hit rates as in the previous example gives: value. The degree of certainty is called-the confidence coefficieni and @O i s specified by the researcher. Common values are 95% and 99%. 0 As,an example, consider a binomial study with 100 trials that w Hit rate Power P(33 or m ore hits) a) results in 35 hits. Using the normal approximation, one would ex- 6 . 0 0 0.30 .289 - pect the proportion of hits in the sample to be within 1.96 standard - 4 00 0.33 .538 4 deviations of the true hit raie 95% of the time. The appropriate cO W 0.40 .939 W standard deviation for the proportion P of hits is X 0 0 0.50 .9998 0 Thus, a 95% confidence interval for the true hit rate is found by 5 W W --adding- --b- trae't*Lng-l-.9&-ofthese--sta-ndard-dzvia-tio-ns-to-ttre -proF---- 0 0 Now there is a more reasonable chance for a successful study, al- 0 0 portion of hits observed in the sample. The resulting interval in this 6 0 W though it is still only 29% even if the true hit rate is 30%. 0 case is 0.35 - 0.09 to 0.35 + 0.09, or 0.26 to 0.44. This tells us that C 4 0 0 For studies in which the null hypothesis does not involve a single with a fair amount of certainty (95%), the true hit rate is covered 0 value, it can be more difficult to compute power because it is not so 0 by the interval from 0.26 to 0.44.' For the same proportion of hits 44. easy to specify a;reasonable alternative. In these cases, it is still pos- in a study with 1,000 trials, the interval would be from 0.32 to 0.38. 44. sible to look at the p value that can be expected if psychic function- The larger the sample size, the shorter the width of the interval. ing were to occur at specified levels for the sample size planned. For Consider two studies designed to test H.: P = .5: example. McClenon and Hyman (1987) conducted a remote-viewing Study I Study 2 study with eight trials, one for each of eight subjects, and used the Z 3.60 2.40 prefere n tial-ra n king method of Solfvln, Kelly, and Burdick (1978) p value .0004 .0164 on the subject rankings. Each subject was asked to rank-order eight n 1,000 100 316 - The journal of Parapsychology Which study provides more convincing evidence that there is a strong effect? In keeping with the results of Rosenthal and Gaito (1963) discussed earlier, most people would say that the first study shows a stronger effect, both because the p value is smaller and be- -cause it is based on a larger sample. In fact, the opposite is U-Ue. number of hits for the two studies are 557 (55.7%) and 62 e 121/6), respectively; the smaller study had a higher hit rate. The 0 confidence intervals for the hit rates in the two studies are % ies we W.53 to 0.59) and (0.53 to 0.72), respectively, so in both studi e relatively sure that the hit rate is at least 53%, but in the second ;r Study it could be as high as 72% whereas in the first it is probably Bo higher than 59%. In studies with huge sample sizes, confidence intervals inake it M Ident that an infinitesimal p vaiiie does not correspond to an ef- ct of large magnitude. For example, consider a study based on E c100,000 trials and designed to test H,,: 11 = .50. Suppose there were t0,500 hits. Then z = 3.16, and the p Value is 7.9 X 10'. But what cdoes this mean in practical terms? A 95% confidence interval for the ue hit rate is from 0.5019 to 0.5081. Thus, it appears that tile true ot rate is indeed different from 0.50, but reporting the results in ;@Jiis way makes it clear that the magnitude of.the difference is very @mall. The reader can decide whether an effect of this size has any uneaning in the context of the experiment. a) 6 In summary, confidence intervals are preferable to hypothesis 4@6 , -sts for the following reasons: 00 1. They show the magnitude of the effect. 02. They show that the accuracy of the conclusion is highly de- 0 endent on the sample size. 2 03. They remove the focus from decision making, which is arbi- 0 grary at best because of sample size problerns. 4. They highlight the distinction between statistical significance -9nd practical significance. 4. 5. They allow the reader of a research report to come to his or he]- own conclusion. Meta-Analyses Meta-analytic techniques may be viewed by some parapsycholo- gists as the solution to studying tile issue of replication. Even though these techniques can address the replication issue in* useful ways, Replication vs. SignVicance 317 they also contain some dangerous pitfalls. For example, both Hy- man (1985) and Honorton (1985) used '@'vote-counting" in their mem-analyses Of the ganzfeld data base. In other words, they tallied the number of significant studies in the data base. This procedure inherits all of the problems associated with the original determina- tion of whether a study was "significant" in the first place. A series of Studies, each with low power, may all be determined to be non- significant, when the combined data may lead to an extremely sig- nificant result. Conversely, a series of studies based on large samples may all be significant, but the magnitude of the effect may be very small. A vote-count showing that most studies are significant could I mislead researchers into believing that there was a large effect. The concept of effect size was introduced to account for tile fact that individual study results are highly dependent on sample size. .,Estimating the effect sizes For a series of studies and seeing whether they are similar is a useful way of studying replication. However, examining only the effect size for an individual study does not give any indication of the accuracy of the result. This should be done in conjunction with some estimate of the accuracy of the result, Such as it confidence interval. Bayesian Methods Many statisticians believe that the conceptual framework of hy- PoLhesis testing and interval estimation is philosophically incorrect. Rather, they start by assigning- prior probabilities, based on subjec- tive belief, to various hypotheses, and then combine these "priors" with the data to compute final or "posterior" probabilities for the hypotheses. This is called the Bayesian approach to statisti' cs. An in- LroducLion to the ideas of Bayesian analysis can be found in Berger and Berry (1988) or Edwards, Lindman, and Savage (1963). A more technical reference is Berger (1985). Berger and Berry (1988), in a recent article in American Scientist, discussed the use of Bayesian methods instead of classical methods: The first step of this demonSLration is to calculate tile actual probability that the hypothesis is true in light of the data. This is the domain of Bayesian statistics, which processes data to produce "final probabilities" ... for hypotheses. Thus, the conclusion of a Bayesian analysis might be that the final probability of H is 0.30. The direct simplicity of such a statement compared with the convo- luted reasoning necessary to interpret a P-value is in itself i poLenL it'- > -0 0 < M CL 0 N ip M ID (D 44, 00 0 > 0 0 -4 00 W X 0 cc) 0 W 0 0 I 4.. 318 The journal of Parapsychology gument for Bayesian methods. Nothing is free,.however, and the cle- gantly simple Bayesian conciusion requires additional input. To obtain the final probability of a hypothesis in light of the experimental data, it is necessary,Lo specify the probability of the hypothesis before or apart rrom dic cxpci-inicnLal dam. M Where does this initial probability come from? The answer is simple. I 0 It must be subjectively chosen by the person interpreting the daut. A < Pei-son who doubts the hypothesis initially might choose a probability oF M CL 0.1; by contrast, someone who believes in it might choose 0.9. (p. 162) -n -0 They then provide an example of testing the hypothesis H: P = .5, whe re P is the proportion of hits expected in a binomial experi- F ment. Suppose that in 17 trials there are 13 successes (76.5%). Tlicn CM the p value is .049, twomtailed. Unless, of course, the experiment was n M designed to stop at the fourth failure instead of at the 17Lh trial. "0 Then the p value, With the identical'data, would only be .021. Such 0 (4 problems arise with classical methods, but not with Bayesian meth- 0 ods. 44. Using the Bayesian approach, suppose that one's p@rior belief 00 M Mthat H is true is 50%. If H isn't true, the prior belief is that the true 0 value of P is equally likely to be anywhere between 0.5 - c'and 0.5 + c (where c is sorne constant), but could not possibly be Farther X than that from 0.5. The choice of c represents prior opinion about the strength of the effect, if there is one. Choosing c = 0.1 (the CD a) effect isn't likely to be very strong even if it exists) results in a final 6 probability of 0.41 for H (given that there were 13 successes in 17 0 4 Lrials), whereas choosing c = 0.4 rcsulLs in a probabiliLy of 0.2 1 for 00 W H. In other words, the final degree of belief in I-I is dependent on X one's prior belief about the strength of th& effect. It also depends 0 0 CA) on prior opinion about the veracity of H, and on the observed data. 0 ne reason that Bayesian methods are not more widely used is 0 0 that they are often difficult to apply. Another reason is that re- C4 0 searchers are uncomfortable-'with having to specify subjective de- 0 grees of belief in their hypotheses. This approach makes particular sense for parapsychology, however, because most researchers have strong opinions"about the probability that psi is real, and these opin- ions play a central role in how psi researchers and critics evaluate the evidence. Posterior probabilities in Bayesian analyses are a furic- don of both the prior probabilities and the strength of the evidence; it may be informative to formalize these opinions and to see how much evidence would be needed to increase the posterior probabil- icy of a psi hypothesis to a non-negligible level when the prior prob- abiliLy %V,15 C105C LO ZCI-o. Replication vs. sigw@icance 319 REFERENCES BAKAN, D. (1967). On method: Toward a reconstruction of psychological investi- gation. San Francisco: Jossey-Bass, Inc. > 0. (1985). Statistical decj@ivn thewy and Bayesian analysis. Now M York: Springer-Verlag. -1 BERGER, J. 0., & BERRY, D. A. (1988, Marcli-April). Statistical analysis aQ M the illusion of objectivity. American Scientist, pp. 159-165. FL BERGER, J. 0., & DELAMPADY, M. (1987). Testing precise hypotheses. Staimm tical Science, 2(3), 317-334. 0 BERKSON, J. (1938). Some difficulties of interpretation encountered in tl;e ipplication of the chi-square tcst.Journal of the American Stelliaical Ass," ir ciation, 33, 526-542. 2) 0 COOVFR, J. E. (1975). Experiments in psychicial research. New York: Arne Press. (Originally published 1917) 0 DRUCKMAN-, D., & SWETS, J. A. (1988). Enhancing human performance. WaslE ington, D. C.: National Academy Press. 0 EDWARDS, W., LINDMAN, H., & SAVAGE, L. J. (1963). Baycsian statistical irk ference for psychological research. Psychological Reuiew, 70, 193-242. 080 FisjiER, R. A. (1926). The arrangement of field experiments. Journal of 11@; Ministry of Ag7iculture of Great Britain, 33, 503-513. I-TANSE.N. G. P., Scm.rnz. M.J.. & TART. G. T. (1984). BiNiography, remot > viewing research, 1973-1981. In R. Targ & K. Harary, The mind (pp. 265-269). New York: Villard Books. HONORTON, C. (1984). How to evaluate and improve the replicability parapsychological effects. In B. Shapin & L. Coly (Eds.), The repealabili 19 problem in parapsychology (pp. 238-255). New York: Parapsycliologg Foundation, Inc. HONORTON, C. (1985). Meta-analysis of psi ganzfeld research: A rcspons@ to Hyman. Journal of Parapsychology, 49, 51-91. 0 CA) ._---HyNrAN@--R @198-5).-Th-e-ganzfefcl --psi --experimerfu- A--critical-appraisal. J'our- 0 nal of Parapsychology, 49, 3-49. 0 0 MCCLENON, J., & HYMAN R. (1987). A remote viewing experiment conw clucted by a skeptic and a believer. Zetelic Scholar, Nos. 12/13, 21-33. 00 PALMER, J. (1985). An evaluative report on the current stat S of parapsyo u LL chology. U.S. Army Research Institute for the Behavioral and Sociat. Sciences, Alexandria, VA. RAO, K. R. (1984). Replication in conventional and controversial sciences. In B. SHAPIN & L. COLY (Eds.). The repeatability problem in parapsychology, (pp. 22-41). New York: Parapsychology Foundation, Inc. RmNE, J. B., & PRA-rr, J. 0. (1957). Parapsychology: Frontier scie7ice of Ilse mind. Springfield IL: Charles C. Thomas. RHINE, J. B., PRA-1-l', J. G., STUART, C. L., Smri,H, B. IM., &- GREENwOOD, J. A. (1940). Extramsensory perception after sixty years. Boston; Bruce Hum- phries. Iq Cl) W 00 t- W 00 C14 0 LL > 2 CL CL Cl) a) 00 I- 9 w a) 00 9M-6 yo '5712VS vixtofyDO jo ?C71s'llaal U's 0.17sliv)s jo uo.zs-zaig -0:) pul! LII!WSD.10,1 '110:)s :-11 'M@,IAUDJO .2j!l Fix k2opljlk,@j -(Ss -0 'd 'OMMIN12 61) -ZOP-969 '09 'aO7OV-7-(yd'D-Ivd',fO pmcnof -.)Ap:)adsx?d s,umpsnels V :;Dleq.)p P[QJzuL'S OLI-L '(9861) Tj -f,sMLLfj -SSDJl XIISlaAlUfl Qgp!aqwr-o :;)Sp!.iqweo -ggs-mq puv vi2syngH :@7uwj 0 -.IdL?U?t Idplt?L A@S-1-DA.L -V D@ ':)IAOI' LL . s <1 Ul mo '9.[DqLunu Ilms jo mul @qi ul PIP9 -(926 1) -Cf 'NVN3NHVX q "V 'ANS-daAg 0 L- JZ).qzll.qK@Cj.tO,U?2?20s uv3u9zay -VIJO 7-Ou.Lnof ioj CL CL ADu Z)UIOS -(8461) 'S -Cl -NOICINna 's "i :a 'x-li-X 'NIAA-B4 Yll)ltllv -V -11 B1111M.-DADA UO '(91,C)I) 'r --I "-l.)VAVS RF-FF 195i k2O7O'l-,kTIjO JOU.1nof jE:)i2ojc)iqDAsd Aq D:)ue:) SIDA.)IJO 11 .OJJVI@) "MI.I.NAN0@1 J. (.70707j9k.T(IT)j-V,j JO Approved For Release 2003/04/18: CIA-RDP96-00789ROO3100030001-4 Paranormal Communication "Error Some Place! by Charles Honorton Review of the ESP controvcrsy traces debate from statistical and methodological issutes to the a priori critique and the paradigm of "normal science." Asked his opinion of ESP. a skeptical psychologist once retorted, "Error Some Placel" I believe he was right, but for the wrong reasons. Western science has always been ambivalent toward the mental side of reality, and it is perhaps not surprising that the occurrence of "psychic" phenomena is one of the most controversial topics in the history of science. The first serious effort toward scientific examination of psi claims was undertaken by the Society for Psychical Research (SPR), founded in London in 1882 for the purpos_e of "making an organized and systematic attempt to investigate the large group of phenomena designated by such terms as mesmeric, Psychical, and spiritualistic." The SPR leadership included many distinguished scholars Of the period, and similar organizations quickly spread to other countries, including the American Society for Psychical Re- search, founded in New York in 1885 under the aegis of William James, who himself took- an active role in early investigations of mediumistic communications. These turn -o f-the-ce n tury investigators focused much of their attention on allLhenticating individual cases of spontaneous experiences suggestive of psi communication. While a great deal of provocative material was care- fully examined and reported (e.g., 18), the limitations inherent in the case study approach prohibited definitive conclusions. However thoroughly au- thenticated, spontaneous cases Cannot provide adequate assessment of such potential sources of contamination as chance coincidence, unconscious in- ference and sensory leakage, retroactive falsification, or deliberate fraud. Charles Honorton is direcEoT of rew2rch in the Division of Par2psycho)ogy and Psychophysics.' DeparLment of Psychiatry, Mairnonides Medical Ccnter, Brooklyn, N.Y. Approved For Release 2003/04/18 CIA-RDP96-'00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO31,00030001-4 Journal of Communicatio-", Wintt-r 1973 Earl), experimental approaches pririarily duction of drawingsat a distance (61). were obtained, the experimental cmclitions random sclection of target (stimulu!) adequate with respect to the possitility otherwise. Neither the spontaneous case sit made much impact upon the scientif comment from prominent period s, the Fellows of the Royal Society, nc proclaimed Helmholtz, "would lea( thought from one person to anothei nels of sense." Thomas Huxley df some of the early SPR investigation idle gossip of old women. involved the "telepathic" re:40- while often striking corresponden:es did not usually provide @or_ material, and were not always to:_Ai), of sensory leakage, intentionn@ or Jies nor the early experimental eFc IIrts : community, though they dre\k- critical ientists. "Neither the testimony of 311 even the evidence of 'my own sens4<."_-- me to believe in the transmission of- independently of the recognized ch--.n- Iined an invitation to participate inz-,-, , saying he would sooner listen to id I e Th-e rudime-nts of an experimental methodo1bg), for testing psi were s-uggested th'ree centuries ago by Fi on. 104 In Sylva Sylvartin?, a work pul@lished posthumous]),, Bacon discu @td ..experiments in consort, monitory' touching transmission of spirits rnd I _ . !T@ forces of imagination." He suggested that "the motions of shuffling C21C' or casting of dice" could be used ',to test the "binding of thoughts. .1 ." The experiment of biridine of tho@ghts should be diversified and tirie to the full; and you are to note whetl@ er it hit for the most part though not-- always" (2). The application of probability itheory to the assessment of cleviati-ons from theoretically expected chancei outcomes was introduced to psychiiuL_ research in 1884 by the French Nobel laureate, Charles Richet, in exF.)en. ments involving card-guessing. Thl popularity of card-guessing as an ex- perimental methodology was great] influenced by the work of J. B. Rl@iine and his associates at Duke Universitiy in the early 1930s. ' Rhine (50) dev%sft@@ a standard set of procedures around a simplified card deck contaiAin@-. s 'z s randomi ed e uences of five geometric forms (circle, cross, wavy lilne@, i 6rc ul-ee . These "ESP cai square, and ds" were prepared in packs of 25, lan( e @ac i ' r @un @hrough the pack was a (>ciated with a constant binomial prol s@ ability of 1/5, since subjects were 'ot given trial-by-trial feedback. Pr -)vid- ing the experimental conditions were adequate to eliminate illicit sen!sor cul>es, recording errors, and rationlal inference, statistically significant 4 d(- partures from binomial chance exF,@ctation were interpreted as indicating extrasensory communication. Initially, "telepathy" tests conslIsted of having a subject in one r-oor attempt to identify the order of Oic cards as they were observed a I ,.agent" in another room. In "clair@oyance" tests, the subject attempteid to "guess" the order of the cards dire@ tly, as they lay concealed in an opaql. Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Pwartortnal COMM(MiCOUGff / "Error Some Place!" container or in another room, Without an agent. "Precognition" tests, introduced somewhat later (59), required the siibiect to make anticipatory guesses of the card order befo)-c the pack was shuffled or otherwise random- izCd. Rhine introduced the term "ESP" in his first major report on the Du@c University work in 1934 (50). He reported a. total of 95,7724 card-guessing trials, carried out with a wide variety of subjects and under a wide range of test conditions. The results as a whole were astronomically significant, though informal exploratory trials were indiscriminately pooled with those carried out under more carefully controlled conditions. The best-conLrolled work during this period was the Pearce-Pratt distance series of clairvoyance tests (58). in which the subject, Pearce, located in one building, attempted to identify the order of the cards as they were handled, but not viewed, by Pratt, the experimenter, located in another building. The level of accuracy obtained in this series of 1,850 trials was associated with a probability of 1 0@22. As a stimulant to experimental research, Rhine's work had unprcce- dented influence. For the first time a common methodology was adopted and employed on a large scale by a number of independent and widely separated investigators. For the first time, also, the scientific community was confronted with a body. of data, collected through conventional meth- ods, which it could no longer ig-nore-nor too hastily accept. The wide- scale adoption of the card-guessing methodology was accompanied by a plethora of critical articles, challenging almost ever), aspect of the evalua- tive techniques and the experimental conditions. During the period be- tween 1934 and 1940, approximately 60 critical articles by 40 authors ap- peared, primarily in the psychological literature. 'While card-guessing is no longer the primary methodology in experimental parapsychology, the questions which arose over its use are of equal relevance to the more sophisticated approaches used today. The first rnajor issue concerned the validity of th-- assumption that the p,robability of success in the card-guessing experiments was actually 115. If chance expectation is other than 1/5, the significance of the observed deviations would obviously be in doubt. This issue was quickly resolved by mathdmatical proof and through empirical "cross-checks," a form of control series in which responses (guesses) were deliberately compared with target orders for which they were not intended (e.g., responses on run n, matched with the target sequence for run n..). Empirical cross-checks were reported for 24 separate experimental series involving a total of 12,228 runs (M5,700 individual trials). While the actual experimental run scores (e.g., guesses on run n) compared to targets for run n1) were highly sig- nificant and yielded a mean scoring rate of 7.23/25, the control-cross-check Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 CIA-RDP96-00789RO031!00030001-4 Journal of Coinmunkatio", Winter 1975 scores were in all cases nonsignifican ;t, with a mean &coring rate of 5.04 (48). Several critics questioned the ap@licability of the binomial distribution as a basis for assessing the statistical@ significance of ESP card-guessing data. Willoughby (79) proposed the use @f an empirical conir-ol scries, but latier Nvithdre-@v the suggestion after compAring the two mctlicds (79*,. Alternati've methods of deriving the probable 'error and recommendations for usi!ng the empirical standard deviation w@re also proposed and later withdrawn (21, 22). Concern over this issue d#nished and was g!nera)]y abandoned following the publication of a large chance control series involving half a million trials and demonstrating! close approximation to the binom'ial- Model (12). Another question arose about whether the binomial model provides sufficient approximation to the nor-mal distribution to allow use o no7al probability integral tables for determination of significance lev Is (17). StMrL and Greenwood (73) showed that when the no-mal 6istribution is used as an approximation to the @inomial model, discrepp.-icies are i,m-_ portant only with cases of borderline significance and few tr;als. The use of the binomial criticallratio (z) to evaluatt the sig-nificance' of the ESP card-guessing deviations was generally approved b%. professional statisticians (6, 20). Fisher (10), ho,,wever, commented that high levels of- statistical significance should not be:accepted as substinites for independent replication. In another vein, Huntington (20) asked, "If in?@hematics has successfully disposed of the hy ,pothOsis of chance, what has psychology to say about the hypothesis of ESP:` I The moflt frequently expressed methodological as tF-- possibility of sornt' comem W form of "semory leakage," giz4ng the ESF subjeq e-nough inform-ation about the targets to acc I6unt for significant, extrachance res-ults As earl), as 1895, two Danish: psychologists, Hansen and Lehmann (16), reported that with the aid of @arabolic reflectors-subjects could detec- digits and other material silently concentrated upon by an agent. In thes--- experiments, the subject and ageni sat with their heads close to the foci of two concave mirrors. While the.agent concentrated. on the number, h- made 2 special effort to keep his I ps closed. Under these conditions, th- subjects were frequently successful in identifying the number. These results were interpreted by Hansen aitd @thmann as supporting the hypothesz-1 of "involuntary whispering." Thei utilization of subtle sensory cues W; demonstrated in a careful investigatilon by S. G. Soal of a stage 11 telepathist (66). There were also reports, such as the case of "Ilga K.," a mentally- retarded Latvian child who could r@acl any text, even in a foreign lang@ag when someone stood behind her.'! reading "silently." Experiments 'wit.. dictaphone recordings revealed that "Ilga" was responding to very s@ight auditory cites (3). Approved For Release 2003/04/18 CIA-RDP96-00789RO031'00030001-4 Approved For Release 2003104118 : CIA-RDP96-00789ROO3100030001-4 Journzi of CO-Mrntmication, Winter 1975 able to the ESP hypothesis made 71.5 percent more errors of commission (increasing ESP scores), while those who were unfavorable to the ESP hypothesis 1`112de 100 percent more errors of omission (decreasing ESP scores). Murphy (37) reported an analysis of 175,000 trials from experiments reporting Positive evidence for ESP and found only 175 errors (0.10 per- cent). Greenwood (12) reported only 90 recording er-rors in rechecking his 500,000-trial control study, of which 76 were errors of omission. Some critics also alleged that improper selection of data could account for experimental successes. This could be done in several ways: (a) selection of subjects; (b) selection of particular blocks of data out of larger samples; (c) selection of one of several forms of analysis; and (d) selective reporting of particular studies. The qpestions raised have sometimes b@ten stated cynically in the form, "Parapsychologists must run 100 subjects before they find one with 'ESP'." As if in defense against this charge, a number of the reported studies specifically stated that all of the data collected were included in the' analysis (see 43, pp. 118-124, Table 12) Concerning selection of subjects, Warner (76) suggested two criteria: first, results of "poor" subjects must be included up to the point when they are discontinued since it does not matter how man), trials a giNen subject makes as long as all of the trials (for all subjects) are included; second, exclude all preliminary trials (for both "good" and "poor" sub- jects) and use preliminary screening studies to select "good" candidates for fonnal work, These criteria were generally endorsed by the chief critics of the period (e.g., 23). The question of post hoc selection of analyses wa 's not a point of serious concern in the period between 1934 and 1940, though it is relevant to the assessment of Some of the pro'cess-oriented investigations reported more recently. The question of whether nonsignificant studies were withheld from publication involves an issue which is of great toncern to the be- havioral sciences as a whole (70, 81) and one which is difficult to accurately assess since there is no way of knowing how many studies may have be@tn withheld from publication because their results failed to disconfirm the null hypothesis. Several studies of American Psychological Assc>ciation publication poli._ cies (4, 70, 81) indicate that experimental studies in Vneral are more likely to be published if the null hypothesis is rejected at the conventional @05 and .01 alpha levels than if it is not rejected. These studies also indi ' that a negligible proportion of published studies are replications. Boza th and Rol>erts (4), in a survey of 1,334 articles from psychological journ Is, found that 94 percent of the articles involving statistical tests of sig-nifi ILC t rep,orted rejection of specific null hypotheses; only eight articles (Iess"@ an I percent) involved replications of previously published studies. With respect to the implications of such selection for the ESP hypothelsis, there are two partial answers. First, considering the degree of critical in er--,,. cst which prevailed in the 1930s, it seems unlikely that nonsignificant fi@d- ings would have been repressed during this period;@ second, the high levels 10.9 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 i aranvrmat Cmmimicalion "Errv-r Somt Plac !P It is clear that at least some of the early ex@ ]oratory series reported in :p Rhine's monograph were open to criticism for @inadequate controls against sensor), cues. While Rhine did not base major 'conclusions on such poorly controlled data, inclusion of thern in his mono I aph provided a read), targe't for critical reviewers and sidetracked discussion away from the better con- trolled work, such as the Pearce-Pratt series, 4ich was not susceptible to explanation by sensory cues. Defects in an early commercial printing of ESP cards were reported by several investigators (18, 25). It was found thaLithe cards were warped and could under certain conditions be identified from the back. This discovery circulated widely for a time as @n explanation 1of all successful (i.e., statis- tically significant) experimental series. The parl,apsychologists retorted that defective cards had not been employed in any @f the experiments reported in the literature and that, in any case, they ccluld not account for results from studies involving adequate screening with such devices as opaque envelopes, screens, distance, or work involvingithe precognition paradigm in which the target sequences were not gener@ted until after the subject had made his responses (53, 54, 72). By 1940 nearly one million experimental tri:ls had been reported under conditions which precluded sensory leakage. These included five studies in which the target cards were enclosed in opaque sealed envelopes (41, 45, 46, 54, 59), 16 studies employing opaque screeni (7, 8, 11, 19, 83, 34, 35, 38, 41, 42, 44, 45, 46, 59, 71), ten studies involvinj separation of subjects and Largets in different buildings (50, 51, 52, 53, 341 32, 8, 77, 61, 60), and two studies involving precognition tasks (59, 75). These data are summarized in Table 1. The results were independently signiricant in 27 ofthe 33 experi- ments. e agreement that the better- By the end of the 1930s th re was gener controlled ESP experiments could not be accbunted for on the basis of sensory leakage. The hypothesis that significant -extrachanc@" deviatioris in ESP experi- menLs might be attributable to motivated scoriAg errors was investigated in several studies. In one investigation (26), 28 obse'rvers recorded 11,125 mock ESP trials. Of these, 126 (1.13 percent) were mlli s recorded. Observers favor- Table 1: ESP icard-guessing experiments (199-19@9) excluding sensory cues- Method Studies N (Trials) Mean/25 P< "Clairvoyance" paradigm, stimuli in sealed, opaque envelopes 5 129,775 5.21 4.@x'10-11 ;7 "Clairvoyance" paradigm, stimuli concealed by opaque screens 16 497,450 1 5.44 10,-31 3 /, 03 Distanceb 10 164,475 5.37 15"00 Precognition paradigme 2 115,330 5.15 2.95,X 10@1 5,0 References given in text. b Includes work with both "telepathy" and "clalrvoyahce" paradigms' Stimuli generated after subjects made their responies 7 Approved For Release 2003/04/18 CIA-RDP96-00789ROO31,00030001-4 O-C Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Journal of Communication, Wiritcr IY73 Two recent examples, one ini-ol%,ing cancer research (7-1) and the oth in%-ol%,ing parapsychology (57), scrve to remind us of tile Importance cross-validation in the assessment of any experimental finding. In bo(h cases, it should be added, tile fraudident acts were detcc(ed in-hoilse, the researchers tllcmscl%,es. The point is that in the final anal@sis an czpc mental finding is of valtic and is to be takcri seriously on]y to the eixtent that it leads to further inquiry. To regard an), cxl)crimcj)L as ail cild in itself is to remove it from the domain of experimental science. It is ob'Yio that hypothetical construct, such as ESP, cannot be validated b)l ai, isolated experiment, no matter how well controlled it might be. Inde. pendent replication is a necessary prerequisite. The claim that psi phenorn-cna op,@rate outside the framework of physLcal probo I i has been a major source, of- priori arguments against acceptance of !ESP. It has been suggested that to accept ESP requires the rejectign - physics. This is absurd, and it is worth noting that such arpimem@ have r, usually been advanced and defended by psychologists rather than phN i i IsIcists. The debate over the incompatibility of physics and ESP has bee@ri conducted almost exclusive]), within the framework of nineLeenLh-CenLur%- deterministic physics, wherein the ultimate constituent of physical reali was still believed to I>e solid matter. Inasmuch as modern microphysics h_- exorcised the material out of matter and deals with processes which on our macrophysical level of sensory perception are ever), bit as erratic ar anomalous as ESP, the a priori claim that ESP violates specifiable ]aws physics can no longer be considered to be of more than historical interest. ESP and other psi phenomena, while no longer incompatible wi physics, are not yet accounted for by physics; but dien, neither are tI. more familiar processes of memory and conscious experience. Indeed, the transformation of "raw' feels" into conscious experience is no less a proble-, for the neurophysiologists of today than it was for the speculative pllil( opliers of classical antiquity. Sir John Eccles, among others, has mpentcd-ly, warned, "I'Ve should not pretend that consciousness is not a mystery." The ESP controversy illustrates several features of the paradig-r@l at. view of science developed by Thomas Kuhn (28). Normal science, accordin@ to Kuhn, is essentially a clean-up operation, constrained by a bro; theoretical framework, or paradigm, which defines the boundaries of legi mate inquiry. Paradigms are scientific world views which provide cohe4'ence and structure and detennine the (ypes of questions to be posed of na!tiir- as well ais the manner in which answers are sought. Normal science is tlll@" a process of paradigm-articulation, rather titan of discovery. Withio, the paradigm structure of non-nal science, observa@ions which conflict with 112 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 CIA-RDP96-00789ROO31.00030001-4 c place!" the wewh ryA implicalions. @,Ypfccl, nllis! -!Ing. - some. crj116sm,i,`%1,,..4i!@Qff aP4-,4i1j! n imuk :mbly"hed siu@@,rs. ng.; challrnge',-, rer -firn@ ar_---if 0", 4a, nd mous irrvv,w .-a b;g-c@rvd -rtt Ory milrat Ancornpat.111-@-: 4pt O-C th6OrV_.,, ifi@ Ithafj( is ih6 @Ore parsimani- @s xW jV, nc, real nbtno,r ttd, My opnkmC P 4 If .statlsticaleirc-',,@.a-,_i q.-iijitt- 1i of @;CA@Ory np'prla ute il u f@,: rtc@ chance res-Au ;,V@z cre"'W Ownde,.*@ 'Aw. @4 J,, neldlty nbmorinal rd@j!igj 7 the eMuingd,@f wiliq 5 As M-kell, Ind le "Out" 2' 1 11@,_ @(S "t is compleut., W, tt latc-, in an 6 1 oi i nvestkMe., @jkpi @w -i @;, , mar! Vt?*)' SIM-` j4 . cTi his Wit, V@ "OtOl S. au d oq, i%, %) Becausf., of tt%e fi ]v e,,,jj _71 distante w Anil a uw PrA t was CIO= r&6 af f Tth con, u r a H , --,W-, 4 A e u, Qc h +7,@t 'N s k I 'fin &tN 4.ccide Oici u ch WJS e ib t W;as that. r1%s1_Vt VMKI11- T"4 jnz(5r11.yV* It ilky'. Ofr1_1(mud-4 10 M Pwclh;z@, i 11.1% t,- I ta Approved For Release 2003/04/18 CIA-RDP96-00789ROO3,100030001-4," Approved For Release 2003/04/18 : CIA-RDP96.-OO789ROO3100030001-4 0j Communkofj", )J1jjjteT 1975 experimental ESP stuAles (1934-1939) Table 2: Ereakdownp,@, _N N studies reported s:ignificant (P < .01) % signif. W, ouk@ grQWb---- 17 15 88 33 20 61 Non-Duke I @0:,_ 70 total 35 Includes all English uage studies Involving assessment of statistical significar -lang -,of data, 1934-1939 inclusive,, V@; x1 (Duke vs. non7Duke @Ignlflcant vs. nonsignificant) 1.70 (1 df) 7 '4 A.. -sequences have bieen "doctored" prior to publicadon in order to@ remo,. tVA K -to Spencer Brown, Artain nonrandom feztures; this practice,-gccorcling the use of standa m4kes such,'sequences - nonrandom and. invalidates' "sign i fica nce @ests; (b): the _ku@ce of some random, numb@e:r sequences invol% - , . randomizing machines which utilize the unpredictability of human be wheil' ekaininecl@ f8r ' 'microscopic variation@.@ Such variation, s2--- havior :A@ pencer Brown ma ' be@ "redictable enough to account for observed ano: S y p C 0 ecl TSUItS T P It falies in random sequences, as well as some of the significant I in ESP guessing experiments; (c) Spencer Brak,@fn Produces evidence to show that anomalous (significant departures from probftbility theory can 7 -bbnined by, matching columns of random numbers (39). -A detaiW examination of these points'was i4ndertaken by Scott, (64). " T, Vith respect to "cloctored" sequences, Scott showed that the maximu rsequences would not affect interpretatio---. error due to rejected (edited) . of results which are m0re'thifi marginally significant and could, in fact, -increase the, .likelihood of making a Type 11 error. On the hypothesis th ESP resu lts are due to s6ffie kind of hyper-regularity affecting both t , Aarget' sequence, and the 4'esponse (guess) sequence simultaneously and -similarly Scott makes1the point that this would lead to the expectation similar results from 'niateh'in.4,any set of humanly produced random _quences. The cross-check, type"of control series and the Greenwood chanLC con trot'se s ries de cribed earlier demonstrate that this is not the case. The anomalie's re r @ecflb@ Sp@ncer Brown (68), obtained by arbitrarily matchi po t lot random numbers, have been ctiticized on the basis of post h columns (40) and illustrate not that there are fundamental defects in prob- bility theory,' but i4therthat significant deviations from chance c i -an _occ6r,in anydata. where hypotheses and analyses are not specified in ad% :n@ mo@t recent.phase of tht@ ESP controversy centers on the hypothesis @of inve@tigator firau'& This argument was most for ented ir -i cefully pres 1ead article; in Science, entitled "Science and tile Supernatural," by G. Price (47), who began %@ith'ihe following observations: I h Be ievers in s c ic enamena appear to have won a decis y p P xo@ viclo7 y and virlualIXsilenced opposition. . . . This victory is the rej-_ I an iinpressive amount of careful experimentatian and intelligent of, , argurnektation. Against all this evidence", almost the only defe Approved For Release 2003/04/18 :'t[A-RDP96-607'89ROO3100030001-4 Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 P4ranornjal C&mmunini6o-n "ErT*-r Sarme of significance attached to some of the reported IIESP investigations would necessitate postulating astronomical numbers of! "chance" trials in order to dilute the overall deviations to chance. To take one example, consi@`cr the Pearrc-PrRtc series of 1,850 tTi-,11S which yielded p = 10-22. As Soal and Bateman (60) pointed out, it is difficult to belie@'e that 10111 (ten thousand million) sets of 1,850 trials could have possibly been carried out between 1934 and 1910 (or, for that matter, since 1940). But, as Soal and Bateman suggest, ". . . if we posit this absurd estimate as an tipper limit (with overall chance totals], that would Still give us odds of 101,0 ... against the supposi- tion that the Pearce-Pratt results were a run of p' I k c@re cuc The possibility of obtaining significant "extra an e" results by stopp-ng an experimental sciies at "favorable" points was! also raised (9, 31). While this "optional stopping" hypothesis was generally Og-reed to be of sig-nificarice only in cases of marginally significant results, ilt led to the adoption of I several procedural modifications: specification of ithe total number of trials in advance of data collection, or accumulation of data in blocks of pre- determined size. i The possibility was raised by several critics that hand-shuffled cards may display a tendency to stick together or otherw4 produce patterr,5 wh:ch could produce spurious results (24, 82). While tHe cross-check type, of con- trol series, described earlier, failed to reveal anly evidence of patterning, there was a general trend away from hand shuffl g in the later publis-ed studies, which utilized tables of prepared rando@ numbers as a basis for generating target sequences. There was--and is (e.g., 15)-a rather wides d belief that most of the evidence supporting the ESP hypothesis ori ed in the Duke Uni- versity studies and that most independent replicaEby,other investigators were nonconfirmatory. A survey of the published. literature between 1934 and 1940 fails to support this claim. Table 2 sho@s all the published exp-.ri- mental reports during this period which providFd statistical treatment of the data. Inspection of this table reveals that @ imajority (61 percent) of the outside replications report significant result,s (p < .01) and that the proportion of significant studies was not significAntly greater for the Duke University group (X2 1.70, 1 d@. By 1940,! the active methodological controversy UW Oller. The issues raised were, for the most part, legitimate, and investigators modified their procedures to safeguard their res@ults from methodological criticism. The major issues raised since 1940 centfr on alleged anomalies in probability theory and the hypothesis of widespread investigator fraud. Spencer Brown (68, 69) has suggested that statistically significant card- guessing studies provide evidence, not of extrasensory modes of communica- tion, but of fundamental defects in probabilit theory. He makes three criticisms of random number senuences: (2@ ntibli-qherl r;inrlnm nitrnkpr Approved For Release 2003/04/18 CIA-RDP96-00789ROO31.00030001-4 Approved For Release 2003/.04/18 CIA-RDP96-00789ROO3100030001-4 llaranor"wt Comrntmication "Error Sotnr Plnce:- the paradigm are seldom made: anomalies ire ignored. WhC11 the anomalies become sufficiently persistent that they can no longer be ignored, thcv are hot])- disptited. Eventually, @a nc%%- paradigm is tentatively ClIcCied which attracts a gi-0111) of adherents, and a period of crisis cnsiies 'which Kiihn calls a paradigm clash. In thi s review I have focused at sonic Icn.-th on the period of the 1930s, not because it provides the best available evidence for ESP or the best understanding of the processe.@ iii,iderlying its operation-it does neither-but rather because it was during this period that the Major substantive mc1hodological issues were raised and to a large extent con- sensually resolved. Since 1940, well over 10,000 journal pages devoted to parapsychological research have been published, and at least 250 experi- mental studies have been reported. The methodological foundations of the research have gradually diversified, enlarging and enriching the scope ,o( inquiry and providing a basis for more sophisticated study. Automated testing equipment has replaced card-guessing in forced-choice ESP tasks, and quantitative methods have been devcloped for the objective assessment of psi interactions in nongUessing tasks. Psychophysiological techniques, permitting determination of psi-optimal organismic states, have been in- tro-duced. and utilized in conjunction- with experimental methods more close]y approximating the conditions 'under which psi interactions occur in vivo. More important, parapsychological investigators have to a large extent shifted their attention away from the "proof-oriented" approach, which can only reaffirm the presence of anomaly, toward systematic at- tempts to identify the antecedent conditions necessary for the occurrence and detection of psi interactions, the delineation of positive attributes, and the-sLudy of individual differences. Only through the pursuit of such "process-oriented" research can we ever hope to achieve the goals of control, assured replicability (or at least predictability), and eventual understanding. REFERENCES ElAmeric2n Psychological Association. "ESP Symposium at the Americ2n Psycholokical Aswiation." Journal of Parapsychology 2 ,1938. pp. 247-272. 2. Bell, K "Francis Bacon, Pioneer in P2172PSYChOlogy." Internalional Journal of Pnra- psychology 6. 1%4, pp. 199-209. 3. Bender, H. "The Case of 11ga K.: Report of 2 Phenomenon of Unusual Perception." . Journal of Parapsychology 2, )938, pp. 5-22. 4. Bozarth, J. D., and R. R. Roberts. "Signifying Significant Significance." American Psychologist 27. 1972, pp. 774-775. 5. Bridgman, P. W. "Probability, Logic, 2nd ESP." Science 123. 1956. pp. 15-17. 6. Camp, B. H. "Statement tinder 'Notes.' " Jounial of Parapsychology 1, 1937, p. 305. 7. Carpenter, C. R., 2nd H. R. Phalen. "An Expcrimcnt in Card Guessing." Journal of Pirrapsychology 1, 1937, pp. 31-43. 8. CrtimbL2ugh. J. C. An Experimental Study of ExtTa-scnsory Perception. M.A. Thesis, 1 1938, Southern Methodist University Library. 9. Feller, W. "St2tiSOC21 Aspects of ESP." Journal of Parapsychology 4, 1940, pp. 271-298. 10. Fi.4w.'R. A. Letter to J. B. Rhine, cited in (50), p. 45. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 CIA-RDP96-00789RO031:00030001-4 journal of Co-mmunication, Winter 1975 11. Cibson, E. P. "A Study of Comparitivc Pcrform3ncc in Several ESP Proccdurcs-" -rholo jourrial of Parapj) gy 1. 1937, pi@. 264-275. 12. GTCCnW00d, J. A. Analysis of a Larg@ Chance Control Scrics of ESP Data." lourq aI of Parnl).i)-rhologi, 2. 1938, pp. 138-1@6. B. Gurney. E., F. Myers. and F. Podmorc, Pharitasins of 1he Living (2) vols.). Lnndan: TrubnCT and Co., lq86. 14. Hanscl, C. F. Kf. "E%:pcrimcnts on Tclepadiv in Children." Britiih joury1c, of Sfc:@Jliral Psycholngy 13, IcKO. pp. 175-17R. 15. Hanscl, C. E. M. E@P-A Scientific &,,alualion. New YoTk: ScribncTs, 1466. 16. Hamm F. C. C., and A. Lchmann.l"Obcr Unwillkfirlichcs Fliktcrn." Philos,).r)h:ral Sludies 17, IR95, pp. 471-550. 1 17. Heinlein, C. P.. and J. H. Heinlein. "Critique of the Prcmi,.cs and Stmistical Wth- odology of ParapsycholM." journal of PsYchology 5. 1938, pp. 155-1*. 18. Hcrl>crL, C. V. C. "Experiment in Evra-scn@ory Perception." Journal of the Socirly fm Psych iro I RejeaTch 30, 1938, pp. 215-21 S. 19. Humphrey, B. M., and J. A. Clark. "..' Comparison of Clairvoyant and Chance Nfatch- ing." Journal of Parapsychology 2, 19 @8. pp. 31-37. 20, Huntington, E. V. "Is It Chancc or UP)" American Scholar 7, 1938, pp- 201-210. 21. Kellogg, C. E. "Dr. J. B. Rhine and :_'xtra-scns0TY Perception," Journal of Abrorrl.41 and Social Psycholog,; 31, 1936, pp. )W-193. 22. Kellogg, C. E. "New Evidence Q) fo Extra-mensory Perception." Scienlihc Afcnf@ly, 1937. pp. 331-341, 1 23. Kellogg, C. E. Critical comments in (431), pp@ 22F@-238. 24. Kennedy, J. L. "A Methodological Review of Extra-scnsory Perception. Psycholog,-cal Bulleiin 36, 1939, pp. 59-103. 25. Kennedy, J. L. "The Visual Cues f@lom the Backs of the ESP Cards.- Journal of Psychology.5, 1938, pp. 149-153. i 26. Kennedy, J. L., and H. F. Uphoff. "Experiments on the Nature of Extra-scnsof)' Perception: 111. The Recording E rOT CritiCiSM of Extra-chancc ScoTes." journal of Parapsychology 3, 1939, pp. 226-2 15. 27. Kennedy. J. L. Critical comments-in (4p), pp. 218-220. 28. Kuhn, T. S. The Structure of Scient c Revolutions. Chicago: University of Chic@go Press, 1962. 29. Lemmon, V. XV. "The Role of Select ion in ESP Data." Journal of Parapsychology 5. 1939. pp. 104-106. i 30. Lemmon, V. W. Critical comments in 43), pp. 222-228. 31. Leuba, C. "An Experiment to Test t@lc Role of Chance in ESP Research.- journal of Parapsychology 2, )9158, pp. 217-221. i 32. hiacFarland, J. D. Shown Between Experimenters by Subjcclts." journal of Parapsychology 2, 1938, p 160-170. 33. N12ff2rland, J. D., and R. W. George. "Extra-senson. Perception of Norm2l Iand Distorted Symbols." Journal of Paraf sychology 1, 1937. pp. 93-101. 34. Martin, D. R. "Chance and Extra-:hancc Results in C2rd-matching." journJ Parapsychology 1, 1937, pp. 185-190. 35. Martin, D. R., and F. P. Stribic. "Studies in Extra-sensory Perception: 1. An AnaJ@sis of 25.000 Trials; and 11. An Analy0s of 2 Second Series of 25,000 Tri2ls-" Journal of Poropsyehology 2, 1938, pp. 23-30:1287-295. 3,6. Mcchl, P. E., and M. Scrivcn. "Com atibilitv of Science and ESP." Science 123, 1956, pp. 14-15. 37. Murphy, G. "On Limits of Recording rrors." In (1), pp. 262-266. 38. Murphy. G, and E. Taves. "Covariance Methods in the Comparison of Extr;t-sc-n- Tasks." Journal of Parapsychology 3, 1939. pp. 38-78. 39. Oram. A. T. "An Experiment with Random Numbcrs." Journal of the Soc-icfy for Psychical Research 37, 1954, pp. 369-377. 40. Or2m, A. Ti Correspondence. journo Iof (he Society for Psychical Rcicarch U. 1 %5 pp. 143-144. Approved For Release 2003/04/18 CIA-RDP96-00789ROO31,00030001-4 Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 Paranor"tal Co-mmunication / "ErT&r S&rne Mace!" 41. Pratt, J. G. "Clairvoyant Blind M;ttching." Journal of Parnpsychningy 1. 1937, pp. 10-17; 2nd "The N%lork.of Dr. C. Hilton Rice in Extra-scnsory Pci-ccption, pp. 259- 259. 42. Pratt, J. G., and M. Price. "The Exp-crimentcr-subject Relationship in Tcsik for F.@P." Journolof Parapsychology 2, 1938. pp. 94-94, 43. Pratt, J. C.. J. B. Rhine. B. Nf. Smith, C. E. Stu2rt, and 1. A. Greenwood. Extrasenjory Perception After Sixt). Years. Boston: Bruce Humphries, 19V). (Origin2lly published in 1940 by Henry Holt, New York.) 44. Rratt, J. G., and J. L. Woodruff. "Size of Stimulus Symbols in Extra-scnsory PcTcep- tion." journal of Parapsychology 3. 1939, pp. 121-158. 45. Price, M. "A Comparison of Blind and Scting Subjects in ESP Tests." Journal of Parapsychology 2, 1938. pp. 273-2136. 46. Price, M., and M. PcgTam. "Extrasensory Perception Among the Blind." Journal of Parapsycholog)- 1. 1937, pp. 14-9-155. 47, Price, G. R. "Science and the Supernatural." Science 122, 1955, pp. 359-367. 48. Price, G. R. "Where is the Definitivc Experiment?" Science 123, 1956. pp. 17-18. 49. Price, G. R. "Apology to Rhine and Soal." Science 175, 1972. pp. 359. 50. Rhine, J. B. Extra-senior), Perception. Boston: Bruce Humphries, 1964. (Origin3ily I published in 1934.) 51. Rhine, 1. B. "Telepathy and Clairvoyance in the Normal and Trance States of a . Medium." Character and Personalit.v 3, 1934, pp. 91-111. 52. Rhine, J. B. "Some Selected Experiments in Extra-sensory Perception." Journal of Abnormal and Social Psychology 31. 1936. pp. 216-228. 53. Rhine, J. B. "The Quevion of Sensory Cues and the Evidence." Journal of Para- psychology 1, 1937, pp. 06-291. 54. Rhine, J. B. "ESP Tests with Enclosed Cards." Journal of Parapsychology 2, 1938, pp. 199-216. 55. Rhine, J. B. "CommenLs on 'Science and the Sup-ematural.'" Science 123, 1956, pp. 11-14. 56. Rhine, J. B. "The Experiment Should Fit the Hypothesis." Science 123, 1956, p. 19. 57. Rhine, J. B. Comments. Journal of Parapsychology, 1974, p. 38. 58. Rhine, J. B., and J. G. Pratt. "A Review of the PeaTce-Pratt Distance Series of ESP Tests." journal of Parapsychology 18, 1954, pp. 165-177. 59. Rhine, J. B., B. M. Smith, and J. L. Woodruff. "Experiments Bearing on the Pre- cognition Hypothesis: 11. The Role of ESP in the Shuffling of Cards." Journal of Parapsychology 2, 1938, pp. 119-131. 60. Riess, B. F. "A Case of High Scores in Card Guessing at 2 Distance." Journal of Parapsychology 1, 1937, pp, 260-263. 61, Shulman, R. "A Study of Card-guming in Psychotic Subjects." journal of Parapsy- chology 2, 1938, pp. 96-107. 62. Sinclair, V. Mental Radio, with a Preface by Albert Einstein. Springfield, Ill.: Charles C. Thomas, 1962. (Originally published in 1930.) 63. Shapiro, A. "Review of ESP-A Scientific Evaluation by C. E. M. Hansel." International Journ.al of Clinical and Experimental Hypnosis 16, 1968, pp. I 64. Scott, C. "G. Spencer Brown and Probability: A Critique." Journal of the Society for Psychical Research 39, 1958, pp. 217-234. 65. Soal, S. G. "On 'Science and the Supcmatural.' " Science 123, 1956, pp. 9-11. 66. Soal, S. G, and F. Bateman. Modem Expeiiments in Telepathy. New Haven: Y21C University Press, 1954. - Recent Experiments in Telepathy. 67. S021, S. G., and H. T. Bowden. The Mind Readers. New Haven, Conn.: Yale University Press, 1954. 68.,Spcncer Brown, G. "Statistical Significance in Psychical Rescarch." Nature 172, 1953, pp. 154-156. 69. Spencer Brown, G. Probability and Scientific inference. New York: Longmans. Green. . 3957. 70. Sterling, T. C. "Publication Decisions and Their Possible Effeas of Inference Drawn 115 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 CIA-RDP96-00789ROO31,00030001-4 jo,urnal of Co-mmunication, Wint- 1973 from Tests of Significancc-Or @icc Versa." Journal of the Ame-ricarl Statistical socio6on 54. 1959, pp. 30-34. 1 71. Stuart, C. E. "The Effect of Rat of Movement in C2T(! Matching Test.; of Ex sensory Perception." Journal of 41-psychology 2, 1958. pp. 171-IS3. 72. Stuart, C. E. "A Review of C t Proposed Hypotheses AlLcm2tivc to Exit-2-scn! er Perception.--- Journal of Abno,;l and Sorinl Psycholory 33, 1938. pp. 57-70. 73. Stuart, C. E., and J. A . Greenwood. "A RevieW of Criticisms of the N12thcmzt Evaluation of ESP Data." journal of Parapsycholop 1, 1937. pp. 295-3N 74. Time, April 29, 1974. 75. TyrTel), G. N, M. "Some Expc ri en 's in Undiffcremiate,@ Ext]72-scnsory Pcrcrpd( journal of the Society for Psychii1cIResearch 29, 1935, pp. 52-M. -rncr, L. "The Role of Luck F Data." journal of Parapsychology 1, 1937, 76. W2 _SP 8+-92. 77. Warner, L. -A TW Case." Journ 1 of Parapsychology 1. 1937, pp. 23+-238. 78. Willoughby. R. "Critical Comment: The Use of the Prcb2blc Error in Evalua Clairvoyance." Character and Personality 4, 1935, pp. 79-60. 79. Willoughby, R. "A Critique of Rhine's 'ExLra-sensoTy Perception.' " journal of normal and Social Psychology 30,11935, pp. 199-207. 80. Willoughby. R. Critical comments In (43), pp. 220-221. 81. Wolins, L. "Nce-dcd: Publication f Negative Results." American Psychologist 14, 1 P. 598. 11 82. NVolfic, D. L. "A Review of the Work on Extra-sensory PCMCpLiorl." American jou of Psychictry, 1938, pp. 943-955. Approved For Release 2003/04/18 CIA-RDP96-00789RO03iOO030001-4 journal of Parapsycliology, Vol. 53, December 1989 "FUTURE TELLING": A META-ANALYSIS OF FORCED-CHOICE PRECOGNITION EXPERIMENTS, 1935-1987 By CHARLES I l0N0RT0N AND DIANE G. 'FERRARI ABSTRACT: We report a meta-analysis of forced-choice precognition experiments published in the English-language parapsychological literature between 1935 and 1987. These studies involve attempts by subjects to predict the identity of target stimuli selected randomly over intervals ranging from several hundred milli- seconds to one year following (fie subjects' responses. We retrieved 309 studies reported by 62 investigators. Nearly two million individual trials were contributed by more than 50,000 subjects. Study outcomes are assessed by overall level of sta- tistical significance and effect size. There is a small, but reliable overall effect (z = 11.41, P = 6.3 X 10-"). Thirty percent of the studies (by 40 investigators) are significant at the 5% significance level. Assessment of vulnerability to selective re- porting indicates that a ratio of 46 unreported studies averaging null results would be required for each reported study in order to reduce the overall result to nonsig- nificance. No systematic relationship was found between study outcomes and eight. indices of research quality. Effect size has remained essentially constant over the survey period, whereas research quality has improved substantially. Four moder- ating variables appear to covary significantly with study outcome: Studies using subjects selected on the basis of prior testing performance show significantly larger effects than studies using unselected subjects. Subjects tested individually by an experimenter show significantly larger effects than those tested in groups. Studies in which subjects are given trialmby-trial or run-score feedback have significantly larger effects than those with delayed or no subject feedback. Studies with brief intervals between subjects' responses and target generation show significantly stronger effects than studies involving longer intervals. The combined impact of these moderating variables appears to be very strong. Independently significant ou(coines are observed in seven of the eight studies tising selected subjects, who were tested individually and received trial-by-trial feedback. Precognition refers to the noninferential prediction of future events. Anecdotal ChlillIS Of "future telling" have occurred throughm Out human history in Virtually every Cult-Ure and period. Today such This work was funded by SRI International and the John F. Fetzer Foundation. We wish to thank our PRL colleague George P. Hansen, who is primarily responsible for retrieving the studies used in the meta-analysis. Wearc grateful to Edwin C. May, Jessica Ults, and to live allollylliolls reviewers at SRI for valliable comments oil all earlier draft ol'this report. Valliable collinlents were also made by 1"'phraill) Schechter and by three anonymous referees. Tile division ofauthorship responsibility is as folm lows: Honorton is responsible for the design of the meta-analysis, definition of study coding criteria, the actual analyses, and the report itself. Ferrari coded the individual research rcpolts ill consultation with I lonortoll and/or I 1:111sell. 0 M CL -n 0 M 2) CO) M 1%) 44- 00 0 > 4 00 (D C? 282 Thr Journal of Parap.@TcholoAry Cklillis arc gencrally belicvcd to be based oil kiclors; Such as delusion, irl-atiOllality, alld SUI)CI-StitiOus thinking. The concept of' precogni- Lion runs counter to accepted notions of' causality and appears to ccinflict with current scientific theory. Nevertheless, over the past Qf-century a substantial number of experiments have been re- p(ated claiming empirical support for the hypothesis of' precoglli- tiji. Subjects in f'orced-choice experiments, according to many re- 13(9fLS, have correctly predicted to a statistically significant degree the idThtity (or order) of target stimuli randomly selected at a later - N LlW. Sle perf'orined a ineta-analysis of f'orced-choice precognition ex- pginierlLS published ill the Lnglish-language research literaLul-C be- t4en 1935 and 1987. FOUr ma'or questions were addressed J tljwugh this meta-analysis: (1) Is there overall evidence for accurate ta0 identification (above-chance hitting) in experiniciltal precog- Bet I niWn studies? (2) What is the 1-ilagiliLude of'Lhe overall precognition ef:B@ct? (3) Is the observed effect related to variations in methodo- lo""cal quality that Could allOW a 11101-C colivelit lot lal cxplallatioll@ (1) 9J D" precognition performance vary systerriatically With potential ing varlabics, Such ;is diFf"CreliccS ill subjCcl poptilaliolls, conditions, experimental setting, knowledge of' results, and tiio!! interval between Subject response and target generation? a) 6 DELINEATINGTHE DOMAIN -4 R4wlezlnl (f S/11(h(PS W Psychological research is still academically taboo, and it is ar& that have escaped publication. Our retrieval of studies for this mga-analysis is therefore based on the published literature. Tile sti2lies include all forced-choice precognition experiments appear- inj@in the peer-reviewed English-language parapsychology journals: Joismal of Parapsychology, journal (and Proceedings) of the Society for Psychical Research,journal of the American Sociely.for Psychical R(uearch, European Journal of' Parapsycholo@7 (Including the Research Letter ol" the. Utrecht University Parapsychology Laboratory), and abstracts of' pcer-rcvieNved papers presciited at Parapsychological Associ;16011 meetings published in Research in Parapsychology. Criteria for Inclusion 0111- I-cvic%v is to fixed-feligill slildles ill Which Signifl- cance levels and effect sizes based on direct hittinp- can be calcu- A A11WI-Aludys' (f I,(),-(-rd-Ch0ic,, Preco 1w iii'm Expri 'mcnis 283 is W H 7 lawd. stlidKIS 11sing oillcollic v;il-l;il)l(..s ollicr 111:111 d]rccI llittilig, Such IS I-Lill-SCOI-C VariallCe MId displacement ef'f'eCLS, are illClUded only if' the report provides relevant information on direct hits (i.e., number of trials, hits, and probability of a hit). Finally, we exclude studi(r, conducted by two investigators, S. G. Soal and Walterj. Levy, who;§ work has been unreliable. 0 Many published reports contain more than.one experiment (I experimental unit. In experiments involving multiple condition?.- i ificance levels and effect sizes are calculated for each condi i sign] I itiola _11 Outcome Measures M 0 S"A"nificance level. Significance levels (z scores) were calculated fdi each study from the reported number Of trials, hits, and probabilit, of' success using the normal approximation to the binomial distrE bution with continuity correction. Positive z scores indicate abovC8 44. chance scoring, and negative z scores reflect below-chance scoring, _L EV, ect size. Because most parapsychological experiments, partip ularly those in the older literature, have used the tria.1 rather that) the subject as the Sall1plilIg unit, we use a trial-based estimator cj@ ef'fect size. The effect size (ES) for each study is the z score dividet' by the square root of the number of trials in the study.' a) General Characterstics of the Domain 6 -4 WC 10CatCd 309 SLLl(liCS ill 1 13 SCJKII-@ItC IMIAIC:16011S. 'I'llese SLUZ ies were contributed by 62 different senior authors and were pug lished over a 53-year period, between 1935 and 1-987. Considering were conducted, it is not surprising that the studies are verydivers45 The database comprises nearly two million individual trials an .9. more than 50,000 subjects. Study sample sizes range-from 25 to 297,060 trials (median = 1,194). The number of subjects range@ from I to 29,706 (median = 16). The studies use a variety of metllP-- 0dologics, raliging 1'rom guessing ESI' cards and other card symbols to autornated random number generator experiments. The domain encompasses diverse stil)'ect populations: the most frequently used Elsewhere (Honorton, 1985), we have used the effect size index Co .hen's h (Cohen, 1977), and one referee has asked that we explain why we are now using zJN". The answer is that h 'and z1N 112 yield virtually identical results, and ZIN" is. computationally simpler. For the present sample of 309 precognition studies, the nican differcilce between the two indiccs is.00047,and the standard deviation of the difference is .026: 1(308) = 0.312, p = .756, two-tailed. The correlation betwecli the two indices is .97. 284 7'hejoarnal (Y'Paral).sychology 1 ABLE OVERALL SIGNIFICANCE LEVEL AND EFFECT SIZE Z ES Mean 0.65 0.020 SD 2@68 0.100 Lower 95% confidence estimate 0.40 0.011 Combined Z I IA 1, P 6.3 X 10 "Fail-safe N" = 14,268 I(ES) 3.51, 308 df, p = .00025 population is students (in approximately 40% of the studies); the least frequently used populations are the experimenters themselves and animals (each used in about 5%' of the studies). Though a few studies tested subjects through the mail, more typ- ically subjects were tested in person, elther Individually Or ill groups. Target selection methods included no randomization at all (studies using "quasi-random" naturalistic events), informal methods includ- ing manual card-shuffling or dice-Lhrowing, and formal methods, primarily random number tables or random number generators. The Lime interval between the subjects' responses and target gellm eration varied from less than one second to one year. OVERALL CUMULATION Evidence for an overall effect is strong. As shown in the top part of Table 1, the overall results are highly significant.' Lower bound (one-talled) 95% confidence estimates of' the rnean z scorc and ES are displayed in the bottom portion of Table 1. 'Ninety-two studies (30%) show significant hitting at the 5% level, and significant outcomes are contributed by 40 different investiga- tors. The z scores correlate significantly with sample size: r(307) .156, P = .003. The mean number of trials for significant studies is 34% larger than the mean number of trials for nonsignificant stud- ies. 'The statistical analyses presented here were performed using SYSTAT (Wilk- inson, 1988). When I tests are reported on samples with unequal variances, they are calculaled using dic scparalc vari:111ccs wilhin group% for dic C1.1-or and degrecs of' freedom following Brownlee (1965). Unless otherwise specified, p levels are olie- tailed. Combined z's are based on Stouffer's method (Rosenthal, 1984). A Nleta-linedysis ly']"Urcrd-Chuice Precognition E 1-riments 285 0 0 @0 0 0 0 .0 0 - 0 7- .19 IN 0 0 00 0 0 lc@ '0 0 M 0 0 am -n 0 0 :0 0 0 0 M 0 00 Cn 0 M 0 0. 7. Z 0 0 0 00 44. 00 -0.2 -0.1 0.0 0. 1 0.2 0.3 0.4 0.5 M M 0 Mean effect size > Figurc 1. Mean effect size by investigator. N = 62 Investigators. CD Rephrahon Across Invest.1grators -4 Virtually the same picture ernerges when the cumulation is by00 W Investigator rathel- thall StUdy @IS thC unit ofanalysis, the combinedW 0 z is 12.13, and 23 of' the 62 investigators (37%) have overall out-6 cornes significant at the 5% level. The mean (investigator)effect size W 0 is 0.033 (SD = .093). 0 There is. a. significant difference in the mean ES across investi-W gators, bLIL It IS SUrl)I'lSillgly SIllall: KrUskalmWallis one-way ANOVAO 0 by ranks, X'(6 1) = 8-9,7 1, P = .034. The effect is clearly not due to-L I a few major contributors. lf investigators contributing more than44- three Studies are eliminated, leaving 33 investigators, the combined z is still 6.00 (p = 1.25 X 10") and the mean ES is .028 (SD .091). Figure I shows the mean effect sizes by Investigator. These results indicate substantial cross-investigator replicability and directly contradict the claim of critics such as Akers (1-987) that 286 Dir.journal ql'Paral).@Irhology successful parapsychological 01,1CMIRIS are achleved by oilly a f'Cw investigators. )@Ize Piledrawer Proble?n 0 A well-known reporting bias exists throughout the behavioral I 1 Vences favoring publication of "sign i fican C Studies (e.g., Sterling, 9-959). The extreme view of this "filedrawer problem" is that "the ATurrials are filled with the 5% of the studies that show Type I er- rs, while the filedrawers back at the lab are filled with the 95% of' ie studies that show nonsignificance... - (Rosenthal, 1984, p. 108). Recognizing the importance of' this problem, the Parapsychological a i i i g i ssociation in 1975 adopted an official policy against selective re- 9)rting of positive results." Examination of the parapsychological lit- gature ShOwS that 11011SIgnifiCant results are fi-CCILiently J)Llbll.ShCd, &d, in the precognition database, 70% of the studies have reported k)nsignificant rcstilis. Neverdicless, 75% ofilie precoglifiloll stil(IRIS L Wre published before 1975, and we must ask to what extent selec- Re publication bias could aCCOLInt For the cumulative effects we ob- F,rve. The central section of' Table I uses Rosenthal's (19,44) statistic to estimate the InIniber Of unreported Studies with z soores averaging zero that would be necessary to reduce the known tabase to nonsignificance. The filedrawer estimate indicates that eer 46 unreported studies must exist for each reported suldy to fflduce the cumulative outcome to a nonsignificant level. X A different approach to the filedrawei - problern is described by awes, Landman, and Williams (1984; personal C01111111.1111CA1011 Mrve analysis, like Rosenthal's "fall-safe N," is based oil noi (Wrve assUllIPLIOlIS. Their null hypothesis is that z scores above some Ck (Mitical level (e.g., z = 1.65, 1.96, etc.) are randomly sampled from 90, 1) above that critical level. The alternative to the null hypothesi is if"that, because there is some real effect, the distribution of Z's is shifted to the right of 0 and the z's will be larger than predicted by the null. For a Critical level of'z = 1.65, the expected mean z is 2.06 and the variance is .14. In the precognition database, there are 92 studies with z's > 1.65. Their average is 3.61, not 2.06 as predicted -'AiiaMes hidicate no si ni6cant difference in the niagnitlide of reported 'SIII(IN, 9 oulconles beforealid after 197.5. Thc mcm, F.S I'Orstllclic@s pl-ic),- I() 1975k 0M21 (Sh = .099), alld for Slodies reported thereafter the mcan is 0.017 (SD = .106): 1(307) = 0.28. p = .782. two-tailed. A Alpfa-Aiialvlvis qf Forred-Choice Precogin',ion Exl)eriillews 287 by the litill hypothesis. Since (lie. variance of' the normal truncated above 1.65 is .14, the test z (Using the Central Limit Theorem) com- paring 3.61 to 2.06 is 39.84 [1.55 divided by (.14/92) 112] . Here, p is V1111.1ally zero. Similar results are found with cut points of 1.96, 2-33, > and 2.58. On the basis of these analyses, we conclude that the cumulative n 0 significance of the precognition studies cannot satisfactorily be ex- < M plained by selective reporting. CL -n 0 I Ouri,ipiz Ri--i)UCTION M ID Cn Although the overall z scores and effect sizes cannot reasonably M 1%) be attributed to chance, inspection of the standard deviations in 0 0 I able I indicates that the study Outcomes are extremely heLerogem W neous. Given the diversity of methods, subject populations, and 44- Other Study Features that characterize this research domain, this is --L 00 not surprising. The study outcomes are In fact extremely heterogeni@ous. Al- 0 though a major objective of this metamanalysis is to account for the i> Variability across Studies by blocking oil d'I'I'el-CIICCS ill Study quality, X 0 procedural features,' and samp'ling characteris tics, the database clearly contains extreme outliers. The z scores range from - 5.1 to a) 19.6, a 25msigma spread! The standardized index of kurtosis (ga,) is 6 0 9.47, suggesting that the tails of the distribution are much too long -4 00 for a normal distribution. CD X We eliminated the extreme outliers by performing a "10 percent o 0 --- ---------- -- zsLcore@sjliarnett & Lewis, 19781. This involves W studies with z scores in the upper and lower 10% o eliminating f the o distribution, and results in an adjusted sample of' 248 studies. 'rhe 0 W trimmed z scores range from - 2.24 to 3.21 (g,, 1. 1). -The rem c) 0 vised z scores and effect sizes are presented in Table 2. 0 Elimination of extreme outliers reduces the combined z scores by approximately one half, but the outcomes remain highly significant. rwenLy-five percent of' the studies (62/248) show ov'erall significant hitting at the 5% level. Lower bound confidence estimates show that the mean z's and effect sizes are above 0 at the 95% confidence level. EllmlnaLl 'oil of'outliers reduces the total number of investigators fron-i 62 to 57, but the results remain basically the same when the analyses are based oil illVeStlgaLOl-S I-adler than studies. The coin- bined z is 6.84; 18 of the 57 investigators (31.6%) have overall.sig- 288 TheJournal of Parapsychology TABLE 2 SIGNIFICANCE LEVEL AN1-) EFFEGT SIZE FOR TRIMMED SAMPLE Z ES Mean 0.38 0.012 SD 1.45 O@065 > Leuer 95% confidence est' imate 0.23 0.005 Combined z 6.0 2, p 1. 1 0 x 10' M i(ES) = 2.90, 2,17 dJ' p = . 002 -n niPtcant outcomes at the 5% level. The mean (invesugator) ES is 0. 00 0 (SD 0 5). 1 Wor the trimmed sample, the difference in ES across invesuga- tog is not significant: Kruskal-Wallis one-way ANOVA by ranks, X'2Z6) = 59.34, p = .355. 11' investigators contributing more than thime studies are eliminated, leaving 37 investigators, the combined 0 - z iWstill 5.00 (p = 3.0 x 10-) and the mean ES is 0.022 (SD .0-l'). Figure 2 shows the Illeall effect SIZe by illVeStlgatoll. @@Thus, elimination of the outliers does not substantially affect the coRlusions drawn from our analysis of the database as a whole. Tl-@;re clearly is a nonchance effect. In the remainder of this report, weTuse the trimmed sample to examine covariations in effect size a4a variety of methodological and other study features. T STUDY QUAI-1-1-Y @;ecause target Stimuli in precognition experiments are selected on%.after the subjects' responses have been registered, precognition Studies are usually not vulnerable to sensory leakage problems. Ot?tr potential threats to validity must, how6er, be considered. L T1,2 problem of variations in research quality remains a source of 5 corcroversy in meta-analysis. Some meta-analysts advocate elirninat- ill OW ClUillity Studies W11CI-CaS OUICI'S I-CC0111111CIld CIIII)il-iCally ac- 99 cesong the impact of variations in quality on study outcome. Rosen- that(1984) I30111tS Out that the practice of' discarding Studies is equivalent to assigning them weights of zero, and he recommends weighting study z scores in relation to ratings of research quality. Study Quality Criteria Ideally, the assessment of study quality should be performed by knowledgeable specialists who are blind to the study outcomes. In A Meta-Analysis of Forced-Choice Precogwition Experiments 289 (b n 0 0- 0 0 :0 0 1 0 .0 0 g, 0: 0 Ob > 0 0 0 0 0 M C) 0 CL ho 00 -n 0 0:00 0 0 0 0 M 10 0 lu 7 0 0 CO) 0 M 1 0 1%) 7- 0 :m 0 0 0 W 0 0 44- 0 0 @ @ I... @ I _L r7_rT_r1_T_r_7 00 -0.20 -0.15 -0.10 -0.05 0.00 0.05 0,10 0.15 0.20 m m 0 Mean effect size > Figure 2. Mean effect size by investigator For trimmed sample. N 57 in- vestigators. a) practice, this is usually not feasible, particularly Ivhen, as in the presm 0 4 ent case, large numbers of studies are involved. For our analysis of co 'W study quality, statistical and methodological variables are defin.ed X 0 and coded in terms of procedural descriptions (or their absence) in the research reports. This approach-was used in an earlier meta- analysis of psi ganzfeld research (Honorton, 1985), and it led to study quality ratings that were generally in agreement, r(26) = .766, (04 0 p = 10', with independent "flaw" ratings by an outside critic (Hy- 0 0 marl, 1985). _L One point is given (or withheld) for each of the following eight 4. criteria: ecification of sample size. Does the investigator preplan the num- SP ber of trials to be Included in the Study or IS the study vulnerable to the possibility of optional stopping? Credit is given. to reports that explicitly's pecify the sample size. Studies involving group testi 'ng, in which it is not feasible to specify the sample size precisely, are also 290 Thejournal of Parapsychology given credit. No credit is given to studies in which the sample size is either not preplanned or not addressed in the experimental re- port. Preplanned analysi@. Is the method ofstaLiAiC@11 analysis, including > mthe outcome (dependent variable) measure, preplanned? Credit is M Igiven to studies explicitly specifying the form of analysis and the 0 - Checki-ng. 011C Point is 21110ttCCI For automated checking of' @11 matches between target and response, and another for duplicate checking of hits. C" I 0 Study Quality Analysis 0 4 00 CD Each study received it quality weight between 0 and 8 (mean X 0 3.3, SD = 1.8). "ATe find no significalit relationship between study 0 W(JUillitY 111d E'S: ?-(2,1( d. -;..@ __ __ __ _L _(i7Lail( o for study outcomes to correlate positively With study quality has the 0 oconsequence that the quality-weighted z score of' 6.26 is slightly W 0 larger than the unweighted z of' 6.02. Table 3 shows the correlations 0 0 between effect size and each of the eight individual quality meas- 4@ures.` The mean effect sizes by quality level are displayed gr1aphi-, cally in Figure 3. The correlation between ES and study quality is also nonsignificant for the un- trimmed sample of 309 studies: r(307) = -.060, p = .289. The qualitymweightcd z score is 7.38: p = 2.32 X 10". However, three ofthe individual quality ineasures are significantly related to performancc. Controls and duplicate checking COITChRe significantly positively with ES, and randomization correlates significantly negatively with ES. These correlations appear to be due to a few studies with z scores that are extreme outliers (z > 7). When the 10 studies with z > 7 are eliminated, tile signifi- cant correlations hetwcen (piality and ES disappe;ir. A Meta-Analysis of Forced-Choice Precognition Experiments 291 TABLE 3 CORRE.LATIONs BETWEEN EFFECT SIZE AND QUALI 1-1- IMEASURES Quality measure 1'(-946) Sample size specified in advance .100 0 Preplanned analysis .001 < M Randomization .011 a Colm-ols .058 -n 0 AutomaLed recording .169 1 Duplicate recording .047 ;0 Automated checking .136 M M Duplicate checking .078 0) U) Quality Extremes Is there a tendency for extremely weak studies to show larger effects than exceptionally "good" studies? Analysis on the extremes _'L of' the quality ratings indicates that this is not the case. 00 This analysis, based on the untrimmed sample of 309 studies, 0 uses Studies with quality ratillgS Outside the interquai-Lile range of' 5F the rating distribution (median = 4, Q, = 2, Q, = 5). There are 56 X "low-quality" studies (ratings of 0-1) and 35 "highmquality" studies 0 6 (ratings of 6-8). The high-quality studies have effect sizes that are a) I not significantly lower than the low-quality studies; the ES means 0 0 are 0.017 (SD = 0.063) and 0.037 (SD 0.137), for the low- and 4 high-quality studies, respectively: 1(82) -.92, p .358, two- 00 W tailed. 0 W . ........ ... WA_ 0 ])recognition ES is not significantly related to source of pUblica- X2 tion: Kruskal-Wallis one-way ANOVA, (4). = 0. 7 8, p .942. 0 However, the sources of publication differ significantly in study 2 quality: Kruskal-Wallis one-way ANOVA, X (4) = 17.19, p = .002. This is due largely to the lower quality of studies published in the Journal of the Society for Psychical Research and in Research in Parapsy- chology. Study Qualily,in Relation to Year of Publication Precognition effect size has remained constant over a half-cenm Lury of' res(&trch, even LhOUgh the 1Y1CLhOdOlOg1G1l qLUdiLy Of' 292 The journal of Parapsychology > 45 0 =C (D CL 4 - 63 -n 0 0 -1 3 - 35 U 2 - 41 2) U) M 1%) 0 0 w 0 8 40k. 00 0 -0.10 -0.05 0.00 0.05 0.10 > Mean effect size Figure;R. Precognition effect size in relation to study quality, with 95% con- fidenca5imits. N = 248 studies. 6 search9pas improved significantly during this period. The correlam tion b@ween ES and year of publication is -.071: t(307) = - 1.25, p = .2g3, two-tailed. Study quality and year of publication are, how- ever, I@sitively and significantly correlated: r(246) = .282, p = 2 x 10', go-tailed. Cri8cs of parapsychology have long believed that evidence for parap#hological effects disappears as the methodological rigor in- crease.gThe precognition database does not support this belief'. "REAL-TIME" ALTERNATIVES TO PRECOGNITION Investigators have long been aware of the possibility chat precog- iition effects could be modeled Without assuming either time rever- ;al or backward causality. For example, outcomes from studies with U ice 4 Jllela-Analys' of Forced-Cho' Prec(@g-nilion Experinients 293 targets based on indeterminate random number generators (RNGs) ,_]A I, XTf- L@ a e .4 ue on tile RIN"Y-a psychok-metic (PK) effect-rather than information acquisition concerning its fu- ture state. In experiments with targets based on prepared tables of rand0n] numbers, the possibility exists that the experimenter or other randomizer may be the actual psi source, unconsciously using "real-time" ESP combined with PK to choose an entry point in the random number sequence that will significantly match the "sub- ect's" responses. While the latter possibility may seem farmfetched, J it cannot be logically eliminated if one accepts the existing evidence for contemporaneous ESP and PK, and it has been argued that it is less farmfetched than the alternative of "true" precogm tion. Morris (1982) discusses models of experimental precognition based on "real-time" psi alternatives and methods for testing "true" pi-ccogn ILion. In general terms, these methods constrain the selec- tion of the target sequence so as to eliminate nonprecognitive psi intervention I In the most common procedure, attributed to Mangan (1955), dice are thrown to generate a set of'nunibers that are mathm eniatically manipulated to obtain an entry point in the random num- ber table. This procedure is sufficiently complex "as to be appar- ently beyond the capacities of the human brain, thus ruling out PK because the 'PKer' would not know what to do even via ESP" (Mor- ris, 1982, p. 329). Two features of' precognition study target determination proce- dures were coded to assess "realmtime" psi alternatives to precogni-_ tion: method of determining random number table entry point and use of Mangan's method. Methods of eliminating "real-time" psi alternatives have not been used in studies with random number generators and have only been used in a small number of studies involving randomization by hand- shuffling. These analyses are therefore restricted to studies using random number tables (N = 138). Method (?I'Deiervaning RN7' Enhy PoinI The reports describe six different methods of obtaining entry points in randoin number tables. If the study outcomes were due to subjects' precognitive functioning rather than to alt ,ernative ps.i modes on the.,part of the experimenter or the experimenter's as- SiSUIII(S, IhCre should be no diffCrenCe III Inean effect Size across tile various methods used to determine the entry point. Indeed, our analysis indica"tes that the study effect sizes do not vary syster"ti- 294 7711C.10717-77ol or Paral).@Iirhology A ilfria-AnalYs?"S ()f FoJ-C(1(J-(,hOi(-P 295 cally as a fillic(ion of' illethod of delcrilluillig tile. enlry point: Krus- kal-Wallis one-way ANOVA by ranks: X"(5) = 7.32, p = .198. > Use of Mangans Method Selected 25 0 We find no significant difference in ES between studies using m complex calculations of the type introduced by Mangan to fix tile CL random number table entry point and those that do not use such -n 0 cAlculations: t(45) 0.38, p = .370, two-tailed. M 2) MODERATING VARIABLES 0 M 0 The stability of precognition study Outcomes over a 50-year pe- 0 riod, which we described earlier, is also bad news. It shows that Ill- 0 vestigratol-s ill this area haVC Yet to (ICVCIOI) SUHiCiellt. Understanding 44- -,of the conditions underlying tile occurrence (or detection) of' these Coef'f'ects to reliably increase Lhell- Magnitude. We have identificd f'OUl' Ovariables that appear to covary systen-laLically with precognition ES: T>(l) selected versus unselected subjects, (2) individual versus group @Utesting, (3) feedback level, and (4) time interval between subject re- 0 Osponse and target generation. a) The analyses use the raw study z scores and effect sizes; we 6found that this results Ili unif'ormly more conservative estimates of' 0 4relationships with moderating variables than when the analyses are 00 Gbased on quality-weighted z scores and effect sizes. OSelected Versu.5 Umelected Sid?jects W - - --------- - ---- -------- 0 0 Our meta-analysis identifies eight subject populations: unspecl- 0 Wfied subject populations, mixtures of several different. popu lilt ions, Oanimals, Students, children, "Volunteers," experilnellLel-(S), and se- -,'lected subjects. 44- Ef'f'eCL Size magnitude does not vary significantly across diese eight subject populations: Kruskal-Wallis one-way ANOVA, X'(7) = 10.90, p = .143. Effect sizes by subject population are displayed in Figure 4. However, studies using subjects selected on the basis of prior performance in experiments or pilot tests show significantly larger effects than studies using unselected subjects. As shown in Table 4, 60% of the studies with selected subjects are significant at the 5% level. The illean z score for these studies is 1.39 (SD = 1.40). J'he E'S is significantly higher f'Or select ed-su 1:jects studies than f`01- StA.11(1- > Exptr 12 0 Volunteer - 26 < (D R. CL Children I - 31 -n 0 C 0- sludelits - lo"i I : t@ M Animals @3 110 2) 0 M Mixed 28 Unspec - 18 44- 00 0 - 0.06 -0.04 -0.02 0.00 0.02 0-04 0.06 0.08 0.10 > I Mean effect size Figui-c -1. Precognition efTect size by subject I)OI)LIkitiOn, with 95% confl- I I (D a) I dence limits. N = 248 studies. 0 0 -4 ES ies With unselected subjects. The I test of the difference in mean 00 W_ . is equivalent to a point-biserial correlation of' .198. 0 Does this difference result From less strill ent controls in studies 0 W with selected subjects? The answer appears to be "No." The average quality of' studies with selected subjects is higher than studies using 5 , (4 TA13m;. 4 SI-ILECTED VF.RSUS UNSEl@l..c-rEl) suluEc-l-s 44- Selected Unselected. N Studies 25 223 Combined z 6.89 4.04 Studies withp, < .05 60% 21% Mean ES .051 .008 SDI 1 .075 .063 1(124 G) 3. 16, p 0 0 1 296 The journal of Parapsychology TABLE 5 INDIVIDUAL VERSUS GROUP TESTING Individual GI-OU13 > N studies 97 105 -0 Combined z 6.64 1.21) Studies with p < .05 30% 19% 0 < Mean ES .021 .004 (D CL SD,,, .060 .066 -n 0 1 (200) 1.89, p .03 unselected subjects: 1(27) = 1.51, p = .142, twomtailed. This result appears to reflect a general tendency toward increased rigor and co M more detailed reporting in studies with selected subjects. 1%) to Individual Versus Group Testing 44. Subjects were tested in groups, Individually, or through the niail. imen- co Studies in which subjects were tested individually by an experi 0 ter have a significantly larger mean ES than studies involving group testing (Table 5). The t test of' the difference is equivalent to @1 poiliLmbiSerial Cor- relation of .132, favoring individual testing. Of the studies with subm CD - a)Jects tested individually, 30% are significant at the 5% level. 6 The methodological quality of studies with subjects tested indi- 11 1 0 4 vidually is significantly higher than that Of studies involving group 00 CD testing: 1(137) = 3.08, p = .003, two-tailed. This result is consistent with the conjecture that group experiments are frequently con- Co ducted as "targets of opportunity" and may often be carried out hastily in an afternoon without the preparation and planning that 0 0 go into a study with individual subjects that may be conducted over W I period of weeks or months. Thirty-five studies were conducted through the mail. In dicse studies, subjects completed the task at their leisure and mailed their 44. responses to the investigator. These correspondence studies yield outcomes similar to those involving individual testing. The coni- bined z score is 2.66, with a mean ES of 0.018 (SD = .082). Ten correspondence studies (25.7%) are significant at the 5% level. Eleven studies are unclassifiable with regard to experimental set- ting. 4 A Meta-Analysis of Forced-Cholce Precogwition Experiments 297 TABLE 6 FEED13ACK RECEIVED BY SUBJECTS Feedback of Results None Delayed Run score Trial-by-trial N studies 15 21 21 47 > Combined z -1.30 2.11 4.74 6.98 M Studies with p < .05 0.0% 19.0% 33.3% 42.6% 0 Mean ES -.001 .009 .023 .035 < (D SD,, .028 .036 .048 .072 CL -n 0 Feedback M Ill I A significant positive relationship exists between the degree of cft feedback subjects receive about their performance and precognitive M r,3 effect Size (Table 6). Sub ect feedback information is available for 104 studies. The j se studies fall into four feedback categories: no feedback, 'delayed 44. feedback (usually notification by mail), run-score feedback, and 00 trial-by-trial feedback. We gave these categories numerical values between 0 and 3. Precognition effect size correlates .231 with feed- 0 back level (102 df, p = .009). Of the 47 studies involving trial-by- > @0 trial feedback, 20 (42.6%) are significant at the 5% level. None of I the studies without subject feedback are significant. CD Feedback level correlates positively though not significantly with CP researc h quality: r(102) = .173, p = .082, two-tailed. Inadequate 0 0 randomization the most plausible source of potential artifacts i I n 4 00 I studies with trial-by-trial feedback. We performed a separate analy- CD X sis on the 47 studies in this group. Studies using formal methods of randomization do not differ significantly mean ES from'those CA) L 1 with informal randomization: t(15) = 0.67, p = .590, two-tailed. 0 0 Similarly, studies reporting randomness control data do not differ 0 CA) significMitly III ES f'roill those not including randoniness controls: 0 0 1(42) = 0.79, p = .436, two-tailed. 0 44. Tinie Interval The interval between the subject's response and target selection ranges from less than one second to one year. Information about the time inter"val is available for 144 studies. This information, h@w- 298 77trJourn(d t!I'Mir(ip.@yrliology A 111riel-Alla/YP'S (?fForced-Choice Prrrog:lilion 299 selccted sul:jects: r(l 22) = - .235, /) = .009, two-tailcd. Studies with selected subjects show a nonsignificant positive relationship between ES and time interval: 7-(18) = .077, .745, two-talled. Although correlations Significant (Z > Montlis 7 IS not the difference between these two 1.24), this suggests that the origin of the decline over time may be 0 Motivational rather than the result. of some intrinsic physical bOLindm < M Weeks 17 ary condition. The relationship between precognition ES and feed- - back also Supports this coi 'ecture. Nevertheless, arIV finding SLig- 13 1 n 0 5 Days 15 geStilig potential boundary Conditions oil the phenomenon should be vigorously pursued. Hours 0 35 Co . Influence of Moderating Vaiiables 111i Combinatlo7l M 1%) 0 Q Mi1ILItCS The above analyses examine the impact of' each moderating var- iable in Isolation. In this final set of analyses, we explore theii-Joint 0 Seconds 20 Influence on precognition performance. For this purpose, We iden- 44. tify two subgroups of studies. One Subgroup is characterized by the 00 Millisec 31 use of" selected subjects tested individually with trial-by-trial feed- 0 back. We refer to this as the Optimal group (N = 8 studies). The > second group is characterized by the use of unseleCLed subjects 0.06 -0.04 -0.02 0.00 0.02 0.04 0.06 0.08 11) tested in groups with no feedback. We refer to this as the Suboptimal group (N = 9 Studies). CD a) Meat) effect size The Optimal studies are Coll i ributcd by four Independent inves- 6 Ugators and the Suboptimal studies are contributed by two of the 0 4 same four investigators. All of the Optimal studies involve short pre- 00 CD Figure 5. EfTect size by precognition inierval, with 95% confidence litnits. cognition time Intervals (millisecond interval); the SubopLinial stud- X N = 144 StUdies. ies involve longer intervals (intervals of weeks or months). All of the Optimal studies and 5 of' the 9 Suboptimal Studies use RNG meth- 0 s ever, is o ten imprecise. ur ana vsis o tie I signi icanL y in average sam- odolOgy. The two gi C 0 not c I ei T] 0 precognitive ES and time interval is therefore firnited to seven broad ple size. The mean Study quality for the Optimal group is signifi- 'D 8 interval categories: milliseconds, seconds, minutes, hours, days, cantly higher than that of the Suboptimal studies: Optimal mean = O 0 weeks, and months. (Effect sizes by precognition interval are dism 6.63, SD = 0.92; Suboptimal mean = 3.44, SD = 0.53; t(10) = L I played in Figure 5.) 8.63, p = 3.3 X 10-', two-tailed. 44. Although it is confounded with degree of' feedback, there is a The combined impact of the moderating variables appears to be significant decline in precognition ES over Increasing temporal dis- quite strong (Table 7). Seven of the 8 Optimal studies (87.5%) are tance: r(142) = -.199, p = .017, twomtailed. The largest effects OC- independently significant at the 5% level, whereas none of the Sub- cur over the millisecond Interval: N = 31 studies, combined z = optimal studies are statistically significant. All four investigators con- 6.03, mean ES = 0.045, SD = .073. The smallest effects occur over tributing studies to the Optimal group have significant outcomes.' periods ranging from ;I month to ,I year: N = 7, combined z = ' * 1 ' 0.53, mean ES = 0.001, SD = .0,19. 309 AudiCS, diel-C al-C a (OU11 oF 17 Optimal Studies. ' III thC 1.111trillimed sample of Interestingly, the decline of' precognition perf'01,111ance Over in- The nicall ES is 0A 17 (SD = .154), wid the combined z is 15.8-1. The percentage of independently significant studies is virtually the same as it is in the trimmed sample: creasing temporal distances results entirely froill studies usili,@,,, 1111- 15 ofthe 17 studies (88.2%) are significant. > 0 < M CL -n 0 M 2) Co M 44. 00 0 > I X CD CY) 6 0 -4 00 (D. X 44. 300 The Joumal of Paral5sychology TABLE 7 IMPACT OF MODERATORS IN COMBINATION "Optimal" studies "Suboptimal" studies N studies 8 9 > Combined z 6.14 -1.29 Studies Nvitli P < .05 87.5% 0.0% * Mean ES .055 .005 < * SD,., .045 .035 CL -n t(I 5). = 2.6 1, p 0 1 0 r = .559 M These results are quite striking and suggest that future studies co pecially reliable effects. M combining these moderators should yield es 1%) SUMMARY AND CONCLUSIONS 44. 00 Our meta-analysis of forced-choice precognition experiments 0 confirms the existence of a small but highly significant precognition T> effect- The effect appears to be replicable; significant outcomes are I W reported by 40 investigators using a variety of methodological par- adigms and subject populations. W The precognition effect is statistically very robust: it remains a) 6 highly significant despite elimination of studies with z scores in the 0 4 upper and lower 10% of the z-score distribution and when a third 00 W of the remaining investigators-the major contributors of precog- nition s 1 1 tudies-are eliminated. Estimates OfLhe "filedrawer" problem and consideration of' para- psychological publication practices indicate that the precognition ef- cc) f'eCL cannot plausibly be explained oil the basis OF Selective J)"blica- W tion bias. Analyses of precognition effect sizes in relation to eight 0 0 measures of research quality fall to support the hypothesis that the observed effect is driven to any appreciable extent by methodolog- 44. ical flaws; indeed, several analyses indicate that inedlodologically sli- perlor studies yield stronger effects than methodologically weaker studies. Analyses of' parapsychological alternatives to ])recognition, d- though limited to the subset of studies using random number tables, provide no support for the hypothesis that the effect results.,,frorn A Meta-Analysis of Forced-Cholce Precognition Expe7iments 301 rhe operation of con Lern poraneOLIs ESP and PK at the time of ran- Jornization- Although the overall precognition effect size is small, this does -lot imply that it has no practical consequences. It is, for example, )f' the sarne order Of Magnitude as effect sizes leading to the early crinination of'several major medical research studies. In 1981, the 1@ational Heart, Lung, and Blood Institute discontinued its study of )ropranolol because the results were so favorable to the propranolol reaLluent. that It would be unethical to continue placebo treatment Kolata, 1981); the effect size was 0.04. More recently, The Steerin 9 omnilttee of the Physicians' Health Study Research Group (1988), n a widely publicized report, terminated its study of the effects of ispirin in the prevention of heart attacks for the same reason. The ispirin group suffered significantly fewer heart attacks than a pla- :ebo control group; the associated effect size was 0.03. The most important outcome of the metamanalysis the ication of several moderating variables that appear to covary sys- ematically with precognition performance. The largest effects are )bserved in studies using subjects selected Oil the basis of prior test )erforinance, who are tested individually, and who receive t-rial-by- rial feedback. The outcomes of studies combining these factors con- rast sharply with the null outcomes associated with the combination )f group testing, unselected subjects, and no feedback of results. Be- ause the two groups of studies were conducted by a subset of the arne investigators, it is unlikely.that the observed difference in per- ormance is due to experimenter effects. Indeed, these outcomes inderscore the importance of carefully exarnining differences in ubject populations, test setting, and so forth, before resorting to acile "explanations" based on psi-mediated experimenter effects or he "elusiveness of psi." The identification of these moderating variables has importan't nplications for our understanding of the phenomena and pr9vIdes clear direction for future research. The existence of moderating arlables indicates that the precognition effect is not merely an nexplained departure from a theoretical chance baseline, but ather is an effect that covaries with factors known to influence lore familiar aspects of human performance. It should now be pos- ble to exploit these illodel-ating factors to increase the magnitude Ild reliability of' precognition effects ill 11CW StUdies. tows, G. (1987). Parapsychology is science, but its findings are inconclus,lVe. Bellavioral'and Brain Scioices, 10, 566-568. > M 0 < M CL -n 0 M co M 44. 00 0 > @0 0 W T -4 00 W W 0 0 W 44. 302 ThrJouinal (@f Pmeip.@ychology (I III cla-A nalysis oj P'orced-Cho ce P)-ecugwition Exper'nients 303 BAIZNF-I 1, V., &- LFWIS. (1978). Outlier% in statisliced dai(i. N(,%N, York: j()1in .1. 11, (19,12). Fvidencc ol, pl-ccognilioll ill ille covariation of'salielice I I IN V. Wiley & Soils. ratios. Journal ofl'arapsychology, 6, 111-143. BROWNI.F.E., K. A. (1965). Statistical theory Wid tnelhodole@gy in S0111101 alld engi- Ni(:()[,j. F, & CARINGTON, W. (1947). Some experiments in willed dle-Lhrow- twering. New York: John Wiley &- Soils. ing. Proceedings Socia'v Jilt- Pvilchical Rrvwrch. 48, 16.1 - 175. > -0 COHEN, J. (1977). Statistical power analYsisfin- the behavioral sciences. New York: Ti-ioui,Ess, R. H. (1949). A corriparative study of" performance in three psi Academic Press. Ulsks. Journal qfllarap,@ycholqgy, 13, 263-273. 0 DANVES, R. M., LANDMAN, J., & WILLIAMS, J. (1984). Reply to Kurosawa. Allier- BASTIN, E. W., & GREEN, J. M. (1953). Some experin-lents in precognition. < M ican Psychologist, 39, 74-75. Journal of Parapsychology, 17, 137-143. CL -n HONORTON, C. (1985). Meta-analysis of' psi ganzk@.ld resear(li: A response to MCMAIIAN, E. A., & IiATE..S, E. K. (1954). Report offurther Marchesi exper- 0 llynianjournal of Pmwpsychobgy, 49, 51 -92. illiclitsJournal of.1'arapsychology, 18, 82-92. N I sal. HYMAN, R. (1985). The ganzfield psi experiment: A critical rai journal MANGAN, G. L. (1955). Evidence of'displacement in a precognition testjournal (f Parap.Vchology, 49, 3-50. (fl'arap.qchology, 19, 35-14. M 2) KOLATA, G. B. (1981). Drug 161.111d to help heart attack Survivors. Science, 214, Clsis, K. (1955). Precognition over time intervals of' one to thirty-three days. CO) @ 74 - 7 7.5. M journal (!f Parapsychology. 19, 82-91. ' 1,3 MANGAN, G. L. (1955). Lvidcnce of'displacellient ill a precognition tes(jollrnal Para- NIEUSEN, W. (1956). Ali exploratory precognition experinient.juunial qj (f Parapsychol(@gy. 19, 35-14. p.%ycholqry, 20, 33-39. MOPRis, R. L. (1982). Assessing experiniental Support fbi- true precognition. NIF-1-SEN W (1956). Mental states associated with success in precognition. jwur- 44. Journal of Parapsychology, 46, 321-336. naI O@Parapsychology, 20, 96-109. RoSENTHAL, R. (1984). Meta-analytic proceduresfor social research. Beverly Hills, FAHLER, J. (1957). ESP card tests with and without hypnosis. Journal of Para- 00 CA: Sage. psychology, 21, 179-185. 0 STEERING COMMI'l-ITE.01:111k. PHYSICIANs'HE.Al:11 I STUDN'RE-SEARCI I GROUP. MANGAN, G. L. (1957). An ESP experiment with dual-aspect targets involving (1988). Preliminary report: Findings froni the aspirin component of, tile one trial a dayjournal of Parapsychology, 21, 273-283. X ongoing Physicians' Health Study. New Englandjournal ofIvIedicine, 318, ANDERSON, M., & WHITE, R. (1958). A survey of work on ESP and teacher- 262-264. pupil attitudes..1,ournal of Parapsychology, 22, 246-268. (D STERLING, T. D. (1959). Publication decisions and their possible effects on 01) ' NASH, C. B. (1958). Correlation between ESP and religious value. Journal of lhe 6 inferences drawn froni tests of significance-or vice versajam-nal oj Parapsychology, 22, 204-209. 0 Ainerican Statisfical Association. 54, 30-34. ANDERSON, M. (1959). A precognition experiment comparing time Intervals of -4 00 WII.KINS0N, L (1988). SYSMT: Fvansion, ll.: ;I f6v days and one yeaijournal (f] 'a rajj.@ycholvgy, 23, 81-89. ANDERSON, M., & GREGORY, E. (1959). A two-year program of tests for clair- ' CHRONOLOGICAL USTING OF STUDIES IN TVILTA-ANALYSIS VOYMICC :111d prcCogllil.ioll with it CkISS 0I'PLIbliC SC11001 PLIJASjournal of 23, 149-177. .. ...... ......... 0 RINGTON, W. (1935). Preliminary experiments in precognitive 9 .SIIIgjw1r- NASH, C. B. (1960). Can precognition occur diametrically? Journal of Parapsy- nal of the Socielyfor Psychical Research, 29, 86-104. chology, 24, 26-32. RHINK, J. B. (1938). Experiments bearing on the precognition hypothesis: 1. 0 FRELMANJ. A. (1962). An experiment in precognition.jaurnal of Parapsychology, 0 Pre-shuffling card calling. Journal of Parapsychology, 2, 38-54. 26, 123-130. 0 -- RHINE, J. B., SMrn-i, B. M., & WoonRut:i:, J. L. (1938). Experiments bearing RIIINE,J. B. (1962). The precognition ofconiputer numbers in a public test. 4- on the precognition hypothesis: II. The role of ESP in the shuffling of' Journal (f Parapsychology, 26, 244-251. cards. Journal of ParapsycholqAy, 2, 119- 131. Ryzi., M. (1962). Training the psi fliculty by hypnosis. Journal of the Society jor HUMPHREY, B. M., & PRAYF, J. G. (1941). A comparison of' five ESP Lest Psychical Research, 41, 234-252. proceduresJournal of Farapsychology, 5, 267-293. SANDERS, M. S. (1962). A comparison of verbal and written responses in .a RHINE, J. B. (1941). Experiments bearing upon the precognition hypothesis: precognition experiment. journal of Parapsychology, 26, 23-34. III. Mechanically selected cards. Journal of Parapsychology, 5, 1-57. FREEMAN, J. (1963). Boy-girl differences in a group precognition test. journal STUART, C. E. (1941). An analysis to determine a test predictive of extra-chance of Parapsychology, 27, 175-181. scoring ill card-calling telLS.J0111-nal (f Pay-upsychology, 5, 99- K)7. RAO, 11L R. (19613). Studies ill tile pref@YCIILial effect: 11. A kIlIgUage ESP Lest HUMPHREY, B. M., & RHINEJ. B. (1942). A confirmatory study of salience ill involving precognition and "intervention."Journal of Parapsychology, 27,147- precognition (esisjournal qf Ilarap.@ychology, 6, 190-219. 160. 304 ThrJournal of Parap.@Tchology FREEMAN,.J. (1964). A precognition test with a high-school science club.Journal (Y'I`aral),@vchology, 28, 214-221. FREEMAN, J., & NIFI-SFN, W (1964). Precognition score deviations as related to anxiety levels.Journal ofFarapsychology, 28, 239-249. SCHMEIDLER, G. (1964). An experiment on precognitive clairvoyance: Part 1. The main results. journal of Parapsychology, 28, 1-14. FREEMAN, J. A. (1965). Differential response of' the sexes to contrasting ar- > rangements of ESP target material. Journ-a! ojr Parapsychology, 29, 251-258. Osis, K., & FAHLER, J. (1965). Space and time variables in ESP Journal of the American Socielyfor Psychical Research, 59, 130-145. 0 < FAHLER, J., & Osis, K. (1966). Checking for awareness of hits in a precognition M CL experiment with hypnotized subjects. Journal of the American Society for -n Psychical Research, 60, 340-346. 0 N FREEMAN, J. A. (1966). Sex diff@rences and target arrangement: High-school booklet tests of precognition. Journal of Parapsychology, 30, 227-235. Rom-itzs, 1). E (1966). Nega(ivc and posiliveafli-cl and F,S11 1-1111-scorc Journal ofFarapsychology, 30, 151-,159. CO) (D Roc;ms, D. R, & CARITNTF.R, J. G. (1966). The decline of' variance of' ESP 1%) scores within a testing session. Journal of Parapsychology, 30, 141- 150. 0 0 BRIER, B. (1967). A correspondence ESP experiment with high-I.Q. sul)iects. W 0 Journal 31, 1,13-1,18. 4_4@1 BUZBY, D. E. (1967). Sub ect attitude and score variance in ESP tests.Journal j _L 00 ofFarap@ychology, 31, 43-50. . .BUZBY, D. E. (1967). Precognition and a test of sensory perception. Journal of 0 Parapsychology, 31, 135-142. 1> FREEMAN, J. A. (1967). Sex diffin-ences, target arrangement, and primary men- X tal abilities. Jou I of Parapsychology, 31, 271-279. 0 rna o HONOR-rON, C. (1967). Creativity and precognition scoring level.journal of CD W Parapsychology, 31, 29-42. 6CARPENTER, C. (1968). Two related studies on mood and precognition run- 0 4 score variance. Journal of Parapsychology, 32, 75-89. D TREDON, E. (1968). ESP experiments with inice.Journal of cco DUVAL, R, & MON W Parapsychology, 32, 153-166. 0 FEATHER, S. R., & BRIER, R. (1968). The possible effect of the checker in W _L precognition teSLS.Journal of Parapsychology, 32, 167-175. 0 FREEMAN, J. A. (1968). Sex differences and primary mental abilities in a group 0 W precognition test.Journal of Parapsychology, 32, 176-182. C) NASII, C. S., & NASII, C. B. (1968). EffeCL OfLargCt SCICC6011, field CICIXII(lenCe, C-) and body concept on ESP performance. Journal ofParapsychology, 32, 248- 257. 4.. RHINE, L. E. (1968). Note on an infbrmal group Lest of ESP. Journal of'ljara- psychology, 32, 47-53. RYZL, M. (1968). Precognition scoring and attitude toward ESP.Journal ofBara- psychology, 32, 1-8. RyzL, M. (1968). Precognition scoring and attitude. journal of' Parapsychology, 32, 183-189. /1 11,11WI-1111(d.Im'S (?f Forrrd-CholCe PrecoKnItion .105 CARPENTER, .1. G. (1969). Further study on a mood ad 'jective check list and ESP run-scorc variance.journal oJ Parapsychology, 33, 48-56. DUVAL, R, & MONTREDON, E. (1969). Precognition in mice: A confirmation. Journal of Parapsychology, 33, 71-72. FREEMAN,J. A. (1969). The psi-differential effect in a precognition test.journal ofFarapsychology, 33, 206-212. FRI:I-'_%IAN,J. A. (1969). A precognition experiment with science teachers.Journal o(Parapsychology, J33, 307-310. JOHNSON, M. (1969). Attitude and target differences in a group precognition test.Journal of Parapsychology, 33, 324-325. MONTREDON, E., & ROBINSON, A. (1969). Further precognition work with mice. Journal of Parapsychology, 33, 162-163. S(;iumi)-r, H. (1969). Precognition of a quantum process. Journal of Parapsy- chology, 33, 99-108. BFNm--1R, H. (1970). Differential scoring of an outstanding Subject on GESP and ckth voyanccjouroal iY'11a)-apsyeholo&,y, 34, 272-27.3. Fiu.-E.MAN, I t @1. A. (1970). Sex differences in ESP response as shown by he Free- iiian picture-figure test.journal of'Parapsychology, 34, 37-46. FRi--.EmAN,J. A. (1970). Ten-page booklet tests with elementary-school children. Journal (?f Forrzp.@Trhology. 34, 192-196. FREEMAN,J. (1970). Shift in scoringdireCLIOn WithjUllior-high-SChool Students: 1 - ) r@sl' A SLI III Inary. Journal ofllara y,hology, 34,275. FREEMAN, J. A. (1970). Mood, personality, and attitude in precognition . tests. Journal of Parapsychology, 34, 322. HARALDSSON, E. (1970). Subject selection in a machine precognition test.journal of Parapsychology, 34, 182-191. HARAI.D.SSON, E. (1970). Precognition of a quantum process: A modified rep- liGition. Journal of'Parapsychology, 34, 329-330. NiE.LSE.N, W. (1970). Relationships between precognition scoring level-and mood. Jaurnal of Parapsychology, 34, 93-116. H. (1970). Precognition test with a high-school gl-OLIP.JOUrVal Of Parapsychology, 34, 70. BELOFF, J., & BATE, D. (1971). An attempt to replicate the Schmidt findings. Journal of1he Socielyfor Psychical Research, 46, 21-31. 1 HONORTON, C. (1971). Automated forced-choice precognition tests with a "sen- -481. sitive." Journal of the Anierican Societyfor Psychical Research, 65, 476 E. D. (1971). An ESP test from Apollo 14.Journal ofl`araf)s)@chology, 35, 89-107. SCHMIDT, I I., & PANTAS, L. (1971). PSi ICSIS 'With PSyCh010giGdly C(1LIiVIlcnL conditions and internally cliff@rent machines.Journal ofFarapsychology, 35, 326-327. STANFORD, R. G. (1971). Extrasensory effects upon "memory." Journal of the Ainerican;Societyfor Psychical Research, 64, 161-186. STKILBhR(;, B' (1971). Investigation of' the paranormal gifts of' the Dutch selisidvb Lida T.Journal ofl'arapsychology, 35, 219-225. > 0 < M CL -n 0 _N CO) M 1%) 0 0 0 44. 00 0 > 0 _0 CD T_ 0 0 -4 00 C) 0 5 C-) 0 I 44. 306 The.fou,77zal (?1'Paral).sychology A Meta-Anal.yvis (fForced.-Choice Precognition ExI)erinientv 307 THoui.i--.Ss, R. H. (1971). Experiiiients oil psi self-trailling Willi Dr. schillid('s pre-cognitive apparatus. Journal of the Socielyfor Psychical Research, 46, 15- 21. > HONORTON, C. (1972). Reported Frequency,of dream recall and ESR Jourlial of the Anierican Socielyfor Psychical Research, 66, 369-374. JOHNSON, M., & NoRDBECK, B. (1972). Variation III Llle SCOI-ing bClIaVi0r 01'a 0 11 psychic" subject. journal of Parapsychology, 36, 122-132. KE.LLY, E. F., & KANTHAMANI, B. K. (1972). A subject's ell'orts toward vollillLary CL -n control. Journal of Farapiychologyy, 36, 185-197. -0 H., & PANTAS, L. (1972). I'si tests with internally diffi@rent inachines. Jounial of Farapsychologay, 36, 222-232. CRAIG, .1. G. (1973). The eff@c( of' contilig(.11cy oil precogilifloll Ill (he raf. (D 2) Research in Parapsychology 1972, 154-156. U) (D FREEMAN, J. A. (1973). The psi quiz: A new ESP test. Research in Parapsychology 1%) 1972, 1.32-134. 0 0 ARTLEY, B. (1974). Confirmation of the smallmrodent precognition work.,lournal of Parapsychology, 38, 238-239. 44- HARRIS, S., & TERRY, J. (1974). Precognition in a watermcleprived Wistar rat. 00 Journal of Parapsychology, 38, 239. RANDALL, J. L. (1974). An extended series of ESP and PK tests with three 0 English schoolboys. Journal of the Societyfor Psychical Research, 47, 485-494. 1> EYsENCK, H. J. (1975). Precognition in rats. Journal of Parapiychology, 39, 222- 227. HARALDSSON, E. (1975). Reported dreani recall, precognitive dreanis, and ESE Research in Parapsychology 1974, 47-48. 6 HONORTON, C., RAMSEY, M., & CABIBBO, C. (1975). Experimenter effects in -4 extrasensory perception.journal of the American Socielyfor Psychical Research, 00 (D 69, 135-149. KANTHAMANI, H., & RAO, 1-1. 1-1. (1975). RCSp0IISe tell(ICIlCiCS alld StiIIII.Illis 0 0 structilre.Journal of Parapsychology, 39, 97-105. 0 0 psychology, 39, 363-365. 0 , W I'FRRY, 0 .1. C., & FlAkRIS, S. A. (1975). Precoglillion if) W;I1erMdCpI.iv(.d rals, 0 Research in Parapsychology 1974, 81. L J. W., & HAIGHT, .1. (1976). Psi experiments with rats.finirnal (f Para- _2 DAVIS, psycholqgy, 40, 54-55. JACOBS, J., & BREEDERVELD, H. (1976). Possible influences ofbirdi order oil ESP ability. Research Leiter (11a ra psychology Laboratory, University of' Utrecht), No. 7, 10-20. NF.vji.i.F., R. C. (1976). Some aspects of'precognition teSLizig. Research In Para psychology 1975, 29-31, DRUCKER, S. A., DRFwFs, A. A., & RuBIN, L. (1977). ESP in relation to cognitive developmerit. and IQ in young children. Journal ofthe Aintrican Society fi)r Psychical Research, 71, 289-298. HARAI.D.SSON, E. (1977). ES13 and the definise niechallisill test (DMT): A further validation. Europeanjournal of Parapsychology, 2, 104-114. SARGEN-r, C. L. (1977). An experiment involving a novel precognition task. Journal of Parapsychology, 41, 275-293. Bil-'ItMAN,D.J.(1978).Testiiigtlie";iclvancedw~ive"hypotliesis:AnatLeMpted > replication. Eumpeanjournal of'J'arapsychology, 2, 206-212. BRAUD, W, (1979). Project Chicken Little: A precognition experiment involving o the SKYLAB space station. Europeanjournal ofParapsychology, 3, 149-165. < (D HARALDSSON, E., &JOHNSON, M. (1979). ESP and the defense mechanism test CL (DMT) Icelandic study No. III: A case of the experimenter effect? European n 0 Journal of Parapsychology, 3, 11-20. O'BRIEN, . .1. T. (1979). Ali examination ofthe. checker eff@ct. Research ill Para- psychology 1978, 153-155. CLEMENS, D. B., & PHILLli,s, D. T. (1980). Further studies of precognition In co mice. Research in Parapsychology 1979, 156. (D HARALDSSON, E. (1980). Scoring in a precognition test as a function of the 1%) frequency of reading on psychical phenomena and belief in ESR Research C4 Letter (Parapsychology Laboratory, University of Utrecht), No. 10, 1-8. 0 44. SARGENT, C., & HARLEY, T. A. (1981). Three studies using a psi-predictive _'L Lralt variable questionnaire. Journal of Parapsychology, 45, 199-214. ca WINKELMAN, M. (1981). The effect of formal education on extrasensory abil- 0 iLies: The Ozolco studyjournal of Paral5sychology, 45, 321-336. > NASH, C. B. (1982). ESP of present and f'UtLIre targets. Journal of the Socielyfor @0 Psychical Research, 51, 374-377. Ti-fALBOUPNE, M., BELOFF, J., & DELANOY, D. (1982). A test for the "extra- (D verted sheep versus introverted goats" hypothesis. Research in Parapsychology C" 6 1981, 155-156. 0 CRANDALL, J. E., & HITE, D. D. (1983). Psi-missing and displacement: Evidence -4 00 fbl- illlpl-Opel-ly focused psi? J0111-tial of the zlliiericaii Socielyfor PsychicalRem CD X search, 77, 209-228. 0 0 inary findings. Research in Parapsychology 1982, 103-105 0 JOHNSON, M., & FIARALDSSON, 1-. (1984). The Def'rise Me e chanism 71@st as a 0 0 predictorcifESPscores: IcelandicSLudies IV and V.Joiti,?ialoj']'a)-ap5y(hology, W C) 48, 185-200. Ti:mwz, W (198,1). ConipLiLer-based long-distance LSI': Ail exploratory eX_ amillaLioll (RB/PS). Research in Parapsychology 1983, 100- 101. HESELTINE, G. L. (1985). PK success durinc, structured and nonstructured b 49, 155-163. RNG operation.journal (f Paraj)sychology, HARALDSSON, E., &JOHNSON, M. (1986). The Defense Mechanism Test (DMT) as a predictor ol'ESP perlbrinance: Icelandic studies VI and VII. Research ill ParapsXhology 1985, 43-44. VASSY, Z. (1986). Experimental study of complexiLy dependence in precogni- tion.Journal qf Parapsychology, 50, 235-270. Iq Cl) Cl) a) co a) co Cl) 0 0 04 0 LL 0 CL CL 'OmtO(P-l''""Id 69@- xO@l *0 d -wo:) L, Ul QDUELUJ()jjz)d 13a )LU[ImIP;).l PUI! UoIIIU20:)QJd -(4861) -L) 'NO.I.MONOH Oojoipk@jazzqj fo Ivu-mof x1j RK, Juiki-, @J I -- 1&', , ".. - - - - - - 0 M am -n 0 N ip M M 1%) W 0 44- _16 00 M M 0 > ;b 0 4 00 CD _16 0 0 0 _16 I 4@6 PSI COMMUNIGATION IN THE GANZFELD EXPERIMENTS WFIT] AN AU'YONIA'YED *YES'l*1N(; SYS'YEN1 AND A COMPARISON WITH A META-ANALYSIS OFEARLIER S'FUDIF,8 By CHARLES HONORTON, RICK E. BERGER, MARIO 1. VARVOGLIS, MARTA QUANT, PATRICIA DFIRR, EmIRAINI 1. SCHFGHTFR, AND DIANE C. FERRARI ABSTRACT: A coniputer-controlled testing sysicill was llsed ill 11 (7XIx-rillwills oil ganzfeld psi communication. Theautoniatc(l ganzfeld system controls target Selection and preSCIIL,16011, SUbJeCtS' blilid-judging, and data recording and storage. Video- taped targets included video segments (4ynamic targets) as well as single images (static targets). Two hundred and forty-one volunteer sub 'jects completed 355 psi ganzfeld sessions. The subjects, oil a blind basis, correctly identified randoinly se- lected and remotely viewed targets to a statistically significant degree, z = 3.89, p .00005. Study outcomes were homogeneous across the I I series and eight different experimenters. Performance on dynamic targets was highly significant, z = 4,62, p - -0000019, as was the difference between dynamic and static targets, p = .002. Suggestively stronger performance occurred With friends than with unacquainted sender/receiver pairs, p = .0635. The automated ganzfeld study outcomes are com- pared with a meta-analysis of 28 earlier ganzfeld studies. The two data sets are con- sistent on four dimensions: overall success rate, impact of dynamic and static targets, effect of senclerfreceiver acquaintance, and prior ganzfeld experience. The combined Z for all 39 studies is 7.53, P = 9 x 10- Research on psi communication in the ganzfeld developed as the result of, earlier rescal-ch sliggestlog 111:11 psi 1,1111cliolling is Fre- quently associated with internal attention states brought. about This work was supported by theJaines S. McDonnell Foundation of' St. Louis, Missouri, and by the John E. Fetzer Foundation of Kalamazoo, Michigan. We wish to thank Marilyn J. Schlitz, Peter Rcjcewicz, and Rosemarie Pilkington for their help in recruiting participants; Daryl J. Bem of Cornell University and Donald McCarthy of St. Johns University for helpful comments oil an earlier draft of this paper; Edwin C. May of SRI International for performing theaudio spectrum analysis; and Robert Rosenthal of Harvard University for sliggestions concerning data analysis. We also wish to thank several PRL colleagues who contributed in var- ious ways to the work reported here: Nancy Sondow for assistance in the preparation relaxation exercise and instruction tape that was used throughout, and George Hansen and Linda Moore who served frequently as lab senders. Hansen also pro- vided technical assistance and conducted a data audit resulting in the correction of' several Inillor cri-ors [hat appeared ill a versioll of' this report presented at the 32nd Annual Convention of the Parapsychological Association. Finally, we thank the 241 volunteer participants for providing us with such interesting data. 100 T11r.lournal of Paral).@vrhology Psi Coninninication in. thr Gai4ru 101 through dreaming, Ilypnosis, llic(litat loll, and similar naturally 0C_ tries-oll a J)JIlIcf bIsis-to identify the actual target From among curring or artificially induced states (Braud, 1978; Flonortoil, 1977). four possibilities. A physically isolated sender (Se) views the target This generalization, based on converging evidence from sPonta- -and attempts to communicate salient aspects of it to R. Two exper- neous case studies, clinical observations, and experimental studies, imenters (Es) are usually required. One E manages R, elicits R's ver- > led to the development of a low-level descriptive model of psi func- bal report of ganzfeld imagery (mentation), and supervises R's blind > tioning, according to which, internal attention states facilitate psi dem judging of the target and decoys; a second E supervises Se, and ran- M 0 tection by attenuating sensory and somatic stimuli that nornially domly selects and records the target. 0 M mask weaker psi input (Honorton, 1977, 1978). This "noise-reduc- We developed an automated ganzfeld testing system ("autoganz- mo -n tion" model thus identified sensory deprivation as a key to die fre- feld") to' eliminate potential methodological problems that were m -n 0 quent association between psi communication and internal attention ' in earlier ganzfeld studies (Honorton, 1979; Hyman & -01 states, and the ganzf eld procedure was developed specifically to test Honorton, 1986; Kennedy, 1979) and to explore factors associated M the impact of perceptual isolation on psi performance. 7- with successful performance. The system provides computer control 2) CO) Fifteen years have passed since the Initial reports of 13si conl- of target selection and presentation, blind judging, subject feedback, U) M municaLlon ill the ganzfeld (Bralld, Wood, & Bratid, 1975; ' and data recording and storage (Berger & Honorton, 1986). A com- M 1%) 0 Honorton & Harper, 1974; Parker, 1975). Dozens of additional psi pu ter-con trolled videocassette recorder (VCR) accesses and auto- 1%) 0 0 W ganzfeld studies have appeared sihce then, and the success of the mat cally presents target stimuli to Se. A second E is required only 0 C_4@ 0 44. paradigm has triggered substantial critical interest. Indeed, there is for assistance in target selection The system includes an experimen- 0 44. at least one critical review or commentary for every ganzfeld study tal design module through which E specifies the sample size and 00 reporting significant evidence of psi communication (Akers, 1984; 00 0 Alcock, 1986; Blackmore, 1980, 1987; Child, 1986; Druckman & status of a new series. 0 T> Swets, 1988; Harley & Matthews, 1987; Harris & Rosenthal, 1988; The system was designed to enable ftirther assessment of factors - > @u Honorton, 1979, 1983, 1985; 1-16vchtiann, 1986; Hyman, 1983, identified with successful performance in earlier ganzfeld studies. X 0 1985, 1988; Hyman & Honorton, 1986; Kennedy, 1979; McClenon Differences in target type and sender/receiver acquaintance seem to 0 W , 1986; Palmer, 1986; Palmer, Honorton, & Utts 1989; Parker & be particularly important. Significantly better performance occurred a) 6 , Wiklund, 1987; Rosenthal 1986; Sargent 1987; Scott 1986; in studies using dynamic rather than static targets. Dynamic targets 0 0 4 , , , Stanford, 1984, 1986; Stokes 1986; Utts 1986) contain multiple images reinforcing a central theme, whereas static 0 -4 00 CD . , , Of the many controversies spanning the history of parapsycholog- targets contain a single image. Also, studies permitting subjects to co CD X 0 ical inquiry, the psi ganzfeld domain is unique in three respects. have friends as their senders yielded significantly superior perfor- ;0 0 0 W First, the central issue involves the replicability of a theoretically mance compared to those requiring subjects to work,with laboratory 0 W t e special abilities-of Exceptional -sen dei 1-z-fe1,21--?vf-eta-- M 6 0 0 dividuals (Honorton, 1977). Second, meta-analytic techniques have Analysis" in the Results section.) 0 W been used to assess statistical significance, effect size, and potential The autoganzfeld system uses both dynamic and static targets. w 0 0 threats to validity (Harris & Rosenthal 1988; Honorton 1985; The dynamic targets are excerpts from films; static targets irfclude 5 , Hyman, 1985, 1988; Rosenthal, 1986). Third, investigators and crit- art work and photographs. Receivers may, if they choose, bring ics have agreed on specific guidelines for the conduct and evaluation friends or family members to serve as their senders; a session setup 44. of future psi ganzfeld research (Hyman & Honorton, 1986). module registers the sender type and other session information. In this report, we present the results of the I I autoganzfeld The Automated Ganzfeld Testing System series conducted between the inauguration of the experiments in February, 1983, and September, 1989, when funding problems- Psi ganzfeld experiments typically involve four participants. The required suspension of the PRL research program.' We focus on subject (or receiver, R) atterripts to gain target.-I-clevalit. imagery 'Thisarticle conforms to the reporting pidelines recommended by Hyman and wilile ill tilt, galizfeld; Following (I)c galizFeld/illiagcry period, R I lonortoll (1986). Becatusc of* tiie size of tiii, &,taimne, iiowem-, it i% not practicil to j v @i f iut vj r aTapsycnotogy (1) evidence for psi in the autoganzfeld situation, (2) the impact of dynamic versus static targets, (3) the effects of' sender/recciver ac- (4) the impact of' prior ps; gamdeld -x-i"ence, an' quaintance. h" ... -- @ lj@ I U a comparison of' these four factors with tile outcomes of earlier nonautomated psi ganzfeld experiments. Our findings on demo- graphic, psychological, and target factors will be presented in later reports. > MSu@jects 0 < The participants are 100 men and 141 women ranging in age M CL from 17 to 74 years (mean = 37.3, SD = 11.8). This is a well- -n educated group; the mean formal education is 15.6 years (SD 0 12.0). m Our primary sources of recruitment include referrals from col- leagues (24%), media presentations concerning PRL research (23%), U) friends or acquaintances of PRL staff (20%), and referrals from M IQ other participants (18%). 0 0 Belief in psi is strong in this population. On a seven-point scale where "I" indicates strong disbelief and "T' indicates strong belief 0 44. in psi, the mean -is 6.20 (SD = 1.03); only two participants rated 00 their belief in psi below the midpoint of the scale. Personal experi- MM ences suggestive of psi were reported by 88% of the subjects; 80% 0 reported ostensible telepathic experiences. Eighty percent of' the > @U participants have had some training in meditation or other tech- oniques involving internal focus of attention. cp Participant Orientation 0 0 4 00 Initial contact. New participants receive an information pack be- to fore their first session. The information pack includes a 55-item per- sonal history Survey (Participant Information Form [PIF]; Psycho- physical Research Laboratories, 1983), Form F of the Myers-Briggs 00 Type Indicator (MBTl; Briggs & Myers, 1957), general information 0 W about the research program, and directions for reaching PRL. Par- 0 0 ticipants usually return the completed questionnaires before their 2 first session. However, if new participants are scheduled on short 4- notice, they either complete the questionnaires at PRL or, in a few cases, at home after the session. include the data in an appendix to the report. Instead, We Will Supply the data to qualified investigators in a Lotus-compatible, MS-DOS computer disk file. There is a small fee to cover materials and mailing. Address inquiries to theJournal. Psi Communication in the Ganzfeld 103 Whenever possible, new participants are encouraged to come in for it preliminary orientation session, prior to their first PRL ganz- feld session. The orientation serves as a "get acquainted" session for participants and the PRL staff, and introduces participants to the PRL program and facility. Participants who avail themselves of this option generally complete the MBTI and PIF questionnaires during tile orientation session. We inforril new pal I'LiCipailLS that they may bring a friend or family member to serve as their sender. When a-0 participant chooses not to do so, a PRL staff member serves asm 0 sender. We encourage participants to reschedule their session rather < than feel they must come in to "fulfill an obligation" if they are not M CL feeling well. -n 20 en they -1 Sess* n orientation. We greet participants at the door wh 0 arrive and attempt to create a friendly and informal social atmos- phere. Coffee, tea, and soft drinks are available. E and other staff U 0 members engage in conversation with R during this period. When U) a laboratory sender is used, time is taken for sender and recei M ver to 1,j 0 become acquainted. 0 If the participant is a novice, we describe the rationale and back- C_4@ 44. ground of the ganzfeld research, and we seek to create positive ex- 0 his infoi- 00 pectat'ons concerning R's ability to identify the target. T I mation is tailored to our perception of the needs of the individual participant, but it generally includes four elements: (1) a brief re- view of experimental, clinical, and spontaneous case trends indicat- >1 X ing that ESP is more readily detected during internal attention states such as dreaming, hypnosis, and meditation (Honorton, 1977), to (2) the notion that these states all involve physical relaxation and (P 0 functional sensory deprivation, suggesting that weak ESP impies- 0 -4 sions may be more readily detected when perceptual and somatic 00 noise is reduced, (3) the development of the ganzfeld technique to 0 test this no'semreduction hypothesis, and (4) the long-term succes's of 00 1 1 tile ganzfeld techniClUe as a means of facilitating psi commupi.cation '_2 in tinselected sub' 0 Jects. 0 0 We encourage "goal orientation", and discourage excessive "task orientation" during the session; this is especially emphasized with 00 participants who appear to be anxious or overly concerned about 0 t1lCil- ability to Succeed in tile ganzfeld task. We discourage partici- 46. pants from analyzing their mentation during the session, and tell them that they will have an opportunity to analyze their mentatl6n during the judging procedure. They are encouraged to adopt the role of an o'utside observer of their mental processes during the ganzfeld. Again, this is emphasized with those who appear an?dous 104 77he journal of Parapsychology about their performance; they are advised to relax, follow the taped instructions, and to simply allow the procedure to work. We inform participants that they may experience various types of correspond- ence between their mentation and the target; they are told that they Psi Communication in the Ganzfeld 105 industrial Acoustics 4_Corp 1205A Sound Doublc wall with 4" a SonexAcoustical Padding and acousti- > may experience direct, literal correspondences to the target, but that /_7_11N Isolation Room I door > tor- they should also be prepared for correspondences involving di SENDER @\s I RECEIVER tions or transformations of the target content, cognitive associations, 0 0 < < and similarities in emotional tone. Finally, we orient new particl- M M CL CL pants to where Se and E will be located during the session. -n n 0 0 SCALE 13's cquipmcnt consoic __00 EXPERIMENTER 5 ft ME'l-HOD M M 2) 2) co U) Layout and Equipment M M 0 R and Se are sequestered in nonadjacent, sound-isolated and electrically shielded rooms. Both rooms are copper-screened, and Figure 1. Floor plan of experimental suite. 01, 44. 44- are 14 ft apart on opposite sides of E's monitoring room, which pro- 00 vides the only access. R and Se remain isolated in their respective judging period associations. There is two-way intercom communicO 'I R completes the blindjudging procedure. 0 rooms unti I cation between E and R. One-way audio communication from R t@) R's room is an Industrial Acoustics Corp., IAG 1205A Sound- Se allows Se to listen to R's ganzfeld mentation. > Isolation Room, consisting of two 4-inch sheetrock-filled steel panels. The two panels are separated by a 4-inch air space, for a Receiver Preparation total thickness of one foot. a) I The inside walls and ceiling of Se's room are covered with 4-inch R sits in a comfortable reclining chair in the IAC room. Se keep= 0 6 SonexlfRl acoustical material, similar to that used in commercial R company while E prepares R for visual and auditory ganzfelcQ 4 00 00 broadcast studios. A free-standing Sonex-covered plywood barrier stimulation. Translucent hemispheres are taped over R's eyes witlio- W (5 ft wide by 8 ft high) positioned inside the sender's room, between Microporel!@D tape. Headphones are placed over R's ears. A clip@org 0 Se's chair and the acoustical door, blocks sound transmission microphone is fastened to R's collar. A 600-watt red-filtered floodwo through-th-e--door-&aTne.--Figure--"h-ows--th-42~~o-of-ptati 0f-tlYe-ex-------- light, located appr6ii-rii~t-eT~-6--ff-iii-fr-6n-t--&f-W~-fic--e,is- adjusted i 0 0 Perilnental 1100111S. intensity until R reports a comfortable, shadow-free, hoinogeneomg 0 E occupies a console housing the computer system and other visual field. White noise level is similarly adjusted; R is info.rmet:18 equipment. The computer is an Apple 11 PIUS with two disk drives, that the white noise should be as loud as possible withouL being an a printer, and an expansion chassis. The computer peripherals in- noying or uncomfortable. The ganzfeld light and white noise inten--.' 4. clude a real-time clock, a noise-based random number generator sity are adjusted from E's console after R and Se are sequestered in 44. (RNG), a Cavri Interactive Video lriterfaccl@', ail Apple ganie pad- their respective rooms. dle, and a fan. Other equipment includes a color TV monitor, the VCR used to access and display targets, and three electrically iso- Sender Preparation lated audiocassette recorders. One audiocassette recorder presents audio stimuli (prerecorded relaxation exercises, session instructions, Se sits iWa comfortable reclining chair in the sender's room. Se and white noise). Another plays background music during the ex- faces a coloi:'TV monitor, wearing headphones. During the session, perimental setup. The third records R's ganzfeld mentation and Se can hear R's mentation report through one headphone;,if dy- 106 The journal of Parapsychology namic targets are used, Se hears the target audio channel through the other headphone. Series Manager Setup Procedures E accesses the autoganzfeld computer program through the Se- nes Manager software. Series Manager is a password-protected, menu- driven control program. It provides the only means through which an experimenter may specify parameters for the series design, reg- ister new participants in the series, set up a session, and run a ses- sion. The Series Manager menu is accessed through entry of a private (and nonechoing) password. -Series design. A valid series design must. exist before sessions can be run in an experimental series. This is done through the seyies Manager "design" module. The design module prompts E to specify the type of series (pilot, screening, or formal), the number of participants, the maximum number of trials per participant, the total number of trials per series, and the series name. There is no provision for changing the series design once it is accepted by E. Design parameters are saved in a disk file; they are passed to the experimental program at the beginning of the session. Participant registration. When R is new to a series, E accesses "Participant Registration" from the Series Manager menu before the session. E is prompted to enter R's name and identification number. The module verifies that the maximum number of participants specified in the design is not exceeded. (An error message appears if an attempt is made to register more participants than are speci- fied in the design; then, control is returned to the Series Manager menu.) Session setup. E then selects "Session Setup" from the Series Man- F ager tnemn, @, is promptcd to enter R's narric and the prograin vcr- i5es that R has not already completed the maximum number of trials specified in the design module. (An error message appears if a participant has completed the number of sessions allowed for the series or has not been properly registered; control is then rettirned to the Series Manager menu.) E enters Se's name and the sender type: lab, lab friend, or friend. Lab senders are PRL staff members whose acquaintance with the participant is limited to the experi- ment. Lab friend refers to PRI. staff senders who have some social acquaintance with R outside the laboratory. P'riend senders are friends or family members of the participant. Finally, E enters the ganzfeld light and noise intensity levels and his or her initials. E then leaves Psi Communication in the Ganzfeld 107 the monitoring room while another PRL staff person supervises tar- @Y_et selection. Targets The system uses short video segments (@ynamic targets) and still pictures (static targets) as targets. Dynamic targets include excerpts from motion pictures, documentaries, and cartoons. Static targets _0> include art prints, photographs, and magazine advertisements. M There are 160 targets, arranged injudging sets of four dynamic 0 or four static targets. The sets were constructed to minimize simi- CL larities among targets within a set. The targets are recorded on four -n one-half-inch VI-IS format videocassettes; each videocassette con- 0 tains 10 target seLs (5 dynamic and 5 SLaLic). A signal recorded on an audio track of each videocassette allows computer access of the (D targets. Target display time-to Se during each sending period and 2, U) to R during the judging period-is approximately one minute; (D blank space added to briefer targets insures that the VCR remains in play mode for the same length of time for all targets. Preview packs. The video display format of the autogarfzfeld tar- 0 gets does not permit simultaneous viewing of the entire target set 00 during the judging procedure as is done in many nonautornated mm ganzfeld studies. Each target set is therefore accompanied by a pre- 0 view pack contai ing brief excerpts of all four targets in the set; this T> n gives R a general impression of the range of target possibilities. R views the preview pack at the beginning of the Judging procedure; J it runs approximately 30 sec. a) -4 Target Selection 00 CD X Thc t.111SCL SCICCLor (TS) is a PRI_ staff nicinlicr who li,6 no C'On- 0 tact with either E or R until after the blind-judging procedure. TS (04 _L is needed to load the videocassette containing the target into the c) 0 VCR. TS is informed which of the-four videocassettes contains the 0 target, but remains blind to the target's identity. If Se is a staff 0 member, Se serves this role; otherwise, a staff member not involved 0 in the session serves as TS. (In the latter case, Se and R are seques- tered in their respective rooms before TS enters the monitoring room.) The Se@ies Manager program prompts TS to press a key on the computer k .eyboard. A program call to the hardware RNG obtains the target:value (a number between 1 and 160) and stores it in, com- 108 The journal of Parapsychology Psi Communication in the Ganzfeld 109 puter memory.' The program determines the target set and video- MentationlSending Procedures cassette number from the Larget value. The videocassette number is displayed oil (fie Illorillor, and TS is prompted to insert it Into Llic Receiver mentation report. Aftcr Llic i-claxaLion exercise and in- VCR. The prograrn verifies that tile correo videocassette has been ' structiolls, R listens to tile wilite n0Ise through Ile;,d P I`10 Iles for 30 > inserted and clears the monitor screen; if tile videocasseLte is not t minULes. R reports whatever thoughts, and feelings occur i > I nr M correc , an error message prompts TS to insert the correct video- . the ganzfeld. The inentation report is monitored bv E and Se frornm I 0 < cassette.. TS places a cardboard cover over the VCR's front. panel to conm the, r respect., ve rooms. The mentation recorded, a report ndO is tape M EL ceal the digital counters and VU meters. Finally, TS leaves the mon- E takes detailed notes for review from R prior tojudging. M CL -11 0 itoring room with the three remaining videocassettes knocking Target presentation and sender procedures. A Cavri Video Interface -n , three times on the monitoring room door as a signal for E to I-CLUI-11 automates computer access and control of targets from a JVC BR- 0 i . 640OU VGR. An electronic v ' , deo swiLcher selectively routes the X M Relaxation Exerci@ses and Ganzfeld Instructions r text mode) to three color TV mon-'U' video output (VCR or compute ' ' ID U) itors, one each for E, R, and Se. E s and R s monitors remain Ina) U) M R and Se undergo a 14-min prerecorded rehlXabon exercise be- computer text mode until theJudging period. During each of the M IC, ' Fore the inentation/scilding period.. This.provides a unique Shared six sending periods, Se s TV monitor IS Switched from computer V experience for R and Se before the ESP task. The relaxation exer- text to CR mode. At the beginning of each sending period, Se's monitor displays 44- Use includes progressive relaxation exercises and autogenic phrases 44- the prompt, "Silently communicate the contents and meaning of the 00 (Jacobson, 1929; SIILIILZ, 1950). Ganzfeld instructions are recorded target to [R's first namel." Se views the target and attempts to com- 00 after the relaxation exercise. The instructions and relaxation exer- municate its contents to R. Se inentally reinforces R for- target- 0 > cise are delivered in a slow, soothing but confident manner with related associations and mentally discourages R when the mentation > ocean. sounds in the background. The style of presentation is similar is unrelated to the target. 0 to a hypnotic induction procedure. The ganzfeld instructions to R, CD which are also heard by Se, are as follows: judging Procedure a) a) 6 0 -4 During this experiment we want you to think out loud. Report. all of tile images thoughts and feelin s that ass th h i d 6 After the mentation period, E turns off the ganzfeld light and 0 1 -4 00 , , roug n your m . Do not g p cling to any of them. just observe them as they go by. At some point reads back R's mentation fr the sessio notes. R remains ganz- co om n during the session, we will send you the target information. Do not try feld during the mentation review to minimize any abrupt shift 'in X 0 ' ' to anticipate or conjure up this information. just give yourself the sug- state. E s and R s TV monitors are switched into VCR mode by the 0 W ..... ....... ... ... ------------- - - mstio-n-,rightryaw-in e 0 n1a klng@aa@;Wi S at t e information ------- -- uter--whicfr-21scf-prompts~-e-to-"SitentIT-di-r-ect-fR's-fi-rst Limp 0 0 will appear in consciousness at the appropriate time. Keep your eyes to select the target that you saw." Se's TV monitor remains blank CD I . 0 W open as much as possible during the session and allow your conscious- 0 (computer mode) during this period. 0 0 ness to flow through the sound you will hear through the headphones. R removes the eye covers and views the preview pack'Fro' I m 0 0 One of us will be monitoring you in the other room. Now get as conim their respective rooms, R and E then view the four potential targets fortable as possible, release all conscious hold of your body, and allow (tile actual target and three decoys), which are presented in one of 4- it to relax completely. As soon as you begin observing Your mental 131-oc- ('Out- random sequences. R, viewing each candidate, associates to the esses, SLarL thinking out loud. Continue to share your dioughts, images, item as though it were the actual target, describing perceived simi- and feelings with us throughout the session. larities between the item and the ganzfeld mentation. While R as- sociates to each candidate, E points out potential correspondences that R may.,.4ave overlooked.," R views any of the target candidates 'Ali exception oc(jir-s ill tric 1%vo t ;Irgel comparlson seties (Scrics 10I mid '302). as often as desired before proceeding to the Judging task. See pp. 112-113. This applies to Pilot Series 3, Novice Series 103-105, and to Experienced Series 110 77zejournal of'Parapsychology A 40-point rating scale then appears on R's TV monitor. The scale is labelled 0% on the left and 100% on the right. Using a com- PULCI--galile paddle to move a pollitCl- horizontally ZICI-OSS tile rating scale, R indicates the degree of similarity between his ganzfeld men- tation and each potential target. E and Se view R's ratings on their > monitors. The program checks for ties, and, if they occur, R re-rates the four candidates to obtain unique ratings for each. The program then converts R's raLillgS into ranks. A rank of' I is assigned Lo the 0 candidate R believes has the strongest similarity to his ganzfeld men- M tation; a rank of 4 is given to the candidate R believes is least like CL -n his ganzfeld experience. 0 N Fe-edback and Post-Session Procedures CO After R finishes judging, Se leaves the sender's room and enters M r.J R's room with E. Se reveals the actual target, which the computer C) ZlLltOl1l,ttiC;lll)' displays oil WS TV Inollitor. The session data are writ- 0 1 ten to a floppy disk file. 46 Following feedback, E is prompted to backup the series data disk. The target videocassette is then automatically wound to a po- 00 sition near the center of' the videocassette (Frame 50,000). E selects 0 "Analysis" from the Series Manager nienu. and obtains a hardcopy > printout of the session data file. The printout includes: the file name, R's name and ID number, series type, session number, Se's,. name, E's initials, date and start time, target number, target position cp in the set, R's target ranking, the standardized target rating (z 0 score), target judging sequence, I'll-get naille, L11-get IYJ)c 'llid set 0 4 number, sender type, light and whiLe nolse levels, finish time, and ca optional experimenter's comments. The printout is attached to E's 0 notes on R's mentation and placed in a ring binder containing all 0 W such information for the series. The audio tape of the session is sim- Q L ilarly filed. 0 Experimenters 0 C? Eight Es contributed to the autoganzfeld database. Flonortoll, one of the originators of the psi ganzfeld technique, has conducted psi ganzfeld experiments over a l6myear period. Derr and Varvoglis 201 wid 302, It (loes iiot :il)l)ly to tile (I'llot Sciies 1-2: No6ce Sej@io 10 1 - 102; or Experienced Series 30 1). This practice was initiated because participants frequently Cailcd to identify obvious Co 1-respol Idel ices betiveell their mellultioll alld target elements. Psi Communication lit the Ganzj'eld III worked with Honorton at Maimonides Medical Center and were bv Iiiiii. Berger is primarily r('sponsible for the RIChnical im- plementation of the autoganzf'eld system. He trained i Honorton, Derr, Varvoglis, and Schechter in its use. Honorton trained Quant, Ferrari, and Schlitz in the use of the autoganzfeld system." Series > Altogether, 241 participants contributed 355 sessions in 11 sa ries. To fully address the issue of selective reporting, we inclu4 e A every session completed from the inauguration of the experim -n in February, 1983, to September, 1989, when the PRL facility wes closed. Thus, this database has no "File-drawer" problem (Rosenth 1984). M The studies include three pilot series and eight formal serif. Five of, tile Formal series were single-sessioll studies with novice p;P- ticipants. The remaining three formal series involved experiencig participants. 0 40k- Z Pilot Series CO Series 1. This initial pilot series was conducted during the devg- opment and testing of the autoganzfeld system. It served to test s ing errors , and tem operation, to detect and correct programm' t fille-11111c session timing functions. Nineteen subjects contributed@09 dD- scssions as R's. Seven, including I'RI, still'[' ilicilibers, had pri()r qk- perience as Rs in nonautomated ganzfeld studies at . Maimonigs Lh no prair Medical Center. The remaining 12 Rs were novices wi W@ cified in advanXp; ganzfeld experience. Series sample size was not spe 1 that the system was the series continued until we were satisfied I . _16 erating reliably. Soies 2. This pilot series was designed by Berger in an atter&t 1(as to avert potential displacement effects and subjectjudging prob by having E rather than R serve as judge. R received feedback C81Y to the actual Larget. Four parti I -111ILS contributed to this sel5s. Nine of the planned 50 sessions were completed before Berger's4ke- J);kl-LLll'C I'l-0111 PRL when this series %vils (11scollLIFILICC1. Bcrge@, Schechter, and varvoglis have doctorate degrees ill ps@-Cliolclv- Q11ant holds a 111a.stel's degree ill coollscllillg psychology, an(I Ferrari lias a Gachelors degree in psycho 'I'Ogy. Schlitz has conducted independent gallZield and relliole-v,iewilig I C- search in other laboratories and has a masters degree in anthropology. 110 Piej ouiwal of'Paral)sychology A 40-point rating scale then appears on R's TV irionitor. The scale is labelled 0% on the left and 100% on the right. Using a corn- pul,er-g@linc paddle to 1110VC a poillLel' 1101-MAILally '.ICI'o.Ss the I'ating scale, R indicates the degree of similarity between his ganzfeld men- tation and each potential target. E and Se view R's ratings on their > monitors. The program checks for ties, and, ifthey occur, R re-rates -a the four candidates to obtain unique ratings for each. The program a I then converts R's ratin s into ranks. A rank of I is assigned to the 0 9 < candidate R believes has the strongest similarity to his ganzfeld men- M cL tation; a rank of 4 is given to the candidate R believes is least like n his ganzfeld experience. 0 "N M Feedback and Post-Session Procedures M After R finishes judging, Se leaves the sender's room and enters M R's room with E. Se reveals the actual target, which the computer automatically displays on R's TV moriltor. The session data are writ- ten to a floppy disk file. :Et Following feedback, E is prompted to backup the series data CO disk. The target videocassette is then automatically wound to a po- sition near the center of the videocassette (frame 50,000). E selects 0 "Analysis" from the Series Manager menu and obtains a hardcopy 5; - @O printout of the session data file. The printout includes: the file 0 name, R's name and ID number, series type, session number, Se's "0 name, E's initials, date and start time, target number, target position W I T in the set, R's target ranking, the standardized target rating (z 0 score), 1.11-get Judging sequence, larget 11:1111c, targel typc and set 4 ca number, sender type, light and white noise levels, finish time, and ;U optional experimenter's comments. The printout is attached to E's notes on R's mentation and placed in a ring binder containing all L such information for the series. The audio tape of the session is sini- I'larly filed. Experimenters Eight Es contributed to the autoganzfeld database. Honorton, onie of the originators of the psi ganzfeld technique, has conducted psi ganzfeld experiments over a 16-year period. Derr and Varvoglis 201 wid 302. It (foes not apply to the earlier series (Ililot Series 1-2; Novice Scries 101-102; or Experienced Series 301). 'I'll is practice %vas in it iated because participan is frequently I'ailed to idellfif'y obvious correspondences beLlYeell their Inentiltion and target elements. 11si C0111.711"InIlcallon. III the Gallzj,eld I I I worked with Honorton at Maimonides Medical Center and were 11-:611cd @(.spoiislblc ','()I the technical il - I)N him. Rcrf@cr Is primarov I plemenLaLion of the autoganzfeld System. He trained I71onorLon, Derr, Varvoglis, and Schechter in its use. Honorton trained Quant, 4 Fernirl, and Schlitz in the use of the autoganzfeld system. > Experimental Series n 11 s' 355 sessions 1 (19 Altogether, 241 participants contri < ries. To fully address the issue of selective reporting, we inclucW CL every session completed from the inauguration of the experimentfi in February, 1983, to September, 1989, when the PRL facility WA (:Josc(l. Thus, this database lias no "File-drawer" problem (Rosenth 1984). The studies include three pilot series and eight formal serica Five of the fornial series were single-session studies with novice paJDL ticipants. The remaining three formal series involved experienca participants. 40k. Pilol Series CO Series 1, This initial pilot series was conducted during the del opment and testing of the autoganzfeld system. It served to test s - ]@b_ r errors, and cm tern operation, to detect and correct programming 'buted 70' leCts con 1 tb@ f-II(__IIIIIc session linling functions. Nineteen stibi tr scssions as Rs. Seven, Including PRI, stal"t' members, had prior ctRl- ni(@s pcrience as Rs in rionautomated ganzfeld studies at Maimo Medical Center. The remaining 12 Rs were novices with no proor (a not specified in advang, ganzfeld experience. Series sample size was the series continued until we were satisfied that the sys@em was (U- I . - . -L erating reliably. 0 dcsigne(l by Berger in ;in atte t Serit's 2. This pilot serics -@Is 99 to avert potential displacement effects and subjectJudging problems by having E rather than R serve as judge. R received feedback ORY to Ole actual target. Four participants contribUted to this sertL11S. Nine of the planned 50 sessions were completed before Berger's be- parture fron, PRL when this series was discontinued. I d vill-voglis li@ivc doctorate degrees in psychology. QUIll't Berger, schecliter, all Ims a baclielors dcgree holds a Illa 'sters degree ill cotillselling psyciiology, ancl I-crrari ind renlote-vielving re- ill psycllolpgy. Schlitz has conducted independent gallzf"eld search in:other laboratories and has a masters degree in anthropology. 112 Thrjourtia] qf Parnp,@,vrhology Serics 3. This plot series was a practice series for participants .who completed the allotted number of' sessions in ongoing formal series but who wanted additional ganzfeld experience. This series also includes several demonstration sessions when TV film crews > were present and provided receiver experience for new PRL staff. -a The sample size was not preset. 0 < Novice ("First-7'imers") Series M CL -n The identification of characteristics associated with successful in- 0 itial performance was a maor goal of the PRL ganzfeld project i (Honorton & Schechter, 1987). Except for Series 105, each novice M ser'ies includes 50 ganzfeld novices, that is, participarits With no 2) U) prior gandeld experience. Each novice contributed a single ganz- M r,a feld session. Most novices had not participated in any psi experiment 6 prior to the novice series. Series 101. This is the first novice series. 0 44. Series 102. Beginning with this series, R was prompted after the co mentation period to estimate the number of minutes since the end 0 of the relaxation/instructions tape. 5 Series 103. Starting with this series, Rs were given the option of' @u having no sender (i.e., "clairvoyance" condition). Only four partici- pants opted to have no sender. CD Series 104. A visiting scientist (Marilyn Schlitz) served as E in CY) 6 seven sessions and as Se in six sessions with subjects from The Juil- -'4 liard School in New York. 00 (D Series 105. This series was started to accommodate the overflow of'Juilliard students from Series 104. The sample size wits set to 25. Six sessions were completed at the time the PRL program was susm c) pended. (There were 20 Juilliard students altogether. Sixteen were in Series 104 and four Were in Series 105.) 0 Experienced Subjects Series Series 201. This series involved especially promising subjects. The number of trials was set to 20. Seven sessions by three Rs were completed at the time the PRL program was suspended. Series 301. This series compared dynamic and Static Largets. Sample size was set to 50 sessions. Twenty-five experienced subjects each contributed two sessions. The autoganzfeld program was mod- ified for this series so that each R would have one session with dym Psi CommunirritI077 7.n the Gai.i -@fe Id 113 I nainic targets and one session with static targets. Subjects were in- formed of' this only after completing both sessions. Series 302. Th is series used a single dynamic target set (Set 20). ln earlier series, Target 77 ("Tidal Wave Engulfing Ancient City") had an especially strong success rate while Target 79 ("High-Speip Sex Trio") had never been correctly identified. We made two pfc- gram modifications for this series. The target selection ('Rando 1- izc") rolitilic was niodific.cl to select only till-gets itISet 20, and tl VCR tape-centering routine was modified to wind the videotape LH a randomly selected position between frame numbers 85,000 ar4 95,000. The second modification insured that E could not be cue;b perhaps unconsciously, by the time required to wind the tape fro9t its initial position to the target location. 2) co The study involved experienced Rs who had no prior experien& with Set 20. Each R contributed one session. Participants were ur"a aware of the purpose of the study or that it was limited to one targa set. The design called for the series to continue until 15 sessiolg were completed with each of the two targets of interest. Twentymffi@ sessions were completed when the PRL program was suspended. C* 0 Statistical Analysis > Except for two pilot series, series sample sizes were specified im advance. Our primary hypothesis was that the observed succe SF rate-the proportion of correctly identified targeLs-would reliablg exceed the null hypothesis expectation of .25. To test this hypothg esis, we calculated the exact binomial probability for the obsqrvel@ 111-1111ber of' direct lilts (ranks of' 1) with p = .25 and q = .75. 0 the basis of the overwhelmingly positiv outcomes of earlier studie*a we preset alpha to .05, onemtailed. We also tested two secondary hypotheses, based on @atterns 0 success in earlier psi ganzfeld research. These are: (1) thatclyfiamic) targets are significantly superior to static targets, and (2) that per4o forniance is significantly enhanced when the sender is it friei-id of R" s compared to when R and Se are not acquainted. We. initially planned to test these hypotheses by chimsquare tests, a trial-based analysis. However, a consultant (Dr. Robert Rosenthal) suggested that a I test using the series as the unit would be a more poweTful test of these hypotheses, and we have followed his recommendation. The remai@ing analyses are exploratory. 'The statistical analyses in this report were performed using SYSTAT,(Wilkin 114 Thejournal of Parapsychology TABLE I OUTCOME BY SERIES Series Series type N subjects N trials N Hits- % Effect size (h) z I Pilot 19 22 8 36 .25 .99 > 2 Pilot 4 9 3 33 .18 25 3 Pilot 25 36 10 28 07 . .22 0101 Novice 50 50 12 24 -.02 -.30 6102 Novice 50 50 18 36 .24 1.60 CLI 03 Novice 50 50 15 30 .11 .67 -n 0104 Novice 50 50 18 36 .24 1.60 1105 Novice 6 6 4 67 .87 1.78 E0 I Experienced 3 7 3 43 .38 .69 M301 Experienced 25 50 15 30 .11 .67 C0302 Experienced 25 25 16 64 .81 3.93 M Overall 241 355 122 34 .20 3.89 ,4Vote. The z scores are based on the exact binomial probability with p .25 Rand q = .75. 00 RF-suLTs 0 Overall Success Rate X 0 Ganzfeld hit rate. There were 241 participants, who contributed 9355 autoganzfeld sessions. The 122 direct hits (34.4%) yield an exact inomial p of .00005 (z = 3.89). The effect size, Cohen's h (Cohen, ,-JI977), is .20. The 95% confidence interval (CI) is a hit rate from 0 %3.0% to 39%. Because this level of accuracy would occur about one Ckiine in 20,000 by chance, we reject. the 11LIll hypothesis. (See Table 0 Success rate by series. Of the I I series, 10 yield positive outcomes. oThe mean series effect size is .29, SD = .29, 1 (10) = 3.32. 0 Homogenei@, of effect sizes. Traditionally, psi investigators have C"been preoccupied by whether there is a significam nonzero effect. n equally important issue, however, is the size of the effect. There is a growing tendency among behavioral scientists to define replic- ability in terms of the homogeneity of effect sizes (Hedges, 1987; son, 1988). When t tests are reported on samples with unequal variances, they are calculated using the separate variances within groups for the error and degrees of freedom following Brownlee (1965). Combined zs are based on Stouffer's method (Rosenthal, 1984). Unless otherwise specified, p levels are one-tailed. 1@si Communication in the Ganzfeld 115 TABLi.: 2 OUTCOME By EXPERIMENTER Experimenter N trials N Hits % Effect size (h) QUallt 106 38 36 .24 Honorton 72 27 38 .29 > Bevg(@r 53 18 34 .20 Derr 45 12 27 .05 -1 Varvoglis 43 11 26 .03 0 < schechiel. 14 5 3 6 .23 M CL Ferrari 15 9 60 .72 -n schli(z 7 2 29 .08 0 Rosenthal, 1986; Utts, 1986). Two or more studies are replicates of-M ID one another if their effect sizes are homogeneous. We assess theu) M homogeneity of effect sizes across the I I series by performing a chi-,,) SCILlare homogeneity test cornparing the effect size for each serieso with the weighted mean effect size (Hedges, 1981; Rosenthal, 1984).-0. The I'01-1111_11a is: 0 44. 00 X 2(k 1) Ni(hi - T)2, 0 > 1 where k is the number of' studies, Ni is the sample size of the Itl lb study, and the weighted mean effect size is: Nihi -4 00 Ni The test shows that the series effect sizes are not significantly non'O homogeneous: X' = 16.25, 10 (if, .093. 11omogeneily oj'Outconze by Experimenter 4.. Eight Es contributed to the autoganzfeld database. (See Table ' 2.) All eight experimenters have positive effect sizes. A chi-square ho- mogeneity, test, using the mean effect sizes for each E weighted by sample size, indicates that the results are homogeneous across ex- perlinenters: x' = 7.13, 7 df, p = .415. 116 Thrjounial ofllaral).@I,rhole)Ar), TABLE. 3 GANZI-A-AA) SUCCESS IN RELAI-R)N i-() Ntiromot ()j, Sl.ISsl()Ns No, of sessions as receiver 1 4 + N subjects 183 23 24 11 N trials I 8@3 W 72 5.1 0 Hits 53 19 31 19 < M % Hits 29 41 43 35 CL -n Effect size (h) .09 .34 .38 .22 0 N Subject-Ba5ed AnalyslS Cn Seventy-six percent of the participants (N 183) contributed a M single session as R. Fifty-eight Rs contributed multiple sessions. Par- 10 0 ticipants with multiple sessions either had direct hits or strongly 0 W suggestive target mentation correspondences in then- first session. 0 40- (See Table 3.) Success rate by subjects. To test the consistency of ganzfeld perfor- CO . . mance across participants, we use the standardized ratings of' the 0 target and decoys (Stanford's z scores; Stanford & Sargent, 1983) as 1> the dependent variable. Stanford zs are averaged for participants with multiple sessions. Direct lilts and Stanford zS are IlighIN col-l-c- CI a lated. In this database, N (353) is .776. The mean Stanford z for the W T 241 participants is .21 (SD = 1.04), and t (240) = 3.22 (p = .00073). 0 The 95% CI is a Stanford z froni .08 to .35. The effect size (Colien's 0 4 d; Cohen, 1977) is .2 1. (The effect size for subjects is nearly identical 00 W to the trial-based effect size, h = .20.) Thus, there is a general ten- X 0 clency for participants to give higher ratings to the actual target. 0 (A than to th decoy T4-l:IA-eM-S- +S@+IOt- e i,and_diie__s@--@- --af-these--expe 0 attributable to exceptional pCrfOrlIlaJICC by ;I few Outstanding SLIb- 0 - o jects. W 0 Dynamic Versus Static Targets 4@ The success rate for dynamic targets is highly significant. There are 190 dynamic target sessions and 77 direct hits (40%, h = .32; exact binomial p = 1.9 X 10", z = 4.62). The hit rate for static targets is not Significant (165 trials, 45 lifts, 27%, h = .05, p = .276, Z = .59). Using the series effect size as (lie outcome variable and target type as the predictor variable, the poinL-biserial correlation (r,) between ganzfeld performance and target type is .663, t (17) Psi C11111111111licallem ill the C(Inz ',Ill 1 17 ,/I TABLE 4 SENDERAZECEIVER PAIRING Sender as: Lab Lab friend Friend -0 -0 N trials 110 66 1,15 0 N hits 46 24 52 M % Hits 33 36 36 CL -n Effect size (h) .18 .24 .24 0 z 2.01 1.93 2.83 1 1) .023 .026 .0023 3.65, P = .002.'3 The 95% Cl for dynamic targets is a hit rate frorrj@ '4% to 47%. The CI for static targets i 3 s from 21% to 34%. ThusO our hypothesis concerning the superiority of dynamic targets ls@O_ 0 strongly supported. 401- CO SenderlRecelVer Painng 0 > Receivers are more successful with friends than with laboratory' senders, although the difference is not statistically significant. TheO litillibcl. of, sessions in this analysis is 351 becatisc foUr SUbJCCLSW opted to have no sender. The best performance occurs with friendT 0 senders. Sessions with laboratory senders, @110101_lgh significant, haveo 4 the lowest success rate. (See Table 4.) 00 CD Using series effect sizes as the unit of analysis and sender type;G @ls 1.1le pi-c(lictor val-Mble (conibining lab fricrid and friends), rp isZ5 r 7 -rhe --ssi ns-- -imth-(4- M &_ 95%,_CI--fbr--se 0 W - ----- Lfd,---) friends is a hit rate from 33.3% to 47%. For hib senders, the CI liso .0 from 18.3% to 41.8%. Thus, although the effect of sender type is c4 not statistically significant, there is a trend toward better resuft@ with friends. Separate effect sizes were obtained for the dynamic and static target sessions of each series. Since Series 302 used dynamic targets only, the analysis is based on 11 dynaillic target effect sizes and 8 static target effect sizes; two static targetserics (105 and 201) had extremely small sample sizes (2 and 3 sessions, respectively). A similar procedure is Ilsed ill [lie analyses of'sender/receiver pairing and experienced vcrsus liovicc subjects. 7,1,111-e' .e series involving laboratory "Cliders. were clinlillatcd fi-oln this analysis be- callse of, extremely sinall sample sizes. These include Series 2 (it = 2), Series 105 (n = 2), and Series 201 (it = 1). Thus, the point biserial correlation is based on I I series with friends and 8 series with laboratory senders. 118 TheJournal qf Parapiyrholo@gy Psi' Communication in the Gan@feld 119 Gaii@feld ExI)erie2ice TWO IlLilldl-ed 111d Cighteell l);tl'tlCil);llltS lKld Lilcil' fit-A expert- ence as ganzfeld receivers in the autoganzfeld series. (This includes the 5 Novice Series 10 1 - 105 and 12 novices in Series 1.) For all but 24 (11%), their initial autoganzfeld session provided their first ex- > perience as participant in any parapsychological research. Of the 218 novices, 71 (32.5%, h .17) correctly identified their target (ex- 0act.binomial p = .0073, z 2.44). M Participants with some ganzfeld experience contributed 137 CL -n trials and 51 hits (37%, h = .26, p = .001, z = 3.09). When series 0 effect sizes are used as the unit of analyslis and prior gmizfeld ex- I)CI-MICC is Used its tile IM'edictm- variable, r, is .078, t, (10) --- 0.25, P = .4 1. The 95% Cl fbl@ 110viCCS is a hit. rate f'roin 25.5% to 119.5%. The C1 for experienced participants is from 29% to 50%. U) M Participation by PRL Laboratory Staff 44- For completeness, we report the contribution of laboratory staff as subjects in this database. PRL staff members contributed 12 ses- 00. M Msions as R. These sessions yield 3 hits (exact binomial p = .50; h 0.00). ;0 0White Noise and Ganzfeld Illumination Levels a) The mean white noise level (it' arbitrary units of' 0-7.5) is 2.97 (SD = 1.77). As measured from the headphones, the mean noise 4 CO level is approximately 68 dB. The mean light intensity (arbitrary W Munits of 0-100) is 73.8 (SD = 26.1). Preferred noise and light 111M 0 0tensity levels are highly correlated: r = .569, 1 (353) = 12.99. _L Neither noise nor light intensity is significantly related to ganz- 00 feld performance. The point-biserial correlation between hits and 0 C4 noise level is -.026, 1 (353) 0.18, p = .6.91, two tailed. For ligl)l. CO intensity, rp is -.040, 1 (353) -0.76, p = 449, two talled. Q RANDOMNLSS TUSTS The adeqmwy of' Mil w;ls ;t 111;601' SOL11-CC 01'(11sagi-cc- ment in two meta-analyLic reviews of' earlier psi ganzfeld research (Honorton, 1985; Hyman, 1985). In this section we document the adc(placy of' out- randoillizal loll procedure ;1ccording to guidelines agreed on by Hyman and Honorton (1986). 7 Global Tests of Random Number Generator Full-range frequency analysis. As described earlier, autoganzfeld "argets are selected through a prograin call to the IZNG for values within the target range (1-160). The number of experimental ses- sions (N = 355) is too small to assess the RNG output distribution for the full range, so we performed a large-scale control series to 0 < test the distribution of values. Twelve control samples were col- M lected. These included five samples with 156,000 trials, six samples -n widi 1,560 1.6@11S, @llld 011C S;tllll)lc Of 1,560,000 trials. The 12 I'CSLIIL- 0 ing -clu-square values were coniparcd to a Cili-S(ILIaN dist.1-ibUtiOll ;0 with 155 df, using the Kolmogorov-Smirnov (KS) onemsample test. 2@m I M The KS test yields a twomtailed p = .577, indicating that the RNG Ill) used in these experiments provides a uniform distribution of values M throughout the full target range." Test offirequency distribution for Set 20. We used a single target set (Set 20) in Series 302. We repeated the frequency analysis in a C) 4 40,000-trial control sample, restricting target selection to the four __ L target values within Set 20 (Targets 77-80). A chimsquare test of the 00 diStlAbIlLiOll of' tar ets Within Set 20 shows that. t lie RNG produces 9 0 a uniform distribution of the target values within the set: x' = 3.19, > 3 df, P = .363. 7'rsts ifthe Ex1m)-imental RNG Usage C) Each autoganzfeld session required two RNG calls. An RNG call 0 -4 at the beginning of the session determined the target; another, 00 (0 made before the judging procedure, determined the order which M the target and decoys were presented for judging. C) W Distribution of targets in the experiment. A chi-square test of fhe dis- _L C) iribilljoll of, Values withill tile largct sets shows that file targets were C) C) selected uniformly f roin alliong Lhe f'OUl- possibilities Ivithill each set; ZZ 0 X with 3 df is 0.86, p = .835, C) C) Distribution qfjudging order. A chi-square test of thejudging order-L indicates that the targeLs were uniformly distributed among the four4. possible judging sequences: the X' with 3 df is 1.85, p = .604. 'One of th6 Im-view pack elements for Set 6, containing Targets 21-2-1, was daniaged. This required filtering the RNG calls in the experiment and Control tests to bypass the 6maged portion of the videotape, leaving the targets in Pool 6 unused. for the-full-range analyses reported here, Lhere are 155 df rather than 159. 120 ThrJournal oJ 1'aralj,@ychology Summa7y The randoinness tests demonstrate that the RNG used for target selection in these experiments provides an adequate source of ran- dom numbers and was functioning properly during the experi- > ments. 0 < M EXAMPLES OF TARGFT-MENTATION CORRESPONDENCES CL _n 0 In this section, we present sonle examples of' correspoildcliccs I ;Ubetween targets and ganzfeld mentation. Although conclusions can- Snot he drawn fi-oin qtialitative data, this 111aterial Should 11ol. be ig- M 2) nored. It constitutes the raw data on which the objective statistical (A Mevidence is based, and may provide important insights concerning ghe underlying process. These examples are excerpts from sessions 3A subjects' ganzfeld mentatiorl reports, identified by them during C'the blind judging procedure as providing their basis for rating the 44. "@target. 00 @Targrt 90, Static: Dall's "ChriSt Crnc@fird.- 5@rries 1. IarliciPaid 11): 77. Rank = 1. zsrore = 1.07. I think of'guides, like spirit guides, leading ine and I colne into like a court with a king. It's quiet.... It's like heaven. The king Is something CP like Jesus. Woman. Now I'm just sort of summersaulting through 0 heaven .... Brooding.... Aztecs, the Sun God .... High priest .... 0 4 Fear .... Graves. Woman. Prayer.... Funeral .... Dark. Death .... 00 LD Souls .... Ten Commandments. Moses...." (Wargel 77, Dynamic: 77idal wave engu@big ancirnl ci@, 'r Vt W -I'- (Lill --@ - --F - ) w-V e-(7 - - - rtlir lrans@ IT, in T on Greek nillholoAg. A huge lidal wave cravhi@v j §to the shore. 7'he scene shifts to a cen'ier courtyard of an ancient Greek there is a statue in the center, and buildings with Greek columns around gly Be periphery. People are running to escape conmniption by the lidal wave. fater rushes through the buildings, destroying the columns and the statue; ftople scurry through a stone tunnel, just ahead of the engu@flng water; debrisfloats through the water. Series: 1. Participant ID: 87. Rank = 1. z score = 1.42. The city of Bath comes to mind. The Roman%. The reconstruction of the baths through archaeology. The Partherion. Also getting sort of' buildings like Stonehenge but sort ofa cross between Stonehenge aild the Parthenon. The Byzantine Empire. The Gates of' Thunder. The I Psi Communication in the Ga7lZje'ld 121 Holy See. Tables floating about.... The number 7 very clearly. Thatjust popped Out Of 110WI-ICI-C. It reminds me it bit of one of the first Clash albums, however. The Clash, '7wo Sevens" I think it was called, I'm not sure ...... [The target was number 77.1 Series 302. Participant ID: 267. Rank = 1. z score = 2.00. V V -5 0 ... A big storm over New York City. I'm assuming it's New York City. (D CL ; No, its San Francisco.... A big storm and danger. It looks so beau tiful -n but I'm getting the sense of danger from it.... It's a storm. An earthm 0 quake ...... X M Target 6.3, Dyna-mic: Horses. Front the fibit, "The Lathe of Heaven. " An (D 0 overhead view offive horses galloping in a snow storm. The camera zooms U) M in on the horses as they gallop through the snow. Vie scene shifts to a close- up of a single horse trotting in a grassy meadow, first atnormal speed, then 0 0 in slow-motion. The scene shVts again; the same horse Irolling slowly W 0 through empty city streets. Series: 101. Participant ID: 92. Rank = 1. z score = 1.25. 00 1 keep going to the mountains .... It's snowing.... Moving . again, 0 Illis lime lo (Ile lel'l, spinning lo ilie lel'i .... Spinning. Uke on :I carousel, > @13 horm-S. I forses oil :1 caroliscl, a circos_. 'J'arget 46, Dynamic: Collapsing Bridge. Newsreel footage of the collapse of Q0 a) a bridge the 1940s. The bridge is swaying back and Iforth and up and down. 6 Light posts are swaying. The bridge collapses from the center into the water. 0 4 Series: 101. Participant ID: 135. Rank = 1. z score = 1.94. 00 "...SoliledlifIg, SOIDC Vertical object bending or swaying, almost some- 0 ir!g ill-the ig snf p .............. 161.... Some kind of ladder-like structure but it seenis to be.almost blowing in the wind. Almost like a ladder-like bridge over some kind of chasm that's waving in the wind. This is.not vertical this is horizon- tal.... A bridge, a drawbridge over something. It's'like one of those old English type bridges that opens tip front either side. The middle part ' ening. There was a flash of an old comes up. I see it opening. It's op English stone bridge but then back to this one that's opening. The bridge is lifting, both sides now. Now both sides are straight up. Now it's closing again. It's closing, it's coming down, it's closed. Arc, images of arcs, arcs, bridges. Passageways, many arcs. Bridges with many' Tmget 1.3 ? 7, Static: "Working on a Watermelon Farni. " Thi5 painting shows a black man with his back to the picture; his suspenders form a Vmshape 122 TheJournal of Parapsychology around his shoulders. A dog is in front of'the nian; there are waterinelons between the dog and the mar.. The man fiaces a dirt path with watermelon patches on either side. On the 1@/l side, another man. pvvht.,s a whir1barrov? filled with huge watermelons. Series: 101. Participant ID: 105. Rank = 2. z score = 0.98. ..a small lamb, very Soft, OULSide. Small, playful.... I see it 'V' > M shape.... An apple.... I see a kitchen towel with a picture on h. Apple M seeds or a fruit cut in half showing the seeds, A LOM;Ito Or ;111 ;11)1)le. <2 The fruit was red on the outside.... I thought of watermelon its in a M watermelon basket. Thinking of kids playing on a beach. Little kids -n playing with balls that are bigger than they are and buckets that are 0 three-quarters their size.... I had a thou ht of going through a tunnel, 9 not the kind of tunnel You see on Earth but the type of tunnel described when someone dies." U)Target 64, Dynamic: 1920s Car Sinking..Proin thefitlin "Ghost Sloiy." The M ,,jcene depicts the murder of a young blonde woman by three young men in ghe 1920s. The men are all wearing suits; one of the men is weari .ng a into a @,redora hat that is turned up in the back. The men push an old car ' 4:1ake. The camera shifts between closemups of theirfacial expressions, and the ,--,'car, as it slowly sinla into the wafrr. Thr womansface and hand, ap/mar the car's large rectangular rear window; she silently screams out for help. QThe car disappears beneath the water as the sequence ends. >Series: 102. Participant ID: 154. Rank = 1. z score = 1.45. @0 C1 Girl with a haircut.... Blond hair.... A car. .. Tlic back of* some- W one's head.... Someone running to the right .... Someone on the right in a brown suit ... and a fedora hat turned up very much in the 6 0 back.... Fedora, trench coat, dark tie.... A tire of a car. The car's going 4 ca to the left. An old movie.... I'm picturing an Edward G. Robinson movie.... Big roundish car like 1940's. Those scenes from the back win- 0 dow. Bumping once in a while up and down looking through the back 0 W s probably a big screen in back of the _L window you could see that it wa 0 car and Lhe car's standing still aCLUally.... I think it's .1 movie I saw. 0 0 They're being shot at and shooLing at the window and then the girl gets W 0 shot.... Girl with the blonde haircut.... Someone walking in a suk, 0 brown Suit.... ICS die 1940's again, 30's maybe. Except it. looks like it's 0 _L in color. Something red, blood ... blood on someone's lap.... A dead person all of a sudden.... A big mouth opened. Yelling, but no sound.... Two people running near a train.... Dressed in 1920 type suits with balloony pants, like knickers.... A big, old-fashioned white cor with a flat top. 1920's, 30's ...... 'Target 107, Static: SlaUted-Glass Madon-na with Child. This 'I'S a stained- glass window depicting the Virgin Mary and Chri.@l child. Psi Communication in the Ganzfeld 123 Si-)-11,s: 102. Parlicipaid 11): 183. Rank = 2. zscore = 0.61. "Some kind of' it house, structure.... Some kind of' wall or building. Something with the sky in the background. Thinking of';, bell. A bell Structure. Something with a hole wi Ith the light I coming through the hole.... Like a stained glass window like you see in churches." Target 19, Static: Flyi 'ng Eagle. An eagle with outstretched wings is about to kind on a perch; its claws art? extended. The eagle's head ' white and its is ivings and body (ire black. 1'04. Series: Participant ID: 316. Rank = 1. z score = 2.00. black bird. I see a dark shape of a black bird with a very A pointed beak with his wings down .... Almost needle-like beak.... Something that would fly or is flying ... like a big parrot with long feathers on a perch. Lots of feathers, tail feathers, long, long, long .... Flying, a big huge, huge eagle. The wings of an eagle spread out .... The head of an eagle. White head and dark feathers ..... Fhe bottom ol'a bird ...... Target 144, E@Tnainic: Hell. From the film "Altered States." This sequence depicts a psychedelic experience. Evetyllt'ig is linted red. 17 -The rapidly shifting scenes include: A man screanzz .ng; many people in the midst of fire and in 70 in a man screaming * on iwlafi n tank; people ' agony; a large Sun with a corona around it; a mass crucifixion; people jumping off a precipice 1 in the midst offire, smoke, and molten lava; spiralinz crucifixes. There is a j closemup of a lizard's head, slowly opening its mouth, at the end of the se- quence. Series: 104. Part' ipant ID: 321. Rank = 1. z score @ 1.49. Zc - I just see a big X. A big X.... I see a tunnel in front of me. It's like a tunnel of smog or -it tunnel of smoke. I'm going down it.... Im going down it at a pretty fast speed.... I still see the color red, red, red, red, red, red, red, red.... Ah, suddenly the sun.... The kind of cartoon Suit you see when you can see each pointy spike around the sphere.. stepped on a piece of glass and there's a bit of blood coming Out of my foot.... A lizard, with it big, big, big head ...... ic Target 148, Slat Three Unumial Planes. Three sinall airCraft fiving in Jormation. The planes are while and have swept-back wings; their landingm gear is extended. A winding road is visible below. Series: 104. Participant ID: 322. Rank = 2. z score @ 0.39. Ajet plane.... A 717 on the way to Greece. Blueskies. Sou Tids like it's g0iilg,b@ighel'.... I think I'm back on the plane again. I never used to be afraid of flying until recently.... They need better insulated jets, SOUndprool' like these roonis. They could use these colilrortable seats, too. And the leg i-oorn. The service isn't b;1d eiLhe1-....SLill Cilli't. get the > MM R < M 0M -n _0% lb W M IC%)) 0 W , 0 4 -'L 00 > a) 1 0 -04 CCDO M 0 5 8) 5 W (D 0 0 121 Thr.1mirntil (!I'Parap.@ychology Psi Conimmill'ation in 1he Ganz- I 125 feeling of being in an airplane out of' my mind. Flying over Greenland Bay- A lion.... Highways .... Lion, see a lion.... Tornado.... Bal- and Iceland when I went to England.... Feels like we're going higher ' loon.... Face niask.... City .... Leaning Tower of' Pisa .... Long hall- and higher.... Descending. It seems we re descending.... Big airplanes way,cloorway.... Long road. Long, long desert road ...... flying over wiLh people like me staring clown.... Flying around ill a piece of' tin.... Feel like I'm getting a G-force. Maybe I alli taking of'f' ' Target 22, Dynamic: Spiders. Froin the documentaly "Life on Earth." A > Sure feels like it. Feels like we're going straight. it)).... I always, f`eel like spider Is zveavi?kq its web. The spiders long kg-s sprint, up and down re- > when I'm on the plane going home, Ijust hope that. plane makes it past peatedly, weaving strands of the web. The body of the spider is constantly in 0 the Rocky Mountains ...... motion, and bounces up and down. A close-up shows one of the veins of the 0 M CL Target 10, Static: Santa and Coke. This is a Coca-Cola Christmas adfroill web be'71g.stretched out by the spider. Various views of the web. M CL -rl 0 the 1950s, showing Santa Claus holding a Coke bottle in his left hand; three c Series: 301. Parti ipant ID: 146. Rank = 2. z score = 0.65. -n 0 buttons are visible on Santa's suit. Behind Santa and to his left, is a la7igge Now visual patterns more like a spider web and the color. And then -1 . bottle cap with the CocamCola logo leaning against an. orna-mritted Chrislill(u like the form of die veins of a windmill.... Something like a spider web (D 2) free: again. A spider web. A pattern that instead ofa spider web it looks like (D 2) U) M Series: 104. Particil)(ml ID: 332. 1?ank = 1. zscore = 1.14. basket weaving.... Ali image ofthe way soine children were able to (to U) M something like flying when I was a child though I never had one. It was 1%) ... There's a man with a dark beard and he's got a sharp face.... a-forgotten what it was called-a pogo stick or ajump stick, something -rhere's another man with a beard. Now there's green and white and in which you Jumped up and down and you could hop quite a distance J 44. he's 'In bushes and he's sort of colonial. He looks like Robin Flood and by doing so.... I have kinesthetic images all over as in vigorous motion 0 44. he's wearing a hat.... I can see him from behind. I can see his hat and expressed in flying or jumping on this sort of spring stick that I men- 00 he has a sack over his shoulder.... Window ledge is looking clown and tioned.... Vigorous motion. It's as though I were trying to combine re- 00 0 there's a billboard that says 'CocamCola' on it.... There's a snowman laxation with participating in an image of something very vigorou's.... 1 0 5 again and it's go( a carrot For a nose and thi-ce black buttons coming really feel carried away by these images of vigorous activity without > ;b down the front..... There's a white beard again. There's a man with a being able to localize this activity as to what it is ...... @13 0 white beard.... There's ;in old man with a beard ...... (D Target 108, Static: 7"wofire ealers. A youngfire eater, lit theforeg-round, W Target 70, Dynamic: Dancing in NY City Streets. From the film "The Wiz. facing to the right of the picture, blows a huge flame out of his mouth. In Thespan OfYI-1107t)-paved bridge ovrr a body (?f Watrr nnd autontobile - f 1 eqfic IS the backAround there ' anotherfire rater. A g-roup of people are watching -4 CO is , visible in the opening scene; the New York City skyline is in the back- on the lej? side of'the picture. 4 00 ground. A hot-air balloon flies overhead. The scene shift@ as Dorothy (Diana Series: 301. Participant ID: 146. Rank = 1. z score = 1.71. Ross), her dog Toto, the Lion, Tin Man, and Scarecrow dance along the 0 0 -idzz'@ su. pD * _rLing_ arches i@ behind 1hem 11w__ChjysLer. brigge;_ one -of- the b7 p ... I keep having images of flames now and then..:.The sound re- @ 0 W __ - 0 Bulldin@q- is in the background. At the end qf the sequence, the characters nid@ me )Ffl 0 ames t es again .... In these new images ' .mX 0 0 0 dance in front of'a painted backdrop oJ'an old-Ja'5hioued building. the fire takes on a very menacing meaning .... Rather mounfainous ' 0 0 W sticking tip of bare rocksjust as though they had come from a regently W 0 Series: 105. Participant ID: 336. Rank = 1. z score = 1.40. formed volcano. Volcanos of'course get back to Lhe fire, extreme heat. 0 1 had an image of a volcano with molten lava inside the crater. Molten "Big colorful hot air balloons.... White brick wall.... Ocean.... People lava running down the side of the volcano.... Cold. Written out there 4. .walking before my eyes. Several people.... A dog. Hot air balloon .... behind the visual field and thinking how it contrasts with my images of a nightclub singer.... Back of a woman's head, short curly hair .... flames. Although my images of flames didn't actually include much real. Water.... Balloon, big balloon.... Yellow.... Very tall building. Look- feeling ofheat. I didn't haveany iniagery (.)f'lieaL in connecLion Iviththe ing clown at a city. Leaving a city, going tip .... Faces. An arc.... ' flames. just abstract thought of flames.... Now I think of the water as Water .... A woniari's face.... Cars, f reeway .... A rock-ii-roll sull. ' ;t W;ly Of I)Uttillg OUt flaines. Suddenly, I was biting lily lit). Biting lily chanting .... Architecture. Ajester s hat.... geometrical figures, designs. lip as thougf; lips had something to do with the imagery and I see lips ... Yellow chocolate bar. Water- Going down into water, deep down.... out in front,o'f me.... And the lips I see are bright red, reminding me Man with long golden hair and sun glasses.... The Bay, San Fran( -:isco ofthe flame@ imagery earlier. And then a bright heart such as Valentine's 126 ',rhe.fournal qf ParapgcholoAry Candy ill the shape of' a heart. The chillailloll flavored candies that I remember as a child having at Valentine's. Red color. . . This red as ill the cinrianion candy is a deep very intense red. And similarly ror the flames. Ali(] now I see dic word 'red ....... 7'arget 94, Dynamic: Haikgg Gliders. Vie sequence shows a skier on a V- g Ahaprd hang .glldrr. 'Ilir %kirr %utiv% high up (thom, %now rovr@rd limuntallIA and a pine Jores'. A I the end, the skier lands on a mountain slope and skis away. The sequence is accompanied by Pachelbel's Canon. Series: 301. Partic* ant ID: 188. Rank = 1. z score = 1.26. lp ... Some kind of' V shape, like all open book.... I get some niourim tain.... Some kind of bird with a long wing.... The shape of an upside down W'.... Ski, something about skiing came to me.... Some kind of' a body like ;ill oval shape of' a body with wings oil top of' it ill a 'V' shape. Another 'V' like it wing shape.... Something with wings.... Again the shape or ;ill unibi-clia came into my mind. A butterfly Shape ...... Targel.80, Dynamic: Bugs Bunny in Space. In this cartoon, there is a close- up of the lower part Of a cigarmshaped rocketship and the supports holding ZI up. The rocket assembly slides over to the launching pad, directly above Bugs Bunny's underground patch. The scene shifts to the underground patch, as Bugs Bunny climbs up the ladder leading out of his patch. Unm knowingly, he climbs up throu@gh the interior of the rocketship. 77he rocket's supports pull away and then it takes qff into space. The rocket's nose cone spins as Bugs Bunny appears through the top and he sets the Earth recede rapidly in the distance. As the sequence ends, Bugs Bunny is hit in the belly by a comet. Series: 302. Participant ID: 292. Rank = 1. z score = 1.48. ... Space crart....'l-he solar system. The underside ora helicopter or a submarine or some kind of fish that you're seeing rrom under- lleath. ... Sort orbeing underneath it. Sort orbeing underneath it.... A very strange image like a cartoon character, animated character. With his mouth open kind of.... Like a hypodermic needle or a candle or this shaft like thing with the a pointed top again.... missiles flying.... An aerial perspective.... Firijust kind of editing here I think. I'm really hoping all this rocketship kind of imagery isn't because or I lie noise. I reel like I'm ill a rocketshil) or something ..... I'llat irilage or [lie ship going into the belly orthe mother ship ...... COMPARISON OF STUDY OUTCOMES WITH -ANALYSIS GANZFELD META In this section, we compare the automated ganzfeld study outm collies with [lie results of, earlier gailzfeld StUdieS, SL1lulll:lriZCd ill It 11st Communication In the Ganzj'rld 127 Tmu.i,. 5 GONIPARISON OF OVE.RAU. PKRFOPMANG.F. IN Au"roNIATED GANZFE.1.1) AND MFTA-ANALYSis DATA SETS Outcome N val-i'lble Daiabase studies Mean SD t df P z scores Meta-analysis 28 1.25 1.57 0.33 25 .748 AutoganzFeld I 1 1.10 1.14 Effect size's (h) Meta-analysis 28 .28 .46 0.14 28 .892 AULogarizreld 11 .29 .29 Note. 1-he p values are two-tailed. nicta-atialysis (1-loriorton, 1985). We compare the two databases on four dimensions: (1) overall success rate, (2) dynamic versus static targets, (@@) sender/receiver pairing, and (4) novice versus expert- enced subjects. Overall Success Rate To assess the consistency of results, we compare the 11'auto- anzfeld series to the 28 studies in a meta-analysis of earlier ganzm feld studies (Honorton, 1985, Table Al, p. 84), using direct hits as the dependent variable. The outcomes of the two data sets are con- sistent. Both display a predominance of positive outcomes: 23 of the 28 studies in the meta-analysis (82%) and 10 of the I I autoganzfeld series (91%) yield positive z scores. The mean autoganzfeld z scores and effect sizes are very similar to those in the meta-analysis. (Seq Table 5.) Combined Estimates of Ganzfeld Success Rate Because the z scores and effect sizes for the automated ganzfeld are consistent with the original set of 28 studies in the meta-analysis, a better estimate of their true population values may be obtained by combining theni. Positive outcomes were obtained in 33 of the 39 studies (85%); the 95% CI is from 69% to 99%. Table 6 shows a stern-and-leaf frequency plot of the z scores (Tukey, 1977). Unlike other methods of displaying frequency distributions, the stem-and- leaf plot retains,the numerical data precisely. (Turned on its side, the stern-and-4eaf plot becomes a conventional histogram.) Each number includ6s a stem and one or more leaves. For example, the stem I is followed by leaves of 6,6,6,7,7,7, representing z scores of l.6,l.6,l.6,1.7,l.7,l.7. Ill the display, the letter "IT' ideliLifleS the > M -1 0 < M CL -n 0 5 2) co M 1%) 44. - 00 MM 0 CP 4 00 0 0 0 0 W C) L 128 TheJournal of Parapsycholc@gy TABLE 6 DISTRIBUTION 01: Z SCOREIS Stem Lea f > Minimum Z -1.97 1. 97 Lower hinge 0.25 0 -0. 85 Median z 0.92 < (D -0. 33 Mean z 1.28 CL 0. H 222224 Upper hinge 2.08 -n 0. M 6667777999 Maximum z 4.02 0 I . 666777 SD 1.44 2. H Oll Skewness (g-,) 0.05 U 2. 8 Kurtosis (g2) = -0.37 U) 3. 01124 Combined (Stouffer) z = 7.53 M 3. 9 0 4. 0 0 44. upper and lower hinges of the distribution, and "M" identifies its median. The z's range from - 1.97 to 4.02 (mean z 1.21, SD = 00 m m 1.45), and the 95% Cl is a z from .76 to 1.66. 0 The combined z for (lie 39 studies is 7.53 (1) 9 X 10"). Rosenthal's (1984) file-drawer statistic indicates that 778 additional studies with z scores averaging zero would be required to reduce the significance of' the combined ganzfeld database to nonsignif ica lice; CD that is a ratio of 19 unknown studies for every known study. T 0 A steni-and-leaf display of' the effect sizes is shown in Table 7. 0 4 I'lle effect sizes rallge From -.93 to 1.11 (incan h .28, SD = .11). 00 (D The two most extreme values on both sides of' the distribiaJon are X 0 outliers. The 95% Cl is all h between .15 and .4 1; the cqUiVidelit. hit 0 rate is from 31.5% to 44.5%. 0 0 Dynaniw* Versus Static Targets 0 The use of video sequences as targets is a novel feature of the auLoganzfeld database. However, a comparable difference ill Lai-get type exists in the earlier ganzfeld studies. Of the 28 direct hits stud- ies in the meta-analysis, 9 studies (by three independent investiga. tors) used View Master stereoscopic slide reels as targets (Honorton, 1985, Studies 7-8, 16-19, 21, 38-39). Static targets (single pictures or slides) were used in the remaining 19 studies by seven independent investigators (Studies 1, 2, 4, 10-13, 23-31, 33- 34, 41-42). Like the autoganzfeld video sequences, View Master tarm gets present a variety of' images reinforcing a central target theme. Psi Communication in the Ganzfeld 129 TABU: '7 DISTRIBUTION OF EFFEC-r SIZES (COHEN'S h) Stem Leaf -.9 3 -.4 0 0 OUTSIDE VALUES < (D Minimum h 0.9-9m -.3 1 Lower hinge o. I (rn -.1 0 Median h 0.2A -.0 51 Mean h 0.20 M .0 7779 Upper hinge 0.4 IU 1 H 002888 Maximum h 1.4,g .2 M 1334 SD 0.4 IM .3 11144777 Skewness (g,) 0.2V) .4 H 01113 Kurtosis (g2) 2.4 9CO4 .5 7 0 .7 3 44. .8 17 00 OUTSIDE VALUES 0 4 To compare the relative impact of dynamic and static targets incD I a) the autoganzfeld and meta-analysis, we obtained point-biserial cor-1 0 relations for each data set using target type (static or dynamic) asc) -4 the predictor variable and the series effeu size, coliens h, as tht,0.0 011I.C0111c variable. We test. the difference between the two correla-;0 0 tions using Cohen's q (Cohen, 1977). Dynarnic targets vield signifi-a '_'O 1 W I; es eT-effect--sv --irr g is .409, 1 (26) = 2.28, p = .015; and for the autoganzfeld, @ks re-o 0 ported above, r, is .663. The two correlations are not s.iknificantlyw different (q = .36; z = 1.14). Therefore, we combine the-two clatao . 0 Sets to obtain a better estimate of the relationship between effect size-, and target type: r, = .439, t (45) = 3.28, p = .002. The 95% CIS4 are 24% to 36% for static targets and 38% to 55% for dynamic tar- gets. Thus, the CUMUlative evidence strongly indicates that dynamic targets are more accurately retrieved than static targets. SenderlReceiver Pairing A similar analysis compares the effects of sender/receiver pairing in the two' databases. Studies in the meta-analysis did not routinely 130 The journal of Parapsychology provide detailed breakdowns regarding sender/receiver pairing. Seri der/recei ve I- pairing in the nieLa-analysis can only be coded ac- cording I whether subjects could bring friends to serve -is their ... 1 -0 scilder ol lv(-Ic I-CS111(tcd lo I.Ibmatory 1,, 17 snid Ics, by six independent investigators, subjects were free to bring friends (I-Ionorton, 1985, Studies 1-2, 4, 7-8, 16, 23-28, 30, 33-31, 38- 39). Laboratory-assigned senders were used exclusively in the re- ma;nl'ng 8 studies, by four independent investigators 'Studies 10- 12, 18-19, 21, 29, 41). (Three studies using chiii-voyance proce- dures and no senders are excluded from this analysis.) For- the au- toganzfeld studies, we calculated separate effect sizes for each series by sender type (combining lab friend and friend for comparability with the meta-analysis). III the nieta-analysis, r,, (23) Is A03; larger effect sizes occurred in studies where friends could serve as sender- 2.11, p = .023). For the autoganzfeld, as reported above, rp is .363, in the same direction. The two correlations are very similar (q .05; z = 0.14) and are combined to give a better estimate of the relationship between sender/receiver pairing and ganzfeld study outcome: rP = .38, 1 (42) = 2.66, p = .0055. The ()5% GIs are 20% to 31% for tinacquaillied sendcr/recelver pairs and 3d. I'Y,, (o 19.2'Yo for friends. Thus, the sender/receiver relationship does have a sig- nificant impact on performance. Effect of Prior Ganzfeld Experience The meta-analysis includes 14 studies, by nine independent in- vestigators, in which novices are used exclusively (Honorton, 1985, Studies 2, 4, 8, 10-12, 16-18, 23-24, 31, 41-42). Experienced or- mixed samples of novice and experienced subjects are used in the remaining 14 studies, by four different invest .igators (Studies 1, 7, 19, 21, 25-30, 33-34, 38-39). Studies using experienced subj ,jects were more successful than those limited to novices; the point-biserial correlation between level of experience and effect size is .229, t (26) 1.20, p = .12. For the autoganzf'eld studies, as reporLed above, rp is .078. The two correlations do not differ significantly (q = .155; z = 0.40), and the combined r,, is .194, t (38) = 1.22, p = .105. The respective 95% CIs are 24.5% to 44.5% for- novices and 35.5% to 48% for experienced subjects. The 95% CIs for these comparative analyses are shown graphi- cally in Figure 2. The bottom two rows are Cls for the overall hit rates in the meta-analysis and autoganzfeld, respectively. The next Psi Communi'cati'on in the Ganzfeld 131 0 V CZ MeLa:Exper AuLo:Novicc > AuLo:SR=Fr - --------------- '0 MeLn:SR=Fr - -1 0 < M AuLo:SR=1.ab- CL -n N1cLwSR=1.ab - 0 AuLo:TGT=SLa- McU:TGT@SLa- M ID AuL,o:TGT=Dyn, :J CO M @- MeLa:TGT=Dyn . Autogatiz. 44- 00 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 02 0.4 0.6 0.8 1.052 > Effect size (h) (D Figure 2. Comparison of autoganzFeld and meta-analysis 95% confidenceT limits. Abbreviations are defined as follows: Meta = nieta-allaNsis studies,O 0 Auto = ;lUtoniated ganzf'eld Studies, Dyn = dynamic iargets, Sta = staLic-4 00 Uirgets, Lab = hlbo rato ry senders, I@I- = sender is friend or acqUaintancew of' receiver, Novice = no prior ganzfeld experience, @Ixper = prior gainz-@O feld experience. 0 two rows give the Cls for dynamic targets in the two data sets, ando SO oil. 0 0 0 4.. DISCUSSION We now consider- various rival hypotheses that might account for- the experimental outcomes, and the degree to which the automated ganzfeld experiments, viewed in conjunction with the earl.lu psi 132 Thrjoumal (?f Paral).@yrhology Psi Conommication In the Gaiiql-I'l I lili ganzfeld studies, constitute evidence for psi communication. Finally, iments. Nevertheless, to totally exclude any possibility of subliminal in lye consider directions for future research suggested by Lllcsc filid- cueing, we modified the equipment. Additional testing confirmed ings. that this modificaLlon cffectivOy eliminated all Icakage. This was > Rival Hypotheses formally confirmed by an audio spectrum analysis, covering the fre- M I quency domain between 475 Hz and 15.2 kHz. The critical question I > M 0 Semo7y Cues. Only Se knows the identity of the target until R of course, is whether performance on dynamic targets diminished , , -0 -1 CL finishes the automated judging procedure. If' Se is not a PRL Staff' .1fter th is Illo(fification. The answer i's no: in f .-Ict, perfornla lice ill]- 0 < -n a staff member not otherwise involved in the session sum member proved. Before the modification, the direct hit rate on dynamic tar- M CL o -1 , pervises target selection. In either case the target selector knows gets was 38% (150 trials, 57 hits, h = .28, exact binomial p -n 0 , only which videocassette contains the target. The target selector .00029, z = 3.44); the 95 % CI was from 3 1 % to 45 %. Following the -1 M 2) leaves the monitoring room with the renlaining three target tapes modification, the direct hit rate was 50% (40 trials, 20 hits, h = .52, ip CO M after knocking three times on the monitoring room door, signalling exact binomial p = .00057, z = 3.25) with a 95% Cl from 37% to i M 11 1%) E to return. Since the target selector only knows the videocassette t rate for all targets-static and dynamic-after 63%. The direct h 0 M 0 0 number, variations in knocking cannot communicate any useful in- the modification was 44% (64 trials, 28 hits, h 39, exact binomial W a formation to F.. The cardboard cover over the VCR eliminates any 0 = .00082, z = 3.15). ' 44. visual cues to E regarding the position of the videotape or the activ- Randomizai '0 n. As Hyman and Honarton (1986, Ip. 357) have " W 00 ity of the VU meters (which are active when the target is dynamic Because ganzfeld experiments involve only one target pointed out, 0 44. _1 and has a soundtrack) selection per session .... the ganzfeld investigator can restrict his or Z 00 0 . Sensory transmission from Se to R during the ganzf'eld session is her attention to a fi-equency analysis allowing assessment of the de- " > b sound-attenuated rooms. eliminated by having R and Se in separate We have docu- gree to which targets occur with equal probability. 0 ; 0 , lf either participant leaves their room before R's ratings have been merited both the general adequacy of the RNG used for target se- i > CD registered in the computer, the session is unconditionally aborted. I lection and its proper functioning dur ng the experiment. The videotape target display systern prevents potential handling Data selection. Except for two pilot studies, the number of partic- o cues during the judging procedure. Computer registration of R's ipants and trials were specified in advance for each series. The pilot ) a 6 4 00 target ratings and autoniated Feedback after the session prevents (lie or formal Status of each series was similarly specified in advance and C) (0 X possibility of cheating by Se during feedback, raised by Hyman recorded on disk before beginning the series. We have reported all i i -4 00 0 0 (1985). trials, includ )l ng series, using the automated ganz-. ng pilot and ongo CD X Feld system. Thus, there is no "file-drawer" problem in this data- 0 0 0 0 0 clear that our hypothesis concerning the superiority of dynamic tar- Psi ganzfeld success rate is similar for pilot and formal sessions. 6 W 0 gets over static targets was receiving substantial confirmation. Bem The proportion of hits for the 66 pilot sessions is .32 (h = .16, p 0 6 0 0 cause dynamic targets contain auditory as well as visual information, .129, z = 1. 13). For the 289 formal sessions, the proportion correct W 0 4 we conducted a SL1pJ)1e1'1le1lt.ary test Lo assess die J)OssibiliLy Ofaudim is .35 (h .22, p = .000 1, z = 3.7 1). The difference is not signifi- 0 5 - tory leakage from the VCR soundtrack to R. With the VCR audio cant: X2 0. 11, 1 df, p = .734. _L 4 set to normal amplification, no auditory signal could be detected ' If we assume that the remaining trials in the three unfinished .. s headphones, with or without white noise. When an exm through R ' 3eries would yield only chance results, these series would still be sta- ternal amplifier was added between the VCR and R s headphones tistically significant (exact binomial p = .009, z = 2.36). This would and with the white noise turned completely off, the soundtrack inclu- reduce the overall z for all 11 series from 3.89 to 3.6 1. Thus could sometimes be faintly detected. It is unlikely that subjects could , sion of the three incomplete studies does not pose an optional stopm have detected any target audio signal with the normal VCR ampli- ping problem'. ". fication and white noise; as we have reported, there is no correlation Multiple anaiysis. Informal examination of recent issues of several between ganzfeld success rate and white noise level in these exper- American Psychological Association journals suggests that correction 134 Die./ourral of Parafz@ychology for multiple comparisons is not ;I common practice in more conven- donal al-CaS of' psychological il)(11.1iry. Nevertheless, half'01, I lylliall's (1985) 50-pagc critique of' earlier psi ganzFeld research [0(@Ilsc(l on issues related to multiple testing. In L11C present case, advance spcc- ification of, the primary hypothesis and method of' analysis prevents prohlellis 111volving 1111 11111pic allalysis or 11111111ple Illdi(cs ill oul Icst of' the overall psi ganzfeld effect. Our direct lilts analysis is actually less signiFicant than either the stim of- ranks inediod (z = 4.04, p = 10-5) 2.7 X Or SLarif'ord's z scores (I = 4.53, 354 dj@' p = 4.1 X 10-6). In addition to the priniarv hypothesis, how(.Vel-, Is() (S,(,( two secondary hypotheses concerning the of' tarprel IV])(- alld sender/recciver pairing oil psi performance, and we have presented atory analyses as well. Our Results section in- several purely explor chl(ICS 1r; sigllifi(all(c ICSIs IlIvolvilig psi pelifolillalicc as 111C dcpCll- dent variable, and the p values cited are not ad Listed for multiple comparisons. Of' the 15 Significance tests, 9 are associated With P < .05. The Bonferroni multiple comparisons procedure provides a conservative method of adjusting the alplia level when several si- Intiltaneous tests of significance are performed (Holkind & (',open- haver, 1988; Hyman & Honorton, 1986; Rosenthal & Rubin, 1984). When the Bonferi-oni adJUSLIlICIlt is applied, Six of the little indlivid- ually significant outcomes remain significant; these are: the overall hit rate, the sul)ject-based analysis using Stanford z scores, the dif- ference between dynamic and static targets, the dynamic target hit rate, and the hit rate for experienced subjects. Although the relationship between psi performance and sender type is not independently Significant ill the auLoganzf'Cld, the cor- relation coefficient of .363 is close to that observed the meta analysis (r = .403), and the combined result is Significant. The Cu- Imitative evidence, (herefore, does Support (lie conclusion [hill the sell der/receiver relalionship is a significant moderator of, gallZI,(.1d psi performance. Secun'ty. Given the large number of subjects and the significance sing sub*ects as the unit of analysis, sub'ect. decep- of the outcome it, 9 .1 tion is not .1 plausible explallat loll. Thc autoillated gallZ1,C]d prolocol has been examined by several dozen parapsychologists and behav- ioral researchers from other fields, including well-known critics of parapsychology. Many have participated as subjects, senders, or ob- servers. All have expressed satisfaction with our handling of security issues and controls. In addition, two experts on the simulation ofpsi ability have ex- amined the autoganzfeld system and protocol. Ford,Kross has 6-een Psi Com7nunication M. the Ganfld 135 1)1,()(,(,Ssl()Il;tl Illellialist for over 20 years. I Ie Is the author of' inally _ cies in nientlalist periodic its and has served as SecretarvtI'reas- art;, I -, I I Urcr of, (lie Psych*c Entertainers Association. Mr. Kross has provided its with the following statement: "In my professional capacity as a Im-111.111st. I llavc r(wiewed PsvchoplIN'sical Research Laboratories' @t,utoniaLed ganzFeld systeni and found it to proVide excellent SeCL1- I- LY against deception by subjects" (personal communication, May, I L . J 1989). We have received silli'lar col"InclitS f ronl Daryl Beiii, Pro- u' 'e fessor of Psychology at Corn 11 University. Professor Bem is well known for his research in social and personality psychology. He is also it member of' the Psychic Entertainers Association and has per- t'n-mcd I' S as a Ilicil(:11ist. 11c visiled PRI. f,)I- Several 0I' III;lIly )-cal. I I d I ay 's and was a subject in Series 101. 1,11c *.,;Sll(- of, call only be conclusively ad- I II I I. I I I dressed through independent replications. It is, however, worth Irawing atterition to the 13 sessions in which a visiting scientist, Marilyn J. Schlitz, served as either experimenter (N 7, 29% hits, h = .08) or sender (N = 6, 67% hits, h = .36). Altogether, these sessions yielded 6 direct hits (N = 13, 46.2% hits, h = .45). This eff ct size is more than twice as large as that for the database as a e whole. Status qf the Evidence for Psi Communication in the Ganzfeld The automated ganzfeld studies satisfy the methodological I :)-Ilidelines recommended by Hyman and Honorton (1986). The re- 9 stilts are statistically significant. The effect size is homogeneous across I I experimental series and eight different experimenters. moreover, the autoganzfeld results are consistent w th the outcomes of' the earlier, nonautomated ganzfeld studies; the combined z.of 7,53 would be expected to arise by cliance less than one time in 9 trillion. We have shown that, contrary to the assertions of certain critics 1 (Druckinan & Swets, 1988, p. 175), the ganzfeld ps' effect exhibits "consistent and lawful patterns of covariation found in other areas of' III(ILIlry. The auLolllatCd galIZI'CId Studies display the sanie pat- terns of relationships between psi performance and target type, sender/receiver acquaintance, and prior testing experience found in earlier ganzfeld studies, and the magnitude of these relationships is consistent across the two data sets. The impact of target type and sender/receiver acquaintance is also consistent with patterns in spon- taneous case studies, linking ostensible psi experiences to emotion- ally significant events and persons. These findings cannot be ex- 136 TheJournal pf Paraf)sycholo@gy Psi Communication in the Gan@feld 137 plained by conventional theories of coincidence (Diaconis & We urge ganzfeld investigators to use dynamic targets and to de- Mosteller 1989). sign their studies to allow subjects to have the option to have friends , Hyman and Honorton (1986) have stated, or acquaintances as their senders. The similarity of the autoganzfeld and nieta-analysis data sets strongly indicates that these factors are > ... the best way to resolve the [ganzfeld] controversy ... is to await the important moderators of psi ganzfeld performance. If our estimate > outcome of future ganzfeld experiments. These experiments, ideally, of the impact of dynamic and static targets is accurate, a 50-sessionM -1 0 will be carried out in such a way as to circumvent the file-drawer prob- Series using dynarnic targets has approxiniatcly in 84% chance of 0 < M leln, problems of' 111111liple. analysis, and ill(- variolls dei'cos ill ralldolll- < yielding it significant ouLcollie. A comparable Series With Static tar- M CL -n ization, statistical application, and docurnentation pointed out by CL gets has only about one chance in five of achieving significance. -n 0 Hyman. If a variety of parapsychologists and other investigators collm 0 1 tinue to obtain significant resUhS under these conditions, then tile exism tence of a genuine communications anomaly will have been, demon- REFERENCES M lu strated. (pp. 353-354) ALCOCK, J. E. (1986). Comments on the HymanmHonorton ganzfeld contro- M lu U) M - We have presented a series of experiments that satisfy these versy. Journal of Parapsychology, 50, 345-348. co M guidelines. Although no single investigator or laboratory call satisfy AKERS,C. (1984). MeLhodologicalcriticismsof parapsychology. InS. Krippner 0 the requirement of independent replication, the automated ganzfeld (Ed.), Advances in parapsychological research, Vol. 4 (pp. 112- 164). Jefferm 0 0 0 studies are quite consistent with the earlier studies. On the basis of son, NC: McFarland. (4 & HONORTON, C. (1986). An automated psi ganzfeld testing 8 R. E. BFRGER 44. -L the cumulative evidence, we conclude that the ganzfeld effect rep- , , 44. system. In D_ H. Weiner & D. 1. Radin (Eds.), Research in paraf).sychology _1Z 00 resents a genuine communications anoinaly. This conclusion will 1985 (pp. 85-88). MetUChen, NJ: Scarecrow Press. 00 0 either be strengthened or weakened by additional independent rep- BLACKMORE, S. (1980). The extent of selective reporting of ESP.ganzfeld 0 T> lications, but there is no longer any justification for the claim made studies. Europeanjournal of Parapsychology, 3, 213-219. ' ' l > @U by some critics that the existing evidence does not Warrant serious s laboratory, journa a Visit to Carl Sargent BLAGKMORE., S. (1987). A report Of attention by the scientific community. of the Society for Psychical Research, 54, 186- 198. BRAUD, W. G. (1978). Psi conducive conditions: Explorations and interpTem Recommendations for Future Research tations. In B. Shapin & L. Coly (Eds.), Psi and states Of awareness (pp. I 34). New York: Parapsychology Foundation, Inc. 4 00 Recent psi ganzfeld research has necessarily focused on methm BRAUD, W. G., WOOD, R., & BRAUD, L. W. (1975). Free-response GESP per- 4 visual CO i t te induced b i l h W X odological issues arising from the ganzfeld conLroversy. It is essen- a c s menta Forniance during all exper ypnagog y I CD and acoustic ganzfeld techniques: A replication and extension. Journal of X 0 0 tial that future studies comply with the methodological standards 0 the American Society for Psychical Research, 69, 105- 113. 0 (4 c) serious attention be given to conditions associated with successful Alto, CA: Consulting Psychologists Press, Inc. c4 outcomes. BROWNLEE, K. A. (1965). Statistical theory and methodology in science and engi- 0 0 Small to medium effect sizes characterize many research findings neering. New York: John Wiley & Sons, Inc. 0 0 in the biomedical and social sciences (e.g., Cohen, 1977; Rosenthal, CH I LD, 1. L. (1986). Comments on the ganzfeld controversy. Journal of Para- 0 4- 1984). Rosenthal (1986) and ULLS (1986) make it strong case for p.@yrholo@,y, 50, 137-3,14. J. (1977). Statistical power analysis for the behavioral sciences (rev. e .). 44. COHEN more careful consideration of the magnitude of effect in the design , New York: Academic Press. and analysis of future ganzfeld studies. The automated ganzfeld DIACONIS, P., & MOSTELLER, F. (1989). Methods for studying coincidences. studies show a success rate slightly in excess of 34%. Utts's (1986) Journal of the American Statistical Association, 84, 853-86 1. power analysis shows that for an effect of this size, the investigator DRUCKMAN, D., & SwETs, J. (1988). Enhancing human performance: Issues, the- has only about one chance in three of obtaining a statistically signifm ories, and techniques. Washington, DC: National Academy Press. icant result in a 50-trial Lxperirnem. Even with 100 tYialS-al1 1_11111- HARLEY, T." 9C MA-1-11ii-ms, G. (1987). Cheating, Psi, arid tile appliance of sually large sample size in ganzfeld research-the probability of a science: A'reply to Blackmore. Journal of the Societyfor Psychical Research, significant outcome is only about .5. 54, 199-207. " q Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 5, Volume 19, Number 12, December 1989 Plenum Press - New York-London This issue completes Volume 19 FNDPA4 19(12) 1441-1538 (1989) ISSN 0015-9018 AT FOUND IONS OF P YSICS An International Journal Devoted to the Conceptual Bases and Fundamental Theories of Modern Physics, Biophysics, and Cosmology Editor: Alwyn van der Merwe Editorial Board Asim 0. Barut R. Bruce Lindsay Karl R. Popper Peter G. Bergmann Per-Olov L6wdin Ilya Prigogine Nikolai N. Bogolubov Henry Margenau Abdus Salam David Bohm Jagdish Mehra John L. Synge Robert S. Cohen Andr6 Mercier 'Hans-J. Treder Olivier Costa de Beauregard Louis N6el Jean-Pierre Vigier Robert H. Dicke Kazuhiko Nishijima Mikhail Vol'kenshtein Bai-lin Hao James L. Park Carl Friedrich von Weizs5cker Max Jammer Linus Pauling Eugene P. Wigner Brian D. Josephson Rudolph Peierls Chen-Ning Yang @ounding Editors: He knry Margenau and Wolfgang Yourgraut Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Foundations of PhYsics, Vol. 19, No. 12, 1989 Evidence for Consciousness-Related Anomalies in Random Physical Systems Dean 1. Radin' and Roger D. Nelson 2 Received May 6, 1988; revised June 12, 1989 Speculations about the tole of consciousness in physical sysilems are ftequenil 'v observed in the literature concerned with the interpretation of vantum mechanics. While onlY three experimen tat investigations can be found on @,his topic in physics journals, more than 800 relevant experiments have been repor 'ed in the literature of parapsychology. A well-defined boa@y of empirical evidence from this domain was reviewed using meta-anal ,ytic techniques to assess method@logical quality and overall effect size. Results showed effects conforming to ch4nce expectation in control c*onditions and unequivocal non-chance effects in expe@imenial conditions. This quantitative literature review agrees with the findings of Iwo earlier reviews, suggesting the existence of some form of consciousness- related omalyinrandom physical systems. i7 1. INTRODUCTION The nature of the relationship between human conscioqsness and the physical world has intrigued philosophers for millenia. I@ this century, speculations about mind-body interactions persist, often contributed by physicists in discussions of the measurement problem in quan@um mechanics. Virtually all of the founders of quantum theory-Planc@k, de Broglie, Heisenberg, Schr6dinger, Einstein-considered this subject i depth,"' and contemporary physicists continue this tradition. 12-1) Department of Psychology, Princeton University, Princeton, New Jer 08544. Present address: Conte] Technology Center, 15000 Conference Center Drive, .0. Box 10814, Chantilly, Virginia 22021-3808. tPy 213epartment of Mechanical and Aerospace Engineering, Princeton' U ni@ersity, Princeton, New Jersey 08544. 14" 0015-9018/89/1200-1499W.0010 C@ 1989 Plcnurn Publishing Corporation Approved For Release 2003/04118 CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 1500 Radin and Ne6m The following expression of the problem can be found in a recent interpretation of quantum theory: If conscious choice can decide what particular observation I measure, and there. fore into what states my consciousness splits, might not conscious choice also be able to influence the outcome of the measurement? One possible place where mind may influence matter is in quantum effects. Experiments on whether it is possible to affect the decay rates of nuclei by thinking suitable thoughts would presumably be easy to perform, and might be worth doing."' Given the distinguished history of speculations about the role of consciousness in quantum mechanics, one might expect that the physics literature would contain a sizable body of empirical data on this topic. A search, however, reveals only three studies. The first is in an article by Hall, Kim, McElroy, and Shimony, wbo reported an experiment "based upon taking seriously the proposal that the reduction of the wave packet is due to a mind-body interaction, in which both of the interacting systems are changed."(" This experiment examined whether one person could detect if another person had previously observed a quantum mechanical event (gamma emission from sodium-22 atoms). The idea was based on the supposition that if person A's observation actually changes the physical state of a system, then when person B obser- ves the same system later, B's experience may be different according to whether A has or has not looked at the system. Hall et al.'s results, based on a total of 554 trials, did not support the hypothesis; the observed number of "hits" obtained in their experiment was precisely the number expected by chance (277), while the variance of their measurements was significantly smaller than expected (p<0.05).(') The second study is referred to by Hall et al., who end their article by pointing out that a similar, unpublished experiment using cobalt-57 as the source was successful (40 hits out of 67 trials)."') The third study is a more systematic investigation reported by Jahn and Dunne,01) who summarize results of over 25 million binary trials collected during seven years of experimentation with random-event generators. These experiments, involving long-term data collection with 33 unselected individuals, provide persuasive, replicable evidence of an anomalous correlation between conscious intention and the output of random number generators. Thus, of three pertinent exper-iments referenced in mainstream physics journals, one describes results statistically too close to chance expectation and two describe positive effects."') Given the theoretical implications of such an effect, it is remarkable that no further experiments of this type can be found in the physics literature; but this is not to say that no such experiments have been performed. In fact, dozens of researchers have Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003104118 CIA-RDP96-00789ROO3100030001-4 Consciousness in Physical Systems 1501 reported conceptually identical experiments in the puzzling anO uncertain domain of parapsychology. Perhaps because of the insular! nature of scientific disciplines, the vast majority of these experiments ar@ unknown to most scientists. A few critics who have considered this literture have dismissed the experiments as being flawed, nonreplicable or open to fraud,"-" but their assertions are countered by at! least two detailed reviews which provide strong statistical support for the existence of anomalous consciousness-related effects with rando n um tr generators. 1",18) In this paper, we describe the results of a cor prehensIv quantitative meta-analysis which focused on the questions of "'I ethodologi- cal quality and replicability in these experiments. 2. THE EXPERIMENTS The experiments involved some form of microelectro ic random number generator (RNG), a human observer, and a set of instructions for the observer to attempt to "influence" the RNG to generat.- particular numbers, or changes in a distfibution, solely by intention. RNGs are usually based upon a source of truly random events such s electronic noise, radioactive decay, or randomly seeded pseudorandom quences.(19) Feedback about the distribution of random events is often provided in the form of a digital display, but audio feedback, computer gra@hni cs, and a t variety of other mechanisms have also been uised. Some o the RNGs st devices described in the literature are technically sophisticated, the employing electromagnetic shielding, environmental failsafe imechanisms triggered by deviant voltages, currents, or temperature@ automatic computer-based data recording on magnetic media, redundarit hard copy output, periodic randomness calibrations, and so on."'-'01 RNGs are typically designed to produce a sequence of random bits at the press of a button. After generating a sequence of say, 100 random bits (O's or I's), the number of I's in the sequence may be providedlas feedback. In an experimental protocol using a binary RNG, a run mig@t consist of an observer being asked to cause the RNG to produce, in three successive button presses, a high number (sum of I's greater than chanc@ Iexpectation of 50), a low number (less than 50), and a control condition with no direc- tional intention. An experiment might consist of a group o If individuals each contributing a hundred such runs, or one individual contributing several thousand runs. Results are usually analyzed by comparing high aim and low aim means against a control mean or theoritical chance expectation. i Approved For Release 2003/04/18 : CIA-RDP96-00789RO03iOO030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 1502 Radin and Neism 3. META-ANALYTIC PROCEDURES The quantitative literature review, also called meta-analysis, has become a valuable tool in the behavioral and social sciences.("' Meta-analysis is analogous to well-esiablished procedures used in the physical sciences to determine parameters and constants. The technique assesses replication of an effect within a body of studies by examining the distribution of effect sizes."'-") In the present context, the null hypothesis (no mental influence on the RNG output) specifies an expected mean effect size of zero. A homogeneous distribution of effect sizes with nonzero mean indicates replication of an effect, and the size of the deviation of the mean from its expected value estimates the magnitude of the effect, Meta-analyses assume that effects being compared are similar across different experiments, that is, that all studies seek to estimate the same pop- ulation parameters. Thus the scope of a quantitative review must be strictly delimited to ensure appropriate commonality across the different studies that are combined."") This can present a nontrivial problem in meta- analytic reviews because replication studies typically investigate a number of variables in addition to those studied in the original experiments. In the present case, because different subjects, experimental protocols, and RNGs were employed within the reviewed literature, some heterogeneity attributable to these factors was expected in the obtained distribution of effect sizes. However, the circumscription for the review required that every study in the database have the same primary goal or hypothesis, and hence estimate* the same underlying effect. Experiments selected for review ex amined the following hypothesis: The statistical output of an electronic RNG is correlated with observer intention in accordance with prespecified instructions, as indicated by the directional shift of. distribution parameters (usually the mean) from expected values. Because this "directional shift" is most often reported as a standard normal deviate (i.e., Z score) in the reviewed experiments, we determined effect size as a Z score normalized by the square root of the sample size (N), e = ZI.,[N-, where N was the total number of individual random events (with probability of a hit at p = 0.5, p = 0.25, etc.). This effect size measure is equivalent to a Pearson product moment correlation.(") 3.1. -Unit of Analysis To avoid redundant inclusion of data in a meta-analysis, "units of analysis" are often specified. We employed the following method: If an author distinguished among several experiments reported in a single Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 Consciousrwss in Physical Systems CIA-RDP96-00789RO03liOO030001-4 article with titles such as "pilot test" or "confirmatory tes independent statistical summaries, each of these studies quality-assessed separately. If an experiment consisted o conditions comparing different intentions or types of RN data were split into separate units of analysis to allow 1:1 coded unambiguously. In general, within a given reviev largest possible aggregation of nonoverlapping data col single intentional aim was defined as the unit of analysis ( an experiment or study). For each experiment, a Z score was assigned co whether the observed result matched the direction of int@ negative Z obtained under intention to "aim low" was positive score. When sufficient data were provided in a calculated from those data and compared with the repor 1503 or provided is coded and two or more i devices, the results to be d report, the cted under a -reafter called responding to ntion. Thus, a recorded as a report, Z was ed results; the new calculation was used if there was a discrepancy. If o ly probability levels were reported, these were transformed into the corresponding Z score. For experiments reported only as "nonsignificant," a conservative value of Z=O was assigned; if the outcome was reported only as "statisti- cally significant," Z= 1.645 was assigned; and if sample sizelwas not repor- ted or could not be calculated from the information pro@ided, a special code of N= I was assigned. 3.2. Assessing Quality Because the hypothesized anomalous effect is not easily accom- modated within the prevailing scientific world-view, it is particularly important to assess the trustworthiness of each reviewed experiment. Unfortunately, estimating expefimental quality tends to be a subjective task confounded by prior expectations and beliefs."') Est'mates of inter- judge reliability in assessing the quality of research reports, for example, rarely exceed correlations of 0.5.(") We addressed this problem by assigning to each experiment a single quality weight derived from a set of sixteen binary (present/absent) criteria. The first auth:)r coded and double-checked the coding for all studies; the second autho@ independently coded the first 100 studies. Inter-judge reliability for quality criteria was r=0.802 with 98 degrees of freedom. These criteria were developed from published criticisms about random-number generator experiments( 14,15,29-33) and from expert opinion on important methodological considerations when performing studies involving human behavior. (20,34.35) Collectively, these criteria form a measure of credibility by which to judge the reported da.a. The criteria assess the integrity of the experiment in four categories-procedures, Approved For Release 2003/04/18 : CIA-RDP96-00789ROO31,00030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 1504 Radin 2nd Ne6m statistics, the data, and the RNG device-and they cover virtually all methodological criticisms raised to date. They are (I ) control tests noted, (2) local controls conducted, (3) global controls conducted, (4) controls established through the experimental protocol, (5) randomness calibrations conducted, (6)'failsafe equipment employed, (7) data automatically recor- ded, (8) redundant data recording employed, (9) data double checked, (10) data permanently archived, (11) targets alternated on successive trials, (12)data selection prevented by protocol or equipment, (13)fixed run lengths specified, (14)formal experiment declared, (15)tamper-resistant RNG employed, and (16) use of unselected subjects. Each criterion was coded as being present or absent in the report of an experiment, specifically excluding consideration of previously published descriptions of RNG devices or control tests. This strategy was employed to reflect lower confidence in such experiments since, for example, random- ness tests conducted once on an RNG do not guarantee acceptable perfor- mance in the same RNG in all future experiments. As a result, assessed quality was conservative, that is, lower than the "true" quality for some experiments, especially those reported only as abstracts or conference proceedings. Using unit weights (which have been shown to be robust in such applications"") on each of the sixteen descriptors, the quality rating for an individual experiment was simply the sum of the descriptors. Thus, while a quality score near zero indicated a low quality or poorly reported experiment, a score near sixteen reflected a highly credible experiment. 3.3. Assessing Effect Size Assume that each of K experiments produces effect size estimates e of a parameter E, based on N samples, and that each e has a known standard error s. The weighted mean effect size is calculated as e. @ Y_ o,e,ly_ (o,, where wi = lls' = N,, and i ranges from I to K. The standard error of e. is s, = (y, coi) -'/'. A test for homogeneity for the K estimates of ei is given by HK = Zcoi(ei-e.)', where HK has a chi-square distribution with K-1 degrees of freedom.(") The same procedure can be followed to test for homogeneity of effect size across M independent investigators. In this case, e.i and sj are calculated per investigator, and the test for homogeneity is performed as H. = Y_ wj(e.,- e. )', where e., and coj are mean weighted effect size and Ils' per investigator, respectively, e.,w coje. j C0j, and e j ranges from I to M. Hw has M - I degrees of freedom, For a quality-weighted analysis, we may determine e.Q= Y_ (Qjcoiei)1Y_ (Qjcoi), where Qj is the quality assessed for experiment i. The standard error associated with eQ is seQ = (Y_ (Q'coi)1(Y_ Qicoj)') -/'; the test for homogeneity is similar to that described above. Finally, following Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-0078 Cor,sciousness in Physical Systents the practice of reviewers in the physical sciences, (13,14) we del "outlier" studies to obtain a homogeneous distribution of effe reduce the possibility that the calculated mean effect size ir spuriously enlarged by extreme values. The procedure used v If the homogeneity statistic for all studies was significant (a level), the study that would produce the largest reduction i was deleted; this was repeated until the homogeneity statisti nonsignificant. 4. RESULTS On-line bibliographic databases for psychology and pl were searched, as was a specialized database covering par articles, technical reports, conference proceedings and Altogether 152 references were found :from 1959 to 1987. described 832 studies conducted by 68 different inve@ experimental studies and 235 control studies). Fifty-four exI 33 control studies reported only as nonsignificant were assi@ experiments and two control studies coded as (N= I eliminated from further meta-analysis because effect size accurately estimated (this required the elimination of one in reported a single study). Figures I and 2 show the distributi( reported for control and experimental studies, respectively. CONTROL, N - 235 FIT Th SHI 6 cr 3 0 5 -4 _3 -2 -1 0 1 2 Z-SCORES Fig. 1. Distribution of Z scores reported in 235 control studies. Thirty, %ere reported only as "nonsignificant" and were assigned Z scores of spurious spike at Z=O, those 33 studies were recast as normally bounded by ± 1.64, averaging Z = 0. 625 19'12.5 1000300014 1505 ed potential sizes and to y have been s as follows: the p < 0.05 this statistic had become isics journals :)sychological manuscripts. 'hese reports igators (597 @rimental and ed Z=0. Six Z > 0) were @ould not be stigator who is of Z scores ----- 4 5 of these studies To replace the uted Z scores, Approved For Release 2003/04/18 CIA-RDP96-00789ROO3.100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 1506 [2 9 6 01 3 0 Rodin and Nelsoa REG META-ANALYSIS: FREQUENCY OF Z-SCORES EXPERIMENTAL, N 597 THEORY FIT o o o o o SHIF7 ----- -1 -A 2 3 4 2-SCORES Fig. 2. Distribution of Z scores reported in 597 experimental studies. Fifty-four of these studies were reported as -nonsignificant" and were assigned Z scores of zero. As in Fig. 1, those 54 studies were recast as normally distributed Z scores, bounded by ± 1.64, averaging Z= 0. 8 2- .U 0rN .1 (a) CKTI (b) EXPT (c) EXPi (d) EXP H (t) EXPA H 0) OUAL (g) CU& H Fig. 3: Mean effect size point estimates ± I standard error for (a) control studies and (b) individual experiments; (c) mean effect size per investigator, (d) homogeneous mean effect size for experiments, (e) homogeneous mean effect size per investigator, (f) mean effect size for quality-weighted experiments, and (g) mean effect size for homogeneous quality-weighted experiments. N- 591 N 61 N1.90 N 591 N 00 N - 58 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003104118 : CIA-RDP96-00789RO03iOO030001-4 Consciousness in Ph%-sical Systems These results, expressed as overall mean effect sizes, sho studies conform well to chance expectation (Fig. 3a), and tha effects, whether calculated for studies or investigators, devia from chance expectation (Fig. 3b, 3c). To obtain a homogei tion of effect sizes, it was necessary to delete 17% of ind studies (Fig. 3d) and 13 % of mean effect sizes across investig This may be compared with exemplary physical and social s where it is sometimes necessary to discard as many as 45 */c to achieve a homogeneous effect size distribution."') Of ind deleted, 77% deviated from the overall mean in the positive of investigator means deleted, all were positive (i.e., sup experimental hypothesis). 4.1. Effect of Quality 1507 v that control , experimental e significantly eous distribu- vidual outlier itors (Fig. 3e). :ience reviews, of the studies vidual studies direction, and )ortive of the Some critics have postulated that as experimental qual Ity increases in these studies, effect size would decrease, ultimately regressin to th "tr " value of zero, i.e., chance TeSUItS.112,13,15.32'33.38' We tested his conjecture with two linear regressions of mean effect size vs. mean quality assessed per investigator, one weighted with a)j as defined above and the Other weighted with the number of studies per investigator. The calculateo slope for the former is -2.5xJ0-'_+3.2xl0-', and for the latter i-7.6xlO-'+ 3.9xlO-'. These nonsignificant relationships between qu lity and effect size is typical of meta-analytic findings in other fields, (3 40) suggesting that the present database is not compromised by poof experimental methodology. Another assessment of the effect of quality was obtained by comparing unweighted and quality-weighted effect sizes @er experiment (Fig. 3b vs. 3f). These are nearly identical, and the samf is true after deleting outliers to obtain a homogeneous quality-weighted distribution (Fig. 3d vs. 3g), confirming that differences in methodolog@cal quality are not significant predictors of effect size. [u It might be argued that the quality assessment pro Id rte employed here was nonoptimal because some quality criteria arec;ore important than others, so that if appropriate weights were , assigned, the quality-weighted effect size might turn out to be quite di@lerent. This was tested by Monte Carlo simulation, using sets of 16 w0ghts, one per criterion, randomly selected over the range 0 to 6. A quality" weighted effect size was calculated for the 597 experiments as before, Inow using the random weights instead of unit weights, and this process w' s repeated one thousand times, yielding a distribution of plossible quali.1ty ratings. The average effect size from the simulation was 3.18xlO-i I" + 0.15 x 10-4, indicating that in this particular database coded by these !sixteen criteria, Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Radin and Nelsm 1W the probable range of the quality-weighted mean effect size clearly excludes chance expectation of zero. 4.2. The "Filedrawer" Problem Although accounting for differences in assessed quality does not nullify the effect, it is well known in the behavioral and social sciences that non- significant studies are published less often than significant studies (this is called the "filedrawer" problem 121,41-43 @). If the number of nonsignificant studies in the filcdrawer is large, this reporting bias may seriously inflate the effect size estimated in a meta-analysis. We explored several procedures for estimating the magnitude of this problem and to assess the possibility that the filedrawer problem can sufficiently explain the observed results. The filedrawer hypothesis implicitly maintains that all or nearly all significant positive results are reported. If positive studies are not balanced by reports of studies having chance and negative outcomes, the empirical Z score distribution should show more than the expected proportion of scores in the positive tail beyond Z= 1.645. While no argument can be made that all negative effects are reported, it is interesting to note that the database contains 37 Z scores in the negative tail, where only 30 would be expected by chance, On the other hand, there are 152 scores in the positive tail, about five times as many as expected. The question is whether this excess represents a genuine deviation from the null hypothesis or a defect in reporting or editorial practices. This question may be addressed by modeling based on the assumption that all significant positive results are reported. A four-parameter fit mini- mizing the chi-square goodness-of-fit statistic was applied to all observed data with Z>, 1.645, using the exponential Y = I e v@2_ (Ix -1i") vF2 or to simulate the effect of skew or kurtosis in producing the dispropor- tionately long positive tail. This exponential is a probability distribution with the same mean and variance as the normal distribution, but with kurtosis = 3.0. To begin, the null hypothesis of a (0, 1) normal distribution with no kurtosis was considered. To account for the excess in the positive tail, N = 585,000 filedrawer studies were required, and the chi-squared statistic remained far too large to indicate a reasonable fit (see Table 1). This large N, in comparison with the 597 studies actually reported together with the poor goodness-of-fit statistic, suggests that the assumption of a (0, t) normal distribution is inappropriate. Approved For Release 2003/04/18: CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 1508 Radin and NeLm the probable range of the quality-weighted mean effect size clearly excludes chance expectation of zero. 4.2. The "Filedrawer" Problem Although accounting for differences in assessed quality does not nullify the effect, it is well known in the behavioral and social sciences that non- significant studies are published less often than significant studies (this is called the "filedrawer" problem (21.41-43 1). If the number of nonsignificant studies in the filedrawer is large, this reporting bias may seriously inflate the effect size estimated in a meta-analysis. We explored several procedures for estimating the magnitude of this problem and to assess the possibility that the filedrawer problem can sufficiently explain the observed results. The filedrawer hypothesis implicitly maintains that all or nearly all significant positive results are reported. If positive studies are not balanced by reports of studies having chance and negative outcomes, the empirical Z score distribution should show more than the expected proportion of scores in the positive tail beyond Z= 1.645. While no argument can be made that all negative effects are reported, it is interesting to note that the database contains 37 Z scores in the negative tail, where only 30 would be expected by chance. On the other hand, there are 152 scores in the positive tail, about five times as many as expected. The question is whether this excess represents a genuine deviation from the null hypothesis or a defect in reporting or editorial practices. . This question may be addressed by modeling based on the assumption that all significant positive results are reported. A four-parameter fit mini- mizing the chi-square goodness-of-fit statistic was applied to all observed data with Z >, 1.645, using the exponential Y= e - 52 Ox -jul@a) (1) to simulate the effect of skew or kurtosis in producing the dispropor- tionately long positive tail. This exponential is a probability distribution with the same mean and variance as the normal distribution, but with kurtosis = 3.0. To begin, the null hypothesis of a (0, 1) normal distribution with no kurtosis was considered. To account for the excess in the positive tail, N = 585,000 filedrawer studies were required, and the chi-squared statistic remained far too large to indicate a reasonable fit (see Table 1). This large N, in comparison with the 597 studies actually reported together with the poor goodness-of-fit statistic, suggests that the assumption of a (0, 1) normal distribution is inappropriate. Approved For Release 2003104/18: CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO31100030001-4 Consciousness in Physical Systems 15,09 Table 1. Four-Parameter Fit (E:N, N, Mean, sd) Minimizing Chi-S@ uare (10df) Goodness-of-Fit Statistic to the Positive Tail of the Observed Z Scor@ Distribution, for Several Exponential: Normal Ratios' Assumption E: N ratio N Mean sd Chi IL I square P Normal distribution 0 585.000 0 1 57 @,@ 67.84 0 , (null hypothesis) 1 5,300 0 1 20 97 0 2 4,800 0 1 J 67.84 0 3 4,600 0 1 11 49.45 0 10 4,400 0 1 !1 19.69 0 Empirical distribution 0 700 0.145 2.10 23.94 0.008 1 747 0345 1.90 1632 U91 2 757 0.445 1.80 14.21 0.164 3 777 0 445 1 90 226 11 08 0 . . . . 10 807 0.445 1.80 11.08 0.351 'The null hypothesis is tested by clamping the mean at 0 and the standa deviation at 1, allowing N and E:N to vary. The empirical database is addressed by ]owing all four parameters to vary. I Adding simulated kurtosis to a (0, 1) normal distribution by mixing exponential [ Eq. (I ) ] and normal distfibutions in a I : I ratiol reduced N by two orders of magnitude, and ratios of 2:1, 3: 1, and 10: 1 @xponential to normal (E:N) yielded further small improvements. How@ver, the chi- squared statistic still indicated a poor fit to the empirical -1 ata. Applying the same mixture of exponential and normal distributions, but starting from the observed values of N = 597, mean Z score= 0.645, @and standard deviation= 1.601, with the constraint that the mean could,@nly decrease from 0.645, resulted in much better fits to the data. Tab I shows the results. I This procedure shows that the null hypothesis is unvia0e, eve .n after allowing a huge filedrawer. The chi-square fit vastly impr@ves with the addition of kurtosis, but only becomes a reasonably good It when mean and standard deviation are allowed to approximate the emPinical values. The filedrawer estimate from this model depends on a num"r of assump- tions (e.g., the true distribution is generally non-nal, but hassTa dispropor- tionately large positive tail). It suggests a total number of!experimental studies on the order of 800, of which three-fourths have @en formally reported. A somewhat simpler modeling procedure was appliedi to the data assuming that all studies with significant Z scores in either tl@e positive or negative tail arc reported. The model is based on the normal distribution with a standard deviation = 1, and estimates the mean and required to Approved For Release 2003104/18: CIA-RDP96-00789RO03 100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Radin and Nt6w 1510 account for the 152 Z scores in the positive tail and 37 Z scores in the negative tail. This mean-shift model, which ignores the shape of the observed distribution, results in an N = 1,580 and a mean Z score = 0.34. These modeling efforts suggest that the number of unreported or unretrieved RNG studies falls in the range of 200 to 1,000. A remaining question is, how many filcdrawer studies with an average null result would be required to reduce the effect to nonsignificance (i.e., p < 0.05)? This "failsafe" quantity is 54,000--approximately W times the number of studies actually reported. Rosenthal suggests that. an effect can be considered robust if the failsafe number is more than five times the observed number of studies. (21) 5. DISCUSSION Repeatable experiments are the keystone of experimental science. In practice, repeatability depends upon a host of controllable and uncon- trollable ingredients, including factors such as stochastic variation, changes in environmental conditions, difficulties in communicating tacit knowledge employed by successful experim enters,"') and so on. Difficulties in achieving systematic replication are therefore ubiquitous, from experimental psychology J2 1,11) to particle physics. 121.21) Of course, this is not to say that systematic replication is impossible in these or other fields, but it may appear to be extraordinarily difficult when experiments are considered individually rather than cumulatively. In the case of the present database, the authors of a recent report issued by the US National Research Council stated that the overall results of the RNG experiments could not be explained by chance,"" but they questioned the quality and replicability of the research. This meta-analysis shows that effects are not a function of experimental quality, and that the replication rate is as good as that found in exemplary experiments in psychology and physics. Besides the issue of replicability, five other objections are often raised about the present experiments. These are (a) the effect is inconsistent with prevailing scientific models, (b) the experimental methodology is techni- cally naYve, thus the results are not trustworthy, (c) the experiments are vulnerable to fraud by subjects or by experimenters, (d) skeptics cannot obtain positive results, and (e) there are no adequate theoretical explana- tions or predictions for the anomalous effect. These criticisms may be addressed as follows: (a) "Inconsistency with the scientific world-view" is essentially a philosophical argument that carries little weight in the face of repeatable experimental evidence, as suggested by the present and two corroborating meta-analyses."'-"' Approved For Release 2003104/18: CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003104118: CIA-RDP96-00789ROO3100030001-4 Consciousness in Physical Systems Indeed, if the "inconsistency" argument were sufficien@ to discount anomalous findings, we would have ignored'much of the motivation leading to the development of quantum mechanics. (b) The '@naYve method- ology" argument was empirically addressed by the i f` ssessment of methodological quality in the present analysis. No significapt relationship between quality and effect size was found, (c) Fraud postulated as the explanation of the results is untenable as it would @ave required widespread collusion among 68 independent investigators In any case, even severe critics of parapsychological experiments have d;counted fraud as a viable explanation.("' (d)Skeptics often assert that oply "believers" obtain positive results in such experiments. However, a thor@ ugh literature search finds not a single attempted replication of the RNG experiment by a publicly proclaimed skeptic; thus the assertion is not base on verifiable evidence. Furthermore, skeptics who claim to have attempt d replications insist (without providing details or references) that the have never achieved positive results in any of their RNG experiments@ (15,47) Such a claim is itself quite remarkable, as the likelihood of neve obtaining a statistically significant result by chance in series of experi ents can be extremely low, depending on the number of experiments con ucted ' Unfor- tunately, because we cannot determine how many experi ents skeptics have actually conducted, it is impossible to judge the v lidity of this criticism. I Finally, (e) the "no theoretical basis" argument is corrc@t, but it does not support a negative conclusion about experimental obser@ation. There are at present no adequate theories, with the possible exce tion of some interpretations of quantum mechanics,('-'-'-'" that convincin ly explain or predict consciousness-related anomalies in random physical systems. We note, however, that the anomalous effects reviewed in this pa r apparently PC can be operationally predicted under well-specified conditio is. For exam- 's ple, when individuals are instructed to "aim" for high (or low) numbers in RNG experiments, it is possible to predict with some sm 11 degree of confidence that anomalous positive (or negative) shifts o distribution means will be observed. 6. CONCLUSION In this paper, we have summarized results of all knowni experiments testing possible interactions between consciousness and the statistical behavior of random-number generators. The overall effect siz@ obtained in experimental conditions cannot be adequately explained by melthodological flaws or selective reporting practices. Therefore, after consider@ng all of the 00030001-4 Approved For Release 2003104118: CIA-RDP96-00789ROO31 Approved For Release 2003/04/18: CIA-RDP96-00789ROO3100030001-4 Radin and Ne" 1512 retrievable evidence, published and unpublished, tempered by all legitimate criticisms raised to date, it is difficult to avoid the conclusion that under certain circumstances, consciousness interacts with random physical systems. Whether this effect will ultimately be established as an overlooked methodological artifact, as a novel bioelectrical perturbation of sensitive electronic devices, or as an empirical contribution to the philosophy of mind, remains to be seen. ACKNOWLEDGMENT'S This study was supported by major grants from the James S. McDonnell Foundation, Inc. and the John E. Fetzer Foundation, Inc. The authors express their gratitude to Dr. York Dobyns of the Princeton University Engineering Anomalies Laboratory for his assistance with the filedrawer models. REFERENCES 1. R. G. Jahn and B. J. Dunne. Margins of Reality (Harcourt Brace Jovanovich, Orlando, Florida, 1987). 2. B. d'Espagnat, "The quantum theory and reality," Sci. Am., pp. 158-181 (November, 1979). 3. 0. Costa de Beauregard, "S-matrix, Feynman zigzag and Einstein correlation," Phys. Len. 67A, 171-173 (1978). 4. N, D. Mermin, "Is the moon there when nobody looks? Reality and the' quantum theory,- Phys. Today, pp. 38-47 (April, 1985). S. A. Shimony, "Role of the observer in quantum theory," Am. J. Phys. 31, 755 (1963). 6. E. P. Wigner, "The problem of measurement," Am. J. Phys. 31, 6 (1963). 7. U. Ziemelis, "Quantum-mechanical reality, consciousness and creativity," Can. Res. 19, 62-68 (September, 1986). 8. E. J. Squires, "Many views of one world-an interpretation or quantum theory," Eur. J. Phys. 8, 173 (1987). 9. J. Hall, C. Kim, B. McElroy, and A. Shimony, "Wave-packet reduction as a medium of communication," Found. PhYs. 7. 759-767 (1977); p. 761. 10. R. Smith, unpublished manuscript, MIT, 1968. (Cited in Ref. 9, p. 767.) 11. R. G. Jahn and B. J. Dunne, "On the quantum mechanics of consciousness, with applica- tion to anomalous phenomena," Found. Phys. 16, 721-772 (1986)@ 12. J. E. Alcock, Parapsychology: Science or Magic? (Pergamon Press, Elmsford, New York, 1981), pp. 124-125. 13. M. Gardner, Scienee: Good, Bad, and Bogur (Prometheus Books, Buffalo, NewYork, 1981). 14. R. Hyman, "Parapsychological research: A tutorial review and critical appraisal," Proc. IEEE 74, 823-849 (1986). 15. P. Kurtz, "Is parapsychology a science?" In Paranormal Borderlands of Science, K. Frazier, ed. (Prometheus Books, Buffalo, New York, 1981). Approved For Release 2003/04/18: CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 CIA-RDP96-00789RO03 i100030001-4 1513 Consciousness in Physical Systems 16. D. F. Marks, "Investigating the paranormal," Nature (London) 320,11119-124 (1986). 11. C. Honorton, "Replicability, experimenter influence, and Parapsychology: An empirical context for the study of mind," paper presented at the annual m Eing of the AAAS, Washington, D.C., 1978. 1 18. E. C. May, B. S. Humphrey, and G. S. Hubbard, Electronic system perturbation techniques." SRI International Final Report, September 30, 1980. 19. H. Schmidt, "Precognition of a quantum process," J. Parapsychol. 33 99-108 (1969); "A PK test with electronic equipment," J. Parapsychol. 34, 175-181 (1970); "Mental influence on random events," New Sci. Sci. J. 50, 757-758 (1971); "PK tests wi h pre-recorded and pre-inspected seed numbers," J. Parapsychol. 45, 87-98 (1981). 20. R. G. Jahn, "The persistent paradox of psychic phenomena: An engin! ring perspective," cc Proc. IEEE 70, 136-170 (1982); R. D. Nelson, B. J. Dunne, and R. . Jahn, "An REG experiment with large data-base capability, III: Operator-related an malies," Technical Note PEAR 84003, Princeton Engineering Anomalies Research La ratory, Princeton University, School of Engineering/Applied Science, September J984; H. Schmidt, R. Morris, and L. Rudolph, "Channeling evidence for a PK eff I t to independent observers," J. Parapsychol. 50, 1-16 (1986). 21. R. Rosenthal, Meia-Anal 'wic Procedures for Social Research (Sage P blications, Beverly Hills, California, 1984); K. Wachter, "Disturbed by meta -anal v@isr' Scientce 241, 1407-1408 (1998). We may note that Cohen's h, the difference L wee, 'on "I and experimental proportions, is a common effect size measure that might@ have been used in the present study. This was rejected in favor of e, as defined, because s1me of the reviewed studies reported only final p values or only overall Z scores; e was us deemed more useful in the present meta-analysis. r 22. R. L. Bangert-Drowns, "Review of developments in meta-analytic method," Psychol. Bull. 388-399 (1986). 23, A. H. Rosenfeld, "The particle data group: Growth and operations." A"u. Rev. Nucl. Sci. 25, 555-599 (1975). 24. C. G. Wohl et al., Rev. Mod Phys. 56, Part 11, p. S5 (1984). 25. G. V. Glass, "In defense of generalization," Behav. Brain Sci, 3, 394-395 (1978). 26. H. M. Cooper, "Scientific guidelines for conducting integrative review@,_ Rev. Educ. Res. 52, 291-302 (1982). 27. R. M. Dawes, -You can't systematize human judgment: Dyslexia." In @ew Directions for Methodology of Social and Behavioral Science: Fallible Judgment in Behavorial Research, R. A. Shweder, ed. (Jossey-Bass, San Francisco, 1980), pp. 67-78. 1 28. S. D. Gottfredson, "Evaluating psychological research reports: Dimensions, reliability, and correlates of quality judgments,- Am. Psychol. 33, 920-934 (1978J.1 29. C. Akers, "Methodological criticisms of parapsychology," In Advances fn Parapsychologi- cal Research, Vol. 4, S. Krippner, ed. (McFarland, Jefferson, North Car@lina, 1994); "Can meta-analysis resolve the ESP controversy?" In A Skeptic's Handbook @f Parapsychology, P. Kurtz, ed. (Prometheus Books, Buffalo, New York, 1985). i 30. 1. E. Alcock, "Parapsychology: Science of the anomalous or search fo the soul," Behav. Brain Sci. 10, 553-565 (1987). 33. P. Diaconis, "Statisti6a) problems in ESP research," Science 201, 131-1P 0978). 32. C. E. M. Hansel, ESP and Parapsychology: A Critical Reevaluation (P ometheus Books, Buffalo, New York, 1980), 33. R. Hyman, "The ganzWd psi experiment: A critical appraisal," J. POT@psychol. 49, 3-50 (1985). 1 34. T. X. Barber, Pitfalls in Human Research: Ten Pivotal Points (Pergamoo Press, Elmsford, New York, 1976). Approved For Release 2003104118: CIA-RDP96-00789ROO31.00030001-4 Approved For Release 2003/04/18: CIA-RDP96-00789ROO3100030001-4 Radin and Nelson 1514 35. J. B. Rhine, "comments: 'A new case of experimenter Linreliability,'" J. ParapsYchol. 38, 215-255 (1974). 36. R. M. Dawes, "The robust beauty of improper linear models in decision making," Am. Psychol. 34, 571-592 (1979). 37. L V. Hedges, "How hard is hard science, how soft is soft science?" Am. Psychol. 42, 443455 (1997), 38. C. E. M. Hansel, ESP: A Scientific Evaluation (Charles Scribner's Sons, New York, 1966), p. 234. 39. R. Rosenthal and D. B. Rubin, "Interpersonal expectancy effects: The first 345 studies," Behav. Brain Scf. 3, 377-415 (1978). 40. G. V, Glass, B. McGaw, and M. L. Smith, Meta-analysis in Social Research (Sage Publi- cations, Beverly Hills, California, 1981). 41. Q. McNemar, "At random: Sense and nonsense," Am, Psychol. 15, 295-300 (1960). 42. S. Iyengar and J. B. Greenhouse, "Selection models and the rile-drawer problem," Technical Report 394, Department of Statistics, Carnegie-Mellon University (July, 1987). 43, L. V. Hedges, "Estimation of effect size under nonrandorn sampling: The effects of censoring studies yielding statistically insignificant mean differences," A Educ. Stal. 9, 61-86 (1994), 44. H. H. Collins, Changing Order: Replication and Induction in Scientific Practice (Sage Publications, Beverly Hills, California, 1985). 45. S. Epstein, "The stability of behavior, II: Implications for psychological research," Ani. Psychol. 35, 7"06 (1980). 46. D. Druckman and J. A. Swets, eds. Enhancing Human Performance: Issues, Theories, and Techniques (National Academy Press, Washington, D.C., 1988), p. 207. 47. A. Neher, The Psychology of Transcendence (Prentice-Hall, Englewood Cliffs, New Jersey, 1980), p. 147. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Reprintedftom the Proceedings of the Parapsychological Association 33rd Annual Convention, Chevy Chase, ME), Atigust 16-20, 1990, Extraversion and ESP Performance: A Meta-Analysis and a New Confirmation Charles Honorton and Diane C. Ferrari Psychophysical Research Laboratories and Daryl J. Bem Cornell University ABSTRACT--- We report a meta-analysis of research on the relationship between performance in extrasensory perception (ESP) tasks and the psychological trait extraversion. The meta-analysis comprises 60 independent studies, 17 independent investigators, and 2,963 subjects. The overall weighted mean correlation is small (r = .09), significant (z = 4.63, p = .000004), and nonhoiriogeneous. For forced-choke ESP studies (N = 45), the ESP/ extraversion relationship appears to be an artifact of subjects'knowledge of their ESP performance upon theirresponses to the extraversion measure: evidencefor therelationship is limited to studies where subjects completed the ESP task prior to extraversion assessment (N = 18 studies, r =.17, z = 3.51); no evidence for an ESP/extraversion relationship exists in stucLies where extraversion was assessed before the ESP task (N = 16 studies, r = -.02, z = -0.79). The two correlations differ significantly (z = 3.58, p = .00045). For free-response studies, a significant ESP/extraversion relationship exists that is free of this problem: extraversion testing preceded the ESP task in 11 of the 14 free-response studies (r =.21, z = 4.57, p = .000005). The ESP/extraversion relationship is both significant (r = .20, z=4.46,p =,0000083) and homogeneous in the subset of free-response studies involving individual testing (N = 12 studies). The effect is homogeneous across investigators and extraversion scales. We also report a new confirmation of the ESP/ex- traversion relationship using the Extraversion/ Introversion Scale of the Myers@Briggs Type Indicator. The effect size (r = .18, 1 = 2.67, 219 df, p = .008) is very close to the meta-analytic estimate for free-response studies (r = .20) and is homogeneous across the eight experimenters. While the relationship between extraversion and ESP in the forced-cboice studies is probably ardfactual, we conclude that there is a significant ESP/extraversion relationship in the free-response studies, that the relationship is consistent across investigators and scales, and that meta-analysis of parapsychological research domains has predictive validity. THE RELATIONSHIP BETWEEN ESP perform- ance and individual differences in psychological traits has been explored in many studies since the 1940s. Extraversion is one of the most frequently studied trait variables and three narrative reviews of the ESP/ extraversion literature have concluded that ESP performance is positively related to extraversion (Eysenck, 1967; Palmer, 1977; Sargent, 1981). We pre- sent a meta-analysis of the extraversion/ESP litera- ture. The purpose of the meta-analysis is to (1) update earlier narrative reviews of the ESP /extraversion re- lationship, (2) estimate the magnitude of the relation- ship, (3), assess pow5ntial threats to validity, and (4) identify proceduraland other variables that moder- ate the relationship. We will then present a new confirmation of the ESP/extraversion relationsl-dp and compare its magnitude to that estimated from the meta-analysis. The Meta-Analysis Previous parapsychological meta-analyses have focused on evidence for psi functioning in such re- search domains as ganzfeld communication (Honor- ton, 1985), precognition (Honorton & Ferrari, 1989), and studies of the impact of conscious intention on random number generators (Radin & Nelson, 1989). In such cases the effect size index is based on the proportion of hits and the unit of analysis is the trial. In the present case, we are interested in the relation- ship between psi performance and a predictor vari- able, extraversion, rather than overall psi perform- This work was supported by SRI International and by the John E. Fetzer Foundation. We wish to express our appreciation to Edwin C. May of SRI International, who recognized the importance of this work and helped make it possible. We are also grateful to Robert L. Morris, John Palmer and K. Ramakrishna Rao for providing additional information concerning several of the studies used in the, meta-analysis, and to H. Kantbamani for valuable comments during the initial phase of this study. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 2 EXTRA VERSION & ESP: A META -A NA L YSIS & NEW CON@IRMA Approved For Release 2003/04/18: CIA-RDP96-00789ROO i0003"401-4 ance, The effect size index is the correlation coeffi- vant to assessing r@, search quality are described in the cient between the two variables and the subject is the following section. Sampling and statistical features unit of analysis. coded include the, i number of subjects, the subject populationused, e instrument used for measuring Method extraversion, and he inferential statistics reported for testing the rela tipnship between ESP performance Retrieval of Studies and extraversion. 1, We attempted to retrieve all English-language studies of the relationship between extraversion and performance in experimental ESP tasks. The source of studies includes the bibliographies of three narra- tive reviews (Eysenck, 1967; Palmer, 1977; Sargent, 1981) and inspection of the principal English-lan- guage outlets for publication of parapsychological research, including the Jaurnal of Parapsychology, Jour- nal of the American Societyfor Psychical Research, fournal of the Society for Psychical Research, and Research in Parapsychology. In addition, we conducted a com- puter search of Psychological Abstracts using the key- words "extraversion," "extroversion," "introver- sion," "intraversion," "sociability," and "outgoing." In vestigator Definition For the purpose of the meta-analysis, we defined independent investigators as investigators who have not worked with other investigators in the data base. In cases of multiple authorship, studies are identified by the senior author of the earliest publication in the data base. For example, studies by Kanthamani (1966), Kanthamani and Rao (1972) and Krishna and Rao (1981) are all indentified as Kan&mani studies. While Kanthamani and Krishna did not work to_ gether, they share Rao as a coauthor and are therefore considered to represent a single investigator set. Similarly, studies by Humphrey (1945,1951) and by Nicol and Humphrey (1953,1955) are all indentified as Humphrey studies. Laboratory affiliation was not used to identify independent investigators since sev- eral investigators worked in different laboratories and several generations of investigators worked in one laboratory. Procedural Features Besides bibliographical data identifying the inves- tigator, publication source and date, we coded vari- ous procedural, sampling, and statistical features for each study. The procedural features include the type of ESP task (forced-choice or free-response), test set- ting (individual or group testing), ESP mode (telepa- thy, clairvoyance, precognition, mixed), and amount of feedback in the ESP task. Procedural features rele- Criteria for Ass+ing Research Quality Methodological fariables were coded in terms of procedural descriptl@ons (or their absence) in the re- search reports. Thislapproach was used in an earlier meta-analysis of th@ ESP,=feld domain (Honor- ton, 1985), resul in stu N, quality ratings that were t g@it(r26= generally in agreem r .77, p = 10-6) with inde- pendent "flaw" ratings by an outside critic (Hyman, 1985).'rwo sets of cri @ria were used. One set assessed threat @ of the ESP measure. The other set assessed threats tb, the validity of the relationship between the ESP and extraversion measures. The ESP quality 'pnalysis includes four criteria. 0ne point was give@ (or Nvithheld) for each of the following: Control against sen@, ry ImUge. Credit was given to GESP (telepathy) studies if the report specified that each of the followm4 cn*ten*a were met: sender and receiver were locatedl in separate rooms, the test situ- ation prohibited aud@tory or other cues from sender to receiver, the send@r and receiver were monitored by experimenters, ar@d in free-response studies in- volving subject jud*& duplicate target sets were employed. Credit wa@ given to clairvoyance studies if the report specified ithat cues from the targets were i prohibited by means @f physical distance, screens, or opaque packaging. Pr@ognition studies were consid- ered to be immune to sensory leakige Iproblems. Randomization. Stu, @ies received credit if random number tables, rando@m number generators, or me- chanical shufflers wete used to randomize the tar- gets. Studies using infon-nal methods of randomiza- tion such as hand sh4fling and dice throwing, or in I which there was no @ randomization, received no credit. Reports failing to identify the method of ran- domization received credit. Duplicate recording. tudies reporting duplicate re- cording of targets an @responses received creclit. Re- d ports that failed to d 'be their data recording pro- es _@n cedures and those in hich data recording was per- formed by a single ex rimenter, received no credit, pe Duplicate scoring. tudies reporting duplicate I- checking of hits reeceiv c it. Reports that failed to ed describe their scorin p Irocedures and those in which 9 scoring was perform by a single experimenter, re- el ceived no credit, Approved For Release 2003/04118: CIA-RDP96-00789ROO3100030001-4 io,tokToN. FERkARI& BEM Approved For Release 0 3/04/18 CIA-RDP96-00789ROO3100030001-4 Figure 1. stem-and-ioat frequency distribution of cor- relation coefficients between extraversion and ESP per- formance (N 47). 4 4 3 6 2 5 2 1 7 4 1 0 9 2 0 0 0 0 5 8 9 1 0 0 1 2 2 7 8 8 2 0 3 3 8 .3 0 0 4 4 4 5 7 8 8 4 0 6 6 9 9 5 3 4 8 6 6 7 7 8 9 1 We assessed two aspects of methodology that could result in a spurious correlation between ESP scores and measures of extraversion: advance speci- fication of the criteria used to define extraversion and the order in which the ESP task and extraversion measure were administered, A priori definition of extraversion. An inflated rela- tionship between ESP and extraversion scores could occur if the investigator selected an "optimal" extrav- ersion/introversion breakdown after observing the data, without correcting for multiple analysis. We coded studies as to whether the classification of ex- traversion /introversion appeared to be predeter- mined or post hoc. No credit was given for studies using nonstandard classifications (e.g., 1 sd, median split, etc.) unless the report explicitly stated that the method of classification was preplanned. Order ofadministration ofextraversion and ESP meas- ures. A spurious correlation between ESP scores and extraversion could arise if subjects' responses on the extraversion scale were influenced by knowledge of their performance in the ESP task. There is some evidence that subjects' responses to psychological tests may be influenced by feedback concerning their ESP performance (palmer & Lieberman, 1975). We coded studies as to whether the extraversion scale was given before or after the ESP test. This informa- tion was available for 45 of the 60 studies. Meta-Analysis of Correlation Coefficlents We combined correlations across independent studies using the procedures described by Hedges and Olkin (1985) and Rosenthal (1984). All statistics were converted to indices of association; t tests were converted to point-biserials and ph coefficients were computed from 2 x 2 contingency tables.' We esti- mated unreported correlations from the reported p values (with results reported onlyas "nonsigrifficant" obtained in 13 cases set equal to .00). The signs of the correlations were adjusted if necessary to insure that positive correlations reflect positive relationships be- tween extraversion and ESP performance. The corre- lations were pooled across studies butwithin catego- ries based on methodological features associated with the studies. The correlations were transformed to their Fisher's z equivalents, weighted by their df, and averaged. We determined the two-tailed signifi- cance levels and 95% confidence intervals (Cls; Hedges and Olkin, 1985). Finally, we conducted chd- square tests of homogeneity (Hedges and Olkin, 1985; Rosenthal, 1984) and transformed mean zs back to the r metric. The chi-square homogeneity test assesses the con- sistency of study outcomes, providing a quantitative index of replicability. A set of studies are exact repli- cates if their effect sizes are identical. They are homo- geneous if the variability of effect sizes can be ex- plained by sampling error. A significant but nonho- mogeneous effect indicates the presence of moderat- ing variables and homogeneity tests are used to iden- tify moderating variables by subdividing studies into smaller, methodologically similar subgroups (Hedges, 1987). Results We retrieved 60 independent studies contained in 35 publications by 17 independent investigators, The studies were reported over a span of 38 years, be- tween 1945 and 1983. The data base comprises 2,963 subjects. Forty-five studies involve ESP card-guess- ing tasks or similar forced-choice tasks. Fourteen studies employed free-response ESP tasks, and one involved a remote physiological influence task The unweighted correlations range from -0.44 to .91. Figure 1 shows a stem-and-leaf display (Tukey, 1977) of the correlation coefficients. (The 13 studies that were assigned r's of zero because of insufficient information are omitted). Unlike other methods of TWo studies provided only trial-based tests (CRd). Cor- relations were estimated for these studies using a method for estimating effect sizes from critical ratios reported by McCarthy & Schechter (1986). Their method provides an estimate of Cohen's d which we then converted to the r metric. Approved For Release 200.3/04/18 : CIA-RDP96-00789ROO3100030001-4 4 FXTRA VERSION& ESP, A META -ANALYSIS Jk NEW Mi@FIRMA TION Approved For Release 2003/04/18: CIA-RDP96-00789ROO3100030001-4 Table], Summary of Extraversion-ESP Meta-Anolysis Number of Number of Total Mean 95% c1brifidence independent independent number of effect size interval effect sizes investigators subjects (r) Fr m TO z P X 2(k-1) Overall, all studies 60 17 2963 .09 A i .12 4.63 .000004 126.21* Studies with r available 47 14 1853 .14 i .101 A9 5.90 <@000001 115,910 Forced-choice guessing tasks 45 13 2169 .06 .02 -11 2.86 .0042 92.92* Individual testing 21 11 920 @15 .091 .22 4.54 .000006 42.99* Group testing 24 8 1249 @00 -.061 .05 -0.12 .904 37.35 Free-response Imaging tasksa 14 4 612 .20 .12 .28 4.82 .0000015 23.40* Individual testing 12 3 512 .20 .11 .29 4.46 .0000083 15.95 Group testing 2 1 100 .19 -.01 .37 1.83 .067 7.53* Notes. r is the weighted average correlation coefficient (Hedges & Olldn, 1985). k re resents the number of independent effect sizes. X2 is the within group homogeneity statistic (Rosenthal, 1984). ll displaying frequency distributions, the stem-and-leaf plot retains the numerical data precisely. (Turned on its side, the stem-and-leaf plot is a histogram.) Each number includes a stem and one or more leaves. For example, the stem.2 is followed by leaves of 0, 3,3,8 representing r's of .20,23,13, and .28. The meta-analysis is summarized in Table 1. The study grouping is identified in the first column. Col- umns two through four show, respectively, the num- ber of studies, investigators, and subjects. The mean weighted effect size (r) is shown in column five and columns six and seven show, respectively, the lower and upper 95% (CD for the effect size, The cumula- tive z-score and its associated two-tailed p-value are shown in columns eight and nine. The chi-square homogeneity statistic is-presented in the last column. Overctil Results The results for all 60 studies are shown in the first row. The . mean weighted r is 0.09 (z = 4.63, p = .00DO04, two -tailed). The 95% CI is an r from .05 to .12. While significant, the study effect sizes are nonhomogeneous (X259 = 126,21, P <.05). The second row shows the same analysis, omitting the 13 studies that were assigned r's of zero. Forced-choice St udles The ESP/extrave@pion correlations for the 45 forced-choice krt.-) s dies are presented in row three 1@ of Table 1. These studies were contributed by 13 independent investigators and include 2169 subjects. The mean weighted r 6.06 (z = 2.86, p = .0042,95% C1 from .02 to .11). The 2@ C correlations are significantly nonhornogeneous 92.82, p < .05). FC studies involvin@, individual testing. Twenty-one of the FC studies inv Ived individual testing (Table 1, row 4). These studi were performed by 11 inde- pendent investigators,and included 920 subjects. The mean weighted r is .1@ (z = 4.54, p =.000006,95% CI from .09 to -22). The correlations are nonhomogene- Ous (X@20 = 42.99, P < .05). i FC studies involving oup testing. The remaffiing24 FC studies involved cup testing. These studies were conducted by ei ht independent investigators and included 1249 s@bjects. They yield a mean weighted rof -0.00 (z -0.12,-p =.904,95%Cl from - .06 to .05). The group PC studies are homogeneous &23 = 37.35, p >.05). thus, the FC studies involving group testing yielded t@_formly null correlations be- tween ESP performanc and extraversion. Approved For- Release 2003/04/18: CIA-RDP96-00789RO031' 00030001-4 @%qffiyl FR Approved For Release dfe-ffDP96-007,39ROO3100030001-4 Individual versus group testing. Evidence for a rela- tionship between extraversion and forced-choice ESP performance is limited to FC studies involving indi- vidual testing. The difference between the ESP/ex- traversion correlations for individual and group test- ingis significant (Cohen'sq =.14, z = 3.47, p =.DD052, two-tailed). Free-response Studies Randomization. Thirteen FC studies satisfy our cri- teria for randomization. The mean weighted r for these studies is .13 (z = 2.27, p = .023). Thirty-two FC studies failed our randomization criteria; they yielded a mean weighted r of .05 (z = 2.13, p =.033, two-tailed). The ESP /extraversion correlation is non- significantly lower in these studies than in studies using formal randomization methods (Cohen's q .08, z = 1.40, p =.162, two-tailed). The results for the 14 free-response (FR) studies are shown in row six of Table 1. The FR studies Nvere contributed by four independent investigators and included 612 subjects. The mean weighted r is .20 (z = 4.82, p = ODDD015, two-tailed). The 95% Cl is an r from.12 to.28. The overall FR outcomes, while hi @Wy significant, are significantly nonhomogeneous tj 13 = 23.40, p <05). This nonhomogeneity is due to a mod- erating variable, test setting. Free-response studies involving individual testing. Twelve FR studies employed individual testing (Ta- ble 1, row 7), These studies were contributed by three independent investigators and include 512 subjects. The results are both significant and homogeneous. The mean weighted r is .20 (z = 4.46, p =.0000083, 95% CI from .11 to .29, X 2 11 = 15.85 p > .05). Free-response studies involving group testing. Only two FR studies involved group testing (Table 1, row 8). Both studies were contributed by the same inves- tigator. The mean weighted r is.19 (z = 1.83, p =.067, 95% Cl from -.01 to .37). The results are significantly nonhomogeneous 1 = 7.53, p < .05). Free-response versus Forced-choice Studies The mean correlation between ESP performance and extraversion is sigT-dficantly larger in studieswith free-response ESP tests than in those using forced- choice t ", (Cohen's q = .14, z = 3.11, p = .0019, tvvo- tailed). Quality Analysis of Forced-Choice Studies Sensory leakage. Thirty-one of the FC studies de- scribe the use of methods that satisfy our criteria for adequate control against sensory leakage in the ESP task. The ESP /extraversion relationship isnotsignifi- cant for these studies; the mean weighted r is .04 (z = 1.47, p = .142, two-tailed). The ESP/extraversion rela- tionship is significant in the 14 FC studies that are amenable to sensory leakage (r = .14, z = 3.20, p = .0014, two-tailed). The difference between the two correlations is significant (Cohen's q = -.10, z = -2.03, p = .042, two-tailed)., Duplicate recording, Six FC studies employed dupli- cate recording of targets and responses. The mean weighted rfor thesestudies is.31 (z =3.52,p =.00043, two-tailed). The 39 studies without duplicate record- ing yielded a mean weighted r of .05 (z = 2.01, p = .0444, two-tailed). The ESP /extraversion relationsWp is significantly stronger in studies with duplicate re- cording (Cohen's q =.27, z = 3.02, p =.0025). Duplicate checking. Eighteen FC studies reported using methods for duplicate checking of hits. These studies are associated with a mean weighted r of .27 (z = 5.72, p < 10,6, two-tailed). The 27 FC studies that did not use methods for duplicate checking yielded a nonsignificant r of.01 (z = 0.23, p =.59). The difference between the two correlations is significant (Cohen's q =.27, z = 5.20, p < 10-6 . two-tailed). A priori definition of extraversion. T'hirty-seven FC studies satisfied our criteria for a priori definition of extraversion. These studies had a mean weighted r of .05 (z = 2.20, p =.028). The remaining eight FC studies failed our criteria for a priori defirdtionof extraversion (r: = .10, z = 2.00, p = .046, two-tailed). The two correlations do not differ significantly (Cohen's q -.05, z = 0,92, p = .358, two-tailed). Order ofadministration ofextraversion and ESP meas- ures, Evidence for a relationship between forced- choice ESP performance and extraversion is entirely dependent upon the outcomes of studies in which extraversion was measured after the ESP test (N = 18 studies, r =17, z = 3.51, p = .00045). The con-elation between FC ESP performance and extraversion in studies in wNch extraversion was measured before the ESP task is not significant (N = 16 studies, T .02, z = -0.78, p =.782) and the difference between the two correlations is significant (Cohen's q = .19, z = 3.58, p = .00034, two-tailed). This difference is not attributable to methodological features of the two groups such as test setting, the extraversion measure, or ESP mode. Degree of ESP feedback is documented in nine of the 18 studies that measured extraversion after the ESP task and subjects received feedback of their ESP performance in each of these studies (r = .29, z = 4.59, Approved For Release 2003/04/18 : CIA-RDP96-0078'9ROO3100030001-4 6 EXTRAVERSION& ESP. A META -ANALYSIS& NEW C8116ffiRO 1 -4 Approved For Release 2003/04/18: CIA-RDP96 - Table 2. Forced-choice Outcomes by Invesfigator Investigator N studies N subjects r Z P Astrom. -? 1 48 .24 1.63 .103 Casper - A 1 20 .53 i 2.46 .014 Green 2 14B .00 0.00 .500 7 1 108 .00 0.00 so A 1 40 .00 0.00 .500 Humphrey 6 138 .26 2.86 .0042 ? 3 55 .27 I.B4 .0&% A 3 83 .22 1.95 .051 Kanthamani 7 301 .21 3.59 .00033 ? 1 60 .00 On .500 A 3 108 .38 1 4.03 .000056 B 3 133 .02 0.25 .400 McElroy -A 1 31 .00 0.00 .500 Nash 8 207 .14 1.92 .054 ? 2 60 .29 2.22 .026 A 6 147 .08 0.86 .390 Nielsen 3 60 -.04 -0.29 .771 A 2 48 .00 0.00 M0 B 1 12 -.23 -0.69 .755 Sargent 3 95 .17 1.46 .144 2 40 -.02 1-0.12 .548 M 1 45 .31 2.07 .038 Shields - 7 2 99 .30 2.98 .0029 Shrager - B 2 76 .18 1.48 '139 Szczygielski -A 1 17 -.36 1-1.42 .922 Thalboume - B 9 939 -.04 1-1.28 .90 Overall Forced-choice z by investigators 3.49, p =.00048, two-tailed, 1XI, =41.20,p<.05 12 Extraversion measured before ESP test (5 investigators): z by investigators = -0.71, p =1.761, tw(,tailed, X' 4= 3.97, p>.05 Extraversion measured after ESP test (8 investigators): z by investigators = 3,51, p =.00 Ib,45, two-tailed, X2 7 = 17.29, p <05 Notes. r is the weighted average correlation coefficient (Hedges & Olkin, 1985). The letters P and A following the investigator's name indicate whether the extraversion /introversion measure was administed before (@) or after (A) the ESP task, Studies where this information is not available are indicated by a question mark (?). Testing orde@ was mixed M in one study. X2 is the within group homogeneity statistic (Rosenthal, 1984). 1 P = .0000045, two-tailed). Seven of the nine studies where feedback is undocumented were group stud- ies which usually involve delayed feedback or no feedback at all. These studies yield a nonsignificant correlation between performance and extraversion (r = .05, z = 0.64, p =.522, two-tailed) which is signifi- cantly lower than that for the studies known to in- volve feedback (Cohen's q = .25, z = 2.54, p = .011, two-tailed). The relationship between forced-choice ESP performance and extraversion thus appears to be artifactual. Patterns over time. @Me forced-choice studies were reported between 194 and 1982. There is a significant decline in the magnitude of the ESP/extraversion relationship over this eriod(r=-A0,t=-2.89,43df, p=.006,two-tailed). oreover, methodological qual- ity, as assessed inte , s of threats to the validity of the ESP measure, has @not improved over the survey period (r = .01, f = 0.93, p = .976, two-tailed). These results are coWr3 the patterns found in meta- analyses of thr, t@r e I parapsychological domains, i@hicli exhibit constatit effect sizes and significant 3100030001-4 Approved ForRelease 2003104118: CIA-RDP96-00789ROO FERRARI & BEM Approved For Release MMA8: CIA-RDP96-00789ROO3100030001,4 'k methodological improvement over time (Honorton, 1985; Honorton & Ferrari, 1989; Radin & Nelson, 1989). There has been substantial improvement with regard to threats to the validity of the ESP/extraver- sion relationship; more recently reported studies have generally involved administration of the extrav- ersion measure before the ESP task (r = .78, f = 7.02, 32 df, p<10-@. These findings are consistent with the conclusion that the FC ESP /extraversion relationship is artifactual. FC Outcomes by Investigator The FC study outcomes by investigator are shown in Table 2. The order of ESP and extraversion testing is indicated following the investigator's name for investigators with single studies or multiple studies involving uniform testing order. Separate break- downs are given for investigators with studies in- volving different testing orders. Significant outcomes were obtained by four of the 13 FC investigators (31%); using the investigator as the unit of analysis, the overall results are significant (z = 3.49, p =.00048, two-tailed) but nonhomogeneous (X2 12 = 41.20, p < .05). The effect of testing order accounts for the overall significance and nonhomogeneity. The five investigators who measured extraversion before the ESP task have outcomes that are nonsignificant 2 and homogeneous (z = -0.71, p = .761, two-tailed, X 4 = 3.97, p > .05), while the outcomes of the eight investi- gators who measured extraversion after the ESP task are significant and nonhomogeneous (z = 3,51, p = .00045, two-tailed, X 27 = 17.29, p < .05). Thus, the impact of F-SP/extraversim testing order is consis- tent across investigators and is not attributable to idiosyncratic research styles or other characteristics of a single prolific investigator. Quality Analysis of Free-response Studies Sensory leakage. All 14 FR studies satisfied our criteria for adequacy of control against sensory leak- age. Randomization. Nine FR studies satisfied our crite- ria for randomization (r = .38, z = 4.74, p = OD00022, two-tailed). Five FR'studies employed informal ran- domization procedures or failed to document their method of randomization (r =14, z = 2.86, p =.0042, two-tailed). The difference between the two correla- tions is significant (Cohens q = .26, z = 2.86, p = .0042, two-tailed). Duplicate recording. Thirteen of the 14 FR studies employed duplicate recording methods (r =.29, z = 4.62, p =.0000039, two-tailed). Duplicate ched:ing. All 14 FR studies employed du- plicate checking methods. A priori definition of extraversion. Eleven FR studies satisfied our criteria for a priori definitionof extraver- sion (r = .16, z = 3.72, p = .0002, two-tailed). Three studies failed to document their basis of classification (r =.48, z = 4.01, p =.000061, two-tailed). The two cor- relations differ significantly (Cohen's q =.36, z = 2.77, p = 0.0056, two-tailed). Order of administration of exf raversion and ESP meas- ures. The extraversion scale was administered before the ESP task in 11 of the FR studies (r = .21, z = 4.57, p = .000005, two-tailed). The remaining three FR studies failed to report the order inwhich the ESP and extraversion measures were given (r =15, z = 1.64, p = .101, two-tailed). The difference between the two correlations is not significant (Cohen's q = .06, z = 0.62, p = .532). Thus, for the free-response studies, the evidence fora relationship between ESP perform- ance and extraversion is not susceptible to explana- tion in terms of an order artifact. Patterns over time. The free-response studies were reported between 1960 and 1982. Unlike the forced- choice studies, the magnitude of the ESP/extraver- sion relationship has increased over time, though not significantly so (r = .19, t',= 0.66, 12 df, p = .524, two- tailed). The methodological quality of the free-re- sponse studies has also improved over time (r =.36, t = 1.35,12 df, p =.202, two-tailed). Confirmation of Differences in Research Quality in relation to Test Setting Honorton & Ferrari (1989), in a meta-analysis of forced-choice precognition experiments, found that studies involving individual testing were of signifi- cantly higher methodological quality than studies involving group testing(t = 3.08/137 dfp =.003, two- tailed). We have confirmed this finding in the ESP/ extraversion meta-analysis (t = 2.27, 39 df, p = .015, one-tailed). This analysis excludes the 12 precognition studies which overlap with the earlier meta-analysis. Since we have determined that there is no ESP/ex- traversion relationship in the forced-choice studies when the effects of task order are considered, the remaining analyses are restricted to the free-response studies. Approved For Release 2003/04/18: CIA-RDP96-00789ROO3100030001-4 EXTRAVERSION & ESP. A META-ANALYSIS & NEW CON@IRMATION Approved For Release 2003104118: CIA-RDP96-00789ROO3~,100030001-4 Table 3. FreG-response Outcomes by Investigator Investigator N studies N subjects r i Z P Bellis & Morris 1 23 .47 2.26 .0-14 Braud 2 100 .19 1.86 .063 Marsh 1 311 .13 2.30 .00-1 Sargent 10 178 .31 3.95 .00012 z by investigators 5.11, p 3 x 10-7 1 two-tailed X 2 2.51, p > .05 3 Notes. r is the weighted average correlation coefficient (Hedges & Olldn, 1995). X is thl'o within group homogeneity statistic (Rosenthal, 19M). Consistency across investigators Table 3 shows the overall PR results by investiga- tor. Three of the four investigators have significant ESP/extraversion correlations and the results of the fourth investigator (Braud) approach si0ficance. The z by investigator is 5.11, a result that should arise by chance less than one time in3.3 million. The results are homogeneous across investigators 3 = 2.51, p >.05). Although 10 of the 14 FR studies were contrib- uted by one investigator (Sargent), evidence for the relationship between free-response ESP performance and extraversion is not dependent upon that investi- gator. When Sargent's work is eliminated, the results of the three remaining investigators still strongly sup- ports a relationship between ESP performance and extraversion (z = 3.35, p = 0.0008, two-tailed). There- fore, we conclude that the ESP/extraversion relation- ship is consistent across investigators. Extraversion Measures Each FR investigator used a different scale for measuring extraversion. Marsh used the Bernreuter Personality Inventory (Super, 1942); Sargent and his group used the Cattell 16PF (Cattell,Eber & Tatsuoka, 1970); Braud and Bellis & Morris used scales con- structed by the investigators (with no psychometric validation provided). It is impossible to isolate the effects of the instruments for measuring extraversion from the ensemble of procedures and research styles associated with the ffivestigators. All that can be said is that a relationship between extraversion and ESP performance is evident in studies using four different measures of extraversion. Selective In order to assess!j to selective reportii 'Tail-safe N" statisti ported studies avera, reduce the known d Fail-safe N is 140 stu to assume that the selective reporting, it 10 unreported studi each reported study. free-response ESP/e be explained on the I :ie vulnerability of these studies @, we used Rosenthal's (1994) to estimate the number of unre- @ing null outcomes necessary to ata base to nonsigificance. The lies. In other words, if we were )bserved outcomes arise from would be necessary to postulate @s averaging nul 'I outcomes for 17herefore, we conclude that the (traversion relationship cannot @asis of selective reporting. PowerAnalysis The PR mean r of 20 is equivalent to an average ESP scoring advanta0e for extravekts over introverts of 0.4 standard devi4tions. The PR studies average sample size is 44 subj t1s; and the likelihood of detect- ing a correlation of .2 at the five percent significance level with d-ds sample size-the statistical power--is @s i 37 percent (Cohen, 19 7, p. 87). Thus, in a sample of t@7 ' 14 studies, the expec number of statistically sig- ed nificant studies is 5.2; e actual number of significant studies is seven e a t binomial probability, with p = .37 & q = .63 = , one-tailed), Thus, the ob- served rate of signi c t outcomes is consistent with a correlation of .2. Achievement of sta@,tistical significance, assuming a correlation of .2, is' essentially a coin toss with sample sizes less than subjects; a sample size of 180 is necessary to ad-dev , 85 percent power. In the following sec "on, we explore the predictive validity of the ESP/e traversion meta-analysis by comparing the meta-a lytic estimate to the outcome i of a new data set. 3' 00030001-4 Approved For- Release 2003/04/18: CIA-RDP96-00789ROO 11. HONORTON, FERRARIA BEM Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Table 4. ESP/Extravwslon Correlations by ExpGrimenter In the PRL Novice Series Experimenter N Subjects Experimenter El Score Honorton 41 .27 1.71 101 Quant 69 .29 2.38 103 Derr 22 .03 0.68 81 Berger 13 -37 -1.18 115 Va@rvoglis 21 .08 0.32 133 Schechter 7 -.05 -0.10 125 Ferrari 10 -.20 -0.54 133 Schlitz 7 .15 0.92 69 Note. r is the weighted average correlation coefficient (Hedges & Olldn, 19B5). A New Confirmation Extraversion data is available for 221 of the 241 subjects in a series of ESP ganzfeld studies reported by Honorton, Berger, Varvoglis, Quant, Derr, Han- sen, Schechter & Ferrari (1990) and conducted at the Psychophysical Research Laboratories (PRL) in Princeton, N.J. The experimental procedures are de- scribed in detail in the Honorton, et al. (1990) report. extraversion.) The mean EI &core for the PRL subjects is 100.36 (sd = 25,18). ESP Measure ESP performance was measured using the stand- ardized ratings of the target and decoys (Stanford's z-scores; Stanford and Sargent, 1983). Stanford Ys were averaged for subjects with multiple sessions. SubJects The subjects were 131 women and 90 men. Their average age is 37 years (sd = 11.7). Ms is a well-edu- cated group; the mean formal education is 15.5 years (5d = 2.0) and belief in psi is strong in this population. On a seven-point scale where "1" indicates strong disbelief and '7" indicates strong belief in psi, the mean is 6.20 (sd = 1.03). Personal experiences sugges- tive of psi were reported by 88 percentof the subjects; eighty percent reported ostensible telepathic experi- ences. Eighty percent have had some training in meditation or other techniques involving internal fo- cus of attention. One hundred and sixty-three sub- jects contributed a single ESP ganzfeld session and 58 contributed multiple sessions. Extraversion Measure Extraversion was measured using the continuous scores of the Extraversion/Introversion (ED Scale in Form F of the Myers-Briggs Type Indicator (MBTI; Briggs & Myers, 1957). The MBTI was not used in any of the meta-analysis studies. The MBT1 El Scale is constructed so that scores below 100 indicate extrav- ersion and scores above 100 indicate introversion. (For consistency with the meta-analysis, we have reversed the sips so that positive correlations reflect a positive relationship between ESP performance and 11 Results Overall results. The correlation between ESP per- formance and extraversion in the PRL series is signifi- cant (r =.18,219 df, t = 2.67, p =.008, two-tailed, 95% Cl from .05 to,.130). This outcome is very close to the meta -analytic"estimate for free-response studies (r =.20) and the difference between the two correla- tions is nonsignificant (Cohen's .q = .02, z = -0.26, p =.793, two-tailed). Ganzfeld Novices. The results are similar if we re- strict our analysis to the five PRL Novice series with inexperienced subjects who each completed a single ganzfeld session. MBT1 data is available for 190 of the 205 Novices and the mean weighted r for the five series is.17 (z = 2,25, p =.024, two-tailed, 95% Cl from .02 to .31). The ESP/extraversion correlations are ho- mogeneous across the five series (X 24 = 2.88, p > .05). Eleven subjects in the first Novice series (Series 101) completed the MBTI between six and eighteen months after their ESP ganzfeld session and we did not maintain records of their identity. However, the results are essentially the same when this series is eliminated. The mean weighted r for the remaining four Novice series is.19 (z = 2.30, p = .021, two-tailed, 95% Cl from .03 to .34). Outcome by experimenter, Eight experimenters con- tributed to the PRL data base (Honorton, et al., 1990). Table 4 shows the ESP/extraversion correlation by Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 10 EXTRAVERSION& ESP; A META-ANALYSIS A NEW C04 IRMATION r Release 2003/04/18: CIA-RDP96-00789RO0340030001-4 Approved Fo experimenter for the five Novice series. The mean weighted r for the eight experimenters is. 16 (z = 2.09, p = .037, two-tailed, 95% CI from .01 to .30). The results are homogeneous across the eight experi- (X2 menters 7 = 6.43, p > .05). Outcorne in relation to El status of experimenter. It is possible that the rela tionsl-dp between ESP perform- ance and extraversion is moderated by personality characteristics of the experimenter. The last column of Table 4 shows the MBTI El scores for each experi- menter. 01 s (Derr and Scl-ditz) -Ily two experimenter are extraverts. Two others (Honorton and Quant) are borderline introverts. While the above analyses indi- cate that the ESP /extraversion correlation is consis- tent across experimenters, there is a nonsignificant tendency for the relationship to be stronger in the data of less introverted experimenters (r = .47, 6 df, p .235, two-tailed). Combined Estimate of the Relationship between Free-response ESP Performance and Extraversion Combining the new confirmation with the meta- analysis, the overall mean weighted r is .19 (z = 5.50, p = 3.8 X 10, 95% CI from.13 to.26). The "Fail-safe JV'for the combined estimate is 181 studies, or a ratio of 12 unreported studies averaging null effects for each known study. Four of the five investigators have overall significant outcomes and the outcomes are homogeneous across investigators X 4 = 6.03, p > .05). Discussion The Meta-Analysis Forced-choice studies. The meta-analysis challenges the conclusions from earlier narrative reviews of the relationship between extraversion and forced-choice ESP performance (Eysenck, 1967; Palmer, 1977; Sar-@ gent, 1981). The apparent relationship between ex- traversion and ESP performance in these studies ap pears to be due to the influence of subjects' knowl- edge of their ESP performance on their subsequent responses to the extraversion measures. Evidence for a relationship betN@reen ESP and extraversion occurs only when extraversion was measured after the ESP test; no evidence of an ESP/extraversion relationship is found in studies where extraversionwas measured before the ESP task, Evidence for a OnZerO effect in the forced-choice studies is also limi ing ESP tl@ ted to the subset of studies involv- sting pAxedures that were vulnerable to potential sensory Lakage. There is reason to believe, I however, that this'play result from a procedural con- found: six theeight@studies in this subgroup forwhich inforrriation on th@ order of testing is available also involved extraver@ion testing following ESP feed- back. The apparent bi. sing effect of ESP feedback prob- ably, arises from o' e of two possibilities. Awareness of ,success,' or 11 ilure" may lead subjects to later perceive themselv as more extraverted or intro- verted. Or, the poblem may arise from an experi- menter expectancy @effect (Rosenthal & Rubin, 1978), inwhichsubjects re@ond to the investigator's expec- tations that extravtrts are more successful in ESP tasks than introver Obviously, further resea rch will be necessary to cla the problem. The existence of is problem, however, necessar- ily arouses conce over the viability of mported. relationships betw ESP performance and other personality factorsi such as neuroticism (Palmer, 1977). Much of the Jesearch in these areas was con- ducted by the same estigators, and it is likely that similar methods we e used. We believe that conclu- .@e V re sions regarding the relationship between ESP per- formance and other personality variables should be Pe suspended until the' relevant study domains can be examined with res to d-ds problem. Free-response studies. The meta-analysis does sup- port the existence of @ relationship between extraver- sion and free-respo ESP performance. The free-re- @ ot amenable to explanation in sponse studies; to terms of an order ar*act or other identifiable threats to validity. The ove@all correlation of .20 would be expected to occur or@y about one time in 674,000 by chance. Three of the four investigators contributing to Us data base obt@ined significant ESP/extraver- sion relationships, and the fourth investigator's re- sults approach sign@i cance. The correlations are ho- mogeneous across in*estigators, and across the larg- est grouping of studi in which subjects were tested individually. The ta remains highly significant even when 71 percent of the studies, contributed by one investigator, are @liminated from consideration. s' Thus, the relationshi seems to be robust. Estimation of the filedrawer prob em (Rosenthal, 1984), indicates ,@e thatitwouldbeneces ryto postulate 10 unreported studies averaging n@ll results for every retrieved study in order to account for the observed effect on the basis of selective r6ortiniz. Approv Ied For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 HONORTON, FERRARI & BEM I I Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 The Now Confirmatfon The results of the confirmation, involving a new set of investigators and a new scale of extraversion, support the meta-analytic findings and increase their generalizability. The relationship between free-re- sponse ESP performance and extraversionnow spans 833 subjects and five independent investigator teams. The homogeneity of the effect across the eight experi- menters in the confirmatory study further increases our confidence that the effect is replicable and is not dependent upon unknown characteristics of individ- ual investigators. A nonsignificant trend in the data does suggest that the ESP/extraversion relationsl-dp may, to some extent, be moderated by the experi- menter's extravertedness and it may be advisable for future investigators to record and report extraver- sion /introversion scores of the experimenters. The Predictive Volidity of Meta-Analysis Meta-analysis is a powerful tool for summarizing existing evidence. It enables more precise estimation of the significance and magnitude of behavioral ef- fects thm has been possible with traditional narrative reviews, and is useful inidentifying moderatingvari- ables. In the present case, meta-analytic techniques revealed a serious source of bias that had been over- looked in earlier narrative reviews of the ESP /extrav- ersion domain. Moreover, the meta-analysis identi- fied a subset of the domain that is not amenable to the discovered bias and provided an estimate of the mag- nitude of the relationship between ESP and extraver- sion in that subset. Ultimately, the usefulness of meta-analysis will be judged by its ability to predict new outcomes and in this regard we consider the results of the confirma- tion study to be especially noteworthy. The correla- tion between ESP performance and extraversion in the confirmation study is very close to that predicted by the meta-analysis. TI-ds is the second test of the predictive validity of meta-analysis in parapsy- chological problem areas; we have previously re- ported that ESP ganzfeld performance in a new series of studies (Honorton, et al., 1990), closely matched the outcomes of earlier studies in a meta-analysis (Honorton, 1985). Predictability is the hallmark of successful science a *nd these findings lead us to be optimistic concerning the prospect that parapsychol- ogy may be approaching d-tis more advanced stage of development. References Briggs, K. C., and Myers, L B. (1957). Myers-Briggs Type Indicator Fonn F. Palo Alto, CA: Consulting Psychologists Press, Inc. Cattell, R. B., Eber, H. W., & Tatsuoka, M. M. (1970). Handbookfor the Sixteen Personality Factor Question- naire. Champaign, FL: Institute for Personality and Ability Testing. Cohen, J. (1977). Statistical power analysisfor the behav- ioral sciences. New York: Academic Press. (Re- vised Edition.) Eysenck, H. J. (1967). Personality and extra-sensory perception. Journal of the Societyfor Psychical Re- search, 44,55-70. Hedges, L. V., & Olkin, L (1985). Statistical methodsfor Meta-Analysis. New York: Academic Press. Hedges, L. V. (1987). How hard is hard science, how soft is soft science? The empirical cumulativeness of research. American Psychologist, 42, 443-455. Honorton, C. (1985). Meta-analysis of psi ganzfeld research: a response to Hyman. Journal ofParapsy- chology, 49, 51-92. Honortor@ C., Berger, R. E., Varvoglis, M. P., Quant, M., Derr, P., Hansen, G., Schechter, E. I., and Ferrari, D. C. (1990). Psi communication in the ganzfeld: experiments with an automated testing system and a comparison with a meta-analysis of earlier studies. In Research in Parapsychology 1989. Metuchen, NJ: Scarecrow Press. (In press.) Honorton, C., & Ferrari, D.C. (1989) "Future Telling": a meta-analysis of forced-choice precognition ex- periments, 1935-198@.' Journal of Parapsychology, 53, in press. Hyman, R. (1985). The psi ganzfeld experiment: A critical appraisal. Journal of Paiapsychology, 49, 3- 49. McCarthy, D., & Schechter, E. 1. (1986). Estimating effect size from critical ratios. In D. H. Weiner & D. I. Radin (Eds.) Research in Parapsychology 1985. Scarecrow Press, pp. 95-96. Palmer, J. (1977). Attitudes and personality traits in experimental ESP research. In B. B. Wolman (Ed.) Handbook of parapsychology. New York: Van Nos- trand Reinhold. Palmer, J., & Lieberman, R. (1975). The influence of psychological set on ESP and out-of-the-body ex- periences. Journal ofthe American Societyfor Psychi- cal Research, 69,193-213. Radin, D. I., & Nelson, R. D. (1989). Evidence for consciousness-related anomalies in random physical systems. Foundations ofPhysics, 19,1499- 1514. Rosenthal, R. (1984). Meta-Analytic proceduresfor social research. Beverly Hills, CA: Sage. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001,4 12 OffRAVERSION & ESP. A META-ANALYSIS & NEW C04RMATION Approved For Release 2003/04/18: CIA-RDP96-00789RO03@100030001-4 Rosenthal, R., & Rubin, D. B. (1978). Interpersonal expectancy effects: The first 345 studies. Behavioral and Brain Sciences, 3,377-386. Sargent, C. L. (1981). Extraversion and perform- ance inextra-sensory perception' tasks. Personal- ity and Individual Differences, 2, 137-143. Stanford, R. G., and Sargent, C. L. (1983). Z scores in free-response methodology: comments on their utility and correction of an error. Journal of the American Societyfor Psychical Research, 77,319-326. Super, D. E. (1942). The Berareuter Personality Inven- tory: a review of research. Psychological Bulletin, 39, 94-125@ Tukey, J. W. (1977). Exploratory data analysis. Reading, MA: Addison-Wesley. I Studies Used in the Meta-Analysis Some reports contain more than one study. For reports with multiple studies, the number of studies is indicated in brackets following the reference. Ashton, H. T., Dear, P. R., & Harley, T. A. (1981). A four-subject study of psi in the ganzfeld. journal of the Societyfor Psychical Research, 51, 12-21. Astrom, J. (1965), GESP and the MPI measures, jou, nal ofParapsychology, 29,292-293. Bellis, J., & Morris, R. L. (1980). Opennessi, closedness and psi. Research in Parapsychology 1,979, 98-99. Braud, L. W. (1976). Openness versus closedness; and its relationship to psi. Research in Parapsychology 1975, 155-159. Braud, L. W. (1977). Openness vs. closedness and its relationship to psi. Research in Parapsychology 1976, 162-165. Casper. G. W. (1952). Effects of the receiver's attitude toward the sender in ESP tests, Journal o fParapsy chology, 16,212-218. Fisk, G. W (1960). The Rhodes experiment. Linkage in extra-sensory perception by M. C. Marsh. jour- nal of the Societyfor Psychical Research, 40, 219-239 and M. C. Marsh. (unpublished). Linlwge in Extra- Sensory Perception. Unpublished doctoral disser- tatiom Dept. of Psychology, Rhodes University, Grahamstown, South Africa. 450 pages. Green, C. E. (1966). Extra-sensory perception and the Maudsley Personality Inventory. JouTnal ofthe So- cietyfor Psychical Research, 43, 285-286. Green, C. E. (1966)., Extra-sensory perception and the extraversion scale of the Maudsley Personality Inventory. Journal of the Societyfor Psychical Re- search, 43,337. Haraldsson, E. (1070). Psychological variables in a GESP test usin4 plethysmograph recordings. Pro- ceedings of the Parapsychological Association. 7,6-7. i Harley T A & S;gent, C. L. (1980). Trait and state fac@ors irIg ESP performance in the gan- zfeld. Research in Parapsycliology 1979, 126-127. Humphrey, B. M. An exploratory correlation study of perso@ality measures and ESP scores. Journal of Parapsychology, 9, 116-123. [3 studies) Humphrey, B. M. 1(1951). Introversion-extraversion ratings in relation to scores in ESP tests. Journal of Parapsychology, @5, 252-262. Kanthamand, B. K. I i(1966). ESP and social stimulus. Journal ofParaps chology, 30,31-38. Kanthamani, B. K.,,& Rao, K. R. (1972). Personality characteristics of ESP subjects: 111. Extraversion an . Journall of Parapsychology, 36, 198-212, Krishna, S. R., & R o, K. R. (1981). Personality and 'beliefinrelatio to language ESP scores. Research inParapsycholo 1980, 61-63. [2studies] McElroy, W. A., an Bro , W. K. R. (1950 Electric shocks for errors! in P card tests. journal (@Para- psychology, 14, 7-266. Nash, C. B. (1966). Re@lation between ESP scoring level and the Minnes@ta Multiphasic Personality In- ventory. Journal ofthe American SocietyforPsychical Research, 60, 56-6@. [8 studies) Nicol, J. F., & Hump@rey, B. M. (1953). The explora- tion of ESP and 'uman personality. Journal of the An @ety lericart Socr r Psychical Research, 47,133-178. Nicol,J.F., &Hump eyB.M.(1955).Therepeatabil- ity problem in ESP-personality research. Journal of the American Soci4yfor Psychical Research, 49,125- 156. Nielsen, W. (1970). R@'lationships between precogni- tion scoring levell and mood, Journal of Parapsy- chology, 34,93-116', 1 Nielsen, W. (1970). St@dies in group targets: a social sychology class. P roceedings of the Parapsychologi- cal Association, 7,55-57. Nielsen, W. (1970). Studies in group targets: an un- usual high school group. Proceedings of the Para- psychological Ass tion, 7, 57-58. Sargent, C. L. (1978). Hypnosis as a psi-conducive state: a controll replication study, Journal of Parapsychology, 42, 1257-275. [2 studies] Sargent, C.L.(1980). E loringpsiintheganzfeld.New 16 York: Parapsychol' Foundation, Inc. [2 stud- ies] 7, Sargent, C. L, Bartlett H.. J., and Moss, S. P, (1982). Response structur and temporal incline in gan- zfeld free-respons GESP testing. Journal of Para- psychology, 46, 85-110. [2 stud les] Sargent, C. L, Harley,T,. A., Lane, J., & Radcliffe, K. (1981). Ganzfeld -optimization in relation to session duration. Rtl earch in Parapsychology 1980, 82-84. Approved For Release 2003/04118 CIA-RDP96-00789R003100030001-4 HONORTON, FERRARI & BEM 13 Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 Sargent, C. L., & Harley, T. A. (1981). Three studies using a psi-predictive trait variable question- naire. Journal of ParapsycholoM 45,199-214. Sargent, C. L., and Matthews, G. (1982). Ganzfeld GESP performance with variable-duration test- ing. Research in Parapsychology 1981, 159-160. Shields, E. (1962), Comparison of children's guessing ability (ESP) with personality characteristics. Journal ofParapsychology, 26, 200-210. [2 studies) Shrager, E. F. (1978). The effects of sender-receiver relationship and associated personality variables on ESP scores. Journal of the American Societyfor Psyd6cal Research, 72, 35-47. [2 studies] Szczygielski, D., & Schmeidler, G. R. (1975). ESP and two measures of introversion. Research in Parapsy- diology 1974, 15-17. Thalbourne, M. A., Beloff, J,, and Delanoy, D. (1982). A test for the 'extraverted sheep versus intro- verted goats' hypothesis. Research in Parapsychol- ogy 1981, 155-156. [2 studies] Thalbourne, M. A., Beloff, J., Delanoy, D., & Jung- kuntz, J. H. (1983). Some further tests of the ex- traverted sheep versus introverted goats hypothesis. Research iii Parapsycliology 1982, 199- 200. [4 studies] Thalbourne, M. A., and Jungkuntz, J. H. (1983). Ex- traverted sheep versus introverted goats: experi- ments VII and VIII. Journal of Parapsychology, 47, 49-51. [2 studies] Approved For RLilease 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 M 00 00 N 0 LL > 0 < ,jetpu3swd ljzqct@[ PUL SpicH ]:-.)Illo]V Xq tit si irip sx)u;)pq jo (tuDpr3V jrIJc)l)uN ;)'P Xq P110ISS'"I -tIlOD v III p.,)l3liplil Z)Cl Ilpt )tll jo 13011m.-md.lmit Put! klealums :)III jo 11311IN -11 leplilloj I'MOPUN --)'I) @cl P:)l.lOcId s 11 0! n * 11111lum p(II1:)S;)p s;)-mp;):)wcI. znp jo.- owos jo matudo[Z)ADI) alil put: -i;)ded sitp jo imlltl@d;ud a*Lll o I, uoJ9 ILIDIj,)DXQ QAPLI (VAl ll'Pw,@H I)-'III Luz' I -sw-I I UzU I Pu , I -OD LIZ WDLII Po-IQ,%`IsUL' SLq U0110UOH -suoilsonb Buillal PUL, lu,)20:) psfr,l ml Limu'(1-1 .111m (QIII irill xjo@@!j 11 sl @Ictli sQ:)TI?.IojJ!P .113til 1[1! j),)AjOS3-I IOLI DAELI 1101-IOU01-1 I)tW LICLUAI-I 'PLI3 Dtil [11 ppjztzi3B 3ill jo Upwwns DAlImpurnb ojq-enIr,k AlsnoUl-TOUD LIP. oAeLl DA '-4-10M DIJAILlUE-lelOW JIoIIj L1.5noup -uopippe ul pf.Dj _z(M.9, )([I Xq pajmm(m2, mrp )III .9mitirnir,w) uf po-iop!sum Qq oi qztnssi QqI JO MOfAQJ JU;)IIQDX,? ULI DAuq om !;mqjjo jpqi jo ipiso-i I! sV I*siuiod Buijoz)s, cti XIdwis ueLp i;)qi-ez sons -Sl DYIILID[:)S OtIl .2UfPUOIS-IQpLIn PUU 2LIIXJjJUID 01 poloA@p uaoq D.@xq SluEdpri-ied 3qi ieqi icop si il -Q1120JUIP QlApmaisuw lpiuoijo-j-sm jo lopout i@ si iap oiEqop -e ompoad oi iLl.3noLp QAIIEZ)J:) PLIE -Uo mil jo jullotuL, IL'QJS U P@@ISDAUI DAUll siqoqDs olgeo2polmoul X11J."3111 ptw snopos OA%J. -OE-@j MILISpIpuileW *@j 'joiipo sjuiunof oqi @q p wILM11.10 SI 1)[11! (@861) LIOI.IOLIOI-j ISDIAILID N861) LII!LLIXH ('!@J uzoomioq DiLqop PIDJZU-e2 JUDDD.1 oy Xq PDAJDS Ilom uD3q DAEll -ii,,InDp.n?d III (-iinbui j-eDT.2oIotp.-(sci-eird pim I-e.iDLio2 Ui x)UDpS -uorV!:)ifdai jo sS4x)nS,3tp jo StADIA -lzmau jo asnlapp-10j sl1w) uorsnpuw atfj -PD.I,)JJO S! -)DII-DlC@)U!aA!)I!)UD) u pur lp;)ss,,:).,;!p a.,.e szdX) pt,E jrinpawd JO %3SOII)OCUtl JUA!H IP!Sl0AO.IjIlOD L jo t! tilpA paird -tuo:) puic lpalunlEA;l lpazpinuums si)pwi at -Spajja IIEUIS 30 @;uxajqoid pim 43tiopuaildai 2utlif2tam Isaipnis Surppiyqns Iguilsal oldilinEu jo stual -qojd %uptipti! passn2ls!p Dic uo!)L:)!jda-x o) paIL13.1 swiss! 3 Dads Inj -asn ajotu u qjpA zAz!A xaplo ;nuosa[qnorj aql 2uiqsrn.2upsrp luorluarldat: jo idazkuo:) aill uo snzioj I kneju;nuwoz).@w ul -p;)AioAu! s;mssi ?ql UDCLEqs o'l Xjqwitupr pakaas lEql 4n2ole!p paluapo-4sul puE lampnilsuoa lampeax) u sEm ajuqap.I!; .)tlj, -uopEnj!s pjajzut;2 itil u! !sd Jqj Dz)uap!Aa aq') inoqr (qq6j)' UVWXH XLT Puu (SR61) UOI-IOU014 sal-iuTj:) uaami;)q z)luqap @Iqrnjt'A @lp UO kimuattituoz) u st: aadrd stT41 :IDVdLSqV IVHJ.N-,ISO-d I.N-490-d Ag :qjyqH(l (j-jj_j7NVq J,Hl',N,'Ol-LVDT'lJ71N JO HWfIJVNT 111-1J, (TNV SHWICEADONd l)LTA-lvNv-vJ,-,lfW fi) jull-illol, lq Cl) C4) 00 9 CD 00 0 C14 4) 16- 0 U- > p < 318 77tc Journal of'l-'arapsychology study failed to replicate that of Smith. Such errors are made very fr-e@uently in most areas of psychology and the other behavioral scl- crices. >Pseudo-Successful Replications 0 Return now to Table I and focus attention oil cell B, tile cell of' <1. Msuccessful i-eplicaLioll." Suppose that two investigators b0dl rejected CL -n the null hypothesis at p < .05 with both results in the same direc- 0tion. Suppose Further, howevcr, tlll( ill olle study [lie cf'('C(:( %Ize P* N was .90 whereas in the other study the effect size r was only 10, signCficantly smaller than the 7- oF .90 (Rosenthal ,;@ lZubin. 1082a). Ill this CiISC 0LJY IILCYJA-CUltioll is more complex. We have I lictced U) Mhad a successful replication of tile rejection of the null, but we have K) 0not come even close to a successful replication of the effect size. 0 40 "Successful Replication" of Type H Error Cell C of Table I represents the situation in which both studies 0 failed to reject the null llyp0tllCSi.%. Under those coild i nons- invcsn- > gators might conclude that there was no relationship between (lie @Ovariables investigated. Such ZI Conclusion could be very Much Ill er- -01-or, the more So tile lower tile power of' the two studies was I OW (Cohen, 1977). If power levels of the two studies (assuming mediurn .effect sizes in the population) were very high, say .90 or .95, then ;@wo failures to obtain a significant relationship would provide evi- ence that the effect investigated was not likely to be a very large V cl@ffect. If power calculations had been made assu ming a v si-nal. -a- ding -M] Ith n u a t ough not provi idence for the null would ;it least suggest that the size of 8trong ev 9.he effect in the population was probably clulte modest. If sample Sizes of the two StUdies failing to reject the null were Soclest so that power to detect all but the largest effects were low, little could be concluded from two failures to reject except that example', the effect sizes were unlikely to be enormous. For two illm vestiga(ors with N's of 20 and 10, rcspectively, find results not sig- nifiCallL ;It 1) < .05. The effeCt. .51:7.CS P11i (i.e., 7- for ClIC1101.0111011s Var- iables) were .29 and .20, respectively. and boili p's were ;ipproxim.itch- .20. 'I'll(- collibilwil P of" illesc two resillis, llowever, is .035[(zi + z,,)/V7 = z], and the mean effect size in the mid-20's is not trivl;lf (Rosenthal &- lZuNn, 19821)). Gan@leld Debalr-Rosenlhal 319 TABLE 3 CONIPARISON OF TWO SETS OF REPLICATIONS Repficatioll sets A N P (two-tallc(l) Z (1)) Colien's q (z, z,) Coniparing Views of Replication -n 0 M 2) U) M The traditional, not very useful, view of replication modeled in Table I has two primary ch aracteris tics: 1. It focuses on significance level as the relevant sumillary statis- tic of a study. C) it makes its evalization of' Whether replication has becil $uc- -0 cessful ill ;I dichotomous f',ishioll. For exilliple, replicatiolls are suc- cessful if' both oi neither p < .05 (or .01, etc.), @illd they are unsuc- cessi'lil if' one 1) < .05 (or .01, etc.) ;ill([ tile oLher P > .05 (or .01, 0 etc.). Psychologists' reliance on a dichotomous decision procedure co accompan ed by an untenable discontinuity of' credib I llity in rQsults varying in 1) levels lias been Well ClOCUlnellLed (Nelson, Rosenthal, & 196 Rosnow, 1986; Rosenthal & Gaito, 1963, 4)m --- - --- -re-P@ ix--wer-,--mcF@e -uyefu r-%,ReFv%@--OT has two pri- ication success acter mary char, stics: i -e Important summary sta- 0 i 1 1. It focuses oil effect size as the moi . 0 tistic of'a study With only a I-ClMlvely 11111101- lilLereSL ill tile statistical significance level. 2. It makes its evaluation of' whether replication has been suc- cessful in a continuous fashion. For example, two studies are not said to be successful or unsuccessful replicates of each other but, 1-;lJJlCy, file c!eg,.-cc ()f '."Ijiltii@e io I-CphCaLe is .1;J)CCll1Cd. 3 shows t.wo sets of replications. 1@epllcatioii sCL A shows two 1-csillis both r(Jecting (lit' 111111 klit with a diff,crelicc ill cf'f,e(:( sizes of J0 lit tillits of I. ()1- .:15 Ill milts of" Fisher's Z 11-iltiSfol-illitLioll of' 7- (Cohen, 1977; Rosenthal & Rasnow, 1984; Snedecor &'Coch- raii, I @)80), 'rilat diff*creticc, ill units of' r or Flslicl-'s z i-S tile (Icgrce B StUdy I Study 2 Study I Study 2 915 15 98 27 .05 .05 .01 1.96 1.96 2.58 1.34 .20 .50 .26 .20 .55 .27 .27 .35 .00 322 77heJournal qJ'ParapsycholqAry Gan--it'l-lal Delwie-Rosenthal 323 111111tiple (111cstions, intiltiple do.-pendcm variables make good scien- tific sense. However, as both Honorton (1985) and Hyman (1985) -goint out, the use of InUlLiple de))eIldCIlL VilriablCs 111,iy aff'eCt (lie 1@curacy of' the p levels comptited. For cxample, 11' live dependent Rariables are used and one of these is found to show an effect at p .05, it would be misleading to say that an effect has been den-l- rnist.rated at p < .05. That, is because ill(- actiml p of' finding 0111, /5 9gnificant at .05 (or any other chosen level) increases as the number tests inade increases. That Is not a good reasol) to deci-caSe Llic if * . (Tarlety of dependent variables tisc(l, assmiling there is a good tile- Iretical basis for choosing to use each one. Alternate procedures are available. Bonferroni procedures can used to a(fitist f-()I- tile 1111111ber of, tests Illade (lZoselithal & lZilb, 9 111, To overcome the conservatism of' this basic approach and de- 4 ia- ,&ease Type 11 errors, It is possible to weight the dependent vari -les according to then- importance and apply .1 so-called ordered lonfmcrroni procedure (1@osentlial & Rubin, 1984, 1985). Perhaps it. Most Ll.%CfLI1, ll()WCVCI-, to a))])ly specially developed procedlires @hat Integrate all the infOrinafioli fron) all tile dependent variables 6nd obtain only a single overall test of significance and effect. size -estimate. This can be accomplished very easily so long its we have (D opeasonable estimates of the intercorrelations aniong the dependent 1 Sariables (Rosenthal & Rubin, 1986). 4 00 ubdividing Studies W Ail issue discussed in tile gall7fleld debate lias to do with tile SUb- czlivision of studies into substudies as a function of different expert- 0 aliental procedures or individiial Clif-f'Crence variables slicli ;is sex, &ge, degree of'belief in psi effects, and the like (Schilleidler, 1968). s Ion as all the data are preserved and entered into the nieta- 9 .@pnalysis, no harm is done by subdividing. Indeed, subdividing is very useful in the search for moderator variables (Roselldial, 1984). Subdividing could have a very biasing effect on the accuracy of a cited p value if the overall data are subdivided in various ways, significant results are reported f'()I' ()Ile M' lll()I'C sUbmildies, and tile rest of the substudies are "thrown away." In the ordinary more proper application of meta-analytic procedures, however, subdividm ing makes little dill'crence. Consider it IMI exi)(21-illielit. with all overm all nonsignificant effect (p = .13, two-tailed). After tile Study is Over, it is noted that aboill hall'Ille, sllblicc(s were C,avol-ahic (oward psi and 1, that Lhei half wer-e !10L nd -e had been both feniale and male SLIb- .1 SUBDIVISION OF A LARGER ExPERIMENT Belleving SIII&CIS Disbelieving stib.jects Two-tailed p Z Two-tailed p .- Z > Fernales .05 2.0 .62 0.3 1 0 Males .32 1.0 .62 -0.5 < (D Noir: For I he st ndy as a whole. p was . 13 and , was 1.5 bef'orc subdividing. PositivI CL -n z's reflect results in the predicted direction; negative z's reflect results in the unpre- 0 dicted direction. -1 ip jects. Suppose that a subgroup Of Subjects, say female believers, M show a significant psi effect but the remaining groups do not. No U) harm is done by reporting that fact, though an adjustment is useful M 1%) in reporting the obtained p that takes into account ho w many 0 a subgroups were tested. It is essential, however, that the results of 14 significance tests for tile nonsignificant subgroups also be entered 0 4@k into the nieta-analysis. Z 00 Table 4 illustrates the SiLuation; four substudies have been formed, only one of which was significant. When we combine the results of the four substudies, however, we find the overall z to be > [(2.0) + (1.0) + (0.5) + (-0.5)J/V_4 = 1.5, p = .13, two-tailed. Es- @0 0 sentially, subdividing makes little difference so long as no data are W discarded. If a particular substudy showed great promise of evi- dencing psi, not.hIng would prevent the invesLigaLor from conduct- ' Ing new studies using only the preselected experimental conditio ns -4 00 Or types Of Subjects. It Would also be appropriate to conduct a meta- analysis on all the substudies that could be found that met t he 0 0 promising condition. In that case, however, the initial "study of dis- C4 covery" should be entered with an adjustment for the fact that sev- era] tests of significance were computed (Rosenthal & Rubin, 1983, 1984). 0 Flaw Effects and Weighting Replications 4@ There are few flawless studies in the behavioral sciences. Flaws can increase Type I or Type 11 errors, and the wise meta-analyst would do well to note how well Hyman (1985) and Honorion (1985) have searched for and evaluated flaws. For each flaw, it would be desirable to make some estimate of how much difference it made to the otitcome, In the pimcsent debate sonic flaws seenied to inake' ;I difference and others did not. When flaws matter we can adjust for 320 - 77hejournal ofParapsychology of railure to replicate. That both Studies Nvere able to re 'ject (lie 111111 and at exactly the same p level is simply a function of' sample size. Replication set B shows two studies with different p values, one sig- nificant at < .05, the other not significant. However, the two effect size estimates are in excellent agreement. We would say, accord- ingly, that replic;11.1oll set B shows lljol-c successfill replicalloll 111all does replication set A. It should be noted that the values of Table 3 were chosen so that the combined probability of the two studies of set A would be iden- 0 tical to the combined probability of the two studies of set B; (z, + z2)/V!! = z of 2.77, p = .0028, one-tailed. -n 0 The Metrics of the Success oj'Replication IN Once We adOI)t it View Of' tile SUCCeSS 01' repliGIL1011 @IS 21 I'LIIICLiOll M 1( 2) of similarity of effect sizes obtained, we can become more precise in U) (D Our assessments of the success of replication. Figure 1 shows the 0 replication plane" generated by crossing the restilts of' the first study conducted (expressed in units of' tile ef*iecL size r) by the re- Q Suits Of the second Study conducted. All perfect replications, those in which the effect sizes are identical in the two studies, fall on a _L dia onal rising f7rom tile lower left corner (- 1.00, - 1.00) to tile lip- 9 0 per right corner (+ 1.00, + 1.00). The results of replication set B > from Table 3 are shown to fall exactly on the diagonal of successful ;o replication (+.26, +.26). The results of replication set A are shown 0 to fall somewhat above the line representing perfect replication. Fig- 'U (D ure I shows that although set B reflects a more successful replica- a) 6 tion than Set A, the latter is also located fairly close to the line and S is, therefore, a fairly successful replication set as well. 4 , 00 Cohen's q. An alternative to the indexing of the success of repil- W cation by the difference between obtained effect size r's is to trans- 0 form the rs to Fisher's z's before taki d@ @-iis_he4:1s-z- (4- - - kiag-t-h-e Md CMI thus be Used ill Setting ic is istribUtCd lle@lrly 110171171iffy confidence intervals and testing hypotheses about r's, whereas r's W distribution is skewed, ',it)(] tile more so as tile popilhitloll vaille o('r 0 moves further from zero. Cohen's q is especially useful for testing L the significance of' difference between two obtained ef'I'CCL size r's. This is accomplished by means of the fact that + _T,==,--3 is distributed as z, the standard normal deviate (Rosenthal, 1984; Ganzfekl Debate-Rosenthal 1.00 .80-- 60-- .40-- .20-- .00-- 0 U 0 -.40 -.60-- .80-- -1.00 321 - 1.00 -.80 -.60 -.40 -.20 .00 .20 .40 .60 .80 1.00 Set A Irl Set B 0- 0 1 Figure 1. The replication plane. Rosenthal & Rubin, 1982a; Snedecor & Cochran, 1980). When there are more than two effect size r's to be evaluated for their variability (i.e., heterogeneity), the three references above all provide the ap- propriate formula for conTuting lie tejs_L@ j_ fthejj@@@f_4s, ISSUEs RLI-ATM -j-o REm,iCATION Multiple Testing In ganzf*eld studies, in parapsychological research more broadly, Mid, indeed, hl miost areas of behavioral science, it is common that more than one Lest of significance is computed to evaluate a re- search hypothesis., There may, for example, be I set of several de- pendent variables used to Miluate outcome. So long as ti .ie.re are > 0 < M CL _n 0 co M 0 > @4 co CD ;U 0 Q 44- (181 d 'L461 u;)qo- pur DIqmxnDp XIIEnbQ iou Diu -59-,)) jr-.)puPp! imr Imp sOnIPA III` SPD.10tim Olqu _o.ICI MV_T UDDAIDq P,_)UQJDjjjp QLIJ 'rJDA0 71 JO QSPjUPAp-e Dill -(1/1 "Q*i) SISDqIOCIXLJ IIIIU DLIJ JOINIII I)DIDD(NO s)Itly)@m1l) jo Holpodom-I I)OL11.10.1 _s.Lre-ij auisDx2 Dqi (q) pu-e pOuieiclo siiq ix)-tip jo Lroti-Iodoid pQu-ljoj z -Suin). auis:vr DIII (U) UD3Akj;)q -,):)uD.m1j.j!p @)tp si q:milm '?/ s,u;)IIo',)jo simn ui ;).it! g Qlqi?j Lit UmojjS sojrurjj@.SD ons j:)Djj;) Pq_L -(pR -d 'qS6 1) U0110UOH Aq pDziivuiwns sDipnis S1111 io3@np 3LIj uo pasi@q so-im ;),I, jo Xrjdvpji@,_)j-pur-LuDis r smotis c, olqp_l_ -pSO DZIS I,)ajjD RF kD7 0.(7 f7997-Pu"11,97S 'd 'c Cl) 961 '11"WXH) 13.1o put, (8c -d '9861 'U01.10uOH) Up-)q!U sjjr D ;)III io@j ipoq splotl itIOLUDD.121, Is Q:) jo uoilmuils X $37, -m)XII .10 %ISXImw-I!1,)uI )III to "'Itn"'o.1 DA11MULIM11) 00 111mixyl p1w (98(il) tiomotiol-I t- 1sIInsDi Dip jo Lionrigid-toim _aDqi ui iDjj!p AaLp q2noqilV -,DIqef!PA131 cb uoilruijojui Dill jo oiotu Dsn btli Irt-11 Lit z)si:)Qjd Djow z)q oi wOdcll?. m .1o C, Salim a- ppm.) 11,11.11A.To 111.1ol mlos BmSn 'spoiji;)m ipipo Qi1j, -awos inq I, = XjImIS11 I LIJIM 1/1 SI 11!1.11 al@hll% I! LIO %%JD.)IIs' JO XIII!Cll! X -qojd Dqi. 'snql -siOSim lem i jo IDS U jo ino UDSOLI:) ST IDS-Iul 1:)Z)Jjo:) QISUIS 31,11 UXIM X1110 %S0_X)tIs v %0_10131.1 siN IDa-111) .10 potlimif aqj, -;)Iqi?I!P IAU D.tDiA virp iiq ixwp qDiqm joj (F)I, Jo lmol 00 T_ v -111 It() % . -)D( Dol I j6). SOlpilis Rg , ;oip 1% SISAIM11-UIDIII M, 110).10110 le -.MI Isliq IDDilp _10 p0LlIQLU 3qI 'SI)0qjDU.I DAIJ DS31.11 JO DUO .11!111 - -1 Id E 111111 ;)INMI IWLI pDsn s.)ipnis Isom mlimaq -(cRCl 'LlIMIXI-1) qls'al DId1I C) C) XQLp iuqi iDuj iqi _iqj (4.) ODUMY!Uf@(S JO SJOAQ1 Pal-10CID-1 J13LII ISIll'im, CV ol papLj pup ".,ipnis ipili LIT sD-)Ipu-i qmS IPJDAQS pasn S-1011121ISDAUT aujos Imp si -eur sull ul LID,WOK), 10 ws'DIITJD DUO 'NS61 1-101-101-10H) a) III pDm U.-x)q iIq(j,j0 SODIPUr DAIJ 'DJOLl DZIJMT-111.1115, I )[JOM 0 '1qUDjU1.IQ(IX,) pl0j7LIV@i @)tlj .10 )1(101 Z)LII LIO j_(0A1 MXjMW LL _111OLU IL1111-10CIIIII DUOI) Z)Arq U01.10M)l I IRM I11!IUXI 1 > SASA-IVNV-V,I._-IW 31M, 9NI7I-dVWINn(,3 CL CL < "Cq 1 13 ilinq @11-em k.j,-6T '-qU111100) -IDSPI -De peq oq,%i jsijuQps r Xq p.Djurcluto.),)i! D,iOm suon:)11-us", 1-tone.)II&M DIII UQLjM XIuo'XIqITuDdDp pDienild-ai oq pInoD sJOSTTV31 Xllug."?d SUoIIDjIjjS.Uj pDIIj?)3I) .10 AIII!(j1!I!I!AI! Dill @_)IRISDJ) -dv -uourDild;m -to S,IDSIq-V-jj_ J() Di'll D11!')1ICl3-1 oI SO-111HI!j 1110 pD11110d DAMI ;o-xionq Jr.)[ SXIICI DIII.jo slimpnis p3Dpuj D.S.ID111D sD1pj1I:)U pl, X o ) 11011mil -mm:)s ImmIAMIXI '10111o to iou st ii ii?qi @unou Xq DAp:)3dsiQd InjDsn Lit w-DI(lo-id siqi QZ)Eld oi sil (11.)IJ Xmu 11 ImIlInIld.)i .10 xclol 1111 9,11mral ;molaq *19861 1 'LIKITIN x . jo iwi sswrms -v ol oycgf, tuoij Dip-I ssO. )ns ino .4,UFA0JdUJI 01 ILID[CAMbO M (o 1 1) DAO(11! P3113 03.11! 3AU JO 1:);)JJZ IIILUS U1IpDU_I DLII 'pQDpUI -D.)um-todim li,:mpind ou jo D,Ir otli Imp uraw Iou SOop Imp I(XIII!ymcl I! e o S Jaz 2UIpIIIDUI) SIDD@Jja IIEWS II.S110111 UDA- -go- > a 11@ JJIIU ?LII j!UIID31Dj jo ;)ILI C/),o 01 ooo'l J.')A0 .10 1,;V11, )I(IIIII'S )JHIb;)A P[nom if '(()I-) @t Jq0_ U0I)1!I1ICf0d III'MIS DIII .10, JO DIIJUA D Z Is- DICIIIJES yiAk go. > (I linnI110,is Dq PITIOm s1cIw-'I1I? u01II!DIIclD-I .10 %4 XIu0 '(9R61 'INIU050 _0@1 sI.1,11!I 1) sm'),w II?1(;.mA0JjU0-) -)AIJ mo jo UI?Ip,)LU aIII '01 - 3-IDAS .1 jo ;)IIIVA 110111!111(10(1 XII J! 'J,)A,)A%0 -I -(mi -d '4 1 461 't1-)IIo-)) K, YR SOZIS QlCILUIN' (111M go' > (/'It! MUJY!U_@IS J(j I)IIIOAI @JdLUJJJI! U0111nildaim 00 'jo IYA3 ow -),I;)Ak J .10 DtIlM uopiqn(lod atp j! '31dwrxD joq -11mr- 0 0111 I.)Dfa.i oi )lqi? ?(I AIDLInno-i plnom som Olduies isapow Xluo qvAq.) S,)II)IIIS 'JINJIMIS(IIII, ).I,)Ak )ZIS J.Mjj.) 011-11 Q4(il 11.)Ilo')) _mmmg 11!MIS11M@ JO SOIdIDLUdd 01 5iUIp.10:)._)V 'SI!D.11! IINSAD J 8 -0.1](10D M(M.I JDP@Xa Pfn0TIS QM m[Ai isnf;me sDzis i:)DjjD llewS '61, M 7,0* U10.11 I1'A.IDIUI DDLIDI)YUOD %/gg I! IM11? ()I - JO UMIMUI I@ 1111AI Rl- 01 00- LIT0.1.1 AlUo D.&Ilu ol (t) qwk) I.x Z)III 1). pm! ;X)IM111.101.10d UMU1111 JO SIM.IR IMq.IQA0.IIU0.-) DAIJ JO MWAD.1 IU;).-);).l I! LIE io,,l -sazfs I[IMIS X([ 0JI! 'W;).W 1jmI!aq;XD linstDAojiuo.) -Buismcf-ins iou si 'Ds.inoD.jo 'iuLtj -Z)P!S 1113UIS DLII UO NS6 1) u0I-I0u0H Put! (9R6 1) -U1,'1UXH Act POI-10cID-1 1:);)JJD PF)PUES DIR .10 azrq a2r1?.IQA1- OLII 'X.2010LULLI-ID1 QL60 s,uDqoo ul -iDz)jj;) jo apft -[U_4,MU SMOLUAOU@l UP wil op po).1odo.1 sqns.).1 ;)Ip imp oo.1944- xalff '.).1tillmaill 1+@Iztw.@, 3111 X(I j)-3JU1M,U!M O-)(MIN1110D JO ODA,3171 1..3 ,'It) (@8m) (tomotioll pur Q:s)(31) urtll,@Ij 1119floilliv 1@ 0 S72djj,7, 71mitS pim kjjnj@rj@(j uo.zjvx74,?t > -(9861 'uiclmd :7g ImIJUDSON !-K?61 'I"LlIuDsON !RL61 'D`1914@ SmI?Ij -to aipaInsip QIdLuis upy suoisnI:)uo:) psimq ssOl .4 SQ1pms jo ST . I prol Xmil II:)1I<.; -salpms .-Tip-la) xin.1, 11mll ssDIT M sItI59134 0,10ZLIOU inq pD,tomol puv soyms Dlqpjai XIii-ii oi oiD7 jo siLIBIDAI @),\U) III!.) )Ak ';)jdIUI!\,) .10, -SDII)IIII, JO 'I,)Ak 1110 UI SAIII) Dq, A 326 > - 0 CL -n 0 M 2) Co M 0 40- 00 0 > a) 6 0 4 , 00 TABi,iz. 5 STEM-AM)-LEAF PI,OT OF "DIRECT Fin- GANZFEU) STUDIES: GOHl-_N'S IL Swill 1.2 1.1 1.0 .9 .8 .7 .6 .5 .4 .2 .1 M -.0 -.1 -.2 .3 -.4 - .5 -.6 .7 - .8 -.9 8 0 2 2 2 4 12 2 4 4 7 8 2 38 8 77 9 5 0 C) _FffKey_TF9T7Yde__velol:)ed tile sLem-and-leal' plot its a special form f frequency distribution to facilitate the insi -)ection of a I-)al(:h or lata. Each number in the data batch i s inade up of' one stem all([ Sne leaf, but each stem may serve several leaves . Thus, the stem . I followed by leaves of 3, 8, 8 representing the numbers .13, .18, The first digit is the stem; the next digit is the leaf. The stem- and-leaf display functions as any other frequency distribution but the original data are retained precisely. Distribution of studies. F, ronn Table 5 we see that tile distribution of effect sizes is unimodal, with the bulk of the results (80%) falling between -.10 and .58. The distribtition is nicely syrililletri,11, wit.11 tile skewness index (gr, = .17) only 24% of'01,11 reqi6red [Or signif'- I icallce at P < -05 (SlIedecor & Cochran, 1980, pp. 78-79, 1192). The tails of tile distribution, however, are too long for, normality with Ganzj@ld Debatr-Rosenthal 327 kurtosis index -2 = 2.04, p = .02. Relative to what we would expect from a normal , (list ribill ioll, we have silldles that show larger posi- live and largcr llcgalivc CH,Ccl s1zcs Illall Nvotild bc reasoll'able. Ill- deed, the two largest positive effect sizes are significant outliers at p V < .05, and the largest negative effect sizc approaches significance, 0 with a Dixon index of .37 compared to one of .40 for the largest positive effect size (Siiedccor & Cochran, 1980, pp. 279-280, 490). - The total sample of studies is still small; however, if a much larger n 0 sample showed the same result, that would be a pattern consistent " " X M ) and strong neg- with the idea that both strong positive results ( psi 1 ative results ("psimmissing") might be more likely to find their way M 0) into print or at least to be more available to a meta-analyst. U) Distribution of subjects. It is useful to examine the distribution of I l SU III InariMl Studies. It I Alld a so be 1 V( effect sizes obtained LlIC 1 useful to examine the distribution of effect sizes obtained by indi- 1 v within the stild, lilarized. For example, IdUal SLIlJCCLS les SLIIII In it Study With a ineiin h of'.20, is OIC dlStI-ibLILIOI1 of'h flairly normal with 00 centering at .20, or is the distribuLion skewed with the bulk of the 0 Subjects centered closer to zero but with a few subjects earning con- sistendy high values of h? Distribution qf investigators. just as it is useful to examine the dism tribiltion of dic resillts of'stildics and of' slibiects withill studies. It i's also Llsehil to (.X.Illllll(, (hc (list ribut loll oI, resillts oblailwd by 6 ent investigators (Honorton, 1985; Hyrnan, 1985; Rosenthal, 1969, 0 -4 1984). The 28 direct hit studies were conducted by 10 different in- 00 vestigators (Honorton, 1985, 1). 60). FOLII_ investigators conducted only one study each, two conchicted two SUldles each, two condUCLed -con u-dt-ecr-- ------ nine studies. Analysis of variance showed that these 10 investigators d SigilifiCalldy alld fill I)OI-lalltly IH (he aVeragC IllagIlitUde Of' diffcre tile effects they obtained with F(9,18) = 3.81, p < .01, eta = .81. there was little relationship between the mean effect interestin ly , g size obtained by each investigator and the number of studies con- ducted (r = A 1; t(8) @ 0.3 1, p > .70). That different investigators may obtain significantly different rem suits frorn their sul@jects is well known various areas f psycholo 0 gy (Rosenthal, 1966). For example, in such a standard experimental area as eyelid conditioning, studies conducted at Iowa obtained re- sLi.Its in the predicted directioll 911 % of, the Lillie, NVIlcreas those c6nm (ILICIC(I CISCMICI@C 0bllilICd such rcstdis ()ill), G2% of' thc iiine with X2(1) 4.05, p < .05, N 25, r = .40 (Rosenthal, 1966, p. '24; 19rAN S I- - -- H 328 771ir jonriml TABLE 6 STATIS-1-1c.m. SuNcNi.-kRi, oi., -DIRi-.,(:,t* Hn'" GANZFELD STUDIES >Cenlral tendency (cojuw' ii) A Variability UnNveiglited meall .28 Maxilimill 1 1.44 0 Weighted inean .23 Quartile 3 Q3) .42 < M Median .32 Median (Q2) .32 CL Proportion positive Sign .82 Quartile I (Q1) n .08 0 .93 Significanre lests Q-3 - Q 1 .34 Conibined Slouffer Z 6.60 eT: 1. 7 5 (Q3 -- Q I j I t,est of Illean Z 3. 23 S .26 A5 0 Z of' proportion positive 3. 4 0 M 1%) Correlaiion of h 0 With z 0 .86 WConfidence intervals" With rawj .98 Froln TO 44. 80% 17 .39 L 95% .11 .45 0 9W, .04 .52 99.9% -.wi .59 ;b 'Based on N of 28 studies. 0 Suinniarv (?f Steni-and-Leqf Displ ay Table 6 provides a SUITIMary Ofthe sLem-and-le 4 af display of-Fa. 00 ble 5 and some additional useful information al)out cell,ral (-it- CD Mdency, variability, significance tests, confidence intervals, and c orrem C8 lations between Cohen's h and (a) significance level (z) and (b) raw W Ldifference in proportions U). Only a Few Comments are I-C(ILlil-C(l. 0 0 P:ffrcl size. The bulk of the restilts (82%) show j positive eft, 0 ed Wsize where 50% would be expected under the null (p = 0 .0004). The 0mean effect size, h, of .28 is equivalent to having a direct lilt rate of' C) .38 when .25 was expected under the null. The 95% confidence in- terval suggests the likely range of effect sizes to be from A I to .45, equivalent to accuracy rates of .30 to .46 when .25 was expected under the null hypothesis. Significance testing. The overall probability that. obtained accuracy Was better than tile 1(:CLII',ICY CXI)CCtC(,l UIIdCr tile null was @I P of 3.37/10" ;ISsociated W1111 .1 Sto"Frer z of 6.60 (Mosieilcr N'. Blish, 1954; Rosenthal, 197& 1, 1984). File-drawer 072alysis. A combined p as low as that obtained can be used as a guide to tile tolerance level for null results that never found their way into the metamanalytic data base (Rosenthal, 1979, 329 1984). It has long been believed that Studies ["ailing to reach statis- Lical significance may be less likely to be published (Rosenthal, 1966; 0 Sterling, 1959). -17hus it may be that there is a residual of' norisignifi- cant studies lailgUIShilIg III tile investigators' file drawers. With sim- pie calculations, it can be shown that, for tile current studies sum- marized, there would have to be 423 studies with mean P = .50, one-tailed, or z = 0.00 in those file drawers before the overall com- billed 1) would become just > .05, its Honorton (1985) has pointed out. 71tal inany studies unrctrieved seenis unlikely For this specialized area of' parapsychology (Houorton, 1985; Hyman, 1985). Based on expericiICC with MeLa-alialyses III odicr-donlains or research (e.g., interpersonal expectancy effects) the mean z or effect size for non- significant studies is not 0.00 but a value pulled strongly from 0.00 toward the mean z or mean effect size of tile obtained studies (Ro- senthal & Rubin, 1978). Coml)arison. unth ari Earlier Meta-AnalYsis It is instrUCLIve to Compare the results of the gatizfeld resear@h meta-analysis by Honorton (1985) with the results of an older and larger meta-analysis of another controversial research domain- 978). that of interpersonal expectancy effects (Rosenthal & Rubin, I In that analysis, eight areas ofexpectancy effects were summarized; 1 effect sizes (Cohen's d, roughly equivalent to Cohen's h) ranged froin .14 to 1.73 with a grand mean d of' .70. Honoi-Lon'S mean ef- Fect size (h = .28) exceeds the mean d of two Of' tile eight areas ratory (reaction time experiments [d = .171, and studies using labo . interviews [d = .14]). The earliel- ineLamanalysis displayed the distribution of the z's as- sociated with the obtained p levels. Table 7 shows a comparison of the two meta-analyses' distributions of z's. It is interesting to note the high degree of similarity in the distributions of'significance lev- els. The total proportion of significant results is somewhat higher for the ganzfeld studies but not significantly so (X'(1) = 1.07, N 373, p = .30, .05). INTERPRETING -riiL NIETA-ANALYTIG RESULTS AIL110LIgh tile ITSUILS Of tile 111CLa-analysis are clear, tile meaning of these results is open to various interpretations. The most obvious > -1 0 CL -n 0 -1 M K3 0 0 W 0 00 (D T 0 6 -4 00 W ;0 a 0 0 0 8) 0 0 0 a 4. 330 Thejournal qf Parap.@yrholqay TABLE 7 PROPORTION OF STUDIi.:s REACHING G-Rri*i(:Ai. Uvi.:us OFSIGNIFICANCE FOR Two RESEARCH AREAS Expectcd Expcclancy GalizFeld Interval for z proportion research" research" DifIlerence Predicted direc6oll - * 3.72 and above .0001 .07 .04 -.03 * 3.09 and above .001 .12 .18 .06 * 2.33 and above .0 1 .19 .25 .06 * 1.65 and above .05 .36 .43 .07 Not significant - 1.64 to + 1.64 .90 .60 .50 -.10 Unpredicted direction - 1.65 and below .05 .03 .07 .04 'N = 34.5 sni(lics: fly)ij, Ro%enfhal 1Z, Rubin (1978). "N = 28 sw(fies; f"0111 11(niorton (1985). interpretation might be that ill it very low p, and with a fairly irl-l- pressive cff'cct size, tile galizi,eld psi j)jIeljoIllellOII lias been dem- onstraLed. However, there are rival hypotheses that will need to be considered, marly of III(-,n )III ill ille dclallcd (-vallialloll bv Hyman (1985). I'l-ocedural Rival J@jpolhrscs 4 00 Senso7y leakage. A Standard rival hypothesis to the hypothesis of' ESP is that sensory leakage occur-red and that the receiver was 0 0 ary between the sender and recei 0 iver. As early as 1895, Hansen and 0 Lehmann (1895) described "unconscious whispering" in the laboi (4 -a- tory, and Kennedy (1938, 1939) was able to show that senders in 2 telepathy experiments Could give atiditoi . IvCrs L -y (:ties jo tilcir re(:e, 4- (JUiLe unwittingly. IngeniOUS use of' parabolic Sound refleCtol-S ljl@lde this denionsLration possible . Moll (1898), Stratton (1921), and War- ner and Ralble (1937) all gave early warnings on tile dallgCI-S OfUll- intentional cueing (for summaries see Rosenthal, 1965, 1966). he subtle kinds of cues described by these early Workers werejust the kind we have come to look for in searching for cues given off by experimenters diat Illigilt serve j, IllediaLe tile expert lliciftcr ex- pectancy effects f.0tIlId in laboratory settings (Rosenthal, 1966, Gan@feld Debale-Rosenthal 331 By their nature, ganzfeld studies tend to minimize problems of sensory cueing. An exception occurs when the subject is asked to choose which of four (or more) stimuli has been "sent" by another person or agent. When [Ile sallic stillill1i lield originally by the sender are shown to tile receiver, finger smudges or other marks may serve as cues. Honort6n has shown, however, that studies con- trolling for this type of'cue yield at least as many significant effects as do the studies not controlling for this type of cue. Recordi?@kgu e?-rors. A second rival hypothesis has nearly as long a history. Kennedy and Uphoff (1939) and Sheffield and Kaufman (1952) both found biased errors of recording the data of parapsy- chological experiments. III it ineta-arialysis of 139,000 recorded ob- servations in 21 studies, it was found that about 1% of all observa- Lions were in error and that, of the errors committed, twice as many favored the hypothesis as opposed it (Rosenthal, 1978b). Although it is difliCulL to rule recording errors Out OfgallZfeld Studies (or any other kind of research), their magnitude is such that they could probably have only a small biasing effect on the estimated average effect size (Rosenthal, 1978b, 1). 1007). Intentional error. The very recent history of'science has reminded its that even though Fratid in science is not quite of epidernic pro- portion, it must be given close attention (Broad & Wade, 1982; Zuckerman, 1977). Fraud in parapsychological research has been a colistal,it Concern, a Concern I'Milld 10 be JUStifiCd by periodic Ila- grant examples (Rhine, 1975). In tile analyses of Hyman (1985) and Honorton (1985), in any case, there appeared to be no relationship between degree of monitoring of participants and the results of the Stalist*cal Rival H@ypotheses F* -draiver ' tff. Tile problem of biased retrieval of' studi 7le UP I es for ;my illeta-analysis was describcd earlier. Part of' this prob1CIII is ad- T' 1, dressed by the 10-year-old norm of' the Parapsychological Associa- tion of reporting negative results at Its meetings and in ILSjournals (Honorton, 1985). Part of this problem is addressed also by Black- more (1980), who conducted a survey to retrieve unreported ganz- feld studies. She found that @ of her total of 19 studies werejudged significant overall by the investigators. This proportion of significant results (.37) was not significantly (or appreciably) lower than' the proportion of published studies found significant (.43) in Honor- Lon's (1985) meLa-an alysis of' direct llit gallZfeld Studies (X-'(I) 0 44. > 0 '1@ M CL n 0 M CO) M r,.) 0 00 0 5 0 CD 0) 6 4 0* W 0 0 Q 4' 332 TheJournal oj Parapsychology 0.17, 4> = .06. Somewhat similar results were obtained by Sommer (in press). in her analysis of research on the menstrual cycle. She found 61% of the pu@llshed results to be significant corripared to 2(j) 40% of the unpublished studies; = 2.30, p < .065, one-tailed, X .20. The results of the Blackmore and Sommer studies did not differ significantly (z = 0.69). Taken together, these studies provide only modest evidence for a serious file-drawer problem. A problem that seems to be a special case of the file-drawer problem was pointed out by Hyman (1985). That was a possible ten- dency to report the results of pilot studies along with subsequent significant results when the pilot data were significant. At the same time it is possible that pilot studies were conducted Without proill- ising results, pilot studies that then found their- way into tile file drawers. In any case, it is nearly impossible to have an accurate es- timate of the number of unretrieved studies or pilot studies actually conducted. Chances seem good, however, that there Would be fewer than the 123 results ofinvall Z = 0.00 re(juired lo bring Ille overall combined P to > .05. Multiple testing. Each gAnzfeld Study may have more than one dem pendent variable for scoring degree of success. If investigators use these dependent variables sequentially until they find one significant at p < .05, the true p will be higher than .05 (Hyman, 1985). This issue was discussed earlier-; it IS riot an inhereildy intractable one 1 . (Rosenthal & Rubin, 1986). Randomization. Hyrnan (1985) has noted that the target st'llililus may not have been selected in a trul random way frorn the pool of' y potential targets. To the extent that this is the case, the p values calculated can be in error. Hyman (1985) and Honorton (1985) di is- agree over tile fl-C(ILICI)CY ill this Sample of' studics of' improper rall- domization. In addition, they disagree over the magnitude of' the I-ChiLionShip between inadequate randoinization ;in(] study outcome. Hyn-lan felt this relationship to be significant and positive; Holiorton felt this relationship to be nonsignificant and negative. Because the median p level of just those 16 studies using random number tables or generators (z = .94) was essentially identical to thal round For all 28 studies, it seems unlikely that poor randomization procedures were associaled will) 11111ch of"In increase III significancc levc] (I lon- Orton, 1985, 1). 71). Statistical errors. Hyman (1()85) and Horlorton agree that 6 of' tile 28 studies contained statistical errors. However, the median effect size of these studies (h = .33) was very similar to the overall Inediall (h = .32), so that it seems unlikely that these errors had a major Ganzj'eld Debate-Rosenihal 333 effect oil the overall effect size estimate. 01121tting these Six Studies from the analysis decreases the mean h from .28 to .26. Such a drop is equivalent to a drop of the mean accuracy rate from .38 to .37 when .25 is the expected value under the null. A Tentative Inference On the basis of the preceding summary and the very valuable meta-analytic evaluations of Honorton (1985) and Hyman (1985), what are we to believe? It would be easiest to say, "Let's wait until more data have been accumulated From Studies purged of the probm lems noted by Hyman, Horiorton, and others."That is not a realistic approach. At any p6,nt in time somejudgment can be made, and though ourjudgment might be more accurate later on when those more nearly perfect studies become available, the situation for the ganzfeld domain seems reasonably clear. We feel it would be im- pl@jjjsibje to (Mlertlin tile ntd] given tile combined p fi-oill these 28 studies. Given tile various problems (.)I- flaws j)oillLed Out by I lynian and Honorton, the true effect size is almost surely smaller than the mean h of .28 equivalent to a m ean accuracy of 38% when 25% is expected under the null. We are persuaded that the net result of statistical errors was a biased increase in estimated effect size of at least a full percentage point (from 37% to 38%). Furthermore, we are persuaded that filemdrawer and related problems are such that some of the smaller effect size results have probably been kept off ize, we the market. If pressed to estimate a niore accurate effect si might think in terms of a shrinkage of h from the obtained value.of .28 to perhaps an h of .18. Thus, when the accuracy rate expected under- tile null is 1/4, WC [night CSI'lliate tile obtailled accuracy rate to be about 1/3. CONCLUSION Parapsychologists in particular and scientists in general owe a great debt of gratitude to Ray Hyman (1985) and Charles Honorton (1985) foy tbeil- C.11-eful @Ijjd ext.ensive @InalytiC and jlIeLa-alIalyLiC work on the ganzfeld problem. Their debate has yielded an espe- cially high light/heat ratio, and many of the important issues have now been brought out into bold relief. In my commentary on the ganzfeld debate, I focused inost closely on the concept of replication. That seemed appropriate, not 334 Thejournal oj'Parapsycholo@r ,y only becallse of, llic celill-ali1v of, (he probicill of' replic;Iblilly III Ille par .apsychological litel_@ItLlre, but. also because of' the centrallt ,y ot'Llie problem ill Irially Sciences, especKilly whell die sizes sought Ill tile l_)0j)Ulati0I1 are small. The cl'fect, size zero is only a Special case of' the class of sniall effect sizes. In closin- I want only to suggest that parapsychological and other behavioral sciences would be well served to modify their view of the success of' replication Ili the direction of the following newer view: I. A replication SUCCeSSf'1Il to the degree that the second study obtains an effect size similar to the effect size of the first study. 2. Thrcc ()I, more inv"iigations i1rc stlct@cssl'tll of- oIle another to tile extent that the efl: ect sizes are hoillogeneous, 3. Significance testing has nothing to do with success of replica- tion though it can be useful in many ways, including the @lssessljjerlt. of the likelihood of the null given all prior research (weighted as desirecl and as reasonable) and the likelihood of real dif're .1-cliCes among the effect sizes of' two or more Studies. REFERENCE.S BLACKMORE, S. (1980). The extent of' selective reporting of" ESI1 ganzfeld studies. European jou? -nal of Parapsychology, 3, 213-219. BROAD, W., & WADE, N. (1982). Betravers of the truth. New York: Simon and Schuster. COHEN, J. (1977). Stati@dical power analysisfor the behavioralsciences (rev. ed.). New York: Academic Ilress. COLLINS, H. M. (1985). Changing order: Replication and induction in stienlific Ganzje'ld Debate-]? osenihal 335 KENNEDN', J. L., &- IT111014% H. 1,. (Iwlm. 1%xperiments oil tile nature of Cxtra-scilsory perception: 111. Flic recording cl-rol. criticisill of' cxara- chance scorcs. journal offlarapsTchoh, , 3, 226-2,15. Moi.j,, A. (1898). Ilypnolistit (401 cd.). New @'ork: Scribner. MOSTE.I.I.1; 'R, F. M., & BUSH, R. IZ. (1954). Selected (JUMitiLative techniques. Ili G. Lindzey (Ed.), Handbook qj'social pgchologoy: Vol. 1. 7711,01y and inelhod (pp. 289-334). Cambridge, MA: Addison-Wesley. NEI-SON, N., ROSEWITIAL, R., & ROSNOW, R. L. (1986). Interpretation of significance levels and effect sizes by psychological researchers. Ameri- can Psychologist, 41, 1299-130 1. RAO, K. R. (V 85). The ganzfeld debate. Journal of Parapsychology, 49, 1-2. RH I N E, J. B. (1975). Seco nd re po rt o n a case of ex perimenter fraud. journal qf'ParaAvycho1(Pu. .19, 306-'@25, Ros[WHIAk, R, (1965), Qcvei- H;iiis: A casc study of' scielltifi( Illethod, Ill 0. 1)fungst, Clever Hans (pp. ix-xlii). New York: Holt, IRinehart and Winston. ROSENTIIAL, R. (1966). ExI)erintenler (jj@cls in behavioral research. New York: Appleton-Century-Crofts. ROSENTHAL, R. (1969). IIltCrj)crs0Ijll cxpectatioll.,;. in R1. 1-loscialial k R. L. Rosnow (Eds.), ArL@ract in behavioral research (pp. 181-277). New York: Ac;idvinic T)rcss. RoSENTI IAI., R. (I 978a). Combillilig I-CS1111S 01' iIldCj)e[IdeIlt StLldiCS. PS)Tho- logical Bulletin, 85, 185-193. @!"` ROSENTHAL, R. (1978b). How often ai -e our numbers wrong? Anierican Psy- chologist, 33, 1005-1008. Rosr.NTHAi., R. (1979). The "file drawer problern" and tolerance for ruill results. 11@vychological Bulletin, 86, 638-641. RosENTHAL, R. (1984). Meta-analylic procedures for social research. Beverly Hills, CA: Sage. FISK'E, D. W. (1978). The several @Hlds of'generalization. 77he Beh(1740ral and Brain Sciences, 3, 393-394. HANSEN, 1". G. G., & LFIIMANN, A, (1895). Uchcr Unwillkfirliches Fifistern. Philosophische Studien, 11, 471-530. HARRIS, M. J., & ROSENTIlAt., R. (1986). Inteipersonal exl5ectaury efficts and huynan peiforniance research. Report prepared For the National Academy of'Sciences. HONORTON, C. (1985). Meta-analysis of psi ganzfeld research: A response t to Hyman. Journal of Parapsychology, 49, 51-9 1. HYMAN, R. (1985). The ganzfeld psi experiment: A critical appraisaijour- nal of Parapsychol(,@Ary, 49, 3-49. KENNEDY, J. L. (1938). Experiments oil "unconscious whispering." P@vycho- logical Bulletin, 35, 526. (Abstract) KENNE@I)Y, L. (1939). A inediodological review of' extra-sclisory percep- expectancy effects. In A. W. Siegman & S. Feldstein (Eds.), Multichannel inlegrations 'If nonverbal behavior (pp. 105-128). Hillsdale, NJ: Lawrence Erlbatim Associa(es. ROSENTHAL, R., & GAITO, J. (1963). The interpretation of levels of signifi- cance by psychological researchers.journal of Psychology, 55, 33-38. RoSENTHAi., R., & GAITO,.j. (1964). Further evidence for the cliff effect in the interpretation of levels of' significance. Psychological Reports, 15, 570. ROSENTHAL, R., & ROSNOW, R. L. (1984). Essentials of behavioral research: Method,), and data analysis. Nc%v York: NlcGraw-Hill. ROSEN'1'1-lAt,, R., & RutuN, D. B. (1978). Interpersonal expectancy effects: The first 345 studies. 'The Behavioral and Brain Sciences, 3, 377-386. ROSENTHAL, R., & RuBIN, D. B. (1979). Comparing significance levels. of independent studies. Psychological Bulletin, 86, 1165-1168. ROSENTH,u., R., &- RUBIN, 1). B. (1982a). Comparing el'I'ect sizes of' inde- --rende- -fjclies 1'-rholqLr;-1 f@ullelin 92, 500-604. L 4.. 1! 336 TheJournal of Parap,@ychology RosEK-rHAL, R.. & RUBIN, D. B. (1982b). A simple, general purpose display of magnitude of experimental effect.. Journal of Educational Psycholq@y, 74, 166-169. ROSENTHAi,, R., & RUBIN, D. B. (1983). Ensemble-adjusted p values. Psycho- logical Bulletin, 94, 540-54 1. RosENTHAL, R., & RUBIN, D. B. (1984). Multiple contrasts and ordered Bonferroni procedures. Journal of Educational Psychology, 76, 1028- 10,34 RosENTIW., R., & RUBIN, 1). B. (1985). Statistical analysis: Sutinilarizilip evidence versus establishing facts. Psychological Bulletin, 97, 527-529. ROSENTHAL, R., & RUBIN, D. B. (1986). Meta-analytic Procedures for com- bining studies with multiple effect sizes. Psychological Bulletin, 99, 400- 406. SCHMEIDLER, G. R. (1968). Parapsychology. In International Encyclopedia q the Social Sciences (pp. 386-399). New York: MacMillan & Free Press. SFIFFFInt), F. D., KAUFMAN, R. S., & RiIINE,J. 11. (1952). A I'K experimen at Yale starts a controversy. Journal oj'the American Society fin- Psychica Research, 46, 111-117. SNEDECOR, G. W., & COCHRAN, W. G. (1980). Statistical methods (7111 ed.) Ames: Iowa State University Press. SOMMER, B. (ill ])I-CSS). 'rllV filC draNver efTecl and plihficalioll raws ill Illell AMA cycle research. Psychology oj'Women Quarterly. Sj,F,,N(.F, K. W. (1964). Anxiety (drive) level and performance ill eyelid con ditioning. I@s'whologiral Bidlefin. 61, 129-- 139@ STERLING, T. D. (1959). Publication decisions and their possible effects-o inferences drawn from tests of significance-or vice versa. journal the American Statistical Association, 54, 30-34. STRA7rON, C. M. (1921). The control of another person by obscure sign Psycholqgr ical Review, 28, 301-314. TRUZZI, M. (1981). Reflections on paranormal communiCaLiOll: A zetetic perspective. In T. A. Sebeok & R. Rosenthal (Eds.), The Clever Hal phenomenon (pp. 297-309). New York: New York Academy of'Science TUKFN', J. W. (1977). Exploratorv data analysis. Reading, MA: Addison-We ley. WARNER, I_ & RAMI.E. NI. (1917). 'I"ch'ImIliv ill Ow pmrchophysical Libor tory. journal of'J'arapsychology, 1, 44-5 1. ZOCKKKNIAN, If. (1977). 1)(wiani bch;lvior :111d S()(.i;ll colill-ol ill scielicv. I E. Sagarin (Ed.), Deviance and social change (pp. 87- 138). Beverly Flil! CA: Sage. Department of Psychology Harvard University Cambri'dge, MA 02138 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 REPLICATION AND META-ANALYSIS IN PARAPSYCHOLOGY Jessica Utts Division of Statistics University of California, Davis 1. INTRODUCTION In a June 1990 Gallup Poll, 49% of the 1,236 respondents claimed to believe in extrasensory perception (ESP), and one in four claimed to have had a personal experience involving telepathy (Gallup and Newport, 1991). Other surveys have shown even higher percentages; the University of Chicago's National Opinion Research Council recently surveyed 1,473 adults, of which 67% claimed that they had experienced ESP (Greeley, 1987). Public opinion is a poor arbiter of science, however, and experience is a poor substitute for the scientific method. For more than a century, small numbers of scientists have been conducting laboratory experiments to study phenomena such as telepathy, clairvoyance, and precognition, collectively known as "psi" abilities. This paper will examine.some of that work, as well as some of the statistical controversies it has generated. Parapsychology, as this field is called, has been a source of controversy throughout its history. Strong beliefs tend to be resistant to change even in the face of data, and many people, scientists included, seem to have made up their minds on the question without examining any empirical data at all. A critic of parapsychology recently acknowledged that "The level of the debate during the past 130 years has been an embarrassment for anyone who would like to believe that schol ars and scientists adhere to standards of rationality and fair play" (Hyman, 1985a, p.89). While much of the controversy has focused on poor experimental design and Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4. Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 2 potential fraud, there have been attacks and defenses of the statistical methods as well, sometimes calling into question the very foundations of probability and statistical inference. Most of the criticisms have been leveled by psychologists, For example, a 1988 report of the U.S. National Academy of Sciences concluded that "The committee finds no scientific justification from research conducted over a period of 130 years for the existence of parapsychological phenomena" (Druckman and Swets, .1988, p. 22). The chapter on parapsychology was written by a subcommittee chaired by a psychologist who had published a similar conclusion prior to his appointment to the committee (Hyman, 1985a, p.7). There were no parapsychologists involved with the writing of the report. Resulting accusations of bias (Palmer, Honorton and Utts, 1989) led U.S. Senator Claiborne Pell to request that the Congressional Office of Technology Assessment (OTA) conduct an investigation with a more balanced group. A one-day workshop was held on September 30, 1988 bringing together parapsychologists, critics, and experts in some related fields (including the author of this paper)., The report concluded that parapsychology needs "a fairer hearing across a broader spectrum of the scientific community, so that emotionality does not impede objective assessment of experimental results" (Office of Technology Assessment, 1989). It is in the spirit of the OTA report that this article is written. After Section 2, which offers an anecdotal account of the role of statisticians and statistics in parapsychology, the discussion turns to the more general question of replication of experimental results. Section 3 illustrates how replication has been (mis)interpreted by scientists.in many fields. Returning to parapsychology in Section 4, a particular experimental regime called the "ganzfeld" is described, and an extended debate about the interpretation of the experimental results is discussed. Section Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/118 : CIA-RDP96-00789ROO3100030001-4 3 5 examines a meta-analysis of recent ganzfeld experiments designed to resolve the debate. Finally, Section 6 contains a brief account of meta-analyses that have been conducted in other areas of parapsychology, and conclusions are given in Section 7. 2. STATISTICS AND PARAPSYCHOLOGY Parapsychology had its beginnings in the investigation of purported mediums and other anecdotal claims in the late 19th century. The Society for Psychical Research was founded in Britain in 1882, and its American counterpart was founded in Boston in 1884. While these organizations and their members were primarily involved with investigating anecdotal material, a few of the early researchers were already conducting "forced-choice" experiments such as card-guessing. (Forced-choice experiments are like multiple choice tests; on each trial the subject must guess from a small, known set of possibilities.) Notable among these was Nobel Laureate Charles Richet', who is generally credited with being the first to recognize that probability theory could be applied to card-guessing experiments (Rhine, 1977, p.26; Richet, 1884). F.Y. Edgeworth, partly in response to what he considered to be incorrect analyses of these experiments, offered one of the earliest treatises on the statistical evaluation of forced- choice experiments in two articles published in the Proceedings of the Society for Psychical Research (Edgeworth, 1885, 1886). Unfortunately, as noted by Mauskopf and McVaugh (1979) in their historical account of the period, Edgeworth's papers were "perhaps too difficult for their immediate audience" (p. 105). Edgeworth began his analysis by using Bayes Theorem to derive the formula for the Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 3 5 examines a meta-analysis of recent ganzfeld experiments designed to resolve the debate. Finally, Section 6 contains a brief account of meta-analyses that have been conducted in other areas of parapsychology, and conclusions are given in Section 7. 2. STATISTICS AND PARAPSYCHOLOGY Parapsychology had its beginnings in the investigation of purported mediums and other anecdotal claims in the late 19th century. The Society for Psychical Research was founded in Britain in 1882, and its American counterpart was founded in Boston in 1884. While these organizations and their members were primarily involved with investigating anecdotal material, a few of the early researchers were already conducting "forced-choice" experiments such as card-guessing. (Forced-choice experiments are like multiple choice tests; on each trial the subject must guess from a small, known set of possibilities.) Notable among these was Nobel Laureate Charles Richet, who is generally credited with being the first to recognize that probability theory could be applied to card-guessing experiments (Rhine, 1977, p.26; Richet, 1884). F.Y. Edgeworth, partly in response to what he considered to be incorrect analyses of these experiments, offered one of the earliest treatises on the statistical evaluation of forced- choice experiments in two articles published in the Proceedings of the Society for Psychical Research (Edgeworth, 1885, 1886). Unfortunately, as noted by Mauskopf and McVaugh (1979) in their historical account of the period, Edgeworth's papers were "perhaps too difficult for their immediate audience" (p. 10 5). Edgeworth began his analysis by using Bayes Theorem to derive the formula for the Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : ClAmRDP96-00789RO03100030001-4 4 posterior probability that chance was oper ating, given the data. He then continued with an aq,ument "savouring more of Bernoulli than Bayes" in which "it is consonant, I submit, to experience, to put 1/2 both for a and B, " i.e. for both the prior probability that chance alone was operating, and the prior probability that "there should have been some additional agency." lie then reasoned (using a Taylor Series expansion of the posterior probability formula) that if there were a large probability of observing the data given that some additional agency was at work, and a small objecti-ve probability of the data under chance, then the latter (binomial) probability "may be taken as a rough measure of the sought a postefioti probability in favour of mere chance" (p. 195). Edgeworth concluded his article by applying his method to some data published previously in the same journal. He found the probability against chance to be .99996, which he said "may fairly be regarded as physical certainty" (p. 199). He concluded: "Such is the evidence which the calculus of probabilities affords as to the existence of an agency other than mere chance. The calculus is silent as -to the nature of that agency -- whether it is more likely to. be vulgar illusion or extraordinary law. That is a question to be decided, not by formulae and figures, but by general philosophy and common sense" (p. 199). Both the statistical arguments and the experimental controls in these early experiments were somewhat loose. For example, Edgeworth treated as binomial an experiment in which one person chose a string of eight letters and another attempted to guess the string. Since it has long been understood that people are. poor random number (or letter) generators, there is no statistical basis for analyzing such an experiment. Nonetheless, Edgeworth and his contemporaries set the stage for the use of . controlled experiments with statistical evaluation in laboratory parapsychology. One of the first American researchers to use statistical methods in parapsychology was Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 5 John Edgar Coover, who was the Thomas Welton Stanford Psychical Research Fellow, in the Psychology Department at Stanford University, from 1912 to 1937 (Dommeyer, 1975). In 1917 Coover published a large volume summarizing his work (Coover, 1917). Coover believed that his results were consistent with chance, but others have argued that Coover's definition of significance was too strict (Dommeyer, 1975). For example, in one evaluation of his telepathy experiments, Coover found a two-tailed p-value of .0062. He concluded "Since this value, then, lies within the field of chance deviation, although the probability of its occurrence by chance is fairly low, it cannot be accepted as a decisive indication of some cause beyond chance which operated in favor of success in guessing" (Coover, 1917, p. 82). On the next page he made it explicit that he would require a p-value of .0000221 to declare that something other than chance was operating. It was during the summer of 1930, with the card-guessing experiments of J.B. Rhine at Duke University, that parapsychology began to take hold ai a laboratory science. In fact, Rhine's laboratory still exists under the name of the Foundation for Research on the Nature of Man, housed at the edge of the Duke University campus. It wasn't long after Rhine published his first book, atrasensory Perception in 1934, that the attacks on his methodology began. Since his claims were wholly based on statistical analyses of his experiments, the statistical methods were closely scrutinized by critics anxious to find a plausible explanation for Rhine's positive results. The most persistent critic was a psychologist from McGill University named Chester Kellogg (Mauskopf and McVaugh, 1979). Kellogg's main argument was that Rhine was using the binomial distribution (and normal approximation) on a series of trials that were not Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-'RDP96-00789ROO3100030001-4 6 independent. The experiments in question consisted of having a subject guess the order of a deck of 25 cards, with five each of five symbols, so technically Kellogg was correct. By 1937 several mathematicians and statisticians had come to Rhine's aid. Mauskopf and McVaugh (1979) speculated that since statistics was itself a young discipline, "a number of statisticians were equally outraged by Kellogg, whose arguments they saw as discrediting their profession" ( p. 258). The major technical work, which acknowledged that Kellogg's criticisms were accurate but did little to change the significance of the results, was conducted by Charles Stuart and Joseph A. Greenwood and published in the first volume of the Journal of Parapsychology (Stuart and Greenwood, 1937). Stuart, who had been an undergraduate in mathematics at Duke, was one of Rhine's early subjects, and continued to work with him as a researcher until Stuart's death in 1947. Greenwood was a Duke mathematician, who apparently converted to a statistician at the urging of Rhine. Another prominent figure who was distressed with Kellogg's attack was E. V. Huntington, a mathematician at Harvard. After corresponding with Rhine, Huntington decided that, rather than further confuse the public with a technical reply to Kellogg's arguments, a simple statement should be made to the effect that the mathematical issues in Rhine's work had been resolved. Huntington must have successfully convinced his former student, Burton Camp of Wesleyan, that this was a wise approach. Camp was the 1937 President of IMS. When the annual meetings were held in December of 1937 Oointly with AMS and AAAS), Camp released a statement to the press that read: "Dr. Rhine's investigations have two aspects: experimental and statistical. On the experimental side rhathematicians, of course, have nothing to say. On the statistical side, however, recent mathematical work has established the fact that, assuming that the experiments have been properly performed, the statistical Approved For Release @003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 7 analysis is essentially valid. If the Rhine investigation is to be fairly attacked, it must be on other than mathematical grounds" (Camp, 1937). One statistician who did emerge as a critic was William Feller. In a talk at the Duke Mathematical Seminar on April 24, 1940, Feller raised three criticisms to Rhine's work (Feller, 1940). They had been raised before by others (and continue to be raised even today). The first was that inadequate shuffling of the cards resulted in additional information from one series to the next. The second was what is now known as the "file-drawer effect," namely, that if one combines the results of published studies only, there is sure to be a bias in favor of successful studies. The third was that the results were enhanced by the use of optional stopping, i.e. by not specifying the number of trials in advance. All three of these criticisms were addressed in a rejoinder by Greenwood and Stuart (1940), but Feller was never convinced. Even in its third edition published in 1968, his book An Introduction to Probability Theory and Its Applications still contains his conclusion about Greenwood and Stuart: "Both their arithmetic and their experiments have a distinct tinge of the supernatural" (Feller, 1968, p. 407). In his discussion of Feller's position, Diaconis (1978) remarks, "I believe Feller was confused ... he seemed to have decided the opposition was wrong and that was that. " Several statisticians have contributed to the literature in parapsychology to greater or lesser degrees. T.N.E. Grevil1e, devoted much of his professional life to developing statistical methods for parapsychology; Fisher (1924, 1929) addressed some specific problems in card- guessing experiments; Wilks (1965) described various statistical methods for parapsychology; Lindley (1957) presented a Bayesian analysis of some parapsychology data; and Diaconis (1978) pointed out some problem@ with certain experiments and presented a method for analyzing experiments when feedback is given. Approved For-Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 8 Occasionally, attacks on parapsychology have taken the form of attacks on statistical inference in general, at least as it is applied to real data. Spenc 6r-Brown (1957) attempted to show that true randomness is impossible, at least in finite sequences, and that this could be the explanation for the results in parapsychology. That argument re-emerged in a recent debate on the role of randomness in parapsychology, initiated by psychologist J. Barnard Gilmore (Gilmore, 1989; Utts, 1989a; Palmer, 1989; Gilmore, 1990; Palm .er, 1990). Gilmore stated that "The agnostic statistician, advising on research in psi, should. take account of the possible inappropriateness of classical inferential statistics" (1989, p.338). In his second paper, Gilmore reviewed several non-psi studies showing purportedly random systems that do not behave as they should under randomness (e.g. Iversen, Longcor, Mosteller, Gilbert, and Youtz, 1971; and Spencer-Brown, 1957). Gilmore concluded that "Anomalous data! ... should not be found nearly so often if classical statistics offers a valid model of reality" (1990, p. 54), thus rejecting the use of classical statistical inference for real-world applications in general. 3. REPLICATION Implicit and explicit in the literature on parapsychology is the assumption that in order to truly establish itself, the field needs to find a repeatable experi ment. For example, Diaconis (1978) starts the summary of his article in Science with the words "In search of repeatable ESP experiments, modem investigators..." (p. 13 1). On October 28-29, 1983, the 32nd International Conference of the Parapsychology Foundation was held in San Antonio, Texas, to address "The Repeatability Problem in Parapsychology." The Conference Proceedings (Shapin and Coly, 1985) reflect the diverse views among parapsychologists on the nature of the problem. Honorton Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO31000300,01-4 9 (1985a) and Rao (1985), for example, both argued that strict replication is uncommon in most branches of science, and that parapsychology should not be singled out as unique in this regard. Other authors expressed disappointment in the lack of a single repeatable experiment in parapsychology, with titles such as " Un repeatability: Parapsychology's Only Finding" (Blackmore, 1985), and "Research Strategies for Dealing with Unstable Phenomena" (Beloff, 1985). It has never been clear, however, just exactly what would constitute acceptable evidence of a repeatable experiment. In the early days of investigation, the major critics "insisted that 'It would be sufficient for Rhine and Soal to convince them of ESP if a parapsychologist could erform successfully a single 'fraud-proof' experiment" (Hyman, 1985a, p. 71). However, as soon as well-designed experiments showing statistical significance emerged, the critics realized that a single experiment could be statistically significant just by chance. British psychologist C.E.M. Hansel quantifi@d the new expectation, that the experiment should be repeated a few times, as follows: "If a result is significant at the .01 level and this result is not due to chance but to information reaching the subject, it may be expected that by making two further sets of trials the antichance odds of one hundred to one will be increased to around a million to one, thus enabling the effects of ESP -- or whatever is responsible for the original result -- to manifest itself to such an extent that there will be little doubt that the result is not due to chance" (Hansel, 1980, p.298). In other words, three consecutive experiments at p !5 .01 would convince Hansel that something other than chance was at work. This argument implies that if a particular experiment produces a statistically significant result, but subsequent replications fail to attain significance, then the original result was probably due to chance, or at least remains unconvincing. The problem with this line of reasoning is that Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 10 there is no consideration given to sample size or power. Only an experiment with extremely high power should be expected to be "successful" three times insuccession. It is perhaps a failure of the way statistics is taught that many scientists do not understand the importance of power in defining successful replication. To illustrate this point, psychologists Tversky and Kahnemann (1982) distributed a questionnaire to their colleagues at a professional meeting, with the question: "An investigator has reported a result that you consider implausible. He ran 15 subjects, and reported a significant value, t = 2.46. Another investigator has attempted to duplicate his procedure, and he obtained a nonsignificant value of t with the same number of subjects. The direction was the same in both sets of data. You are reviewing the literature. What is the highest value of t in the second set of data that you would describe as a failure to replicate?" (1982, p. 28). In reporting their results, Tversky and kahnemann stated: "The majority of our respondents regarded t 1.70 as a failure to replicate. If the data of two such studies (t = 2.46 and t 1.70) are pooled, the value of t for the combined data is about 3.00 (assuming equal variances). Thus, we are faced with a paradoxical state of affairs, in which the :same data that would increase our confidence in the finding when viewed as part of the original study, shake our confidence when viewed as an independent study" (1982, p. 28). At a recent presentation to the History and Philosophy of Science Seminar at the University of California at Davis, I asked the following question. Two scientists, Professors A and B, each have a theory they would like to demonstrate. Each plans to run a fixed number of Bernoulli trials and then tes t B@: p = .25 versus H.: p > .25. Professor A has access to large numbers of students each semester to use as subjects. In his first experiment he runs 100 subjects, and there are 33 successes (p = .04, one-tailed). Knowing the importance of replication, Professor A runs an additional 100 subjects as a second experiment. He finds 36 successes (p = .009, one-tailed). k Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Professor B only teaches small classes. Each'quarter she runs an experiment on her students to test her theory. She carries out ten studies this way, with the following results: n Number of successes one-tailed 12-value 10 4 .22 15 6 .15 17 6 .23 25 8 .17 30 10 .20 40 13 .18 18 7 .14 10 5 .08 15 5 .31 20 7 .21 I asked the audience by a show of hands to indicate whether or not they felt the scientists had successfully demonstrated their theories. Professor A's theory received overwhelming support, with approximately 20 votes, while Professor B's theory received only one vote. If you aggregate the results of the experiments for each Professor, you ,Aill notice that each conducted 200 trials, and Professor B actually demonstrated a higher level of success than Professor A, with 71 as opposed to 69 successful trials. The one-tailed p-values; for the combined trials are .0017 for Professor A and .0006 for Professor B. To address the question of replication more explicitly, I also posed the following scenario. In December of 1997 it was decided to prematurely terminate a study on the effects of aspirin in reducing heart attacks because the data were so convincing (See e.g. Greenhouse and Greenhouse, 1988; Rosenthal, 1990a). The physician-subjects had been randomly assigned to take aspirin or a placebo. There were 104 heart attacks among the 11,037 subjects in the aspirin group, and 189 heart attacks among the 11,034 subjects in the placebo group (chi-square Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 1.2 - 25. 0 1, P < . 0000 1). After showing the results of that study, I presented the audience with two hypothetical experiments conducted to try to replicate the original result, with outcomes as follows: REPLICATION #1 REPLICATION #2 Heart Attack Heart Attack Yes No Yes No Aspirin 11 1156 Aspirin 20 2314 Placebo 19 1090 Placebo 48 2170 Chi-square = 2.596, p =. 11 Chi-square 13.206, p =. 0003 I asked the audience to indicate which one they thought was a more successful replication. The audience chose the second one, as would mostjournal editors, because of the "significant p-value". In fact, the first replication has almost exactly the same proportion of heart attacks in the two groups as the original study, and is thus a very close replication of that result. The second replication has very different proportions, and in fact the relative risk from the second study is not even contained in a 95% confidence interval for relative risk from the original study. The magnitude of the effect has been much more closely matched by the "non- significant" replication. Fortunately, psychologists are beginning to notice that replication is not as straightforward as they were originally led to believe. A special issue of the Journal of Social Behavior and Personality was entirely devoted to the question of replication (Neuliep, 1990). In one of the articles, Rosenthal cautioned his colleagues: "Given the levels of statistical power at which we normally operate, we have no right to expect the proportion of significant results Approved For Release 20.03/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 13 that we typically do expect, even if in nature there is a very real and very important effect" (Rosenthal, 1990b, p. 16). Jacob Cohen, in his insightful article fitled "Things I Have Learned (So Far)," identified another misconception common among social scientists: "Despite widespread misconceptions to the contrary, the rejection of a given null hypothesis gives us no basis for estimating the probability that a replication of the research will again result in rejecting that null hypothesis" (Cohen, 1990, p. 1307). Cohen and Rosenthal both advocate the use of effect sizes as opposed to significance levels when defining the strength of an experimental effect. In general, effect sizes measure the amount by which the data deviate from the null hypothesis in terms of standardized units. For instance, the effect size for a two-sample t-test is usually defined to be the difference in the two means, divided by the standard deviation for the control group. This measure can be compared across studies without the dependence on sample size inherent in significance levels. (Of course there will still be variability in the sample effect sizes, decreasing as a function of sample size.) Comparison of effect sizes across studies is one of the major components of meta-analysis. Similar arguments have recently been made in the medical literature. For example, Gardner and Altman (1986) stated that the use of p-values "to define two alternative outcomes - significant and not significant. - is not helpful and encourages lazy thinking" (p. 746). They advocated the use of confidence intervals instead, As discussed in the next section, the arguments used to conclude that parapsychology has failed to demonstrate a replicable effect hinge on these misconceptions of replication and failure to examine power. A more appropriate analysis would compare the effect sizes for similar Approved For. Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 14 experiments across experimenters and across time to see if there have been consistent effects of the same magnitude. Rosenthal also advocates this view of replication: "The traditional view of replication focuses on significance level as the relevant summary statistic of a study and evaluates the success of a replication in a dichotomous fashion. The newer, more useful view of.replication focuses on effect size as the more important summary statistic of a study and evaluates the success of a replication not in a dichotomous but in a continuous fashion" (Rosenthal, 1990b, p. 28). The dichotomous view of replication has been used throughout the history of parapsychology, by both parapsychologists and critics (Utts, 1988). For example, the National Academy of Sciences Report critically evaluated "significant" experiments, but entirely ignored "nonsignificant" experiments. In the next three sections we will examine some of the results in parapsychology using the broader, more appropriate definition of replication. In doing so, we will show that the results are far more interesting than the critics would have us believe. 4. THE GANZFELD DEBATE IN PARAPSYCHOLOGY An extensive debate took place in the mid-1980's between .a parapsychologist and critic, questioning whether or not a particular body of parapsychological data had demonstrated psi ,abilities. The experiments in question were all conducted using the ganzfeld setting (described below). Several authors were invited to write commentaries on the debate. As a result,'this data base has beenmore thoroughly analyzed by both critics and proponents than any other, and provides a good source for studying replication in parapsychology. The debate concluded with a detailed series of recommendations for further experiments, and left open the question of whether or not psi abilities had been demonstrated. A new series Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 15 of experiments that followed the recommendations were conducted over the next few years. The results of the new experiments will be presented in Section 5. 4.1 Free-response Experiments Recent experiments in parapsychology tend to use more complex target material than the cards and dice used in the early investigations, partially to alleviate boredom on the part of the subjects and partially because they are thought to "more nearly resemble the conditions of spontaneous psi occurrences" (Burdick and Kelly, 1977, p. 109). These experiments fall under the general heading of "free-response" experiments, because the subject is asked to give a verbal or written description of the target, rather than being forced to make a choice from a small discrete set of possibilities. Various types of target material have been used, including pictures, short segments of movies on video tapes, actual locations, and small objects. Despite the more complex target material, the statistical methods used to analyze these experiments are similar to those for forced-choice experiments. A typical experiment proceeds as follows. Before conducting any trials, a large pool of potential targets is assembled, usually in packets of four. Similarity of targets within a packet is kept to a minimum, for reasons made clear below. At the start of an experimental session, after the subject is sequestered in an isolated room, a target is selected at random from the pool. A sender is placed in another room with the target. The subject is asked to provide a verbal or written description of what he or she thinks is in the target, knowing only that it is a photograph, an object, etc. After the subject's description has been recorded and secured against the potential for later alteration, a judge (who may or may not be the subject) is given a copy of the subject's description and the four possible targets that were in the packet with the correct target. A Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 16 properly conducted experiment either uses video tapes or has two identical sets of target material and uses the duplicate set for this pa-rt of the process, to ensure that clues such as fingerprints don't give away the answer. Based on the subject's description, and of course on a blind basis, the judge is asked to either rank the four choices from most to. least likely to have been the target, or to select the one from the four that seems to best match the subject's description. If ranks are used, the statistical analysis proceeds by summing the ranks over a series of trials and comparing the sum to what would be expected by chance. If the selection method is used, a "direct hit" occurs if the correct target is chosen, and the number of direct hits over a series of trials is compared to the number expected in a binomial experiment with p = .25. Note that the subjects' responses cannot be considered to be "random" in any sense, so probability assessments are based on the random selection of the target and decoys. In a correctly designed experiment, the probability of a direct hit by chance is .25 on each trial., regardless of the response, and the trials are independent. These and other issues related to analyzing free-response experiments are discussed by Utts (1989b). 4.2 The Psi Ganzfeld Experiments The ganzfeld procedure is a particular kind of free-response experiment utilizing a perceptual isolation technique originally developed by Gestalt psychologists for other purposes. Evidence from spontaneous case studies and experimental work had led parapsychologists to a model proposing that psychic functioning may be masked by sensory input and by inattention to internal states (Honorton, 1977). The ganzfeld procedure was specifically designed to test whether or not reduction of external "noise" would enhance psi performance. In these experiments, the subject is placed in a comfortable reclining chair in an Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 17 acoustically shielded room. To create a mild form: of sensory deprivation, the subject wears headphones through which white noise is played, and stares into a constant field of red light. This is achieved by taping halved translucent ping-pong balls over the eyes and then illuminating the room with red light. In the psi ganzfeld experiments, the subject speaks into a microphone and attempts to describe the ta:-,et material being observed by the sender in a distant room. At the 1982 Annual Me-e-ting of the Parapsychological Association, a debate took place over the degree to which the results of the psi ganzfeld experiments constituted evidence of psi abilities. Psychologist and ci-_'tic Ray Hyman and parapsychologist Charles Honorton each analyzed the results of all known psi ganzfeld experiments to date, and reached stri@ingly different conclusions. The debate continued with the publication of their arguments in separate articles in the March 1985 issue of @.he Journal of Parapsychology. Finally, in the December 1986 issue of the Journal of Parapsychology, Hyman and Honorton wrote a joint article in which they highlighted their agreernDats and disagreements, and outlined detailed criteria for future experiments. That same issue contained commentaries on the debate by ten other authors. The data base analyzed by Hyman and Honorton consisted of results taken from 34 reports written by a total of 47 authors. Honorton counted 42 separate experiments described in the reports, of which 28 reported enough information to determine the number of direct hits achieved. Twenty three of the studies (55 %) were classified by Honorton as having achieved statistical significance at .05. 4.3 The Vote-C.ounting Debate S Vote-counting is the term commonly used for the technique of drawing inferences about an experimental effect. by counting the number of significant versus non-significant studies of Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 18 the effect. Hedges and Olkin (1985) give a detailed analysiS of the inadequacy of this method, showing that it is more and more likely to make the wron-2 decision as the number of studies increases. While Hyman acknowledged that "vote-countirc, raises many problems (Hyman, 1985b, p.8)," he nonetheless spent half of his critique of :he ganzfeld studies showing why Honorton's count of 55% was wrong. Hyman's first complaint was that several of the studies contained multiple conditions, each of which should be considered as a separate study. -Using this definition he counted 80 studies (thus further reducing the sample sizes of the individual studies), of which 25 (31%) were "successful. " Honorton's response to this was to invite readers to examine the studies and decide for themselves if the varying conditions constituted separate experiments. Hyman next postulated that there was selection bias, so that significant studies were more likely to be repor-ted. He raised some important issues about how pilot studies may be terminated and not reported if they don't show significant results, or may' at least be subject to optional stopping, allowing the experimenter to determin@ the number of trials. He also presented a chi-square analysis that "suggests a tendency to report studies with a small sample only if they have significant results" (Hyman, 1985b, p. 14). but I have questioned his analysis elsewhere (Utts, 1986, p. 397). Honorton refuted Hyman's argument with four rejoinders (Honorton, 1985b, p.66). In addition to reinterpreting Hyman's chi-square analysis, Honorton pointed out that the Parapsychological Association has an official policy enc@ouraging the publication of non- significant results in its journals and proceedings, that a large number of reported garizfeld studies did not achieve statistical significance, and that there would have to be 15 studies in ithe Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 19 "file-drawer" for every one reported to cancel out the observed significant results. The remainder of Hyman's vote-counting analysis consisted of showing that the effective error rate for each study was actually much higher than the nominal 5 %. For example, each study could have been analyzed using the direct hit measure, the sum of ranks measure, or one of two other measures used for free-response analyses. Hyman carried out a simulation study that showed the true error rate would be .22 if "significance" was defined by requiring at least one of these four measures to achieve the .05 level. He suggested several other ways in which multiple testing could occur, and concluded that the effective error rate in each experiment was not the nominal .05, but rather was probably close to the 31 % he had determined to be the actual success rate in his vote-count. Honorton acknowledged that there was a multiple testing problem, but he had a two-fold response. First, he applied a Bonferroni correction and found that the number of significant studies (using his definition of a study) only dropped from 55 % to 45 %. Next, he proposed that a uniform index of success be applied to all studies. He used the number of direct hits, since it was by far the most commonly reported measure and was the measure used in the first published psi ganzfeld study. He then conducted a detailed analysis of the 28 studies reporting direct hits and found that 43 % were significant at .05 on that measure alone. Further, he showed that significant effects- were reported by six of the 10 independent investigators, and thus were not due to just one or two investigators or laboratories. He also noted that success rates were very similar for reports published in refereed journals and . those published in unrefereed monographs and abstracts. While Hyman's arguments identified issues such as selective reporting and optional Approved ForRelease 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 20 stopping that should be considered in any meta-analysis, the dependence of significance levels on sample size makes the vote-counting technique almost useless for assessing the magnitude of the effect. Consider for example the 24 studies where the direct hit measure was reported and the chance probability of a direct hit was .25, the most common type of study in the data base. (There were 4 direct hit studies with other chance probabilities and 14 that did not report direct hits.) Of the 24 studies, 13 (54 %) were "nonsignificant" at a .05, one-tailed. But if the 367 trials in these "failed replications" are combined, there are 106 direct hits, z = 1. 66, and p .0485, one tailed. This is reminiscent of the dilemma of Professor B in Section 3. Power is typically very low for these studies. The median sample size for the studies reporting direct hits was 28. If there is a real effect and it increases the success probability from the chance .25 to an actual .33 (a value whose rationale will be made clear below), the power for a study with 28 trials is only .181 (Utts, 1986). It should be no surprise that there is a "repeatability" problem in parapsychology. 4.4 Flaw Analysis and Future Recommendations The second half of Hyman's paper consisted of a Meta-Analysis of Flaws and Successful Outcomes" (1985b, p. 30), designed to explore whether or not various measures of success were related to specific flaws in the experiments. While many critics have argued that the results in parapsychology can be explained by experimental flaws, Hyman's analysis was the first to attempt to quantify the relationship between flaws and significant results. Hyman identified 12 potential flaws in the ganzfeld experiments, such as inadequate randomization, multiple teits used without adjusting the significance level (thus inflating the significance level from the nominal 5 %), and failure to use a duplicate set of targets for the Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 21 judging process (thus allowing possible clues such as fingerprints). Using cluster and factor analyses, the 12 binary flaw variables were combined into three new which Hyman named General Security, Statistics and Controls. Several analyses were then conducted. The one reported with th-I most detail is a factor analysis utilizing 17 variables for each of 36 studies. Four factors eme--ged from the analysis. From these, Hyman concluded that security had increased over the yews, that the significance level tended to be inflated the most for the most complex studies, and t@,-,at bmh effect size and level of significance were correlated with the existence of flaws. FoUo,,,;-ing his factor analysis, Hyman picked the three flaws that seemed to be most highly correlated with success, which were inadequate attention to both randomization and documentation, and the potential for ordinary communication betwee-n the sender and receiver. A regression equation was then computed using each of the three flaws a:, dummy variables, and the effect size for the experiment as the dependent variable. From this equation, Hyman concluded that a study without these three flaws would be predicted to have a hit rate of 27%. He concluded that this is "well within the statistical neighborhood of the -^5 % chance rate" (ibid, p. 37), and thus "the ganzfeld psi data base, despite initial impressions, is inadequate either to support the contention of a repeatable study or to demonstrate the reality of psi" (ibid p. 38). Honorton discounted both Hyman's flaw classification and his analysis. He did not deny that flaws existed, but objected that Hyman's analysis was faulty and impossible to interpret. Honorton asked psychometrician David Saunders to write an Appendix, to his article, evaluating Hyman's analysis. Saunders first criticized Hyman's use of a factor analysis with 17 variables (many of which were dichotomous) and only 36 cases, and concluded Chat "the entire analysis Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 22 is meaningless" (Saunders, 1985, p.87). He then noted that Hyman's choice of the three flaws to include in his regression analysis constituted a clear case of multiple analysis, since there were 84 possible sets of three that could have been selected (out of nine potential flaws), and HN man chose the set most highly correlated with effect size. Again, Saunders concluded that "any interpretation drawn from [the regression analysis] must be regarded as meaningless" (ibid, p. 88). Hyman's results were also contradicted by Harris and Rosenthal (1988b) in an analysis requested by Hyman in his capacity as Chair of the National Academy of Sciences' Subcommittee on Parapsychology. Using Hyman's flaw classifications and a multivariate analysis, Harris and Rosenthal concluded that "Our analysis of the effects of f1wX's on study outcome lends no support to the hypothesis that ganzfeld research results are a significant function of the set of flaw variables" (1988b, p. 3). Hyman and Honorton were in the process of preparing papers for a second round of debate when they were invited to lunch together at the 1986 Meeting of the Parapsychological Association. They discovered that they were in general agreement on several major issues, and decided to coauthor a "Joint Communique" (Hyman and Honorton, 1986). It is clear from their paper that they both thought it was more important to set the stage for future experimentation than to continue the technical -arguments over the current data base. In the abstract to their paper they wrote: "We agree that there is an overall significant effect in this data base that cannot reasonably be explained by selective reporting or multiple analysis. We continue to differ over the degree to which the effect constitutes evidence for psi, but we agree that the final verdict awaits the outcome of future experiments mconducted by a broader range of investigators and according to more stringent standards" (Ibid, p. 351). Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003104118: CIA-RDP96-00789ROO3100030001-4 23 The paper then outlined what these standards should be. They included controls against any kind of sensory leak-age, thorough testing and documentation of randomization methods used, better reporting of judging and feedback protocols, control for multiple analyses, and advance specification of number of trials and type of experiment. Indeed, any area of research could benefit from such a careful list of procedural recommendations. 4.5 Rosenthal's Meta-Analysis The same issue of the Journal of Parapsychology in which the Joint Communique appeared also carried commentaries on the debate by 10 separate authors. In his commentary, psychologist Robert Rosenthal, one of the pioneers of meta-analysis in psychology, sumi-narized the aspects of Hyman's and Honorton's work that would typically be included in a meta-analysis (Rosenthal, 1986). It is worth reviewing Rosenthal's results so that they can be used as a basis ofcomparison for the more recent psi ganzfeld studies reported in Section 5. Rosenthal, like Hyman and Honorton, focused only on the 28 studies for which direct hits were known. He chose to use an effect size measure called Cohen's h, which is the difference between the arcsin transformed proportions of direct hits that were observed and expected: h=2x(arcsin@Fp-arcsinrp) One advantage of this measure over the difference in raw proportions is that can be used to compare experiments with different chance hit rates. If the observed and expected numbers of hits were identical, the effect size would be zero. Of the 28 studies, 23 (82%) had effect sizes greater than zero, with a median effect size Approved For Release 2003/04118: CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 24 of .32 and a mean of .28. These correspond to direct hit rates of .40 and .38 respectively, when 25 is expected by chance. A 95 % confidence interval for the true effect size is from . I I to .45, corresponding to direct hit rates of from .30 to .46 when chance is .25. A common technique in meta-analysis is to calculate a "combined z, " found by summing the individual z scores and dividing by the square root of the number of studies. The result should have a standard normal distribution if each z score has a standard normal distribution. For the ganzfeld studies, Rosenffial reported a combined z of 6.60 with ap-value of 3.37 x 10". He also reiterated Honorton's file-drawer assessment by calculating that there would have to be 423 studies unreported to nega.,e the significant effect in the 28 direct hit studies. Finally, Rosenthal aclatowledged that because of the flaws in the data base and the potential for at least a small file drawer effect, the true average effect size was probably closer to .18 than .28. He concluded, "Thus, when the accuracy rate expected under the null is 1/4, we might estimate the obtained accuracy rate to be about 1/3" (Ibid, p. 333). This is the value used for the earlier power calculation. It is worth mentioning that Rosenthal was commissioned by the National Academy of Sciences to prepare a background paper to accompany its 1988 report on parapsychology. That paper (Harris and Rosenthal, 1988a) contained much of the same analysis as his commentary summarized above. Ironically, the discussion of the ganzfeld work in the National Academy Report focused on Hyman's 1985 analysis, but never mentioned the work it had commissioned Rosenthal to perform, which contradicted the final conclusion in the report. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 25 5. A META-ANALYSIS OF RECENT GANZFELD EXPERIMENTS After the initial exchange with Hyman at the 1982 Parapsychological Association Meeting, Honorton and his colleagues developed an automated ganzfeld experiment, that was designed to eliminate the methodological flaws identified by Hyman. The execution and reporting of the experiments followed the detailed guidelines agreed upon by Hyman and Honorton. Using this "autocanzfeld" experiment, eleven experimental series were conducted by eight experimenters between February 1983 and September 1989, when the equipment had to be dismantled due to lack of funding. In this section the results of these experiments are summarized and compared to the earlier ganzfeld studies. Much of the information is derived from Honorton et al (1990). 5.1 The Automated Ganzfeld Procedure Like earlier ganzfeld studies, the "autoganzfeld" experiments require four participants. The first is the Receiver (R), who attempts to identify the target material being observed by the Sender (S). The Experimenter (E) prepares R for the task, elicits the response from R, and supervises R's judging of the response against the four potential targets. (Judging is double- blind; E does not know which is the correct target.) The fourth participant is the lab assistant (LA), whose only task is to instruct the computer to randomly select the target. No one involved in the experiment. knows the identity of the target. Both R and S are -sequestered in sound-isolated, electrically shielded rooms. R is prepared as in earlier ganzfeld studies, with white noise and a field of red light. In a Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04118 : CIA-RDP96-00789ROO3100030001-4 26 non-adjacent room, S watches the target material on a television and can hear R's target description ("mentation") as it is being given. The mentation is also tape-recorded. The judging process takes place immediately after the 30 minute sending period. On a TV monitor in the isolated room, R views the four choices from the target pack that contains the actual target. R is asked to rate each one according to how closely it matches the ganzfeld mentation. The ratings are converted to ranks, and if the correct target is ranked first, a direct hit is scored. The entire process is automatically recorded by the computer. The computer then displays the correct choice to R as feedback. There were 160 pre-selected targets, used with replacement, in ten of the eleven series, They were arranged in packets of 4, and the decoys for a given target were always the remaining three in the same set. Thus, even if a particular target in a set were consistently favored by R's, the probability of a direct hit under the null hypothesis would iernain at 1/4. Popular targets should be no more likely to be selected by the computer's random number generator than any of the others in the set. The selection of the target by the computer is the only source of randomness in these experiments. This is an important point, and one that is often misunderstood. (See Utts, 1989b for elucidation.) Eighty of the targets were "dynamic," consisting of scenes from movies, documentaries and cartoons; and 80 were "static", consisting of photographs, art prints, and advertisements. The four targets within each set were all of the same type. Earlier studies indicated that dynamic targets were more likely to produce successful results) and one of the goals of the new experiments was to test that theory. The randomization procedure used to select the target and the order of presentation for Approved For Release -2003/04/18 CIA-RDP96-00789ROO31.00030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 27 judging was thoroughly tested before and during the experiments. A detailed description is given by Honorton et al (1990. p. 118-120). Three of the eleven series were pilot series, five were formal series with novice receivers, and three were formal series with experienced receivers. Thl- last series with experienced receivers was the only one that did not use the 160 targets. Instead, it used only one set of four dynamic targets in which one target had previously received several first place .ranks, and one had never received a first place rank. The receivers, none of whom had had prior exposure to that target pack, were not aware that only one target pack was being used. They each contributed one session only to the series. This will be called the "special series" in what follows. Except for two of the pilot series, numbers of trials were planned in advance for each series. Unfortunately, three of the formal series were not yet completed when the funding ran out, including the special series, and one pilot study with advance planning was terminated early when the experimenter relocated. There were no unreported trials during the six year period under review, so there was no "file-drawer". Overall, there were 183 R's who contributed only one trial and 58 who contributed more than one, for a total of 241 partidpants and 355 trials. Only twenty three R's had previously participated in ganzfeld experiments and 194 R's (81 %) had never participated in any parapsychological research. 5,.2 Results While acknowledgifig that no probabilistic conclusions can be drawn from qualitative data, Honorton et al (1990), included several examples of session excerpts that R's identified as Approved For Release 2003/04/18: CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 28 providing the basis for their target rating. To give a flavor for the dream-like quality of the mentation and the amount of information that can be lost by only assigning a rank, the first example is reproduced here. The target Nvas a painting by Salvador Dali called "Christ Crucified." The correct target received a first place rank. The part of the mentation R used to make this assessment read: "... I think of guides, like spirit guides, leading me and I come into a court with a king. It's quiet.... It's like heaven. The king is something like. Jesus. Woman. Now I'm just sort of surnmersaulting through heaven.... Brooding..... Aztecs, the Sun God .... High priest.... Fear.... Graves. Woman. Prayer.... Funeral.... Dark. Death.... Souls .... Ten Commandments. Moses ...... (Ibid, p. 120). Over all eleven series there were 122 direct hits in the 355 trials, for a hit rate of 34.4 % (exact binomial p-value = .00005) when 25 % were expected by chance. Cohen's h is .20, and a 95 % confidence interval for the overall hit rate is from .30 to .39. This calculation assumes, of course, that the probability of a direct hit is constant and independent across trials, an 2ssumption that may be questionable except under the null hypothesis of no psi abilities. Honorton et al also calculated effect sizes for each of the eleven series and each of the eight experimenters. All but one of the series (the first novice series) had positive effect sizes, as did all of the experimenters. The special series with experienced R's had an exceptionally high effect size with h .81, corresponding to 16 direct hits out of 25 trials (64%), but the remaining series and the experimenters had relatively homogeneous effect sizes given the amount of variability expected by chance. If the special series is removed, the overall hit rate is 32.1 %, h = .16. Thus, the positive effects are not due to just one series or one experimenter. Seventy one of the 218 trials contributed by novices w :ere direct hits (32.5 %, h = .17), Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003104118: CIA-RDP96-00789ROO3100030001-4 29 compared with 51 hits in the 137 trials by those with prior ganzfeld experience (37%, h = .26). The hit rates and effect sizes were 31 % (h = .1 4) for the combined pilot series, 32.5 % (h = 17) for the combined formal novice series, and 11.5 % (h = .35) for the combined experienced series. The last figure drops to 31.6% if the outlier series is removed. Finally, without the outlier series the hit rate for the combined series where all of the planned trials were completed ,,@-as 31.2 % (h = .14) while it was 35 % (h = .2:) for the combined series that were terminated early. Thus, optional stopping cannot account for the positive effect. There were two interesting comparisons that had been suggested by earlier work and were preplanned in these experiments. The first was to compare results for trials with dynamic targets with those for static targets. In the 190 dynamic target sessions there were 77 direct hits (40%, h = .32) and for the static targets there were 45 hits in 165 trials (27%, h = .05), thus indicating that dynamic targets produced far more successful results. The second comparison of interest was whether or not the sender was a friend of the receiver. This was a choice the receiver could make. If he or she did not bring a friend, a lab member acted as sender. There were 211 trials with friends as senders (some of whom were also lab staff), resulting in 76 direct hits (36%, h = .24). Four trials used no sender. The remaining 140 trials used non-friend lab staff as senders and resulted in 46 direct hits (33 %, h .18). Thus, trials with friends as senders were slightly more successful than those without. Consonant with the definition of replication based on consistent effect sizes, it is informative to compare the autoganzfeld experiments with the direct hit studies in the previous ,data base. The overall success rates are extremely sin-tilar. The overall direct hit rate was 34.4% for the autoganzfeld studies and was 39% for the comparable direct hit studies in the Approved For Release 2003/04/18: CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 30 earlier meta-analysis. Rosenthal's (1986) adjustment for flaws had placed a more conservative estimate at 33 %, very close to the observed 34.4 % in the new studies. One limitation of this work is that the autoganzfeld studies, while conducted by eight experimenters, all used the same equipment in the same laboratory. Unfortunately, the level of funding available in parapsychology and the cost in time and. equipment to conduct proper experiments make it difficult to amass large amounts of data across laboratories. Another auto-anzfeld laboratory is currently being constructed at the University of Edinbur h in C, 9 Scotland, so interlaboratory comparisons may be possible in the near future. Based on the effect size observed to date, large samples are needed to achieve reasonable power. If there is a constant effect across all trials, resulting in 33 % direct hits when 25 % are expected by chance, to achieve a one tailed significance level of .:05 with 95 % probability would require 345 sessions. We end this section by returning to the aspirin and heart attack example in Section 3, and expanding a comparison noted by Atkinson et al (1990, p. 237). Computing the equivalent of Cohen's h for comparing observed heart attack rates in the aspirin and placebo groups results in h = .068. Thus, the effect size observed in the ganzfeld data base is triple the much- publicized effect of aspirin on heart attacks. 6. OTHER META-ANALYSES IN PARAPSYCHOLOGY Four additional meta-analyses have been conducted in various areas of parapsychology since the original ganzfeld meta-analyses were reported. Three. of the four analyses focused on evidence of psi abilities, while the fourth examined the relationship between extraversion and Approved For Release 2003/04/18 : CIA-RDP96-00789RO03 100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 31 psychic functioning. In this section, each of the four analyses will be briefly summarized. There are only a handful of English-language journals and proceedings in parapsychology, so retrieval of the relevant studies in each of the four cases was simple to accomplish by searching those sources in detail and by searching other bibliographic data bases for keywords. Each analysis included an overall summary, an analysis of the quality of the studies versus the size of the effect, and a "file-drawer" analysis to determine the possible number of unreported studies. Three of the four also contained comparisons across various conditions. 6.1 Forced-choice Precognition Experiments Honorton and Ferrari (1989) analyzed forced-choice experiments conducted from 1935 to 1987, in which the target material was randomly selected after the subject had attempted to predict.,what it would be. The time delay in selecting the target ranged from under a second to one year. Target material included items as diverse as ESP cards and automated random number generators. Two investigators, S.G. Soal and Walter J. Levy, were not included because some of their work has been suspected to be fraudulent. Overall Results. There were 309 studies reported by 62 senior authors, including more than 50,000 subjects and nearly two million individual trials. Honorton and Ferrari used z IVn as the measure of effect size (ES) for each study, where n was the number of Bemoulli trials in the study. They reported a mean ES of 0.020, and a mean z-score of 0.65 over all studies. They also reported a combined z of 11.41, p = 6.3 x 10-21 . Thirty percent (92) of the studies were statistically significant at ot = .05. The mean ES per investigator was 0.033, and the significant results were not due to just a few investigators. I Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 32 Quality. Eight dichotomous quality measures were assigned to each study, resulting in possible scores from zero for thelowest quality, to eight for the highest. They included features such as adequate randomization, preplanned analysis, and automated recording of the results. The correlation between study quality and effect size was 0.081, indicating a slight tendency for higher quality studies to be more successful, contrary to claims by critics that the opposite would be true. There was a clear relationship between quality and year of publication, presumably because over the years experimenters in parapsychology have responded to suggestions from critics for improving their methodology. File-drawer. Following Rosenthal (1984), the authors: calculated the "fail-safe AP' indicating the number of unreported studies that would have to besitting in file-drawers in order to negate the significant effect. They found N = 14,268, or a ratio of 46 unreported studies for each one reported, They also followed a suggestion by Dawes et al (1984) and computed the mean z for all studies with z > 1. 65. If such studies were a random sample from the upper 5 % tail of a N(O, 1) distribution, the mean z would be 2.06. In t ;his case it was 3.61. They concluded that selective reporting could not explain these results. Comparisons. Four variables were identified that appeared to have a systematic relationship to study outcome. The first was that the 25 studies using subjects selected on the basis of good past performance were more successful than the 223 using unselected subjects, with mean effect sizes of .051 and .008, respectively. Second, the 97 studies testing subjects individually were more successful than the 105 studies that used group testing; mean effect sizes were .021 and .004, respec@tively. Timing of feedback was the third moderating variable, but information was only available for 104 studies. The 15 studies that never told the subjects what Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-.RDP96-00789ROO3100030001-4 33 the targets were had a mean effect size of -.001. Feedback after each trial produced the best results, the mean ES for the 47 studies was 035. Feedback after each set of trials resulted in mean ES of .023 (21 studies), while delayed feedback (also 21 studies) yielded a me-an ES of only .009. There is a clear ordering; as the gap between time of feedback and time of the actual guesses decreased, effect sizes increased. The fourth variable was the time inten,al between the subject's guess and the actual target selection, available for 144 studies. The best results were for the 31 studies that generated targets less than a second after the guess (mean ES = .045), while the worst were for the 7 studies that delayed target selection by at least a month (mean ES = .001). The mean effect sizes showed a clear trend, decreasing in order as the time interval increased from minut-es to hours to days to weeks to months. 6.2, Attempts to Influence Random Physical Systems Radin and Nelson (1989) examined studies designed to test the hypothesis that "The statistical output of an electronic RNG [random number generator] is correlated with observer intention in accordance with prespecified instructions" (p. 1502). These experiments typically involve RNGs based on radioactive decay, electronic noise, or pseudorandom number sequences seeded with tr-ue random sources. Usually the subject is instructed to try to influence the results of a string of binary trials by mental intention alone. A typical protocol would ask a subject to press a button (thus starting the collection of a fixed-length sequence of bits), and then try to influence the random source to produce more zeroes or more ones. A run might consist of three successive button presses, 6ne each in which the desired result was more zeroes or more ones, and one as a control with no conscious intention. A z score would then be computed for each Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 34 button press. The 832 studies in the analysis. were conducted from 1959 to 1987, and included 235 ti control" studies in which the output of the RNGs were recorded but there was no conscious intention involved. These were usually conducted before and during the experimental series, as tests of the RNGs. Results. The effect size measure used was again z IVn, where z was positive if more bits of the specified type were achieved. The mean effect size for control studies was not significantly different from zero (-1.0 x 10'). The mean effect size for the experimental studies was also very small, 3.2 x 10', but it was significantly higher tha: n the mean ES for the control studies (z = 4. 1). Quality. Sixteen quality measures were defined and assigned to each study, under the four general categories of procedures, statistics, data, and the RNG device. A score of 16 reflected the highest quality. The authors regressed mean effect @size on mean quality for each investigator, and found a slope of 2.5 x 10-1 with standard error of 3.2 x 1.0', indicating little relationship between quality and outcome. They also calculated: a weighted mean effect size, using quality scores as weights, and found that it was very similar to the unweighted mean ES. They concluded that "differences in methodological quality are not.significant predictors of effect size" (p. 1507). File-drawer. Radin and Nelson used several methods for estimating the number of unreported studies (p. 1508-10). Their estimates ranged from 200 to 1000 based on models assuming that all significant studies were reported. They also 61culated the fail-safe N to be 54,000. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO31.00030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 35 6.3 Attempts to Influence Dice Radin and Ferrari (1991) examined 148 studies, published from 1935 to 1987, designed to test whether or not. consciousness can influence the results of tossing dice. The also found 31 "control" studies in which no conscious intention was involved. Results. The effect size measure used was z IA where z was based on the number of throws in which the die landed with the desired face (or faces) up, in n throws. The weighted mean ES for the experimental studies was 0.0122 with a standard error of 0.00062; for the control studies the mean and standard error were 0.00093 and 0.00255, respectively. Weights for each study were determined by quality, giving more weight to high quality studies. Combined z scores for the experimental and control studies were reported by Radin and Ferrari to be 18.2 and 0. 18, respectively. Quality. Eleven dichotomous quality measures were assigned, ranging from automated recording to whether or not control studies were interspersed with the exPlUimental studies. The final quality score for each study combined these with information on method of tossing the dice, and with source of subject (defined below). A regression of quality score versus effect size resulted in a slope of -. 002, with a standard error of . 00 11. However, when effect sizes were weighted by sample size there was a significant relationship between quality and effect size, leading Radin and Ferrari to conclude that higher quality studies produced lower weighted effect sizes. File-drawer. Radin and Ferrari calculated Rosenthal's fail-safe N for this analysis to be 17,974. Using the assumption that all significant studies were reported, they estimated the number of unreported studies to be 1, 152. As a final assessment, they compared studies 11 Approved For. Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 36 published before and after 1975, when the Journal of Parapsychology adopted an official policy of publishing nonsignificant results, They concluded, based. on that analysis, that more nonsignificant studies were published after 1975, and thus "We must consider the overall (1935- 1987) data base as suspect with respect to the filedrawer proble m. Comparisons. Radin and Ferrari noted that there was bias in both the experimental and control studies across die face. Six was the face most likely to come up, consistent with the observation that it has the least mass. Therefore, they examin ed results for the subset of 69 studies in which targets were evenly balanced among the six faces. They still found a significant effect, with mean and standard error for effect size of 8.6 x 10 and 1.1 x 10', respectively. The combined z was 7.617 for these studies. They also compared effect sizes across types of subjects used in the studies, categorizing them as unselected, experimenter and other subjects, experimenter as sole subject, and specially selected subjects. Like Honorton and Ferrari (1989), they found the highest mean ES for studies with selected subjects; it was approximately .02, more than twice that for unselected subjects. 6.4 Extraversion and ESP Performance Honorton, Ferrari and Bern (1990) conducted a meta-analysis to examine the relationship between scores on tests of extraversion and scores on psi-related tasks. They found 60 studies by 17 investigators, conducted from 1945 to 1983. Results. The effect size measure used for this analysis wa;s the correlation between each subject's extraversion score and ESP score. A variety of measures had been used for both scores across studies, so various correlation coefficients were used. Nonetheless, a stem and leaf diagram of the correlations showed an approximate bell 'shape with mean and standard Approved For Release 2003/04/18 CIA-RDP96-00789RO031-00030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 37 deviation of .19 and .26, respectively, and with an additional outlier at r = .91. Honorton et al reported that when weighted by degrees of freedom, the weighted mean r was .14, with a 95% confidence interval covering .10 to .19. Forced-choice versus Free-response Results.' Because forced-choice and free-response tests differ qualitatively, Honorton et al chose to examine their relationship to extraversion separately. They found that for free-response studies there was a significant correlation between extraversion and ESP scores, with mean r = .20 and z = 4.46. Further, this effect was homogeneous across both investigators and extraversion scales. For forced-choice studies, there was a significant correlation between ESP and extraversion, but only for those studies that reported the ESP results to the subjects before measuring extraversion. Honorton et al speculated that the relationship was an artifact, in which extraversion scores were temporarily inflated as a result of positive feedback on ESP performance. Confirnwtion with New Data. Following the extraversion/ESP meta7analysis, Honorton et al attempted to confirm the relationship using the autoganzfeld data base. Extraversion scores based on the Myers-Briggs Type Indicator were available for 221 of the 241 subjects who had participated in autoganzfeld studies. The correlation between extraversion scores and ganzfeld rating scores was r = .18, with a 95 % confidence interval from .05 to .30. This is consistent with the mean correlation of r = .20 for free-response experiments, determined from the meta-analysis. These correlations indicate that extraverted sOjects, can produce higher scores in free-response ESP tests. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 38 7. CONCLUSIONS Parapsychologists often make a distinction between "proof-oriented research" and "process-oriented research." The former is typically conducted to test the hypothesis that psi abilities exist, while the latter is designed to answer questions about how psychic functioning works. Proof-oriented research has dominated the literature in parapsychology. Unfortunately, many of the studies used small samples and would thus be nonsignificant even if a moderate- sized effect exists. The recent focus on meta-analysis in parapsychology has revealed that there are small but consistently nonzero effects across studies, experimenters, and; laboratories. The size of the effects in forced-choice studies appear to be comparable to those reported in some medical studies that had been heralded as breakthroughs. (See Section:5, and Honorton and Ferrari, 1989, p. 301.) Free-response studies show effect sizes of far greater magnitude. A promising direction for future process-oriented research is to examine the causes of individual differences in psychic functioning. The ESP/extraversion meta-analysis is a step in that direction. In keeping with the idea of individual differences, Bayes and empirical Bayes methods would appear to make more sense than the classical inference methods commonly used, since they would allow individual abilities and beliefs to be modelled. Jeffreys (1990) reported a Bayesian analysis of some of the RNG experiments, and showed-, that conclusions were closely tied to prior beliefs even though hundreds of thousands of trialswere available. It may be that the nonzero effects observed in the meta-analyses can be explained by something other than ESP, such as shortcomings in our unde,rstanding of randomness and Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 39 independence. Nonetheless, there is an anomaly that needs an explanation. As I have argued elsewhere (Utts, 1987) research in parapsychology should receive more support from the scientific community. If ESP does not exist, there is little to be lost by erring in the direction of further research; which may in fact uncover other anomalies. If ESP does exist there is much to be lost by not doing process-oriented research, and much to be gained by discovering how to enhance and apply these abilities to important world problems. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 40 REFERENCES Atkinson, R.L., Atkinson, R.C., Smith, E.E. and Bern, D.J. (1990). Introduction to Psychology, 10th Ed. Harcourt Brace Jovanovich, San D :iego. Beloff, J. (1985). Research strategies for dealing with unstable phenomena. In ne Repeatability Problem in Parapsychology (B. Shapin and L. Coly. eds.) 1-21. Parapsychology Foundation, New York. Blackmore, S.J. (1985). Unrepeatability: Parapsychology's only finding. In The Repeatability Problem in Parapsychology (B. Shapin and L. Coly. eds.) 183-206. Parapsychology Foundation, New York, Burdick, D.S. and Kelly, E.F. (1977). Statistical methods in parapsychological research. In Handbook ofParapsychology (B.B. Wolman, ed.) 81-130. Van Nostrand Reinhold, New York. Camp, B.H. (1937). (Statement in Notes Section.) Journal of Parapsychology 1305. Cohen, J. (1990). Things I have learned (so far). American Psychologist 45 1304-1312. Coover, J.E. (1917). Experiments in Psychical Research at Leland Stanford Junior University. Stanford University, Stanford, CA. Dawes, R.M., Landman, J. and Williams, J. (1984). Reply to Kurosawa. American Psychologist 3974-75. Diaconis, P. (1978). Statistical problems in ESP research. Science 201 131-136. Dommeyer, F. C. (1975). Psychical Research at Stanford University. Journal of Parapsychology 39 173-205. Druckman, D. and Swets, J.A., Eds. (1988). Enhancing Human Performance: Issues, Theories, and Techniques. National Academy Press, Washington, DC. Edgeworth, F.Y. (1885). The calculus of probabilities applied to psychical research. Proceedings of the Society for Psychical Research 3 190-199. Edgeworth, F.Y. (1886). The calculus of probabilities applield to psychical. research Il. Proceedings of the Society for Psychical Research 4 189-208. Feller, W.K. (1968). An Iniroduction to Probability neory and Its Applications, Volume 1, 3rd Ed. Wiley, New York. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 41 Feller, W.K. (1940). Statistical aspects of ESP. Journal of Parapsychology 4 271-297. Fisher, R.A. (1924). A method of scoring coincidences in tests with playing cards. Proceedings of the Society for Psychical Research 34 181-185. Fisher, R. A. (1929). The statistical method in psychical research. Proceedings of the Societyfor Psychical Research 39 189-192. Gallup, G.H. Jr. and Newport, F. (1991). Belief in paranormal phenomena among adult Americans. Skeptical Inquirer 15 137-146. Gardner, M,J. and Altman, D.G. (1986). Confidence intervals rather than p-values: estimation rather thah hypothesis testing. British Medical Journal 292 746-750. Gilmore, J.B. (1989). Randomness and the search for psi. Journal of Parapsychology 53 309- 340. Gilmore, J.B. (1990). Anomalous significance in pararandom and psi-free domains. Journal of Parapsychology 54 53-58. Greeley, A. (1987). Mysticism goes mainstream. American Health 7 47-49. Greenhouse, J.B. and Greenhouse, S.W. (1988). An aspirin a day ... ? Chance 124-3 1. Greenwood, J.A. and Stuart, C.E. (1940). A review of Dr. Feller's critique. Journal of Parapsychology 4 299-319. Hansel, C.E.M. (1980). ESP and Parapsychology: A Critical Re-evaluation. Prometheus Books, Buffalo. Harris, M.J. and Rosenthal, R. (1988a). Interpersonal Expectancy Effects and Human Per fionnance Research. National Academy Press, Washington DC. HarTis, M.J. and Rosenthal, R. (1988b). Postscript to Interpersonal Expectancy Effects and Human Performance Research. National Academy Press, Washington DC. Hedges, L. V. and Olkin, 1. (1985). Statistical Methodsfor Meta-Analysis. Academic Press, Inc., Orlando, FL. Honorton, C. (1977). Psi and internal attention states. In Handbook of Parapsychology (B.B. Wolman, ed.) 435-472. Van Nostrand Reinhold, New York. Honorton, C. (1985a). How to evaluate and improve the replicability of parapsychological effects. In 7he Repeatability Problem in Parapsychology (B. Shapin and L. Coly. eds.) Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 42 238-255. Parapsychology Foundation, New York. Honor-ton, C. (1985b). Meta-analysis of psi ganzfeld research: A response to Hyman. Journal of Parapsychology 49 51-9 1. Honorton, C., Berger, R.E., Vargoglis, M.P., Quant, M., Derr, P., Schechter, E.I. and Ferrari, D.C. (1990). Psi communication in the ganzfeldi experiments with an automated testing system and a comparison with a meta-analysis. of earlier studies. Journal of Parapsychology 54 99-139. Honorton, C. and Ferrari, D.C. (1989). "Future telling": A:meta-analysis of forced-choice precognition experiments, 1935-1987. Journal of Parapsychology 53 281-308. Honorton, C., Ferrari, D.C., and Bern, D.J. (1990). Extraversion and FISP Performance: A meta-analysis and a new confirmation. Proceedings of the Annual Meeting of the Parapsychological Association. Hyman, R. (1985a). A critical overview of parapsychology In A Skeptic's Handbook of Parapsychology (P. Kurtz, ed.) 1-96. Prometheus Books, Buffalo. Hyman, R. (1985b). The ganzfeld psi experiment: A critical appraisal. Journal of Parapsychology 49 3-49. Hyman, R. and Honorton, C. (1986). Joint communiqu6: The psi ganzfeld controversy. Journal of Parapsychology 50 351-364. Iversen, G.R, Longeor, W.H., Mosteller, F., Gilbert, J.P., a.nd Youtz, C. (1971). Bias and runs in dice throwing and recording: A few million throws. Psychometrika 36 1-19. Jeffreys, W.H. (1990). Bayesian analysis of random event generator data. Journal of Scientific Exploration 4 153-169. Lindley, D.V. (1957). A statistical paradox. Biometrika 44 187-192. Mauskopf, S.H. and MeVaugh, M. (1979). The Elusive Science: Origins of Experimental Psychical Research. The Johns Hopkins University Press, Baltimore. McVaugh, M.R. and Mauskopf, S.H. (1976). J.B. Rhine's Extrasensory Perception and its background in psychical. research. Isis 67 161-189. Neuliep, J.W. (Ed.) (199Q). Handbook of replication research in the t>ehavioral and social sciences. Journal of Social Behavior and Personality (Special Issue) 5(4). Office of Technology Assessment (1989). Report of a workshop on experimental Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 43 parapsychology. Joumal of the American Society for Psychical Research 83 317-339. Palmer, J. (1989). A reply to Gilmore. Journal of Parapsychology 53 341-344. Palmer, J. (1990). Reply to Gilmore: Round two. Journal of Parapsychology 54 59-61. Palmer, LA., Honorton, C. and Utts, J. (1989). Reply to the National Research Council study on parapsychology. Joumal of the American Society for Psychical Research 83 31-49. Radin, D.I. and Ferrari, D.C. (1991). Effects of consciousness on the fall of dice: A meta- analysis. Joumal of Scientific Exploration 5 (to appear). Radin, D.I. and Nelson, R.D. (1989). Evidence for consciousness-related anomalies in random physical systems. Foundations of Physics 19 1499-1514. Rao, K.R. (1985). Replication in conventional and controversial sciences. In The Repeatability Problem in Parapsychology (B. Shapin and L. Coly. eds.) 22-41. Parapsychology Foundation, New York. Rhine, J.B. (1934). Extrasensory Perception. Boston Society for Psychical Research, Boston. (Reprinted by Branden Press in 1964). Rhine, J.B. (1977). History of experimental studies. In Handbook of Parapsychology (B.B. Wolman, ed.) 25-47. Van Nostrand Reinhold, New York. Richet, C. (1884). La suggestion mentale et le calcul des probabilites. Revue Philosophique 18 608-674. Rosenthal, R. (1984). Meta-Analytic Procedures for Social Research. Sage, Beverly Hills. Rosenthal, R. (1986). Meta-analytic procedures and the nature of replication: The ganzfeld debate. Joumal of Parapsychology 50 315-336. .Rosenthal, R. (1990a). How are we doing in soft psychology? American Psychologist 45 775- 777. 1 Rosenthal, R. (1990b). Replication in behavioral research. Joumal of Social Behavior and Personality 5 1-30. Saunders, D.R. (085). On Hyman's factor analysis. Journal of Parapsychology 49 86-88. Shapin, B. and Coly, L. (Eds.) (1985). The Repeatability Problem in Parapsychology. Parapsychology Foundation, New York. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 44 Spencer-Brown, G. (1957). Probability and Scientific Inference. Longmans Green, London and New York. Stuart, C.E. and Greenwood, J. A. (1937). A review of criticisms: of the mathematical evaluation of ESP data. Journal of Parapsychology 1295-304. Tversky, A. and Kahneman, D. (1982). Belief in the law of small numbers. In D. Kahneman, P. Slovic and A. Tversky (Eds), Judgment under uncertainty: Heuristics and biases. Cambridge University Press, Cambridge. Utts, J. (1986). The ganzfeld debate: A statistician's perspective.: Journal of Parapsychology 50 395-402. Utts, J. (1987). Psi, statistics, and society. Behavioral and Brat .n Sciences 10 615-616. Utts, J. (1988). Successful replication versus statistical significance. Journal of Parapsychology 52 305-320. Utts, J. (1989a). Randomness and randomization tests: A reply to Gilmore. Journal of Parapsychology 53 345-351. Utts, J. (1989b). Analyzing free-response data - a progress report. To appear in Psi Research Methodology: A Re-examination (L. Coly, ed.). Parapsychology Foundation, New York. Wilks, S. S. (1965). N.Y. Statistician 16 (nos. 6 and 7). Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04118 : CIA-RDP96-00789ROO3100030001-4 III MAIN-STREAM PUBLICATIONS one measure of the acceptance of anomalous mental phenomena as a valid area for investigation is the degree to which research papers appear in the main-stream scientific literature. The reports in this section have been selected because they are a representative sample of such papers. The number that appears in the upper right-hand corner of the first page for each publication is keyed to the following descriptions: 9. Targ, R. and Puthoff, H. E., "Information Ttansmission Under Conditions of Sensory Shielding," Nature, Vol. 252, pp. 602-607, (October, 1974). Targ and Puthoff describe a series of experiments with selected individuals, including Mr. Uri Geller, and introduce an anomalous cognition technique called remote viewing. The paper also includes a pilot experiment to investigate the effects of anomalous cognition on the alpha rhythms in the brain. 10. Puthoff, H. E. and Thrg, R., "A Perceptual Channel for Information Transfer over Kilometer Distances: Historical Perspective and Recent Research," Proceedings of the IEEE, Vol. 64, No. 3, pp, 329-354, (March, 1976). Puthoff and Targ.provide a historical review of the pertinent literature and describe over 50 remote viewing (i.e., anomalous cognition) trials. The paper also includes representative examples of remote viewing. 11. Jahn, R. G., "The Persistent Paradox of Psychic Phenomena: An Engineering Perspective," Invited Paper, Proceedings of the IEEE, Vol. 70, No. 2, pp. 136-170, (February, 1982). Jahn describes a replication of remote viewing and extends the distance to over 10,000 kilometers. In addition to an independent overview of parapsychology, Jahn also includes descriptions of a number of anomalous perturbation experiments. 12. Child, I. L, "Psychology and Anomalous Observations: The Question of ESP in Dreams," American Psychologist, Vol. 40, No. 11, pp. 1219-1230, (November, 1985). Professor Child, the then Chairman of the Psychology Department at Yale University, provides a critical review of the anomalous cognition dream studies conducted at Maimonides Medical Center in the early 1970's. Professor Child warns the general psychological research community not to dismiss the body of research and suggests that it should be of wide interest to them. 13. Atkinson, R. L, Atkinson, R. C, Smith, E. E., and Bem, D. J., Introduction to Ps)cholo&, 10th Edidor; pp. 234-243, Harcourt Brace Jovanovich, New York, (1990). Professor Bern included anomalous cognition in a chapter on cc)nsciousness and its altered states in a widely-used introductory text in psychology. Bern provides definitions of terms, a review of the experimental evidence for anomalous cognition, an analysis of the debate over the evidence, and a review of the anecdotal evidence. 14. Walker, E. H., May, E. C., Spottiswoode, S. J. P., and Piantanida, T, "Rsting Schr6dinger's Paradox with a Michelson Interferometer," Physics B, Vol. 151, pp. 339-348, (1988). While not directly related to anomalous mental phenomena, this paper describes an experimental test to determine if consciousness is a necessary ingredient for determining physical reality. The authors conclude that is it not, and thus, this result has implications for anomalous perturbation research. 15. Hyman, R., "Parapsychological Research: A Tutorial Review and Critical Appraisal," Invited Paper, Proceedings ofthe IEEE@ Vol. 74, No. 6, pp. 823-849, (June, 1986). Dr. Hyman is a Professor of Psychology at the University of Oregon in Eugene and has been a long-time critic of and commentator on the field of parapsychology. Hyman reviews the historical experiments and provides a critical analysis of the current research. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Yol. I)/ actober T6 IY14 --- Approved For ReleaE 251 No@ 5476 October 18, 1974 Publishc@d weekly by Macmillan Journals Ltd London 4 Little P--,Stx Strv-1, WC2R 3LF f,lcphone: (01) 836 6633 Telex: 262024 T,Icgrarns: Phusis London WC2R 3LF Washington -National Press Building, DC 20045 Telephone: (202) 737 _"355 Telex: 64280 Editor David Davies Deputy Editor Roger Woodharn Editorial staff Cjillia_nBoucher Colin Norman* John Gribbin Sally Owen John Hall Allan Piper Eleanor Lawrence Miranda Robertson Mary Lindley Fiona Selkirk* Peter Milford Robert Vickers Peter Newmark Mary Wade, John Wilson -wuhinlton office Publishing Director Jenny Hughes Display advertisement enquiries to: London Offict Classffleii advertisement enquiries to: T. G. Scott and Son Ltd, I Clement's Inn, London WC2A 2ED Telephone: (01) 242 6264 and (01) 405 4743 Telegrams: Textualist London WC2A 2ED Subscription enquiries to: Macmillan Journals Ltd, Brunel Road, Basingstoke, Hants, RG21 2XS Telephone: Basingstoke 29242 Publication ad&ess in the United States -The Wm Byrd Press Inc., 2901.Byrdhill Road, Pichmond, Virginia 232,28 Second Class Postage for the USA paid at Richmond, Virginia US Postmaster, please send form 3579' to 'Nature, 711 National Press Buildin& Washington DC 2CO45 Price JE22. per year-excepting USA and Canada (M per year) Registered as a riewspa' per at the .! British Post Office Copyright 0 Macmillan Journals Ltd, October 18, 1974 Cover Picture A hu@ndred years ago Nature was reviewing E. J. Marey's Animal Mechanism (page 518, October 29, 1874). These cumbersome mechanisms Were soon to be replaced by Muy, bridge's zoopraxiscope . camera. On Page .567 we look, at a Muybridge sequence and---a century later-what happens when tber light is.switzbed on. Volume 252 October 18, 1974 Investigating the paranormal 559 For those in peril on the factory floor 560 INTERNATIONAL NEWS 562 NEWS AND VMWS 569 ARTICLES Human reproduction and family plaming: research strategies in developing countries- A. Kessler and C. C. Standley 577 Compositional variation in recent Icelandic tholeiiies and the Kverkf)bll hot spot- G@ E. Sigvoldason, S.. Stein thorsson, N. Oskarison andP. linsland 579 Properties of hybrids between Salmonella phagt P22 and coliphage D. Botstein and]. Herskowitz 584 LETMRS TO NATURE-Physical Sciences Distance to Cygnus X- I @C.-C. Cheng, X. J. H. Phillips and A. M. Wilson 589 High energy radiation from white holes-J. Y. Norlikar, K. Al. V. Appa Roo and IV. Dodhich 590 Spectrum of the cosmic background radiadon between 3 mm and 800 pm- E. 1. Robson- b. G. Yickers, J. S. Hukinga, J. E. Beckman and P. E. Clegff 5@- new solar-terfestrial relationship-G. M. Brown 5@' Rainfall, drought and the solar cycle-C. A. Wood and R. R. Lovert 5C Dynamic implications of mantle hotspots-M. A. Khan ' _( A-type doubling in the CH molecule-R. E. Hommersley and W. G. Richards Draiz-reducing volymm and licuid-column osci)lations-W. D. McComb 11f noise with a low fmquency white noise limit-K. L. Schick and A. .4. Verveen Second Law Jf Thermodyna'mics-D. R. Wilkie 6U Information transmission under conditions of sensory shielding-R. Torff and H. Puthoff 60.' LETTERS TO NATURE-Biological Sciences The stability of a feasible random ecosystem-A. Roberts 607 Objective evaluation of auditory evoked EEG responses-B. McA. _50yers and H. 14. Beaffley 608 Imprinting and exploration of slight novelty in chicks-P. S. Jackson and P. F. G. Bateson 609 Microbial activation of prophenoloxidase from immune inswt larvat-A. E. Pye 610 Elevation of total serum IgE in rats following helminth parasite infection- I E. Jarrett and 11. Bazin 612 Alternative route for nitrogen assimilation in higher plants-P. J. Lea and . B. 1, Miflin 614 Evolution of cell senescence, - atherosclerosis and benign tumours-D. Dykhuizen 616 Insulin stimulates myogenesis in a rat myoblast line-J.-L. Mandel and M. L. Pearson 618 Sickle cell.resisiance to in viyo hypoxia-0. Castro, S. C. Finch and G. Osbaldislone 620 Expr=ion of the dystrophia muscularis (dy) recessive gene in mice-R. Parsons 621 Growth of human muscle spindles in vitro-B. J. Ellion and D. G. F. Harriman 627 Multiple control mechanisms underlie initiation of growth in animal cells- L. J. de Asua and E. Rozengurt 624 Control of cell division in yeast using the ionophore, A.23187 with calcium and magriesiurn-J. H. Duffus and L. J. Patterson 626 Antigen of mouse bile capillaries and cuticle of intestinal mucosa- N. 1. Khromkoya and T. D. Beloshapkina 627 Ultrastructural analysis of toxin binding and entry into mammalian cells- G. L. Nicolson 628 Serum 0oparnine p-nyqroxylase activity in cleveloping hypertensive rau--1 - Iyagarsit, T. Kdro, Y. Numata (Sudo), X Ikuta, If. Uniezerwa, M. Ma(suzaki and r. Takeuchi 630 3zymatic syninesis oz accryicnoune oy a scrotonin-containing neurone irom neux- 'M. R. Hanley, G. A. Cottrell, P. C. Emson and F. Fonnum -631 Approved For Release 2003/04/18 : CIA@RDR96-00789RO031000300014 Approved For Release 2003/04/18 ; CIA-RDP96-00789ROO3100030001-4 Information transmission under conditions of sensoiy sbielding' WE -present results of experiments suggesting the existenct of one or more perceptual modalities through which individuals obtain -information about -their environment,'-although this -information. is not presented to any -known serise. The litera- .lure'--' ;and car observations lead US to Conclude that such .abilities can be studied under laborat,ory -conditions. We have investigated the ability of mr-tain people to describe- graphical material or Temote scenes shielded -against ordinary perception. In addition, we performed pilot studies to determine if clectroencephalographic (EEG) TCCOrdings might indicalt perception of remote happenings even in the absence of corTect overt responses. We concentrated on what -we consider to be our primary responsibility-to resolve under conditions as unambiguous as possible the basic issue of whether a certain class of para- -normal perception phenomena exists. So we conducted our .experiments with sufficient control, utilising visual, acoustic and electrical shielding, to ensure that all conventional paths of =-isory input were blocked. At a times we took measures to prevent.semory leakage and to prevent deception, whether intentionai or unintentional. Our goal is not just to catalogue interesting. events, but to iincover patterns of causeeffect relationships -that lend them- selves to analysis. and hypothesis. in the forms with which we are familiar in'scientific study. The results presented hert constitute a 1irst step towara that goal; we hAve established under known conditions a daTi bas: from which departures as 11 function of physical and psychological variables can be studied in future work. REMOTE PERCEMON OF: GRAPIBC MATERIAL First, we conducted experiments with Mr Tjri'Geller in which we examined hJs ability, while located in an electrically shielded room, to reproduce targ@t pictures drawn by expen, nr-nters - located at remote locatipns. Second, we conducted double-blind experiments with Mr Pat Price, in which We nmsured ".ability to describe remote outdoor scenes many miles from his physical location. Finally, we conducted Pre- Approved For Release 2003/04/18: CIA-RDP96-00789RO031;00030001-4 Approved For Release 2003/04/18 iYalure Yol, 251 October 18 1974 60. liminary tests using EEGS, in which subjoats w= asked to perceive whether a remote light was flashing, and to determine whether a subject could perceive the presr-mm of the light, even if only at a noncognitive level of awareness. : In preliminary testing Geller apparendy demodstrated an ability to reproduce simple pictures (line drawings) which had been drawn and placed in opaque scaled envelopes which he was not permitted to handle. But since each of the targets was known to at least one experimenter in the room with Geller, it was not possible on the basis of the preliminary testing to discriminate between Geller's direct perception of envelop-- contents. and perception through some mechanism involving the experimenters, whethex paranormal or subliminal. So we examined the phenomenon under conditions designed to eliminate all conventiona.1 information channels, overt or subliminal. Geller was separated from both the target material and anyone knowledgeable of the material, as in the experiments of ref, 4, In the'first part of the study a series of 13 separate drawing experiments were carried out over 7 days. No expt-riments are deleted from the results presented hem At the beginning of the experiment either.G-eller or the experimenters entered a shielded room so that from that time forward Geller vrLs at all times visually, acoustically and electrically shielded from p--rsoanel and material at the target location. Only following Geller's isolation from the experi- mentors was a target chostn and drawn, a procedure designed to eliminate pre-experiment cueing. Furthermore, to eliminate the possibility of pre-experiment target forc-ing, Geller was kept ignorant as to the identity of the person se1ccting the target and as to the method of taxget selection. This was accomplishod by the use of three different- techniques: (1) pseudo-random technique of op-ening a dictionary arbitrarily and choosing the first word that could be drawn (Experiments I-L4); (2)- targets, blind to experimenters and subj,.@ct, piepared independentl@ by Spl scientists outside the experimental group (raowin@ Geller's isolation) and providod to the expmimenters durinj the course of the experiment (Experiments 5-7, 11-13); and (3 arbitrary selection from a target pool decided upon in advanci of daily experimentation and designed to provide data concern! ing information content for use m tesung specific hypothese: pperiments 8-10). Geller's task was to reproduce with pe- on paper the line drawing generated at the target location Following a period of effort ranging from a few minutes tc half an hour, Geller either Passed (when h,-- did not feel con hdent) or indicated be was ready to submit a drawing' to tha experimenters, in which case the drawing was collected befor Geller was'permitted to see the tarpt. To prevent x-asory cueing or the t2iget inform.&don, Experiment I through 10 were carried out using a shielded room in SRI's ficilit, for EEG research. The acoustic &nd vintal isolation is proNide, by a double-walled steel roo" locktd by means of an inner anQ outer door, each of which is secured Nith a reffigerator- type lockin, moch&nism. Following target selection when Geller was insid, the room, a oat,-way audio monitor, operating only from the insid to the out@i&, was activated to monitor Geller during his efromN Tle target picz= was never discuss@d by the experimenters after tb, picture was drawn 2.nd brought nmr the shielded room. In ou detailed examination of the shielded room and the protocol used ii these experim-ents, no s-casory leakage has been found- The conditions and results for the 10 experiments carried out in tL shielded room are displayed in Table I and Fig. 1. AA experimew except 4 and 5, were conducted with Geller inside the shielded roorr. In Experiments 4 and 5, the procedure was reversed. For thos; experiments in which Geller was inside the shielded room, the targc location was in an adjacent room at & distance of about 4 m, excep for Experiments 3 and 8, in which tht target locations were, resM: tively, an offic-- at a distance of 475 m and a room at a distance c about 7 m. A-rosponso was obtained in all experiments cxcept Number 5-7. In Experiment 5, the person-to-penon link was eliminate,, by arranging for a scientist outside7 the usual experimentu grox@o to draw a picture, lock it in ther shielded room befor Geller's arrival at SRI, and leave the area- Geller was then @,,l TARGET RESPONSE I z1- RESPONU 2 a C TARGET R WON" TAROET' J Appr6ue,d For TARG9r MESPOMS4E @063/64/18: CIA-RDP96-00789ROO3100030001-4 ........... . TARGET RSSPON" I RESPON&E 2 RESPONSE 3 L TARGET RESPONSE P96-0078AO031 0 URI (;F-Ljr-Q: 6 A TARGET REsPDNzE 11-4 9' @Flg- I Target pictww amd responses drar by Uri Geller tmder shieldod oonditio= - Nature Vol. 251 October 18 1974 eve d For RAIAW 0030001-4 ptIA00DW.6@00789RO031 0 &p@ Geller Location Ta*tlocation Target Figure (month,.day, y-) 8/4173 Shielded room I* Adj@nt room (4.1 m)t Firecracker Ja 2 814M Shielded room I' Adjacent room (4.1 in) Grapes 3b 3 8/5[73 ShicldedTOOM I Office@475m) * Devil Ic 4 8/5173 Room adjacent to Shielded room I Solar system - ld - Welded room I . (K2.m) . 5 8/6173 Room adjacent to SI-delded room I Rabbit No drawing shielded room 1 (3:2 Dal 6 8f7/73 Shielded room I Adjacent room (4.1 rn) Tree No drawing 7 8f7[73 Shieldcd-room'] Adjacent room (4.1 m) Envelope, No drawing 8 818n3 Shielded room I Remote room (6.75 in) Camel le 9 818n3 Shielded room I Adjacent room (4.1 in) Bridge I f 10 8/8173 Shielded room I Adjacent room (4.1 in) Scagu.11 Ig 11 8/9/73 Shielded room 2t Computer (54 in) Kite (00PIPUtCT CRT) 12 8/10173 Shieldod room 2 Computer (54 m) Church @computer memory) .2b 13 8/J 0[73 Shielded room 2 Compuier (54 m) Arrow tl@rough heart _2c (compyt r CRT, zero *EEG Facility sl-iiclded room (ste text), .fPerceiver-unet distances measured in metms. ISRI Radio Systems Laboratory shielded room (see text). by the expezianenters: to *the-shielded roorn an d -asked to draw' the picturt located inside the room. Be said that he. got no clear impression and therefore did Dot submit a drawing.'Tbe elimina- 'tion of the. person-tc,@person link was examined furt@er in the second seric:s of expetrimcnits with this subject. ExperimczLs 6 az'd 7 were carried out while we, attempted to record Geller's EEG during his efforts to perceive the target Pictures. Tbe target pictures were, TCSPCCtiVCIY, -a tree and an envelope. lit found it difficult to hold adequately still for good EEG records, said that be experienced difficulty in getting impressions of the targets and again -submitted no drawings. EXperimez,ts; 11 through 13 were icarried out in SRI's Engin- etring Building, t 'o makt ust of the computer facilities available thcrc@ For these exp'erimenters', Geller was secured in -a double- -walled, coppe--r-screen Faraday cage 54 in down the hall..and around the comer from' the computer room. The Faraday cage provides 123 dB attenuation for plant wave radio'frequency radiation over a range of iS kHz to I GHz_ For m4gnefic fields the attenuation is '69 dB at 15 kHz and decrea@ses- to 3 0 at 60 Hi. FoUoWing G@llcr'.s isolation, -the targeis for thee expe,riments were chosen by computer laboratory personntl Dot othemiste associated with either the experiment or Geller, and the experimenters and subject were kept blind as to the contents of the target. pool. For Experiment 11, a picture of a kite was drawn on the face of a cathode ray tube display screen, driven by the computer's graphics program. For Experiment 12, a pict= of a church was drawn and stored in the memory of the computer. In Experimeni 13, the -target drawing, an arrow through a heart (Fig. 4), w?_s drawn on the face of the cathode ray tube and then the display intensity was tumed off so that no picture was visible. To obtain an independent evaluation of the correlation I>-- tween tar9w and response. data, the experimenters the data for judging on a 'blind' basis by two -SRI scien'tists who were not otherwise associated with the research- For the 10 cases in which Geller provided a response, the judges were asked to match the, resp6nse data with tht corresp*onding target data (without replacement). In those cases in which Geller made more than one drawing as his resp6nse to :the target, all the drawings were combined as a set for judging. The two judges each miatcbed the target data.to the responsie data with no error. For either judge such a correspondence has an apriori probability, under the null hypothesis of no in- formation channel, of P = (10!)-l - 3 x 10-7. A second series of expai=nts was carried out -to determitne whether direct Perception of eny4ope contents was possible withobt son-4 person knowing of the target picture. O;e hundred target picturts of-everyday 6bjecu were drawn by im SRI Mist and'sealed'by oth;tr SRI personnel in double env6lopes containing black @cardboard. The hundred targets were divided randomly into groups of 20 for use in each of the -three, days' experiments. On each of the thrm days of these experimmis, Geller passed, That is, he, declined.to associate any envelope with a drawing that he made', expressing dissatisfaction xith the existence of such a large target pool. On each day he made approximately 12 r4coginisable drawings, which he 'felt were associated with the entire target pool of 100. On reach of the three days, two of his drawings could reasonably be associated with two of the 20 dady targets. On the third day, two of his drawings were YeD- close replications of -two of that day's target pictures. The drawings resulting from this experiment do not depart signific- antly from what would be expected by chance. In a simpler experiment Geller was successful in obtaining information under conditions in which no.persons were know- ledgeable of the iarget. A double-blind experiment was per- formed -in which a singit 314 inch die was placed in a 3 x 4 x 5 in6h itml box.'The box was then vigorously sbatn by one of the experimenters and placed on@: the table, a technique found in control runs to produce a distribution of die faces -differing non- significantly -from'chance. The orientation of the die within the box was unknown lo the experimenters at that time. Geller would then write down which die face was uppermost. The target pool was.known, but the targets were individually pre- pared in a manner blind to all pe@sons involved in the experi- ment. This experiment was performed ter) thes, with Geller passing twice and giving a response eight times. In the eight times in which be, gave' a response, he was correct each time. The distributi'on of responses consisted of thrm 2s, one 4, two 5s, and two 6s. The probability of this occurring by chance is approximately one in 104. In certain .situations significant infom-vation transmission can take Place under shielded conditions. Factors which appear to be important and therefore candidates for future investigation include whether the subiw knows the set of targets in the target pool, the actual number of txrgets in the target pool at any given time, and whether the target is known by any of the experimenters. It has been widely rePorted that Geller has demonstrated the ability to bend metal by parablormal means. Although metal bending by Geller has boen obs4r'ved in our laboratory, we have not been able to combine suc .h observations with adequatrlY controlled experiments to obtain data sufficient to support the paranormal hypothesis. REMOTE VIEWING OP NATURAL TARCETS A study by Osis' led us to deltermine whether a suoJect could describc. randomly chosen geographical sites located several' miles from the subject's position and dernarcated by some Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 TARGET RESPONSE a 2 Fiv, Z,Computer driLwings and resp6nses &@Ynb@.Uri Geller. a, Computer drawing stoi-ed on video display; J@, ooinputer`dr&vnnrstoreonses similar to those obtdined under conditions of direct stimulation. For exampl@, wheii normal subjects am stimulated with a flashing light@ their EEG typically shows -a decrease in the amplitude of the resting rhythm and a drivinj of the brain waves at the frequency of the flashe:91. We hypotho.- sised that if we stimulated one subiect iri this mamer (a sender), the. BEG of another subject in a itmote room with no Ruh Present (a receiver), might show changes in Wpha (9-11 Rz) activity,-and possibly EEG driving similar to that dftfie sender. We informed our subject that at certain times a light was to be: flashed in a sender's eyes in a distant room, and if the subject perceived that event, consciously or unconsciously, it might be evident from @banges in his EEG output. The receiver was Mted in the visually opa@que, acoustically. and el WMIC'Uy shielded dotible,-walled steel room previously described. The sender was seated in a room about 7 m from the receiver. * To find subjects who were responsive to such a remote stimulus, we initi@Uy worked with fo"Ur' female and two malt vblu4t&-z subjequ, all of whom believed that suemss in the experimental situation might be po3sible. Tiie= were'designatod receivm'. The sender@ were either offier subjem or expc-rimenter-s. 'We decided beforehand to Tun one or tA sessions of 36 trials each with each sub*t in this sele-c promduTt, and to do -a more extensive study with any sl:_t whose Ttsul6 were positive. A Grass M2 photostimUlator place@d aoout I m in front of t sender was used to present hash trains of D.- s duration. The receiyt EEG wivity from the occipital region (Oz), referenced to mastoids, was amplified with a Grass SP-1 preamplifier and assc@ i drive, amplifier with a bandpass of 1-12b Hz. The EEG data recorded on =gnetic tape with an Ampex SP 300 recorder. On each trial, a tone burst of fixed frequ:ncy was presented to-b-o sender and receiver and was followed in cne second by either _( train of flashes, or a null flash interval prest-nted to the sender. 7 -1 six such trials were given in an experimen-ul session, consisting null irials-no.flashes following the tone-: 2 trials of flashes at 6 f.p and 12 trials of flashes at 16 f.p.s., all randomly intermixed, dtit mined by entries from a table of random numbers. Each of the_i@ generated an I 1-s EEG epoch. The last 4 s of the epoch %4w se- - i for analysis to minimise the desynchronising action of the %k,-,i cut. This 4-s segment was subjected to Fo-_,rier -analysis on a LINC computer. Spectrum analyses gave no evidehm ofEE-G driving in any rec--?, although in qontrol nins the receivers did exhibit driving - -i pl@ysically stimulated with the, flashes. Du,, of the six subjects si-i initially, one subject (H. H.) showed a consistent alpha blocking effe We therefore under-took further study with this subject. " - Data from seven sets of 36 trials each were collected froF@d subject on three separate days. This corri,-rrises all Ilit data col- :i to date with this subject under the test conditions described L@, The alpha band was identified from average sprx-tra, then scores average power and p-tak'power were obtained froffi individuaL_-trL and subjected to statistical analysis. Of our six #ubjects, H. H. had by far the most monochro-i, EEG specurum. Figure 3 shows an OVeTiayof the three avemg spectra from ont of this subject's 336-trial runs, displav i changes in her alpha activity for the three stimulus condl---)r Mean vaiue,s for the average power and peak power foi TgLbk 3 EEG data for H.H. showing average power and peak!nv in the 9-13 YIz band, as a function or fiash frequency and Flash Frequency Q 6 ' 16 0 6 16 Sender Alvierage Power Peak Power J.L. 94 ..8 94.1 76.8 357.7 329.2 289.6 R.T. 41 '345.5 37.0 160.7 161.0 125.0 No sender (subject * informed) 2@1 35.7 28.2 87.5 95.7 81.7 J.L. 54.2 55.3. 44.8 191.4 170.5 149.3 J.L. M.8 50.9 32.8 24-0.6 178.0 104.6 R.T. 39!.8 24.9 30.3 145.2 74.2 122.1 No sender (subject not 86.0 53.0 52.1 318.1 180.6 202.3 informed) Averages 56 .8 49.9 43.1 214-5 169.9 153.5 ! @ 12 */. - 24 %(P < 0.04) -21 Y.-28%(PLn@o Each entry is an avei-age ovtr 12 trials Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Nature Vol. 251 October 18,1974 of th *61 " 0 ,a We e M m "s in qte a f) Uin ftt§ 0 s. m seven peak power measures and in six out of seven average power me .asures. Note also the reduced effect in the case in which the subject was informed that no sender was present (Run 3). It seems that overall alpha production was reduced for ads run in conjunction with the subject's expressed apprc@ hension about conducting the experiment without a sender. This is in contrast to the case (Run 7) in which the subject was not informed. Siegel's two-tailed t approxi'madon to the nonparametric randomi- sation, (est' was applied to the data from all sets, which included two sessions in which the sender was removed. Average power on trials associated with the occurrence of 16 f.p.s. was significantly less than when there were no flashes (i = 2.09, d.f. - I 18, P < O.D4). The second measure, peak power, was also si"cantly less in the 16 f.p.s. conditions that) in the null condition (t - 2.16, d-f. = 118, P<0.03). The average response in the 6 f.p.s. condition was in the same dircc- don as that associated with 16 f.p.s., but the effect was not statistically significant. Spectrum analyses of control recordings made from saline with a 12 kfl resistance in place of the subject with and without the addi- don of a 10 H2, 50 gV test signal applied to the saline solution, revealed no indications of Rash frequencies, nor perturbations of the 10 Hz signal. These controls suggest that the results were not due to system artefacts. Further tests also gave no evidence of radio fre- quency energy associated with the stimulus. Subjocts were asked to indicate ;heir conscious assessment for each trial as to which stimulus was generated. They made their guesses known to the experimenter via one-way telegraphic communication. An analysis of these guesses has shown them to be at chance, in- dicating the absence of any supraliminal cueing, so arousal as evid- enced by significant alpha blocking occurred only at the noncognitive level of awareness. We hypothesise that the protocol described hem may prove to be useful as a screening procedure for latent remote perceptual ability in the general population, f so-caJlcd paranormal phenomena caj #AJVOOJ9@1-4nd it is our hope that othe laboratories wi,11 inidate additional research to at-tempt ti replicate these findings. This research was sponsored by The Foundation for Paraser sory Investigation, New York City. We' thank Mrs Judit. Skutch, Dr Edgar D, Mitchell of the Institute of Noeti Sciences-as well as our SRI associates, Mr Bonnar Co? Mr Earle Jones and Dr Dean Brown-for support an encouragement. Constructive .suggestions by Mrs Jean Mayc Dr Charles Tart, Univmity of California, and Dr Robei Ornstein and Dr David Gaft of the Langley Porter Neurc psychiatric Institute are acknowledged. RussELL TA.RG HARoLD Pu-rHoFF Edectronics and Bioengineering Laboratory, Statkford Research Institute, Afenlo Park, California 94025 Received March 11; revised July 8, 1914. Pratt, J., Rhine, J. B., Stuart, C., and Greenwood, J., Extr Sensory Perception after Sixty Years (Henry Holt, Nei York, 1940). Soal, S., and Bateman, F., Modern Experiments in Telepath (Faber and Faber, London, 1954). 'Vasillicv, L. L., Experiments in Mental Suggestion (ISMI Pub lications, Hampshire, England, 1963). Mus.so, J. R., and Gmero, M., J. Parapsychology, 37, 13--3 (1973). Osis, K., ASPR Newsletter, No. 14 (1972). Tart, C. T., Physiological Correlates of Psi Cognition, Int. Parapsychology, V, No. 4 (1963). rDean, E. D., Int. J. NeuropsychiLvry, 2 (1966). Ell, D., and Parr, G., Electroenceph'alography : A Symposiur on its Various Aspects (Macmillari, New York, 1963). Siegel, S., Nonparametric Statistics for the Behaviore Sciences, 152-156 (McGraw-Hill, New Yorlc, 1956).. 16 5 Hz 10Hz ....15 Hz Fig. 3 Occipital EEG spectra, 0-20 Hz, for one subject (IT.E.) acting as receiver, showing aniplitude changes.in the 9-11 Hz band as a function of strobc frrqucm@cy. 71iave ca=: 0, 6, and 16 fps, (12 trial averages)..*: CONCLUSION From these experiments. we concludc that: *A channel exists whereby information a .boum remote location can be obtained by means of an as yet unidentified perceptual modality-: $AS with-all. bi6logical-systerns, the information channel appears to be. imperfect, containing noise7 along with-tho signal- *WhBe a quantitative signal-to-noise ratio in theinformation- theoretical sense cannot as yet be determined, ther results of our experiments indicate-that the functioning is at ther level Of useful information tiransfer.. It may ber that remote perceptual"ability is widely distributed in the general population, but,bec:ause the perception is generally below an individual's leve.1 of awareness, it is repressed or noc noticed. For exampJe, t*O of our silbiocts; K H. and P, E.) )iad- 'lot considered themselves to have unusual. per@cptual abiUty before theii7 Participation.'in these.experiments.' : .. Our observationof the phenomena leads us to''Co'nclude; thar Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 PROCEEDINGS OF THE IEEE, VOL. 64, NO. 3, MARCH 1976 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 A Perceptual Channelior Information Transfer over Kilometer Distances: Historical Perspective and Recent Research HAROLD E. PUTHOFF, MEMBER, IEEE, AND RUSSELL TARG, SENIOR MEMBER, IEEE Abstract-Foi more than 100 years, scientists have attempted to determine the truth or falsity of claims for the existence of a perceptual channel xhereby certain individuals are able to perceive and describe remote &ta not presented to any known sense. This,paper presents an , outline of the history of scientific inquiry into such so-called par-anor., mal perception and surveys the current state of the art in parapsycho- logical reseaTch in the United States and abroad. The nature of this perceptual channel is examined in a series of experiments carried out in the Electronics and Biciengine-er-ing Laboratory of Stanford Research Institute. The perceptual modality most extensively investigated is the ability of both experienced subjects and inexperienced volunteers to view, by innate mental processes, remote geographical or technical targets including bufldings, roads. and laboratory apparatus. The ac- cumulated data indicate that the phenomenon is not a sensitive func. tion of distance, and Faraday cage shielding does not in any apparent way degrade the quality and accuracy of perception. On the basis of this research, some areas of physics axe suggested from which a descr-ip- tion or explanation of the phenomenon could be forthcoming. 1. INTRODUCTION "IT IS THE PROVINICE of natural science to investigate nature, impartially and without prejudice" [I). Nowhere in scientific inquiry has this dictum met as great a chal- lenge as in the area of so-called extrasensory perception (ESP), the detection of remote stimuli not mediated by the usual sensory processes. Such phenomena, although under scientific consideration for over a centiry, have historically been fraught with unrebabUity and controversy, and validation of the phe- nomena by accepted scientific methodology has been slow in coming. Even so, a recent survey conducted by the British publication New Scientist revealed that 67 percent of nearly 1500 responding readers (the majority of whom are working scientists and technologists) considered ESP to be an estab- lished fact or a likely possibility, and 88 percent held the investigation of ESP to be a legitimate scientific undertaking (2]. A review of the literatuxe reveals that although experiments by.reputable researchers yielding positive results were begun over a century ago (e.g., Sir William Crookes' study of D. D. Home, 1860's) (31, many consider the study of these phe- nomena as only recently emerging from the realm of quasi@ science. One.reason for this is that, despite experimental results, no satisfactory theoretical construct had been advanced to correlate data or to predict new experimental outcomes, Consequently, the. area in question remained for a long time in the recipe stage reminiscent of electrodynamics before the Manuscript received July 25, 1975; revised November ?, 105. The submission of this paper was encouraged after review of an advance proposal. This work was supported by the Foundation for Parasensory Investigation and the Parapsychology Foundation, New York, NY; the Institute of Noetic Sciences, NPalo Alto, CA; and the National Aero- nautics and Space Administration, under Contract NAS 7-100. The authors are with the Electronics and Bioengineering Laboratory, Stanford Research Institute, Menio.Park, CA 94025. 10 unification b_@uzaht about by the work of Ainpere, 17--aday. and Maxwell. Since the early work, however, we have s-.-@n the development Df information theory, quantum theo-, and neu,ophysiol,-,:cal research, and these disciplines -.-ovide powerful con:-,-Liual tools that appear to bea7 directly on the issue. In fac--. steveral physic'sts (Section V) are now of the opinion that @'n_-Se phenomena are not' at all inconsisterr with the framewo-"1 of modern physics: the often-held vi-- that obser),ations c-' this type are a priori incompatible with i.nown laws is eTrone._-,us in that such a concept is based on th-_ naive realisr-n preval-, before the dtvelopment of ouantum -,,.-.eory. In the emergl7g view, it is aczepted that resea'-ch in t1_1 area can be conclu---e-6 so as to un -cover not just a catalog of inter- esting events, rather pattems of cause--eff-ct relatic--iships of the type ---at lend themselves to analysis and hyp::hesis in the forms with which wt are familiar "n the L)'--vsical sciences. On@= hvi)othesis is That information. transfei under conditions off sensory shield-ng is mediated by exi@tmely low-frequency (ELF) electromagnetic waves, a propos2i that does not seera to be ruled cut by any obvious phyt_-al or biological fac:s. Further, the development of information theory makes it possible to characterize and quantity the performance of a communications channel regardlass of the underlying mechanism. For the pa:- three years, we have had a program @ri the Electronics and Biciengineering .Laboratory of the Stan- ford Research Institute (SRI) to investigate. those factts of human percep:ion that appear to fall outside the range of well- understood percept ual /processing capabilities. Of par:icular interest is a human information-accessing capability that we call "remote viewing." T"s phenomenon pertains ic, the ability of cerL-in individuals to access and describe, by means of mental processes, information sources blocked from ordi- nary perception, and generally accepted as secure agairis7, such access. In particula-7, the phenomenon we have investigated most extensively is the ability of a subject to view remote geozraph- ical locations up to several thousand kilometers distant from his physical location (given only a known person on whom to target).' We have carried out more than fifty experiments under controlled laboratory conditions with several indiNiduals whose remote perceptual abilities have been developed suf- ficiently to allow them at times to describe correctly -often in great detail-geographical or technical material such as build- ings, roads, laboratory apparatus, and the like. As observed in the laboratory, the basic phenomenon appears to cover a range of,subjective experiences variously referzed to 'Our initial work in this area was reported in Nature 14 1, ;Lnd re- printed in the IEEE Commurt. Soc. Newsletter, vol. 13, Jan. 1975, Approved For Re.lease 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 330 PROCCEEDINGS OF THE IEEE, MARCH 1976 Approved For Release 2003104118: CIA-RDP96-00789ROO3100030001-4 Fig. 1 1.1.@ J - /Vo or Airport in San AndreL., Colombia, used as remote-viewing target, along w:ith skatch produced by subjeci in California. in the literature as autoscopy (in zhe med-cal literature); exteri- than fifty experiments with nine subjects carried out in our orization or disassociation (psychologiall literature); simple own laboratory, which represent a sufficiently stable &-.a base clairvoyance, traveling clairvoyance, or 0--Tt-of-body experience to permit testing of various hypotheses conceming th, func- (parapsychological literature); or astral p.-ojection (occult liter- tioning of this channel. Finally, in Section V, we i-7dicate ature). We choose the term "remote -4ewing" as a neutral those areas of physics and information theory that ap-ear to descriptive term free from prio7 associations and bias as to be relevant to an :undtrstanding of certain aspects )f the mechanisms. phenomena. The development at SRI of a successful experimental pro- First, however, we present an illustrative exEmple ge-.-,rated cedure to elicit this capability has evolved to the point where in an early pilot experiment. As will be clew from o@_r later persons such as visiting government scientists and contract discussion, this is not a "best-ever" example, but r--ther a monitors, with no previous exposure to such concepts, have typical sample of the level of proficiency that can be :!ached learned to perform well; and subjects who have trained over a and that we have come to expect in our researca. one-year period have performed excellently under a variety of Three subjects participated in a long-distance experiment experimental conditions. Our accumulated data thus indicate focusing on a series@ of targets in Costa Rica. These s.!bjects that both specially selected and unselected persons can be said they had neveri: been to Costa Rica. In this expe:iment, assisted in developing remote perceptual abilities up to a one of the experimenters (Dr. Puthoff ) spent ten. days taveling level of useful information transfer. through Costa Rica on a combination business/pleasu:e trip. In experiments of this type, we have th-ee principal findings. This information was all that was known to the subject.5 about First, we -have established that it is possible to obtain signifi- the traveler's itinerary. The experiment called for Dr. Puthoff cant amounts of accurate descriptive info.-mation about remote to keep a detailed ritcord of his location and activities, @nclud- locations. Second, an increase in the distance from a few ing photographs of-each of seven target days at 1330 PDT. meters up to 4000 km separating the subject from the scene A total of twelve daily descriptions were collected before the to be perceived does not in any apparent way degrade the traveler's return: six responses from one subject, fiv! from quality or accuracy of p6rception. Final2yi the use of Faraday another, and one from a -third. cage electrical shielding does not prevent high-quality descrip- The third subject who submitted the single response SLPplied tions from being obtained. a drawing for a day @in the middle of the series. (The sLbject's To build a coherent theor-y for the explanation of these response, together with the photographs taken at the @4.te, are phenomena, it is necessary to have a 6--ar understanding of shown in Fig. 1). : Although Costa Rica is a moun-Lainous -what constitutes the phenRmp'na. In this paper, we first briefly country, the subject unexpectedly perceived the trave'Jer at a summarize previous efforts in this.field in SWion 11. We then beach and ocean setting. With some misgiving, he descrYbed an present in Sections III and IV -the results of a series of more airport on a sandy beach and an airstrip with the ocear, at the Approved For Release 2003/04/18 ClAmR DP`96-00789RO03100030001-4 PUTHOFF AND TARG: PERCEPTUAL CHANNEL FOR INFORMATION TRANSFER Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 end (correct). An airport building also was drawn, and shown to have a large rectangular overhang (correct). The traveler had taken an unplanned one-day side trip to an offshore island and at the time of the experiment had just disembarked from a :plane at a small island airport as described by the subject 4000 km away. The sole discrepancy was that the subject's drawing showed a Quonset-hut type of building in place of the rectangular structure. The above description was chosen as an example to illustrate a major point observed a number of times throughout the program to be described, Contrary to what may be expected, a subject's description does not necessarily portray what may reasonably be expected to be correct (an educated or "safe" guess), but often runs counter even to the subject's own expectations. We wish to stress again that a result such as the above is not unusual, The remaining submissions in iNs experiment pro-- vided further examples of exceDent corTes pond en ces betw6en target and response. (A target period of poolside relaxation was identified; a drive through a tropical forest at the base of a truncated volcano was described as a drive through a jungle below a large bare table mountain; a hotel-room target descrip- tion, including such details as rug color, was correct; and so on.) So as to determine whether such matches were simpl .y -that is, could reasonably be expected on the basis fortuitous of chance alone-Dr. Puthoff was asked after he had returned to blind match the twelve descriptions to his seven target locations. On the basis of this conservative evaluation proce- dure, which vastly underestimates the statistical significance of the individual descriptions, five correct matches were ob- tained. his number of matches is significant at p = 0.02 by exact binomial calculation.2 The observation of such unexpectedly high-quality descnip- tions early in our program led to a large-scale study of the phenomenon at SRI under secure double-@Iind conditions i.e., target unknown to experimenters as weLl as subjects), with independent random target selection and blind judging. The results, presented in Sections III and IN7, provide strong evi- dence for the robustness of this phenomenon whereby a human perceptual modality of extreme sensitivity can detect complex remote stimuli. 11-BACKGROUND Although we are approaching the study of these phenomena as physicists, it is not yet possible to separate ourselves entirely from the language of the nineteenth century when'the labora- tory study of the paranormal was begun. Consequently, we continue to use terms such as "paranormal," "telepathy," and the like. However, we intend only to indicate a process of information transfer under conditions generally accepted as secure against such transfer and with no prejudice or occult assumptions as to the mechanisms involved. As in any other scientific pursuit, the purpose is to collect the observables that result from experiments and to try to determine the functional relationships between these'observables and the laws of physics as they are currently understood. The. probability of a correct daily match by chance for any given transcript is p = -T. Therefore, the probability of at least five correct matches by chance out of twelve tries can be calculated from 12 121 P = Z - I (L) .(6) 0.0,2_ f-S 11 (12 1 7 -7 331 Organized research into so-called psychic functiomira began roughly in the time of J. J. Thomson, Sir Oliv--r LLo &e, and Sir William Crookes, all of whom took part in tlle fou.-ding of the Society for Psychical Research (SPR) in 18@2 in England. Crookes, for example, carried out his principal in wesi_@,,a t ions with D. D. Home, a Scotsman who grew up ir. Arne.--ca and returned to England in 1855 [ 31. According to -.he no:tbooks and published reports of Crook-es, Home had -Jemo@:itrated the ability to cause objects to move without tcuchinE them. We should note in passing that, Home, unlike most s-biects, worked only in the light and spoke out in the 5tromg!st pos- sible terms against the darkened seance rooms Topula: at the time [5]. Sir William Crookes was a pioneer in the stud-,, of 6!ctrical discharge in gases and in the development of -iacu-uTr tubes, some types of which still bear his name. Althol-gh eve-vtl-iing Crookes said about electron beams and plasmas was ac.:2pted, nothing he said about the acl-devements of D. D. Flo-e ever achieved that status. Many of Iiis colleagues, who hid not observed the experiments with Home,stated pub-Lcly tha they thought Crookes had been deceived, to wl-dch C-ookes ingrilY responded: Will not my critics 6ve me credit for some amount of c:oT-,l-non sense? Do they no- -imagine that the obvious precaij-.ions, Nviich to them as soon as they sit down to pick _10les i:-. my occur experiments, have occurred to me also in the courst of :m@ Dro- longed and patient investigation? The answer to t---is, as -,: all other objections is. Prove it to be an error, by shc-,Ying v-.ere the error, hes, or ff a trick, by showing how the :xick is @er- formed. Try the experiment fully and fairly. If th-n fzau' bp found, expose it; if it be a truth, proclaim it. Thiy is the :nly scientific procedure, and it is that I propose sieadir/ to) pirsue [3 ]. In the United States, scientific interest hn the parB:aormal was centered in the universities. In 1912, Joh-a Cloo,er [61 was established in the endowed Chair of Psychic-1 Restarch at Stanford University. In the 1920's, Harvard Unversit@ set up research programs with George Estabrooks and 7 . T. -1-oland (71, [8]. It was .in this framework that, in '1930, Villiarn McDougall invited Dr. J. B. Rhine and Dr. Lo-aisa Rhine to join the Psychology Department at Duke Univeriity 19'. For more than 30 years, significant work was carried cut at Thine's Duke University Laboratory. To examine the existtace of pa-ranormal perception, he used the now-famo,.is EST cards containing a boldly printed picture of a star, cross, @quare, circle, or wavy lines. S@ubjects were asked to nEme tht order of these cards in a freshly shuffled deck of twenty-fiv-- such cards. To test for telepathy, an experimenter vould hok at the cards one at a time, and a subject suitably separatez- from the sender would attempt to determine which card vm being viewed. Dr. J. B. Rhine together with Dr. J. G. Pratt carri-d out thousands of experiments of this type under Vdely iarying conditions [101 . The statistical results from these exper@ments indicated that some individuals did indeed posstss a pEranor- mal perceptual ability in that it was possible to obtiin an arbitrarily high degree of improbability by cont@nued -esting of a gifted subject. The work of Rhine has been challenged on miny grounds, however, including accusations of improper handling of statis- tics, error, and fraud. With regard to the statistim, the r-neral consensus of statisticians today is that if fault is to be found in Rhine's work, it would have to be on other than statistical grounds [I I]. With regard to the accusations of fraud, the Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 332 PROCEEDINGS OF THE IEEE, MARCH 1976 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 most celebrated case of criticism of Rhine's work, that of reinforcement, he- developed several outstanding subjects, one G. R. Price [121, ended 17 years after it began when the of whom, Pavel Stepanek, has worked with experimenters accusation of fraud was retracted by its author in an article around the world for more than 10 years. entitled "Apology to Rhine and Soal," published in the same Ryzl's pioneerin :g work came as an answer to the questions journal in which it was first put forward [13]. It should also raised by the 1956 CIBA Foundation conference or. extra- be noted that parapsychological researchers themselves re- sensory perceptiori. The CIBA Chemical Company has annual cently exposed fraud in their own laboratory when they meetings on topics of biological and chemical interest, and encountered it [14). that same year t.bey assembled several prominent rarapsy- At the.end of the 1940's, Prof. S. G. Soal, an English mathe- chologists to have' a state-of-the-art conference on ESP [231. matician working with the SPR, had =ied out hundreds of The conference concluded that little progress would @t made card guessing experiments involvIng tens of thousands of calls in parapsychology research until a repeatable experiment 115). Many of these experimeits wee carried out over ex- could be found; namely, an experiment that different experi- tended distances. One of the@ most notable experiments was menters could repeat at will and tha-, would reliably yield a conducted with Mrs. Gloria Siewart between London and statistically significant result. Antwerp, This experiment gave results whose probability of Ryzl had by 19 62 accomplished thar goal. His primary con- occurring by charlCe were less than 10-S. With the, publication tribution was a decision to interact with the subject as a per- of Modern Experiments in Telepath 'v by Soal and Bateman son, to try to bui.]d up his confidence 2-id ability. His p.-otocol (both of whom were statisticians), it appeared that card guess- depended on "woiki7-,g with" rather than "running" ]-Lis sub- ing experiments produced significant results, on the average. 3 jects. Ryzl's star subject, Pavel Stepanek, has produce@ highly The most severe criticism of aU this work, a criticism diffi- significant results with many contemporary researchers 124) - cult to defend against in principle, is tl@2t leveled by the well- (29 ]. In these experiments, he was able to tell with 60-percent q known British parapsychological critic C. E. M. Hansel [ 171 , reliability whether. a hidden card was green side or white who began his examination of -he ESP hypothesis with the side up, yielding §taiistics of a milEon to one with only a stated assumption, "In view of the a p.-:ori arguments against thousand trials. it we know in advance that telepathy, etc,, cannot occur." As significant as uch results are statistically, the infoTmation Therefore, based on the "a priorf unlikeliihood" of ESP, channel is imperfect, containing nois,- along with the signal. Hansel's examination of the literature c-entered primarily on When considering how best to use si-,:h a channel, one is led the possibility of fraucl. by s-L:bjects or investigators. He to the communication theory concep- of the introduc.-ion of reviewed in depth foui experiments which he regarded as redundancy as a means of coding a message to combat the providing the best evidence of ESP: the Pearce-Pratt distance effects of a noisy channel [301. A p-ototype experiment by series (181; the Pratt-Woodruff (19) series, both conducted RyzI using such techniques has proved to be successfu'. RyzL at Duke; and Soal's work with Mrs. Stewart and Basil Shackle- had an assistant select randomly five groups of thlee digits ton [15], as well as a more recent series by Soal and Bowden each, These 15 digits were then enco6ed into binary form and [20). Hansel showed, in each case, how fraud could have been translated into a sequence of green and white cards ir. sealed committed (by the experimenteTs in the Pratt-Woodrulf and envelopes. By means of repeated calling and an elaborate Soal-Bateman series, or by the subjeczs in the Pearce-Pratt majority vote protocol, RyzI was able after 19 350 calls by and Soal-Bowden experLments). He gave no direct evidence, Stepanek (averaging 9 s per call) to correctly identify all 15 that fraud was committed in these experiments, but said, "If numbers, a result tignificant at p = 10-1s The hit -,ate for the result could have arisen through a trick, the experiment individual calls was 61.9 percent, 11 978 hits, and 7372 misses must be considered unsatisfacto:y proof of ESP, whether or (31 not it isfinally decided that such a trick was in fact used" [ 17, Note Added in Pioof: It has been brought to our attention p. 181. As discussed by Honorton in a review of the field that a simiiar procedure was recently used to transmit without [211, Hansel's conclusion after 241 pages of careful scrutiny error the word "peace" in International Morse Code (J. C. therefore was that these experiments were not "fraud-proof" Carpenter, "Toward the effective utilization of enlrianced and therefore. in principle could not serve as conclusive proof weak-signal ESP effects," presented at the Annual Meeting of of ESP, the American Asso@ciation for the Advancement of Science, Even among the supporters of ESP research and its results, New York, NY, Jan.. 27, 1975). there 'remained the consistent problem that many successful The characteristics of such a channel ran be specified in subjects eventually lost their ability and, their scores gradually accordance with the precepts of communication theory. The drifted toward chance results. This decline effect in no way bit rate associated with the information channel is calculated erased their previous astiono'mical success; but it was a disap- from [30) pointment since if paranormal perception is a natural ability, R = H W - Hy W one would like to see subjecis improving with practice rather where H(x) is the uncertainty of the source message containing than getting worse. symbols with a priori probability pi: One of the first successful attempts to overcome the decline effect was in Czechoslovakia in the work of Dr. Milan RyzI, a chemist with the Institute of Biology of the Czechoslovakian Pi 1092 Pi (2) Academy of Science and aLso an amateur hypnotist [22]. Through the use of hypnosis, together with feedback. and and Hy W is the conditional entropy based on the a posteriori probabilities that a received signal was actually transmitted: 'Recently, some of the early §pal experiments have been criticized -116). However, his long-dist?nce experiments cited here were judged in a doubleblind fashion of the type that escaped the criticism of the 2 Hy W. P U, j) Iog2 pi 0) (3) early experiments.. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 -4 PUTHOFF AND TAAG: PERCEPTUAL CHANNEL FOR INFORMATION TRANSFER 333 pproved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 I For Stepanek's run, with pi = T , pi(j) = 0.619, and an average tirne of 9 s per choice, we have a source uncertainty H(x) = 1 bit and a calculated bit rate of R - 0.04 1 bit/symbol RIT - 0.0046 bit/s. (Since the I 5-digit number (49.8 bits) actuaBy was transmitted at the rate of 2.9 X 10-4 bit1s, an increase in bit rate by a factor of about 20 could be expected on the basis of a coding scheme more optimum than that used in the experiments. See, for example, Appendix A.) Dr. Charles Tart at the University of California has written extensively on the so-called decline effect. He considers that having subjects attempt to guess cards, or perform any other repetitious task, for @;-hich they receive no feedback, follows the classical techniqut for deconditioning any response. He thus considers card Euessing "a technique for extinguishihg psychic functioning in the laborafory" [32). Tart's injunctions cf the mid-sixties were being heeded at Maimonides Hospital, Brooklyn, NY, by a team of researchers that included Dr. Montague Ullman, who was director of research for the hospital; Dr. Stanley Krippner; and, later, Charles Honorton. These three worked together for several years on experiments on the occurrence of telepathy in dreams. In the course of a half-dozen experimental series, they found in their week-long sessions a number of subjects who had d!reams that consistently were highly descriptive of pictorial material that a remott sender was looking at throughout the night. This work is @escTibed in detail in the. experimenters' book Dream Telepathy [33]. Honorton is continuing work of this free-response type in which the subject has no precon- ceived idea as to what -he target may be. In his more recent- work with subjects in the waking state, Honorton is providing homogeneous stimulation to the subject who is to describe color slides viewed by another person in a remote room. In this new work, the subject listens to white noise via earphones arid views an homogeneous visual field imposed through the use of Ping-Pong ball halves to cover the subject's eyes in conjunction with diffuse ambient illumina- tion. In this so-called Ganzfeld setting, subjects are again able, now in the waking siate, to give correct and often highly accurate descriptions of the material being viewed by the sender [34]. In Honorton's work and elsewhere, it apparently has been the step away from the repetitive forced-choice experiment that has opened the way for a wide variety of ordinary people to demonstrate significant functioning in the laboratory, with- out'being bored into a decline effect. This survey would be incomplete if we diid not indicate certain aspects of the current state of research in the USSR. It is clear from translated documents and other sources [351 that many laboratories in the USSR are engaged in paranormal research. Since the 1930's, in the laboratory of L. Vasiliev (Leningrad Institute for Brain Research), there has been an interest in the use of telepathy as a method of influencing the behavior of a person at a distance. In Vasiliev's book Experiments in Mentdl Suggestion, he makes it very clear that the bulk of his labora- tory's experiments were aimed at long-distance communica- fion combined with a forrq of behavior modification; for example, putting people at a distance to sleep through hyp- nosis [36). ve bee Similar behavior modification types of experiments ha n carried out in recent times by 1. M. Kogan, Chairman of the Bioinformation Section of the Moscow Board of the Popov Society. He is a Soviet engineer who, until 1969, published extensively on the theory of telepathic communication [371- [401. He was concerned with three principal kinds of experi- ments: mental suggestion without hypnosis over short dis- tances, in which the perci@ient attempts to identify an object; mental awakening over short distances, in which a subject is awakened from a hypnoti-- steep at the "beamed" suggestion from the hypnotist; and long-range (intercity) telepathic com- munication. Kogan's mai:-, interest has been to quantify the channel capacity of the pa---nornial channel. He finds that the bit rate decreases from 0.' bit/s for laboratory experiments to 0.005 bitfs for his 1000-'%-m intercity experiments. In the USSR, serious comideration is given to the hypothesis that telepathy is mediated 'by extremely low-frequency (ELF) electromagnetic propagati@ri. (The pros and cons ol this hypothesis are discussed --ri Section V of this paper.) In general, the entire field of paranormal research in the USSR is part of a larger one con:.-rned with the interaction between electromagnetic fields and living organisms [41), [42). At the First International Congress on Parapsychology and Psychotronics in Prague, Cztchoslovakia, in 1973, for example, Kholodoy spoke at length about the susceptibility of living systems to extremely low-@tvel ac and dc fields. He described conditioning effects on the behavior of fish resulting from the application of 10 to 100 LW of RF to their tank [43). The USSR take these data senously in that the Soviet safery re- quirements for steady-sta-e n-dcrowave exposure set limits at 10 PW/crn 2, whereas the United States has set a steady-state limit of 10 mW/cm2 [44). Kholodoy spoke also about the nonthermal effects of microwaves on animals' central nervous systems. His experiments -were very carefully carried out and are characteristic of a new @-imension in.paTanormal research. The increasing importance of this-area in Soviet research was indicated recently when the Soviet Psychological Association issued an unprecedented p-osition paper calling on the Soviet Academy of Sciences to step up efforts in this area [451. They recommended that the -newly formed Psychological Institute within the Soviet Academy of Sciences and the Psychological Institute of the Academy of Pedagogical Sciences review the area and consider the creation of a new laboratory within one of the institutes to study persons with unusual abilities. They also recommended a comprehensive evaluation of experiments and theory by the Academy of Sciences' Insti- tute of Biophysics and Institute for the Problems of Informa- tion Transmission. The Soviet research, along with other behavioristically oriented work, suggests that in addition to obtaining overt responses such as verbalizations or key presses from a subject, it should be possible to obtain objective evidence of informa- tion transfer by direct measurement of physiological parame- ters of a subject. Kamiya, Lindsley, Pribram, Silverman, Walter, and others brought together to discuss physiological methods to detect ESP functioning, have suggested that a whole range of electroencephalogram (EEG) responses such as evoked potentials (EP's), spontaneous EEG, and the contingent negative variation (CNV) might be sensitive indicators of the detection of- remote stimuli not mediated by usual sensory .:p1rocesses [46]. Early experimentation of this type was carried out by Douglas Dean at the Newark College of Engineering. In his Approved For Release 21003/04/18 : CIA-RDP96-00789ROO3100030001-4 334 BWf4 OF Approved For Release 2003/04/18 : CIA-RDP96-00789RO03 1OO Search for physiological correlates of information transfer, he used the plethysmograph to measure changes in the bloo@ 0 volume in a finger, a sensitive indicator of autonomic nervous system functioning [471. A plethysmographic measuremen-. j was made on the finger of a subject during telepathy expen- ments. A sender looked at randomly selected target car& L consisting of names known to the subject, together with names unknown to him (selected at random from a telephone book). The names of the known people were contributed by the sub- ject and were to be of emotional significance to him. Dear. < @ found significant changes in the chart recording of fingC7 _J 0 blood volume when the remote sender was looking at those > names known to the subject as compared with those names randomly chosen. Three other experiments using the physiological approach have now been published. The first work by Tart (48 ], a late: work by Lloyd [49], and most recently the work by the authors (4] all follow a similar procedure. Basically,asubject is closeted in an electrically shielded room while his EEG is recorded. Meanwhile, in another laboratory, a second person is stimulated from time to time, and the time of that stimulus is marked on the magnetic-tape recording of the subject's EEG. The subject does not know when the remote stimulus periods are as compared with the nonstimulus periods. With regard to choice of stimulus for oui own experimenta. tion, we noted that in previous work others had attempted. without success, to detect evoked potential changes in a sub. ject's EEG in response to a single stroboscopic flash stimulus observed by another subject [501. In a discussion of that experiment, Kamiya suggested that because of the unknown temporal characteristics of the information channel, it might be more appropriate to use repetitive bursts of light to increase the probability of detecting information transfer [51). There. fore, in our study we chose to use a stroboscopic flash train of I" duration as the remote stimulus. In the design of the study, we assumed that the application 04@ the remote stimulus would result in responses similar to those obtained under conditions of direct stimulation. For example, when an individual is stimulated with a low- frequency (< 30 Hz) flashing light, the EEG typically shows a decrease in the amplitude of the resting rhythm and a driving of the brain waves at the frequency of the flashes [521. We hypothesized that if we stimulated one subject in this manner (a putative sender), the EEG of another subject in a remote room with no flash present (a receiver) might show changes in alpha (9-11 Hz) activity and possibly an EEG driving similar to that of the sender, or other coupling to the sender's EEG [53), The receiver was seated in a visually opaque, acoustically and electrically sh'ielded, double-walled steel room about 7 m from the sender. The details of the experiment, consisting of seven Tuns of thirty-six 10-s trials each (twelve periods each for 0-liz, 6-Hz, and 16-Hz stimuli, randomly intermixed), are piesented in [4). This experiment proved to be successful. The receiver's alpha activity (9-11 Hz) showed a significant reduct 'ion in average power (-24 percent, p < 0.04) and peak power (-28 percent, p < 0.03) during 16-Hz flash stimuli as compared with periods of no-flash stimulus.. [A similar response was observed for 6-Hz stimuli (-12 percent in average power, -21 percent in peak power), but the latter result did not reach statistical significance.) Fig.2 shows an overlay of three averaged EEG spectra from of the subject's 36 t!rial Tuns, displaying differences in alpha activity during the three stimulus conditions. Extensive control pr6cedures were undertaken t6 determine if these THE IEEE, MARCH lz76 5 H2 10 Hz 15 Hz Fig. 2. Occipital EEG freqi:@,ncy spectra, 0-2011z, of one subject (H_--i.) acting as receiver showing 3naplimcle changesin the 9-11-Hz band ii a function of strobe frequency. Three cases: &-. 6-, and 16-liz flBi.-ds (twelve trial averages). results were produced by system artifacts, clectromagne:ic pickup (EMI), or. subtle cueang; the results were negative [-' I . As part of the @xperirnenral protocol, the subject was ash@d to indicate a con.Pcious 1-s-sessment for each trial (via telegrL-2h key) as to the nature c` the stimulus; analysis showed thAie guesses to be at ckance. Thus arousal as evidenced by sign.",- cant alpha blocking occ-zred only at the noncognitive level )f physiological response, Hence the experiment provided dirt,:t ph@siblogical (EEG) evi_'enc@@ of perception of remote stirn-_-@ even in the absene of overt cognitive response. Whereas in our :expenmen--_s we used a remote light flash 2-- a stimulus, Tart [481 in his -work used an electrical shock :o himself as sender, and Lloyd [491 simply told the sender 70 think of a red t 1riangle each time a red warning light 'Aas illuminated wit= his view. Lloyd observed a consisttnt evoked potential in his subiects; whereas in our experimeT.:s and in Tart's, a r :eduction i-n amplitude and a desynch_ron@:a- tion of alpha was obser-i-ed-an arousal response. (If a subj-.,-t is resting in an alpha-dom--inant condition and he is th-a stimulated, for example ;n a_ny direct manner, one will obse-;e a desynchronization anZ* cle--,@rease in alpha power.) We ccn- sider that these combinel-i ressulfs are evidence for the exister._@e of noncognitive awareness of remote happenings and that tht-y have a profound implication for paranormal research. III. SRI INVESTIGATIONS OF REMOTE VIEWING Experimentation' in remoi:e viewing began during stud-.---s carried out to investigate the abilities of a New York arti;t, Ingo Swann, when he expreswcl the opinion that the insie.*s gained during experimenB at SRI had strengthened his abili-y (verified in other research before he joined the SRI program) to view remote I(ocations [54). To test Mr. Swann's ass@.-- tion, a pilot study was set up in which a series of targets from around the globe were supplied by SRI personnel to the ex- perimenters on a double-blind. basis. Mr. Swann's ipparent ability to describe correcrly details of buildings, roa@.S, bridges, and the like indicated that it may be possible for a subject by means of mental imagery to access and descri-be randomly chosen geog7aphical sites located several mi.].-s from the subject's: position and demarcated by some appio- priate means. Therefore, we set up a research program to test the remote-viewing hypothesis under rigidly controlltd icientific conditions. In carrying out this program, we concentrated on what we considered to be. our principal responsibility-to resolve under unambiguous conditions the basic issue of whether or not t1is Approved. Fo Ir Release 2003104118: CIA-RDP96-007 89 ROO 3100030001-4 PUTHOFF AND TARG: PERCEPTVA@ CHANNEL F Approved For Release __ - --b4ii"-8F':'~W2'R'D~bg~d6~69ROO3100030001-4 class of pa-ranormal perception phenomenon exists. At all times, we and others responsible for the overall program took measures to prevent sensory leakage and subliminal cueing and to prevent deception, whether intentional or unL-itentional. To ensure evaluations independent of belief str-uctu:es of both experimenters and jLdges, all experiments were carried out under a protocol, described below, in which target selection at the beginning of experiments and blind judging of results at the end of experiments were handled independently of the researchers engaged in carrying out the experiments. Six subjects, designated S I through S6, were chosen for the study. Three were considered as gifted or experienced subjects (S I through S3), an,--'- three were considered as learners (S4 through S 6). The c priori dichotomy between gifted and learners was based on the, experienced group having been successful in other s-udies conducted before this program and the learners grOUD being inexperienced with regard to paranormal, experimentation. The study consisted of a series of double-blind tests with local targets in the Sari Francisco Bay Area so that several in- dependent judges could visit the sites to establish documenta- tion. The protocol was to closet the subject with an expen- menter at SRI and a-, an agreed-on time to obtair. from the subject a description of an undisclosed remote site bein g visited by a target team. In each of the experiments, on .e of the six program subjects served as remote-viewing subject,. and SRI experimentens served as a target demarcation team at the remote location chosen in a double-blind p:otocol as follows. In each experimem, SRI management randomly chose a target location from a list of targets within a 30-min dri'ving time from SRI; the target location selected was kept blind to subject and experimenters. The target pool consisted of more than 100 target locations chosen from a target-rich environ- ment. ..(Before the experimental series began, the Director of the Information Scien'ce and Engineering Division, not other- wise associated with the experiment, established the set of lo- cations as the target pool which remained known only to him. The target locations were printed on cards sealed in envelopes and kept in the SRI Division office safe. They were available only with the personal assistance of the Division Director who issued a single random-number selected target card that con- stituted the traveling orders for that experiment.) In detail: To begin the experiment, the subject was closeted with an experimenter at SRI to wait 30 min before beginning a narrative description of the remote location. A second ex- perimenter then obtained from the Division Director a target location from a set of traveling orders previously prepared and randomized by the Director and kept under his control. The target demarcation team, consisting of two to fo ur SRI experi- menters, then proceeded by automobile directly to the target without any communication with the subject or experimenter remaining behind. The experimenter remaining with the sub- ject at SRI was kept ignorant of both the particular target and the target pool so as to eliminate the possibility of cueing (overt or subliminal) and.to'allow him freedom in questioning the subject to clarify his descriptions. The demarcation team :remained at the target site for an agreed-on I 5-min period following the 30 min allotted for travel.4 During the obserya- 335 tion period, the remote-viewing subject was ask-d to d,-scribe his impressions of the target site into a tape re-,order ind to make any drawings he thought appropriate. An nfomai com- pariSOD was then made when the demarcation t!im ret-@@rned, and the subject was taken to the site to povide ft@dback. A. Subject SI: Experienced To begin the, series, Pat Price, a former Califorra polic- com- missioner and city councilman, participated 2z a >-ub-@ct in nine experiments, In general., Price's abilit% to d-scribe correctly buildings, docks, roads. gardens, and tht Like, riclud- ing structural materials, color, arnbience, and -_-tivity-often in great detail -in dica Ted the functioning of - rernol! per- ceptual ability. A Hoover Tower target, for txamplt, was recognized and named by name. Nonet-ieless, L gener-i', the descriptions contained inaccuracies as well as -orrect state- A typical example is indicated by the sub, -!cCs d7lwing ments. shown in Fig. 3 in which he correctly describ-Z' a pa:k-Like area containing two pools of water: one recta.--gulai, zO by 89 ft (actual dimensions 75 by 100 ft); the c-her ci:zula-r, diameter 120 ft (actual diameter I 10 ft). He in:orrectl% indi- cated the function, however, as water filtratio-- rar-he- than recreational swimming. (We often observe essendally c-rrect descnptions of basic elements and patterns coL-)IeC, w-:h in- complete or erroneous analysis of funct'on.) Al cr-- bt seen from his drawing, he also 'included some elernmrir-s. si:_4 as the tanks shown in the upper right, that are not -_-esent - the target site. We also note an apparent lef--right j"-Versal,often observed in paranormal perception expei-L-nents. To obtain a numerical evaluazion of the acc-!ra--y c-1 the. remote-viewing experiment, the experLniental rasialts were subjected to independent judging on a blind bats by an SRI research analyst not other-,@--ise associated with -he res-arch. The subject's response packets, which contair-,-d the nine typed unedited transcripts Of The tap-.-record-d naxiltives along with any associated drawings, were unlab -zled an@ pre- sented in random order. While standing at eac- t2z-get loca- tion, visited in turn; the judge was.required to blL@d ramkorder the nine packets on a scale 1 to 9 (best to worstniatch)- The statistic of interest is the sum of ranks assigned -o the lirget- associated transcripts, lower values indicating be-,er ma*ches. For nine targets, the sum of ranks could range from rii-le to eighty-one. The probability that a given sum cf ranks s or less will occur by chance is given by [55) k Pr (s or less) -NTn Y_ 7- (- 1), n) (i I=n 1=0 where s is obtained sum of ranks, N is number of as-zsig-riable ranks, n is number of occasions on which rank-inp were made, and I takes on values from zero to the least positive integer k in (i - n)ln. (Table I is a table to enable easy kpplication of the above formula to those cases in which N = n.) The sum in this case, which included seven direct hits out of :he nine, was 16 (see Table II), a result significant at p = 2.9 X 10-5 by exact calculation. In Experiments 3, 4, and 6 through 9, the subject was se- cured in a double-walled copper-screen Faraday cage. The Faraday cage provides 120-dB attenuation for plane-wave 4 ridio-frequency radiation over a range of 15 kliz to I GHz. The first subject (SI) was allowed 30 min for his descriptions, but For magnetic fields, the attenuation is 68 dB at 15 kHz and it was found that he fatigued and hadlittle comment after the first 15 . min. The viewing. time was therefore teduced to'l 5 min for subjects decreases to 3 dB at 60 Hz. The results of rank ordexjudging S2 through S6. _(jable II) indicate that the usb of Faraday cage electrical Approved For Release 2003104118: CIA-RDP96-00789ROO31000300 1-4 336 :PROCEEDINGS OF THE IEEE, NIAKCH 1@76 Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 _L_J I AV E. PLAY 'ROUNDS 0 DIST. POOL HOUSE SERVICE CONCRETE BLOCK YARD FIRE N STATION PICNIC A_R@A/ (a) (b) Fig. 3. Swimming pool compltx as remote-viewing target. (a) City map of target lo,cation. (b) Drawing by Price (S 1). TABLE I CRITICAL VALUES OF SUMS OF RANKS FOR PREFERENTIAL M ATCJiING Fumber of Prob abili:y (one -taile d) that the i ndicated Sum o f Ranks or L4ss Would Occur by Ch ance Assignable R. nks (N) 0 20 0.10 0.05 0.04 0.025 0.01 0,005 0.002 0.001 O.OOQ5 10-1 IO-s 10-6 10- 4 7 6 5 4 4 5 11 10 9 8 8 7 6 6 5 5@ 6 16 15 13 1 13 12 11 10 9 8 7@ 6 7 22 20 18 18 17 15 14 12 12 11 9 8 8 29 27 24 24 22 20 19 17 16 15 13 11, 9 8 9 37 14 11 30 11 26 14 22 11 17 14 12 10 10 46 42 39 38 36 33 31 29 27 25 22 19 16 13 11 56 51 48 47 45 41 38 36 34 32, 28 24 20 17 12 67 61 58 56 54 49 47 43 41 39 35 30 25 22 Note:.ThiS table applies only to those Special' Cases in which the number of occasions on which objects are being ranked (n) is equal to the number of assignable ranks (N). Each, entry represents the largest number that is significant at the indicated p-level. Source: R. L. Morris [55 1 shielding does not prevent high-quality descriptions from being B. Subject S4: Learner obtained. This experiment was designed to be a replication of our pre- As a backup judging procedure, a panel of five additional vious experiment with Price, the first replication attempted. SRI scientists not otherwise associated with the research were The subject for thIs experiment was Mrs. Heua Harnmid, a asked simply to blind match the unedited typed transcripts gifted professional photographer. She was selected for this- (with associated drawings) generated by the remote viewer series on the basis. of her successful performance as a per- against the nine target. locations which they independently cipient in the EEG experiment described earLier. Outside of visited in turn. The transcripts were unlabeled and presented that interaction, she. had no previous experience with appartnt in random order. A correct match consisted of a transcript paranormal functioning. of a given date being matched to the target of that date. In- At the time we began working with Mrs. Hammid, she had stead of the expected number of I match each per judge, the no strong feelings about the likelihood of her ability to suc- number of correct matches obtained by the five judges was 7,. ceed in this task. ;This was in contrast to both Ingo Swann @6, 5, 3, and 3, respectiv8ly. Thus, rather than the expected who had come to our laboratory fresh from a lengthy and total number'of 5 correct matches from the judges, 24 such apparently successftil series of experiments with Dr. Gertrude Matches were obtained. Schmeidler at City:; College of New York [ 5 61 and Pat Price Approved For Release 2003104118: CIA-RDP96-00789ROO3100030001-4 PUTHOFF AND TARG: PERCEPTUAL CHANNEL FOR INFORMATION TRANSFER 331 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 PEDESTRIAN OVERPASS TARGET ------------- Fig. 4. Subject H:mmid (S4) drawing, described as "some kind cf diagonal trough up in :-1e air." TABLE 11 DISTRIBUT.TON OF RAINKINGs AssIGSED To TRAqSCRIPTS AssocIATED NVITH EACH TARGET LOCATION FOR EXPERIE14CED SUBJECT PRICE (SI) scientific rigor, one of our priinLry tasks as restarchers is to provide an -environment in whic@ the subject -eels safe to explore the possibility of parano-nal perception. With a new subject, we Elso try to stress the aorluniqueness :f the abilio, because from our experience pa:--,normal functicning appears latent abil@ty that all su_:jects artici:late to be Targe: Location (k= 1 Transcripc some can a to Hoover Tower, SLanford 3.1' 1 degree. Baylands Nature Preserve .i Palo Alto 6.-' 1 Because of Mrs. Hammid's arti.-tic backgrount., she was ca- Z 6 pable of drawing and describing visual images that she could PcadLo telescope, Portola Valley . . I not identify in any cognitive or an-lytic sense. When the target Marina, Redwood City 6.8 1 demarcation team went to a niget location -which was a Bridge toll plaza, Fre@ont 14.5 1 6 pedestrian overpass, the subject !aid that she saw "a kind of Drive-in theater, Palo Alto 5. 1 1 trough up in the air," which she indicated in the upper part Arts and Crafts Plaza, Menlo Park 1.9 1 of her drawing in Fig. 4. She Rent on to explain, "If you Catholic Church, Portola Valley 8.5 3 stand where they are standing you will see something like S-1-ing pool cmplex, Palo Alto 3.4. 1 --- this," indicaag the nested squa:es at the bottom of Fig. 4. Total am of ranks 1 As it turned out, a judge standing where she ind-cated would --) (P-1. 9.10 have a view closely resembling w@at she had drawn, as can be t l h i h f h f h oca- s o t seen rom t e targe e accompany otograp ng p tion. It needs to be emphasized, however, thatjL:dges did not who felt that he used his remote-viewing ability in his every- have access to our photographs of the site, wed here for (lay life. illustrative purposes only, but railler they procetcled to each In comparison with the latter two, many people are more of the target locations by list. jhfluenced by their environment and are reluctant under In another experiment, the subject described seeing "an ublic scrutiny to attempt activities, that are generally thought P open bamlike structure with a p@tched roof." She also saw to be impossible. Society often provides inhibition and nega- a "kind of slatted side to the strLcture making Light and dark tive feedback to the individual who might otherwise have bars on the wall."' Her drawing and a photograph of the explored his own nonregular perceptual ability. We all share associated bicycle shed target are shown in Fig. S. (Subjects an historical tradition of "the stoning of prophets and the ate encouraged to make drawing! of anything they visualize burning of witches" and, in more modem times, the hospitaliza- and associate with the remote location because drawings they tion of those who claim to perceive- things that the majority do make are in general more accurate than their verbal description.) not admit -to. seeing. Thereforb in addition to maintaining As in the original series with Price, the results of the nine, , Approved For ikilease 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Disca%tt Rank of Associatad 338 PROCEEDI GS OF THE IEEE, NIARCF 197t Approved For Release 2003/04/18: CIA-RDP96-00789RO03 10003-A - DETAIL OF BICYCLE SHED @,4 \16A Z, Fig. 5. Subject Hammid (S4) response to bicycle shed target described as 2-n open "barn-like building" with "Slats on the sides" and a "pitched roof." TABLE III Again, a: a backup judging procedure, a panel of five 2ddi- DI@rRIBUTION oF RAS KINGS Ass iG NED To TRANSCRIPTS ASSOCIATED tional judzes not otherwise associated with the research wmi EAcii TARGET LocATION FoR LEARNER SUBJECT HAmmn) (S4) were asket simply to blind match the unedited typed :ran- Rank of scripts and associated drawings generated by the remote vitwer, Target Locatlon I C a n C e (; ASsociated T against the nine target locations which they independently an) ranscript visited in nim. A correct match consisted of a transcri-t of Hezzodist Church, Palo Alto 1.9 1 a given dE-,e being matched to the target of that date' In- Nets A%>ditcrlum, Menlo Park 0.2 1 stead of fhe expected number of I match each per j@_!dge, Merzy-go-rcund, Palo Alto 3.4 1 the number of correct matches obtained by the five j,--dges Parl@fn& gs@age, Mountain Viev 8.1 2 was 5, 3, 3, 2, and 2, respectively. Thus, rather than the ex- SRI International Courtyard, Menlo Park 0.2 1 pected tot@_-J number of 5 correct matches from the jiliges, Bicycle shad, Menlo Park 0.1 2 15 such mEtches were obtained, Railroad trestle bvidgep Alto Palo 1.3 2 .Puo9kin patch, Menlo Park .1.3 1 Pedeitrian overpass, Palo Alto 5.0 2 7otst am of ranks 13 experiment series were submitted for independent judging on a blind basis by an SRI research analyst not otherwise associ- ated with the research. "While at each target location, visited in turn, the judge was required to blind rank order the nine unedited typed mariuscripts of the tape-recorded narratives, along with any associated drawings generated by the remote viewer, on a scale I to 9 (best to worst match), The sum of ranks assigned to the target-aWciated transcripts in this case Was 13 ' a result significant@.at-p = 1.8-X 10-6 by exact calcula- tion isce Table I and discdsision)@, and included five direct hits and four sedond ranks (Table 111). 40P96-00789ROOM Approved For Release 2003/04/18: CIA C. Subject: S2and@S3: Experienced Having ccmpleted a series of 18 remote-viewing experiments, 9 each with experienced subject S I (Price) 'and leame- S4 (Hammid), additional replication experiments, four with each subject, we-e canie4 out with experienced subjects S 2 (Elgin) and S 3 (S-@;-ann) and learners S 5 and S 6. To place the judging on a basis comparable to that used with S I and S4, the four transcripts tach of experienced subjects S2 and S3 were com- bined into a group of eight for rank order judging to be Corn- pared with the similariy combined results of the leamers SS and S6. The series with S2 (Elgin, an SRI research analyst) pro*,ided a'further example of the dichotomy between verbal and draw- ing responses. (As with medical literature, case histories often are more ilturnimating than the summary of results.) 'The ex- periment described her tiIe third conducted with this SAO& -4 BICYCLE SHED TARGET L PUTHOFF AND TAB" N N E @6b:p IATION TRANSFER _1108%pcfft@ Welease CIA-RDP96-00789ROO3100030001-4 339 Fig. 6. Subject Elgin,(S2) dawings in response to tennis court target. gabject. It was a demonstration experiment for a government visitor who had heard of our work and wanted to evaluate our e@.perimental protocol. In the laboratory, the subject, holding a bearing compass at Eam's length, began the experiment by indicating the direction of the target demarcation team, correctly to within 50. (In all four experiments with this subject, he has aiways been within .10" of the correct direction in this angular assessment.) The subje,ct then generated a 15-min tape-recorded description and the drawings shown in Fig. 6. In discussing the drawings, Elgin indicated that he was uncertain as to the action, but had the impression that the demarcation team was located at a museum (known to him) in a particular park. In fact, the target was a tennis court lo- cated in that park about 90 m from the indicated museum. Once again, we note the characteristic (discussed earlier) of a resemblance between the. target site and certain gestalt ele- ments of the subject's response, especially in regard to the drawings, coupled with incomplete or erroneous analysis of the significances. Nonetheless, when rank ordering transcripts I through 8 at the site, the judge ranked this transcript as 2. This example illustrates a continuing observation that most of -the correct information related to us by subjects is of a non- analytic nature, pertaining to shape, form, color, @and material rather than to function or name.. experiffients, he dictates two lists for us to -ecord. 0--e list. contains objects that he "sees," but does -ot Ehir.-k a_:-@ Iccate:-, at the remote scene. A second list con-ains o`LJecLS t-at r@t thinks are at the scene. In our evaluatio.--, he h2s made muc-h. progress in this most essential dbility -o sepaate m!-,nor-j and imagination from paranormal inputs. T-his is the 12y to bringing the remote-viewing channel to fraition With Tegird vo its potential usefulness. The quality of transcript that can be gtnerateb by- thi, pro- cess is evident from the results of our most recent expe"meml with Swann. The target location chosen by the usual d-.,uble- blind protocol was the Palo Alto City HalL Swann descrbed ?. tall building with vertical columns and "!e_t in" windovi,:. His sketch, together with the photograph of the site, is shcvn in Fig. 7. He said there was a fountain, "but I don't heL: it-" At the time ,the target team. was at the City H211 ftri:g the experiment, the fountain was not running. lie also) mEie a-n effort to draw a replica of the designs in t:ie pavememt ir front of the building, and correctly indicated 'the number oJ tree-5 (four) in the sketch. For the entire series of eight, four each from S2 amd S3, the numerical evaluation 'based on blind rank ordering of tran- scripts at each site was significant at p = 3.9 X 10-4 a7-d im- eluded three direct hits and three second ranks for t1be iarget- associated transcripts (see Table IV). A second example from this group, generated by S3 (Swann), .indicates the level of proficiency that can be attained' with D. Subjects S5 and S6:.Learners -practice. In the two years'since we first started working with To complete the series, four experiments each were carried Swann, he,has been studying the pio'blem of separating the ex- out with learner subjects S5 and S6, a m2n and woman on. the o&fth stay, 0 The results in this case, takei. as a ternal ,signal -AVpr6Ve~dtF6plRLq4bs'dn2ftS/ff4"8rv 9- 00 0030001-4 TARGET-TENWS COURTS 340 PROCEEDIINGS OF THE IEEE, IMARCH 197( 9nn-31n4118: CIA-RDP96-00789ROO3100030001-4 WX& ux@:@ tub er@k Wl" OtIA i,M4 .4 - a. )ez (_V@_ 0@0% WAX,,@ Fig. 7. Subject Swa=- (S3) response to City Hall target. group, did not differ significantly from chance. For the series of eight (judged as a group of seven since one target came up twice, once for each subject), the numerical evaluation based on blind rank ordering of trariscripts at each site was non- significant at p = 0.08, even though there were two direct hits and two second ranks out of the seven (see Table V). One of the direct hits, which occurred with subject S6 in her first experiment, provides an example of the "first-time effect" that has been rigorously explored and is well-known to experi- menters in the field (57]. The outbound experimenter obtained, by random protocol from the pool, a target blind .to the experimenter wilh th@ subject, as is our standard pro- 'cedure, and proceeded to 1he location. The subject, a math@5- matician in the'computer science laboratory who had no pre- TABLE IV DISTRIBUTION OF RANKINGS ASSIGNED To TRANscRIPTs AssoclATED wiTH EACH TARGET LOCATION Fop, ExPERIENCED SUBJECTS ELGIN (S2) AND SWANN (S3) Sub I ect Target Locatim Distance Rank f Asiocl&ced Transcript S2 BART Station (Transit system), Fremont 16,1 1 S2 Shielded room, SR1, Menlo Park 0.1 2 s2 Tennis court, Palo Alto 3.4 2 S2 Golf course bridge, StAn!ord 3.4 2 S3 City Hall, Palo x1to 1.0 1 S3 Miniature golf course, henlo ?ark 3.0 1 S3 Kiosk in park, Menlo Par-, 0.3 3 S3 B.Ylads X.At.r@ Pr-tr't. pal. Alto 6,4 3 Total sum of ranks 15 (p.3.BYLO-1) TABLE V DISTRIBUTION OFtRANKINGS ASSIGNED To TRANSCRIPTS ASSOCIATED WlTji EACH TARGET LOCATION FOR LEARNER SUBJECTS S5 AND S6 Subject ;Target Locati@n Distance Okm) Pa n)c o f Assc,cL.t.d I Transcript S5 Fades Lan overpass, ralo L.1to , ' 3.0 3 S5 Railro a d trescLe bridit, ?alo Alto *..3 6 S5 Windmill, porcola Valley Z.5 2 S5, S6 White Ploza, $tanford (2) 3.8 1 S6 AirporL,!Palo Alto 5 . 5 2 56 Klosk insPark, Menlo -ark 0.3 5 S6 Boathouse Scanford I Total um of ranks 20 F (I .08, NS) Yiou s experience in remote viewing, began to describe a large square with a fountain. Four minutes into the experiment, she recognized the location and correctly identified it by name (see Fig. 8)@ (It should be noted.that in the area from which the target locations were drawn there are other fountains as well, some of which were in the target pool.) As an ex- ample of the style of the narratives generated during remote viewing with inexperienced subjects and of the part played by the experimenter :remaining with the subject in such a case, we have included the entire unedited text of this experiment as Appendix B. E. Normal and'Paranormal: Use of Unselected Subjects in Remote Viewing , After more than a year of following the experimental pro- tocol described above and observing that even inexperienced subjects 'generated results better than expected, we initiated a series of experiments to explore further whether individuals other than putative "psychics" can demonstrate the remote- viewing ability. to test this idea, we have a continuing pro- gram to carry out additional experiments of the outdoor type with new subjects whom we have no a priori reason to believe have paranormal perceptual ability. To date we have collected data from five experiments with two individuals in this cate- gory: a man and a woman who were visiting government scientists interested in observing our experimental protocols. The motivation for these particular experiments was twofold. First, the experiments provide data that indicate the level of proficiency that can be expected from unselected volunteers. Approved For Release 2003104/18: CIA-RDP96-00789ROO3100030001-4 bj> t3 NO 1040 dVvA. PUTHOFF AN hil, CHA W' ecffior e ease NV3f&ki~':'L44-i~6F~k-bbW9ROO3100030001-4 p - 7_5 AMMM ti t-d Fig. 6. Subject (S6) drawing of White Plaza, Stanford University. Sub- ject drew what she called "curvy benches" and then announced cor- rectly that the place was "White Plaza at Stanford." Second, when an individual observes a succes sful demonstra- tion experiment involving another person as subject, it inevi- tably occurs to hiin that perhaps chicanery is involved. We have found the most effective way to settle this issue for the observer is to have the individual himself act as a subject so as to obtain personal experience against which our reported results can be evaluated: The first visitor (VI) was invited to participate as a subject in a t.hree-experiment series. AM three experiments contained elements descriptive of the associated target locations; the quality of response incre9sed with practice. The third re- RESPONSES OF VISITING sponse is shown in Fig. 9 N@here again the pattern elements in - SCIENTIST SUBJECT the drawing appeared tole a closer match than the subject's analytic interpretation of the target object as a cupola. Fig. 9. Subjeect (VI) drawing of merry-go-round target. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 7@ `7 341 342 PROCEEDINGS OF THE 1EE-E, MARCH 1976 Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 TECHNOLOGY SERIES I TYPEWRITER TARGET &.p ft 1.p J-d,# d04n- - CC put" a, SUBJECT SWANN (S3) RESPONSE 'T' d SUBJECT HAMMID (S4) RESPONSE Fig. 10. Drawings of a typewriter target by two subjects. The second visitor [V21 participated as a subject in two ex- TABLE V1 : periments. In his first experiment, he generated one of the DISTRIBUTION OF P-ANK INGs ASSIGNED To TRANSCRiPT@ AssociATED W17,I higher signal-to-noise results we have observed. He began EACH TARGET LOCATION FOR. VISITOR SUBJECTS VI AND V2 his narrative, "There is a red A-frame building and next to it lank of is a large yellow thing [a tree-Editor]. Now further left Subject Target Location o=) Ty-scrip, there is another A-shape, It looks Ue a swing-set, but it is pushed down in a gully so I can't see the swings." [All cor- vi bridge over stream, Menlo Park E,3 I rect.] He then went on to describe a lock on the front door vi Baylands Nature Preserve, Polo Alto f,4 2 that he said "looks like it's made' of laminated steel, so it V1 Merry-go-rou;nd, Palo Alto @.4 1 must be a Master)ock." (Also correct.) For' the series of five-three from the first subject and two v2 Windmill, portola Valley LA 1 from the second-the numerical evaluation based on blind V2 Apartment svimming po@ol, Mountain View i.1 rank ordering of the transcripts at each site was significant at, -p = 0.0 17. and included t&ee direct hits and one second -rank Total am of ranks for tlie target-associated transcripfs. (See Yable YQ (1-0.017) Approved For Releate 2003/04/18 CIA-RDP96-00789ROO3100030001-4 k@4 .4 VW v4 .PUTHOFF AND AV0bVW`126F44e%LM 1206t)bWit~'tfA2Pb~OW-W789ROO3100030001-4 TARGET LOCATION: XEROX MACHINE (TECHNOLOGY SERIES) 343 TO ADD INTEREST TO TARGET LOCATION EXPERIMENTER WITH HIS HEAD BEING XEROXED Fig. 11. Drawings by three subjects (S2, S3, and V3) for Xerox machine target. Wher. asked to descrit-2 the s: iare at L.pper lef: Of response on the right, subject (V3) said, "There was this predominant light source which might have been a window. and a uorkjing s--rface w`ich might have been the sill, or a working surface or desk." Earlier the subject had said, "I have *,ae feeling that 1r.2re is s:mething Ahoue-ze.' against th! window." Observations with unselected subjects such as those de- Comparisons of the targ=-ts and subject irawin,-s for three of ay be a latent and scribed above indicate that remote viewing rn the multiplt-response cases (the typewn:er, Xerox machine, widely distributed perceptual ability. and video ftrminal) are shown ir. Figs. 10, 11, and 12. As is apparent forn these illustraticns aloi-.e, the experiments F. Technology Series: Short-Range Remote Viewing provide circumstantial evidence for an mform_ation channel Because remote viewing is a perceptual ability, we consid- of useful b:-, rate. This Lncludei experiments "m -which visit- ered it important to obtain data on its resolution capabilities. ing govern.-nent scientistS partiapated as su1jecis (XeroX To accomplish this, we turned to the use of indoor techno- machine and video termirial) tc observe the :)roiocol. In logical targets. general, it appears that use of M71tip]e-,--bject Yes-)onses to a Twelve experiments were carried out with five different sub- single targc provides bet@er si"-to-ncise rati:) than target jects, two of whom were visiting government scientists. They identification by a single L-ndividual. This conclusion is bomt were told that one of the experimenters would be sent by out by the judging described belov. :random protocol to a laboratory within the SRI complex and Given that in general ;he drkwin.-s tonstiturte the most that he would interact with the equipment or apparatus at accurate po-tion of a subject's dt:icription, in the first judging that location. It was further explained that the experimenter procedure a judge was asked simply to t1ind match only the remaining with the subject was, as usual, kept ignorant of the drawings (i.e., without tape transc--ipts) tothetargets- Multiple- contents of the target pool to prevent cueing during question- subject responses to a given targe.- were stapled top-ether, and ing, (Unknown to subjects, targets in the pool were used with thus seven subject-drawing res-.onse packets were to be replacement, one of the goals of this particular experiment. was matched to the seven different targets for which drawings were to obtain multiple responses to a given target to investigate made. The judge did not have Eccess to our photographs of whether correlation of a number of subject responses would the target locations, used for illustration purposes only, but provide enhancement of the. signal-to-noise ratio.) The sub- rather proceeded to each of the urget locations by list. While ject was asked to describe the target both verbally (tape standing at each target location, the judge was required to rank recorded) and by means of drawings during a time-synchronized 6rder the seven subject-drawing reiponse packets (presented in 15-min interval in which the outbound experimenter inter- random order) on a scale I to 7 (best to worst match). For acted in an appropriate' manner with the equipment in the seven targets, the sum of ranks could range from 7 to 49. The target area. sum in this case, which includet I direct hit and 4 second In the twelve experiments, seven targets were used: a drill ranks out of the 7 (see Table VIfj was 18, a result significant press, Xerox machine, video terminal, chart recorder, four- at p = 0.036. state random number generator, machine shop, and type- . In the second more detailed effort at evaluation, a visiting .writer. Three of these wer@ used twice (drill press, video scientist selected at random one of the 12 data packages (a terminal, and typewriter) ."and one (Xerox machine) came up drill press experiment), sight unv-en and submitted it for in- three times.in our random selection procedure. dependent analysis to an engineer with a request for an esti- Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 344 Approved For Release 2003104118: CIA-RDP96-00789RO03 .1006043SOF THE IEEE, MARCH 1916 TARGET: VIDEO MONITOR FOR TEXT EDITING (TECHNOLOGY SERIES) 0@2 Q@ I L`F I L@11 (a) (b) Fig. 12. Draving by nvo subjects of a video monitor target. (a) Subject (S4) dramng of "box @&ith light coming out of it ... painted flat black and in the middle of the room." (b) Second subject (V2) saw a computer termin-i-I with relay racks in tIte background. TABLE V11 TABLE V111 DitTRIBUTION OF P_ANYiNGs AssiGNED TO SUBJECT SUMMAR)-: RE1,10TE VIE"'ING DR.AWINGs AssocIATED WITH EACH TARGET LOCATION N-@@nber -! p-7.Iue, Ra=;, X@erittns 0-det Judginz W, -h natural taq :t-s 51 (expertencet 9 2.9 x 10 -5 S 2and S3 (ex-e.-.; enced) a 4' 3.8 x 10 5:. (learner) 9 6 1.8 x 10 SS and S6 (lea _@ers) 8 0.08 (NS) ';1 and V2 (lea _ers/visitors) 5 0.017 With technology :irgets 52, S31 S4 , V3 12 0.036 S), 54 Type-ziter 2 Total am of ranks tabulation of the 'statistical evaluations of these fifty-one ex- periments with nihe subjects is presented in Table VIII. The overall result, e duated conservatively on the basis of a judging procedure:that ignores transcript quality beyond that mate as to what was being described. The analyst, blind as to necessary to rank order the data packets (vastly underestimat- the target 2nd given only the subject's taped narrative and ing the statistic2l ! significance of individual descriptions), drawing (Fig. 13), was able, from the subject's description clearly indicates the presence of an information channel of alone, to correctly classify the target as a "man-sized vertical useful bit rate. Furthen-nore, it appears that the principal boring machine." difference betwein' expenenced subjects and inexperienced volunteers is not that the latter never exhibit the faculty, but G. Summary ofRemote Viewing Results rather that their n !sults are simply less reliable, more sporadic. 1) Discunion: The descriptions supplied by the subjects Nevertheless, as 6,scribed earlier, individual transcripts from in the experiments involving remote viewing of natural targets the inexperienced group of subjects number among some of or laboratory apparatus, although containing inaccuracies, & best obtained. Such observations indicate a hypothesis were sufficiently accurate to @)ermit the judges to differentiate that remote viewing may be a latent and widely distributed among various targets to the@ deece indicated. A summary perceptual ability. F S,@ J ec t Target Rank of Associated Dra.imgs S3, 54 Drill Press 2 S2, 53, V3 i Xerox machine 2 S4, 112 Video cer.inal I S3 ChArt recorder 2 S4 Random number generator 6 S4 FAchin. shop 3 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 PUTHOFF AND TABAG: PERCEPT@JAL HANNEL TION T Approved For lp i&A'V CI,A-W~i6,-"00789ROO3100030001-4 Fi 001, ,g. 13. Subject (S4) drawing of drill press showing belt drive, st and a "vertical graph that goes up and down." T'hus the primary acWevement'of the SRI program was the elicitation of high-quality remote viewing from individuals who agreed to act as subjects. Criticism of this claim could in principle be put forward on the basis of three potential flaws.. 1) The study could involve naivet6 in protocol that permits various forms of cueing, intentional or unintentional. 2) The experiments discussed could be selected out of a larger pool of experiments of which many are of poorer quality. 3) Data for the reported experiments could be edited to show only the matching elements,-the nonmatching elements being discuded. All three criticisms, however, are invalid. First, with regard to cueing, the use of double-blind protocols ensures that none of the persons in contact with the subject can be aware of the target. Second, selection of experiments for reporting did not take place; every experiment was entered as performed on a master log and is included in the statistical evaluations. Third, data associated with a given experiment remain unedited; all .experiments are tape recorded' and all data are included un- edited in.the data package to be judged and evaluated. 345 In the process of judging-attenipting to matc-L trans-ripts against targets on the basis of the informa:ion ir. the transcripts-some patterns and regularities in the tran@cript descriptions became evident, particularly regaidirg indiN-.dual styles in remote viewing and in the perceptual form of the descriptions given by the subjects. These patt-Lrns an,@ the judging procedure are discussed below, . a) Styles of response: The fifty-one transcripts were taken from aine different subjects. Compari7.g the :ran- scripts of one subject with those of another revea@ed tba, tach pattern tended to focus on certain aspects of the re-note target complex and to exclude others, so that 2ach hail an individual patt--rn of response, like a signature. Subject S3, for example, frequently responde.@ with :opo- graphical descriptions, maps, and architectural fe--cures c` the target locations. Subject S2 often focused on tht behav::r of the remote ex;cerimenter or the sequence of actions ht cL-ried out at the target. The transcripts of subject S-. more :han those of other subjects, had descriptions of the f--al of th_- lo- cation, and experiential or sensory gestalts-for exa:--nple, light/dark eleanents in the scene and indoor/cutdoo7 and enclosed/open distinctions. Prominent features of S I's :ran- scripts were detailed descriptions of what the ta:,get pe--sons were concretey experiencing, secing, or doing-for example, standing on as-phalty black-top overlooking wate-@ look-J-7 at a purple iris. The range cr.' any individual subject's responses was @ide. Anyone might draw a map or describe the mood of tht remote experimenter, but the consistency of each sulojtct's o-;erall approach suggests that just as individua-I descr-.ptians of a directly vie@wedl scene would differ, so these difftrences also occur in remoie-viewing processes. . b) Nature of the description: The concrete descrip-Jons that appear most commonly in transcripts are at the le-vtl of subunits of th'e overall scene. For example, whe:i the 'Urget was a Xerox copy machine, the responses inch;ded (S2) a rolling object (the moving light) or dials and a covey t'rat is lifted (S3), bu: the machine as a whole was not iienrifiei by name or function. In a few transcripts, the subjects correctly id.-ntffied and named the ta:get. In the case of a computer ttrmina]. the subject (V2) apparently perceived the terminal and the elay racks behind it. In the case of targets which were Hcover Tower and White Plaza, the subjects (S I and S6, Yespecti;ely) seemed to identify the locations through analysis of their initial images of the elements of the target. There were also occasional incorrect identifications. Gestalts were incorrectly named; for example, swimming pools in a park were identified as water storage tanks at a wa-.er filtration plant (S 1). The most common perceptual level was thus an intermediate one-the individual elements and items that make up the tar-. get. This is suggestive of a scanning process that takes sa.-nple perceptions from within the overall environment. When the subjects tried to make sense out of these ftagmen- tary impressions, they often resorted to metaphors or con- structed an image with a kind of perceptual inference- From a feeling of the target as an "august" and "solemn" building, a subject (SQ said it might be a library; it was F church. A pedestrian overpass above a freeway was described as a conduit (SQ. A rapid transit station, elevated above the countryside, was associated with an observatory (S2). These responses seem to be the result of attempts to process partial informa- Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 @46 PROCEEDINGS OF 7-4F_ IEEE, MARCH 197@ Approved For Release 2003/04/18: CIA-RDP96-00789ROO3100030001-4 don: similarly, this occurs in other parapsychological exDeri- not be visible to an observer merely st-iding at grourxi: level L Leses ments. These observations are compatible with the hypot and describing what he sees. (In partkular, a subject often .hat information received in a putative remote-viewing mode correctly describes clements'not visible :,D the target de-narca- @i processed piecemeal in pattern form (consistent with 2 low tion team.) Finally, motion is seldom reported; i:-. fact, ,it rate process, but not necessarily requiring it); and the moving objects often are unseen eye7. when nearby static !ZrOTS arise in the processes of attempted integration of the. objects are correctly identified. -4ata into larger patterns directed toward verbal labeling. A comparison of the results of remoTt viewing (a s.-Called When the subjects augmented the verbal transcripts with free-responsee task) with results of forcef-choice, tasks, ---ch as or sketches, these often expressed the ta-rget clerntnts the selection of one of four choices ge-erated by a r-adom -more accurately than the verbal descriptions. Thus the e.-aw- number generator 1581 , reveals the follo@Lng findings. -om a ..izs tended to correspond to the targets more clearly and statistical viewpoint, a subject is more, hKely to describt. with -recisely than the words of the transcript. sufficient accuracy to permit blind ma:,;hing, a remc:e site The descriptions given by the subjects sometimes wen: be- chosen at random than he is to select --orrectly one c-' four ,iond what the remote experimenter experienced, at least :on- random numbers. Our experience with r_tse phenomen- leads iously. For example, one subject (S4) described and Crew us to consider that this difference in i-k performan,-! may belt drive at the top of a drill press that was invisible ev-a to stem from fundamental signal-to-noise --onside ratio m Two -he remote experimenter who was operating the mac*,inc,; principal source ntly art nem- s of noise in the, systerr, ippare, Lriother subject .(Sl) described a number of items bt@ind cry and imagination, both of which cL7 give rise to @@ental i@'irubbery and thus not visible to members of the demarc@:ion pictures of greater clarity than the targe: to be perceivt@. In .zarn at the site. the random number task, a subject can c-!ate a perfect -ental Curiously, objects in motion at the remote site were rL:ely picture of each of the four possible oui---its in his own @magi- mentioned in the transcript. For example, trains crossing the nation and then attempt to obtain the @orrecr answe. by a trestle target were not described, though the reizote mental matching operation. The same is -:ue fo. card g-@:tssing txperimenter stood very close to them. experiments. On the other hand, the su-ject in remo',-- view- Also in a few cases, the subject descriptions we@re inaccu@ate ing is apparently more likely to apprcl-ch the task -ith a .@igarding size of structures. A 20-ft courtyard separating :wo blank mind as fie attempts to perceive --ictoria-1 info=ation _uildings was described as 200 ft wide, and a small shed was from remote locations about which he -nay have no ;,ored -xpanded to a barn-like structure. mental data. c) Blind judging of transcripts: The judging proce_@ure Finally, we observe that most of th-- correct info=ation entailed examining the transcripts for a given experimental that subjects relate to us is of a nonana:,-@-tic narure penlining ,--,ties and attempting to match the transcripts with the --or- to shape, form, color, and material rath-.:- than to func:_--on or Y-ct targets on the basis of their correspondences. The t:an- name. Inconsult ation with Dr. Robert Oi-stein of the La@gley- s-ripts varied from coherent and accurate descriptions to -raix- Porter Neuropsy@hiatric Institute, San Francisco, CA, anf with tures of correspond ences and noncorrespondences. Since the Dr. Ralph Kiernan of the Department of Neurology, S-,---.-..Iford judge did not know a priori which elements of the descriptor14 University Medical Center, Stanford, CA, we have lormed -;.;,ere correct or incorrect, the task was complicated, and Lan- the tentative hypothesis that paxanoirzal functioning may @cripts often seemed plausibly to match more than one ta:--et. involve specialization characteristic of t-t brain's right. hemi- confounding factor in these studies is that some targe: lo- ossibility is derived from I variety of e-dence sphere. This p cations have similarities that seem alike at some level of -)er- from clinical and neurosurgical sources v,-ich indicate that the c@tption. For example, a radio telescope at the top of a aill, two hemispheres' of the human brain at specialized 1-nr dif- t:ie observation deck of a tower, and a jetty on the edge of a ferent cognitive functions. The left he-misphere is pT=-Jomi- tay all match a transcript description of "looking out ov!r a nantly active in verbal and other analyticLl functioning L--d the long distance." A lake, a fountain, and a creek may all rtiult right hemisphere: predominates in spathil and other holistic in an image of water for the subject. Therefore, in sevtral processing [59], [601. Further resta-rch is necessr--7 to cases, even correct images may not help in the conservaccive elucidate the relationship between right hemisphere function differential matching procedure used. and paranorma-I abilities. Nonetheless, we can say FC this According to the judge, the most successful procedure was a point that the remote-viewing results of :he group of P-2bjects careful element-by- element comparison that tested each t:an- at SRI have characteristics in commor. with more f:.-niliar s-ript against every target and used the-transcript descripbons performances that require right hernispneric function. The a-rid drawings as arguments for or against assigning the L-an- similarities includ .e the highly schematic.-.zed drawings of ob- s:ript to a particular target.- In most cases, this resulted in jects in a room or of remote scenes. Vt.-bal identificat@:on of rther a clear conclusion or at least a ranking of probable these drawings is often highly inaccura-.e and the drawings matches; these matches were subjected to the statistical themselves are ffequently left-right rev!rsed relative to the analyses presented in this paper. target configuration. Further, written material generally is 2) Summary: In summary, we do not yet have an under- not cognized. These characteristics ha-,re been seen Ln left standing of the nature of the information-bearing signal that a brain-injured patients and in callosal-sectioned patients. subject perceives during' remote viewing. The subjects com- As a result of the above considerationti, we have learned te monly report that they perceive the signal visually as thcugh urge our subjects simply to describe what they see as opposed they were looking at the object or place from a position in, its to what they.think they are looking at. We have learntd that immediate neighborhood. Furthermore, the subjects' per- their unanalyzed perceptions are almost always a betterguide ctptual viewpoint has mobility in that they can shift their to the true target tha-n their interpretat:ons of the perceived -point of view so as to descri8e elements of a scene that would - data. Approved For Release 2003/ .04/18: CIA-RDP96-00789ROO3100030001-4 PUTHOFF AND TAP PB@*&kySl 4eE@ ftWJT§ ki IZZ Jf,RN )@ Rb bwbb789ROO3100030001-4 IV. CONSIDERATIONS CONCERNING TIME If the authors may be forgiven a personal note, we wish to express that this section deals with observations that we have been reluctant to publish because of their striking apparent in- compatibility with existing concepts. The motivating factor for presenting the data at this time is the ethical consideration that theorists endeavoring to develop models for paranormal functioning should be apprised of all the observable data if their efforts to arrive at a comprehensive and correct descrip- tion are to be succmful, During the course of the experimentation in remote viewing (Section III), subjects occasionally volunteered the informa- tion that they had been thinking about their forthcoming par- ticipation in a remote-viewing experiment and had an image Come to them as to what the target location was to be. On these occasions, the information was given only to ihe experi- menter remaining at SRI with the subject and was urilknown to, the outbound experimenter until completion of the expen'*-- iment. Two of these contributions were among the most accurate descriptions turned in during those experiments. Since the target location had not yet been selected when the subject communicated his perceptions about the target, we found the data difficult to contend with. We offer these spontaneous occurrences not as proof of pre- cognitive perception, but rather as' the motivation that led us to do further work in this flield. On the basis of this firsthand evidence, together with the copious literature describing years of precognition experiments carried out in various other labo- ratories, we decided to determine whether a subject could per- form a perceptual task that required both spatial and temporal remote viewing. It is well known and recently has been widely discussed that nothing in the fundamental laws of physics forbids the appar- ent transmission of information from the future to the present (discussed fuJrther in Section V). Furthermore, there is a gen- eral dictum that "in physical law, everything that is not forbid- den, is required" [611. With this in mind, we set out to con- duct very well-controlled experiments to determine whether we could deliberately design and execute experiments for the sole purpose of observing precognition under laboratory conditions. The experimental protocol was identical to that followed in previous remote-viewing experiments with but one exception. The exception was that 'the subject was required to describe -the remote location during a 15-min period beginning 20 min 'before the target was selected and 35 min before the outbound experimenter was to arrive at the target location. In details as shown in Table IX, each day at ten o'clock one of the experimenters would leave SRI with a stack of ten sealed envelopes from a larger pool and randomized daily, con- taining traveling instructions that had been prepared, but that were unknown 'to the two experimenters remaining with the subject. The subject for.this experiment was Hella Hammid (SQ who participated in the nmie-experiment series replicating the original Price work described earlier. The traveling experi- menter was to drive continuously from 10:00 until 10: 3 0 be- fore selecting his destination with a random number generator. (The motivation for continuous motion was our observation that objects and persons in rapid motion arenot generally seen in the remote-viewing mode of perception, and we wished the trayeler to be a poor target unW he reached his target site.) At end of 30 min of drividg,-the traveling experimenter gener- TABLE IX EXPERINIE14TAL PROTOCOL: PRECOGNITIVE REMOTE VIEWING 347 Time Sz@h.d.le Exparkmenter/Subject Activity 50:00 Outbound experimenter leaves with 10 envelopes (contitning target locations) and random number generator; begins half-hour drive L0:10 .Lth subject in the labratc-y rimenter rem ining E@p :. li It from bje:t a d scri tion of where outbound : . . experimenter will be from 10:45-11:00 1:0: 25 Subject response completed, at which time labor*-o-- part of experiment is over 0: 30 Outbound experimenter obtains random number from . r. ndm n=ber generator, counts down to associated envelope. and proceeds to target location indicated !,:):45 Outbound experimenter remains at target location for 15 minutes (10:45-11:00) ated a random digit from 0 to 9 with a Texas Instruments SR-51 random number generator; while. still in motion, he counted down that number of envelopes and proceeded di- rectly to the target location so as to arrive there by 10:45. He remained at the target site until 11:00, at which time he re- turned to the laboratory, showed -his chosen target name to a secuZity guard, and entered the -experimental room. During the same period, the protocol in the laboratory was as follows. At 10: 10, the subject was asked to begin a descrip- tion of the place to which the experimenter would go 35 min hence. The subject then generated a tape-recorded description and associated drawings from 10: 10 to 10: 2 5, at which time her part in the experiment was ended. Her description was thus entirely concluded 5 min before the beginning of the tar- get selection procedure, Four such experiments were Carried out. Each of them ap- peared to be successful, 'an evaluation later verified in blind judging without error by three judges. We will briefly sum- marize the four experiments below. The first target, the Palo Alto Yacht Harbor, consisted en- tirely of mud flats because of an extremely low tide (see Fig. 14). Appropriately, the entire transcript of the subject per- tained to "some kind of congealing tar, or maybe an area of condensed lava. It looks like the whole area is covered with some kind of wrinkled elephant skin that has oozed out to fill up some kind of boundaries where (the outbound experi- menter) is standing." Because of the lack of water, the dock Y@here the remote experimenter was standing was in fact rest- ing directly on the mud. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Fig. 34. Subject Hammid (S4) described "some kind of congealing tar, or maybe an area of condensed lava that has oozed out to fill up some kindof boundaries." 348 Approved For Release 2003104118: CIA-RDP96-00789ROO310dM66M" IEEE, MARCH 1976 Note that the subject has leamed not to rush into interpreta- tion as to the nature or purpose of the place. This is a result of our cautioning based on the observation that such efforts tend to be purely analytical and in our experience are almost invariably incorrect. If a subject can limit himself to what he sees, he is often then able to describe a scene with sufficient accuracy that an observer can perform the analysis for him and identify the place. The' second target visited was the fountain at one end of a large formal garden at Stanford University Hospital (Fig. 15). The subject gave a lengthy description of a formal garden be- hind a wall with a "double colonnade" and "very well mani- cured." When we later took the subject to the location, she was herself taken aback to find the double colonnaded wall leading into the garden just as described. The third target.was a chil *dren's swing at a small park 4.6 krn from the laboratory (Fig. 16). The subject repeated again and again that the main focus of attention at the site was a "black iron triangle that the outbound experimenter had somehow walked into or was standing on." The triangle was "bigger than a man," and she heard a "squeak, squeak, about once a second," which we observe is@a match to the black metal swing that did squeak. Fig. 16. Subject (S4) saw a "black iron triangle that Hal had somehow walked into" andheard a "squeak, squeak, about once i second." Fig.J 7. Subject (S4) described a very tall structure locE-ed among city streets and covered with "Tiffany-like glass." The final target! was the Palo Alto City Hall (F-,g. 17). The subject described a very, very tall structure :overed N;@th "Tiffany-like glm," She had it located among ciy streets and with little cubes at the base. The building is glass-covered, 2nd the, little cubes are a good match to the small elevator exit buildings located in the plaza in front of the building. To obtain a numerical evaluation of the accuracy of the pre- cognitive viewing@ the experimental results were subjected to independent judging on a blind basis by three SRI scientists who were not otherwise associated with the experiment. The judges were asked to match the four locations, which they 'visited, against the unedited typed manuscripts of the tape- recorded narratives, along with the drawings generated by the remote viewer. The.transcripts were presented unlabeled and in random order and were to be used without replacement. A correct match req: uired that the transcript of a given exptri- ment -be matched with the target of that experiment. All three judges independently matched the target data to the responw data without error. Under the null hypothesis (no information channel and a random selection of descriptions, without re- placement), each judge independently obtained a result signifi- cant at p = (4!)-' = 0.042. For reasons we do not as yet understand, the four transcripts generated in the precognition experiment show exceptional co- @erence and accuracy as evidenced by the fact that all of the judges were able to match successfully all of the transcripts to Approved For Release 2003/'04/18 : CIA-RDP96-00789ROO3100030001-4 Fig, 15. Subject (S4) described a formal garden "very well manicured" behind a double colonnade. PUTHOFF AND RG: PERC@UUA C NNE NWORMATION TRANSFER 1_0 @e'A I TApproved r lease '2011 /18 : CIA-RDP96-00789ROO3100030001-4 the corresponding target locations. A long-range experimental program devoted to the clarification of these issues and inyolv- ing a number of subjects is under way. The above four experi- ments are the first four carried out under this program; Currently, we have no precise model of this spatial and tem- poral remote-viewing phenomenon. However, models of the universe involving higher order synchronicity or correlation have been proposed by the physicist Pauli and the psychologist Carl Jung [621. ACAUSALITY. If natural laws were an absolute truth, then of course there could not possibly be any processes that deviate from it. But since causalitys is a statistical truth, it holds good only on average and thus leaves room for exceptions which must somehow be experienceable, that is to say, real. I try to regard synchronistic events as acausal exceptions of this kind. They prove to be relatively independent of space and time; they rela- tivize space and time insofar as space presents in principle no ob- stacle to their passage and the sequence of events in time is in- verted so that it looks as if an event which has not yet occurred r- were causing a perception in the present. We shall see in the next section that such a description, though poetic, has some basis in modern physical theory. V. DISCUSSION It is important to note at the outset that many contempo- rary physicists are of the view that the phenomena that we have been discussing are not at all inconsistent with the framework of physics as currently understood. Ir) this emerg- ing view, the often-held belief that observations of this type are incompatible with known laws in principle is erroneous, such a concept being based on the naive realism prevalent before the development of modem quantum theory and information theory. One hypothesis, put forward by I. M. Kogan of the USSR, is that information transfer under conditions of sensory shielding is meaiated by extremely low-frequency (ELF) electromagnetic waves in the 300-1000-kin region [37]_ 1401. Ex.perimental support for the hypothesis is claimed on the basis of slower than inverse square attenuation, com- patible with source-percipient distances lying in the induc- tion field range as opposed to the radiation field range; ob- served low bit rates (0.005-0.1 bit/s) compatible with the :information carrying capacity of ELF waves; apparent ineffec- tiveness of ordinary electromagnetic shielding as an attenuator; and standard antenna calculations entailing biologically gener- ;ated currents yielding results compatible with observed signal- tc@noise ratios. M. Persinger, Psychophysiology Laboratory, Laurentian Uni- versity, Toronto, Canada, has narrowed the ELF hypothesis to the suggestion that the 7.8-Hz "Shumann waves" and their harmonics propagating along the earth-ionosphere wave&ide dluct may be responsible. _ Such an hypothesis is compatible with driving by brain-wave currents and leads to certain other hypotheses such as asymmetry between east-west and west- east propagation, preferred experimental times (midnight-4 A.m.), and expected negative correlation between success and the U index .(a measure of geomagnetic disturbance throughout the World). Persinger claims initial support ' for these factors on the basis of a literature search [631, [64]. , On the negative side with regard to a straightforward ELF interpretation as a blanket hypothesis are the following: a) ap- 1. SAs usuaUy understood. \. - 349 parent real-time descriptions of remote activities in sufficient detail to require a channel capacity in all probability greater than that allowed by a conv entional modulation of an ELF signal; b) lack of a proposed mechanism for coding and decod- ing the information onto the proposed ELF carrier; and c) ap- parent precognition data. The hypothesis must nonetheless re- main open at this stage of research, since it is conceivable that counterindication a) may eventually be circumvented on the basis that the apparent high bit rate results from a mixture of low bit rate input and high bit rate "filling in the blanks" from imagination; countefindication b) is common to a number of normal perceptual tasks and may therefore simply reflect a lack of sophistication on our part with regard to perceptual functioning [65); and count erindi cation c) may be accom- modated by an ELF hypothesis if advanced waves as well as retarded waves are admitted [661, (67]. Experimentation to determine whether the ELF hypothesis is viable can be carried out by the use of ELF sources as targets, by the study of para- metric dependence on propagational directions and diumal timing, and by the exploration of interference effects caused by creation of a high-intensity ELF environment during ex- perimentation, all of which are under consideration in our lab- oratory and elsewhere. Some physicists believe that the recoriciliation of observed paranor-mal functioning with modern theory may take place at a more fundamental level-namely, at the level of the founda- tions of quantum theory. There is a continuing dialog, for example, on the proper interpretation of the effect of an ob- server (consciousness) on experimental measurement [68], and there is considerable current interest in the implications for our notions of ordering in time and space brought on by the observation [69], '[70] of nonlocal correlation or "quan- tum interconnectedness" (to use Bohm's term [71 ] ) of distant parts of quantum systems of macroscopic dimensions. The latter, Bell's theorem [721, emphasizes that "no theory of reality compatible with quantum theory can require spatially separated events to be independent" [7731, but must permit interconnectedness of distant events in a manner that is con- trary to ordinary experience [7 '41-[75). This prediction has been experimentally tested and confirmed in the recent experiments of, for example, Freedman and Clauser [69], [701. E. H. Walker and.. 0. Costa de Beauregard, independently Proposing theories of paranormal functioning based on quan- tum concepts, argue that observer effects open the door to the possibility of nontrivial coupling between consciousness and the environment and that the nonlocality principle permits such coupling to transcend spatial and temporal barriers [761, [771. Apparent "time reversibility"-that is, effects (e.g., observa- tiofis) apparently preceding causes (e.g., events)-though con- ceptually difficult at first glance, may be the easiest of appar- ent paranormal phenomena to assimilate within the current theoretical structure of our world view. In addition to the familiar retarded potential solutions f(t - r1c), it is well known that the equations of, for example, the electromagnetic field admit of advanced potential solutionsf(t + r/c)-solutions that would appear to imply a reversal of cause and effect. Such solutions are conventionally discarded as not corresponding to any observable physical event. One is cautioned, however, by . statements such as that of Stratton in his basic text on electro- magnetic theory [78). Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 350 PROCEEDINGS OF TiHE IEZE, MARCH 1976 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 been' The reader has doubtless noted that the choice of the function f(t - r1c) is highly arbitrary, since the field equation admits also asolutionf(t +r1c). This function leads obviously to an advanced time, implying that the field can be observed before it has been generated by the source. The famiUa-r chain of cause and effect is thus reversed and this alternative solution might be discarded as logically inconceivable. However, the application of "logical" causality principles offers very insecure footing in matters such as these ind we shall do better to restrict the theory to retarded action solely on the grounds that this solution alone conforms to the present physical data. Such caution is justified by the example in the early I 920's of Dirac's development of the mathematical description of the relativistic electron that also yielded a pair of solutions, one of which was discarded as inapplicable until the discovery of the positron in 1932. In an analysis by 0. Costa de. Beauregard, an argument is put forward that advanced potentials constitute a convergence toward "finality" in a manner symmetrical to the divergence of retarded potentials as a result of causality [77). Stich phenomena are generally unobservable, however, on the gross macroscopic scale for statistical reasons. This is codified in the thermodynamic concept that for an isolated system entropy (disorder) on the average increases. It is just this requirement of isolation, however, that has been weakened by the observer problem in quantum theory, and 0. Costa de Beauregard argues that the finality principle is maximally operative in just those situations where the intrusion of consciousness as an ordering phenomenon results in a significant local reversal of entropy increase. At this point, further discussion of the subtleties of such considerations, though apropos, would take us far afield, so we simply note that such advanced waves, if detected, could in certain cases constitute a carripr of information precognitive to the event. The above arguments are not intended to indicate that the precise nature of the information channel coupling remote events and human perception is understood. Rather, we in- tend to show only that modem theory is not without resources that can be brought to bear on the problems at hand, and we expect that these problems will, with further work, continue to yield to analysis and specification. Furthermore, independent of the mechanisms that may be involved in remote sensing, observation of the phenomenon implies the existence. of an information channel in the information-theoretic. sense. Since such channels are amenable to analysis on the basis of communication theory techniques, as indicated earlier, channel characteristics such as bit rate can be determined independent of a well-defined physical channel model in the sense that thermodynamic concepts can be ap- plied to the analysis of systems independent of underlying mechanisms. Furthermore, as we have seen from the work of RyzI discussed in Section 11, it is possible to use such a channel for error-free transmission of information if redundancy coding is used. (See also Appendix A.) Therefore, experimentation involving the collection of data under specified conditions per- mits headway to be made despite the formidable work that needs to be done to clarify the underlying bases of the phenomena. V1. CONCLUSION For the past three years we have had a program in the Elec- tronics and Bioengineering J@aboratory of SRI to investigate, .those fac 'ets of human p 'brception that appear to fall outside the range of well-understood perceptual or-processing capa- Approved For Release 2003/04/18 bilities. The pluinary achievement of zhis p;-@)gram has the elicitation of high-qual-ity "rernote viewing--the ability of both experit-'riced subjects and inexperienced volunteers to view, by mt@ns of innate mental process-ts, remote geo- graphical or te:@ltnical targets such as roads, buildings, and laborato,-y appa:@atus. Our accamulated data from over fifty experiments wi:h more than a hadf-dozen s7ibjects indicate the following. z) The phenomeriom is not a stnsitive function of distance ovei a: range of several kilometers. 'o) Faraday cage shielding does n- ',appea,- to degade- the qualir.; or accuracy of perception. c) Most of the conect information that subjects relate is of a r.manalytic nature pertaining -o shape, form, color, and mattlial rather than to funztion or name. (This aspect suggests a hypothesis ihat information transmission under conditiors of sensory slield-ing may tt mediated pri- marily ly the '-..rain's right hernis-.phere.) C, T-he principal difference betv-,en experienced subjects an_@ inexperienced volunteen is no:.that the latter newer exhibit -he faculty, but rather that theij results are simply Dess reliable. (This obserya- tion suggests t1t ' hypothesis that remote viewing may be a latent and wid. " distpbuted, though repressed, perceptual ability.) Although the ?recise nature ol- thzt informat-'Dn channel cou- pling remote evt :nts and humaT: perception ii not yet under- stood, certain toncepts in informat--on thtory, quantum theory, and neu.-@physiological research appea- ro bear directly on the. issue. AU a result, the working assum-.tion aniong re- searchers in tht :field is that f-le phenomenc'n of interest is consistent with -.podem scientific thought, ard can therefore be expected to yield to the scientific znethoC@. Further, it is recognized that communication theo.-y provides powerful techniques, suchas the use of reclu-ndancy coding to improve signal-to-noise 1E:10, which can be e--Tnployed to pursue special- purpose applicafon of the remo--e-sensing chamael independent of an understan@@hig of the underlying mechanisms. We there- fore consider it important to contimue data collection and to encourage othe7ii. to do Hkewis!; iravesligatio,-.s such as those reported here n-dd replication and extension under as wide a variety of rigomAy controlled conditions as possible. APPENDIX A SIGNAL ENHANCEMEN7 IN A PAPANORMAL COMMUNI@ATION CHANNIEL By APPLICATION OF REDUNDANCY CODING Independent of the mechanisms that may be involved in remote sensing, obspervation of the phenomenon implies the existence of an .information chamnel in the information- theoretic sense. As we have seen fiTom the work of RyzI dis- cussed in Sectioa1l,5 it is even possible to use such a (noisy) channel for emor- ree transmission of infoanation if suf- ficient redundarcy coding is used [3 0], [ 3 11. Following is a general procedum that we have ttsed successfully for signal enhancement. We shall assume that the "message" consists of a stream of binary digits (0J) of equal probability (e.g.. binary sort of green/white card3l.as in Ryzl's case, English text encoded as in Table X and sent long distance by strobe light onloff, and so on). To comba channel noise, each binary digit to be sent through the channel requires the addition of redundancy bits (coding). Efficient coding requires a compromise between the desire to maximlie reliability and the desire to minimize re- See also the note: added In proof om the succcuful work done by Carpenter. CIA-RDP96-00789ROO3100030001-4 PUTHOFF AND TARG: PERCEPTUAL CHANNEL FOR INFORMATION TRANSFER Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 TABLEX 5-BIT CODE FOR ALPHANUMERIC CHARACTERS 00000 Y 0 ION) 11111 G,J 1011i 00001 W 01001 R 11110 V 10110 00010 B 0101D 11101 0 10101 X 00011 1 01011 5, X, Z 11100 2 10IDC D 00100 3 0 1 10C '4 11011 4 1001i 00101 5 011011 71KIQ 11010 6 10010 00110 7 Oilit. 11001 8 1000i 0011.1 9 01111 11GOO 10000 Note: AJphabet characters lisied in order of decreasing frequency in Fnglis@a text. See, for example, A. Sinkov [79]. (The low-fre- qi:--ncy letters, X, Z, K, Q, and J, hzve betri grouped with similar chaia.cters to provide space for numerics in a 5-bit code.) In consideration of the uneven dis- tribution of letter frequencies in English text, this code is chosen such that 0 and I have equal probability. z 50 z V) ul ' 0 40 M C) U LU F- V) DECISION 1 DE C @ < L ' 30 Accept "I" as A pt (D h Bit -'79 the Bit Being a LL < 0 V) ' Transmitted 'nsm, td T" ECISION 3 A 0: UJ 1 2 20 C.:lntinue rans-misiion 'a n @- co DECISION 2 :E I,- =) LU Accept "0" as Z 0 cc 10 the Bit Being LU< Transmitted > F- 351 0 L'@ I J I < 7 _J - D 0 10 20 30 40 50 60 70 ;0 2 NUMBER 0:: TRIALS D L) Fig. 18. Enhancement of signal-to-noise ratio by sequential sam:ling procedure (po = 0.4, p I = 0.6, z = 0.0 1, 0.0 1). The question to be addressed is whether, after repeated transmission, a given message bit is labeled a "l" at alow-ate PO commensurate with the hypothesis Ho that the bit in crues- tion is a "0," or at a higher rate T, I commensurate with the hypothesis H, that the bit in ques-.ion is indeed a "I." The decision-making process requires :he specification of four parameters. po The probability of labeling incorrectly a "0" messag@ bit as a "L" The probability of labeling correctly a "C" as a "0" is P = 0.5 + Vlb = 0.6. Therefore, the probab2ity of labeling incorrectly a "0" asa "I" is I - p = 0.4=po. p, The probability of labeling correctly a "I" message bit as a "I," is given by p, = 0.5 -. h = 0.6. a The probability of rejecting a correct identificatior for a "0" (Type I error). We shall take a= 0.01. The probability of accepting an incorrect identific2:ion for a " 1 " (Type Il error). We shall take 0.0 1. dundancy. One efficiera coding scheme for such a channel is ob-,,uned by application of a sequential sampling procedure of the type used in production-line quality control [80]. The adpotation of such a procedure to paranormal communication chEanels, which we now discuss, was considered first by Ta-.zzsch [ 81 ] . The sequential method givts a rule of proce- du.-- for...making one of three possible decis@ions following the rectipt of each bit: accept 1 as the bit being transmitted; reject 1 as the bit being transmitted (i.e., accept 0); or continue tran-,unission of the bit under consideration. The sequential sampling procedure differs from fixed-length coding in that the 'number of bits required to reach a f-mal decision on a message bit is not fixed before transmission, but depends on the results accumulated with each transmission. 'Me principal advantage of the sequential sampling procedure as compared with the other methods is that, on the average, fewer bits per final decision are required for an equivalent degree of reliability. Use of the sequential sampling procedure requires the speci- fication of parameters that are determined on the basis of the following considerations. Assume that a'message bit (0 or 1) is being transmitted. In the absence ofa priori knowledge, we may assume equal probab 'ility (p = 0.5) for the two possibili- ties (0, 1). Therefore, from the standpoint of the receiver, the prc;DabWty of Correctly identifying the bit being transmitted is p = 0.5 because ol chance'alone. An operative remote-sensing channel could then be expected to alter the probability of correct identification to a value p = 0.5 + @, where the param- ettr lp satisfies 0 < 141 1 <' 0.5. (The quantity may be positive or negative depending on whether the paranormal channel restilts- in so-called psi-hitting or psi-missing.) Good psi func- tioning on a repetitive task has been observed to result in * = 0.12, as reported by Ryzl [@lj. Therefore, to indicate the '@esign -procedure, let us %.assume a -baseline psi parameter 41b = 0. 1 and desi .tion system on this basis. gn a communica With the parameters thus specified, the sequential sampang procedure provides for construction of a decision grap-- as shown in Fig. 18. The equations fo: the upper and lower limit lines are d, + S-V 0 =-do + SN here w 1-0 1 log - log 01 d, I - P o do Pi 1 - Po log log Po 1 - Pi Po 1 - pi I -po log - S=- - 1-pi log L1_ I P 0 Po I- pi ih which S is the slope, N is the number of trials, and d, and do are the y-axis intercepts. A cumulative record of rec6rer- generated responses to the target bit is compiled until either Approved For. Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 352 PROCEEDINGS OF THE IEEE, MARCH 1976 Approved or Release 2003/04/18 CIA-RDP 96-00789ROO3100030001-4 1.0 T-r- viewing. The target, determined by random procedure, was White's Plaza, a plaza with fountain at Stanford University 1 (shown in Fig. 8). As is our standard protocol, the experi- 0.8 menter with the subject is kept ignorant of the specific target visited as well as the contents of the target pool. The experi- >- menter's statements and questions are italics. C: 0.6 .0 @0 Today is Monday, October 7th. It is 11.-00 and this is a re- m .0 cu - inote viewing experiment with Russ Targ, Phyllis Cole, and Hal 0 C71 Puth off. In this; experimen t Hal will drive to a remote site 0.4 E - chosen by a random process. Phyllis Cole will be the rernote viewer, and Russ Targ is the monitor. We expect this experi- S 0.2 ment to start at :twenty minutes after eleven and run for fif- < E @ teen minutes. _J LU It is just about:twenty minutes after eleven and Hal should cc '0 0 1 1 1 1 1 be at his target location by now. +0A +0.2 0 -0@2 -0.4 Why don't you tell me what kind of pictures you see and V, (psi parameter) whatyou think he! might be doing or expeTiencirg. The first thing t hat came to mind was some sort of a large, Fig. 19. Reliability curve for sequential sampling procedure (p, 0.4, It was a s uare kind of a s ha e Like Hal was in front of it p@ = 0.6. cc = 0.01, 0 = 0.01). . . p q . ... not a building. or something, it was a square. I don't know the upper or the lower limit line. is reached, at which point a if it was a window, but something like that so that the bottorn decision is made to accept 0 or I as the bit being transmitted. line of it was not at the ground. About where his waist was, at Channel reliability (probability of correctly determining least. That's what it seemed to me. It seems outd oons some-. message being transmitted) as a function of operative psi how. Tree. parameter @ is plotted in Fig. 19. As observed, the sequential Does Hat seem to be looking at that square? sampling procedure can result in 90 percent or greater ]Tb2bility I don't know. The first impression was that ht- wasn't, but I with psi parameters on the order of a few percent. have a sense that.whatever it was was something one might Implementation of the sequential sampling procedure re- look at. I don't know if it would be a sign, but something that quires the transmission of a message coded in binary digits. one might look at.:. Therefore, the target space must consist of dichotomous ele- Canyou tell ifit:is on the ground or vertical? ments such as the white and green cards used in the experi- It seemed verticail. ments by RyzI. I don't have a s@nse that it was part of anything particular. In operation, a sequence corresponding to the target bit (0 It might be on a building or part of a building, but I don't @ or I) is sent and the cumulative entries are made (Fig. 18) until know. There was a tree outside, ibut I also got the impression a decision is reached to accept either a I or a 0 as the bit being of cement. I don't have the impression of very many people transmitted. At a prearrang6@ time, the next sequence is or traffic either. .1 have the sense that he is sort of walking begun and continues as above until the entire message has been back and forth. I don't have any more explicit picture than received. A useful alternative, which relieves the percipient of that. the burden of being aware of his self-contradiction from trial Can you move into where he is standing and try to see what to trial, consists of cycling through the entire message r, _pcti he is looking at? tively and entering each response an its associated graph until I picked up he was touching something-something rough. a decision has been reached on all message bits. The authors Maybe wan-n and rough. Something possibly like cement. have used this technique successfully in a pilot study, but a It is twenty-four minutes after eleven. discussion of this would take us beyond the intended scope of Can you change; your point of view and move above the this paper. scene so you can get a bigger picture of what's there? From the results obtained in such experiments, the channel I still see some trees and some sort of pavement or some- bit rate can be ascertained for the system configuration under thing like that. Might be a courtyard. The thing that came to consi -deration. Furthentiore, bit rates for other degrees of mind was it might be one of the plazas at Stanford campus or reliability (i.e., for other po, p I , a, and 0) can be estimated by something like that, cement. construction of other decision curves over the same data base Some kinds of landscaping. and thus provide a measure of the bit rate per degree of I said Stanford campus when I started to see some things in reliability. White Plaza, but I think that is misleading. In summary, the procedures described here can provide for a I have the sense that he's not moving around too much. specification of the characteristics of a remote-sensing channel That it's in a small 'area. under well-defined conditions. These procedures also provide I guess I'll go ahead and say it, but I'm afraid I'm just putting for a determination of. the feasibility of such a channel for on my impressions; from Stanford campus. I had the impres- particular applications. sion of a fountain.. There are two in the plaza, and it seemed APPENDix B that Hal was possibly near the, what they call Mern Claw. 1@hat is that? REMOTE-VIEWING TRANSCRIPT It's a fountain that looks rather like a claw. It's a black Following is the unedited @ranscript of the first experiment @culpture. And it has benches around it made of cement. with an SRI -volunteer (S6),,a mathematician in the computer' Are there any buildings at t& place you are looking at? Are science laboratory, with- no previous experience in remote there any buildings? You described a kind of a courtyard. Approved For Release 2003104/18 CIA-R DP96-00789ROO31.00030001-4 PUTHOFF AND TARG: PERCEPTUAL CHANNEL FOR INFORMATION TRANS7ER Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Usually at some places there should be a building, large or small that the courtyard is about. Look at the end or the sides of the cour@-yard. Is there anything to be seen? I have a sense that there are buildings. It's not solid build- ings. I mean there are some around the periphery and I have a sense that none of them are very tall. Maybe mostly one story, maybe an CCcasional two Story one. Do you @ave an -v better idea of what your square was that you saw at :he ou;set? No. I coutd hazard different kinds of guesses. Does it seenz par, of this scene? It I think it could be. It could almost be a bulletin board or something with not,ices on it maybe. Or somefaing that people are expected to look at. Maybe a window wi:h thinEs in it that people were expected to look at. What kinZ of trees do you see in this place? I don't know what kind they are. The impression was that. they were shade t-ees and not terribly big. Maybe 12 feet-of trunk and -hen a zertain amount of branches above that. So that the branches have maybe a 12 foot diameter, or some- thing. Not -eal bigg trees. New trees rather Man old trees? Yeah, maybe 5 cr 10 years old, but not real old ones. Is there arything interesting about the pavement? No. It s,--,ems to be not terribly new or terribly old. Not very interesting. There seems to be some bits of landscaping around. Lfttle pprches of grass around the edges and periph- eries. Mayle some flowers. But, not lush. You sawwme benches. Doyou want to tell me about them? Well, tha-.'s my unsure feeling about this fountain. Them was some kind of benches of cement. Curved benches, it felt like. They wert of rough cement. What do -iou think Hal is doing while he is there? 1 have a sense t@iat he is looking at things trying to project them. Looking aE different things and sort of walking back and forth not covering a whole lot of territory. Sometimes standing still while he looks around. I just had the irnpression of him talking, and I almost sense that it was being recorded or something. I don't know if he has a tape :ecorder, but if it's not that, then he is saying some- thing because it needed to be remembered. It's 11:33. He's just probabtv getting ready to come back. - ACKNOWLEDGMENT The authors wish to thank the principal subjects, Mrs. Hella Hammid, Pat Price, and Ingo Swann, who showed patience and forbearance in addition to their enthusiasm and outstanding perceptual abilities. We note with sadness the death of one of our subjects, Mr. Price. We express our sincere thanks also to Earle Jones@ Bonnar Cox, and Dr. Arthur Hastings, of SRI, and Mrs. Judith Skutch and Richard Bach, without whose en- courag6ment 'and support -this work could not have taken place. REFERENCES J. R. Smythles, Ed., Science and ESP. London, England- Rout- ]edge, 1967. [2) C. -Evins, "Parapsychology -What the questionnaire revealed," . New Scientist, Jan. 25, 1973, p. 209. [3) A. Gauld, 77te Founder; of Psychical Research. Now York: Schocken Books, 1968. See also W. Crookes, Researches In the Phenomena of Spirituallsm. - London, England: J. Burns, 1874. 141 R. Tart and H. Puthoft; "Information transmission under condl- tions'of sensory shielding," Nature, Vol. 2.52, pp. 602-607, Oct. 353 18, 1?74. (5) 1). D_ Home, Lights and Shadows of Spiritualism. New York: G.W.Carleton, 1877. . I 16] J. Ccover, Experiments in Psychical Research. Palo Alto, CA: Stanf6rd Univ. Pr@ss, 1917. 171 G. Es-abrooks, Bull. Boston Society for Psychical Research, 1927. See Eam [ 12, pp. 18-19). 181 L. T. Troland, Techniques for the Experimental Study of Telep- athy md OtherAlleged Clairvoyant Processes. Albany, NY, 1928. (9) J. B. 'Rhine, New Frontiers of the Mind. New York: Farrar and Rintiart, 1937. 1101 J. Pritt and J. B. Rhine et al., Extra-Sensory Perception after Sixt) Years. New York: Henry Holt, 1940. (11) C. Smtt, "G. Spencer Brown and probability: A critique," J. So.,@ Psychical Res., vol. 39, pp. 217-234, 1958. [12) G. R. Price, "Science and the supernatural," Science, vol. 122, pp.359-367,1955. [131 -, @Apology to Rhine and Soal," Science, vol. 175, p. 359, 1972. [14] J. B. Rhine, "A new case of experimenter unreliability," J. Pararnychol_ vol. 38, pp. 215-225, June 1974. [15] S. G Soal and F. Bateman, Modern Experiments in Telepathy. Lon6)n, England: Faber and Faber, 1953. [16) C. Sott and P. Haskell, "'Normal' explanation of the Soal- Golftey experiments in extra-sensory perception," Nature, vol. 245, ;p. 52-54, Sept. 7, 1973. [17) C. E M. Hansel, E&P-A Scientific Evaluation. New York: Scribier, 1966. 1181 J. B. Rhine and J. G. Pratt, "A review of the Pearce-Pratt dis- tance series of ESP tests," J. Parapsychol., vol. 18, pp. 165-177, 1954@ [19) J. G. ?ratt and J. L. Woodruff, "Size of stimulus symbols in extra- sensory perception," J. Parapsychol., vol. 3, pp. 121-158, 1939. (20] S_ G. Soal and H. T, Bowden, The Mind Readers: RecentExperi- rnenr: in Telepathy. New Haven, CT: Yale Univ. Press, 1954. [2 11 C. H.-norton, "Error some placel" J. Commun., vol. 25, no. I (Anntnberg School of Commun.), Winter 19? 5. (22) M. Ryzl, "Training the psi faculty by hypnosis," J. Amer. Soc. PsychS:alRes., vol. 41, pp. 234-251, 1962. [231 CIBA Foundation Symposium on Extra Sensory Perception. Bostca, MA: Little, Brown, 1956. [241 M. R3zI and J. Pratt, "A repeated-calling ESP test with sealed c-ards.."-T. Parapsychol., vol. 27, pp. 161-174, 1963.. 1251 -, A further confirmation of stabilized ESP performance in a salemtd subject," J. Parapsychol., vol. 27, pp. 73-83, 1963. 126) 1. Pritt, "Preliminary experiments with a 'borrowed' ESP sub- ject,"J. Amer. Soc. Psychical Res. '-Vol. 42, pp. 333-345, 1964. [27] J. Pritt and J. Blom, "A confirrriatory experiment with 'bor- rowef' outstanding ESP subject," J. Ame)@. Soc. Psychical Res., Vol. 42, pp. 381-388, 1964. 128] W. G. Roll and J. G. Pratt, "An ESP test with aluminum targets," J. Arrer. Soc. Psychical Res., vol. 62, pp. 381-387, 1968. (291 J. PrEtt, "A decade of research with a selected ESP subject: An over-6-w and reappraisal of the work with Pavel Stepanek," Proc. Amer. Soc. Psychical Res., vol. 30, 1973. 1301 C. Stannon and W. Weaver, The Mathematical Theory of Com- munkation. Urbana, IL: Uniy. Illinois Press, 1949. [31] M. Ryzl, "A model for parapsychological communication," J. Parxpsychol., vol. 30, pp. 18-3 1, Mar. 1966. [32] C. Ttrt, "Card guming tests: Learning paradigm or extinction parad4m," J. Amer. Soc. Psychical Res., vol. 60, p. 46, 1966. 33) M. Ulman and S. Krippner, with A. Vaughan, Dream Telepathy. New York: Macmillan, 1973. [341 C. Hcaortoh, "State of awareness factors in psi activation," J. Amer. Soc. Psychical Res., vol. 68, pp. 246-2 57, 1974. 35 J P@roc. 2nd Int. Congr. Psychotronlc Research (Monte Carlo). Cotat@ CA: Int. Assoc. Psychotronic Res., 1975. 361 L. L. Vasiliev, Experiments In Mental Suggesilon. Hampshire, EnilBad: ISMI Publ., 1963. 371 1. M. Kogan, "Is telepathy possible?" Radio Eng., vol. 2 1, p. ?5, Jan. 3966. [381 -, 'Telepathy, hypotheses and observations," Radio Eng., Yol. 22, p_ 141, Jan. 1967. 139) -, 'Information theory analysis of telepathic communication expenments," Radio Eng., vol. 2 3, p. 122, Mar. 19 68. 140 1-, @The information theory aspect of telepathy," RAND Publ., Santa Monica, CA, p. 414S, July 1969. [411 A. S. Presman, Electromagnetic Fields and Life. New York: Plenum, 1970. [421 Y. A. Kholodoy, Ed., Influence of Magnetic Fields on Biological Objerr, JPRS 63038, NTIS. Springfield, VA, Sept. 24, 1974. [43] Y. A- Kholodoy, "Investigation of the direct effect of magnetic fields on the central nervous system," in Proc. Ist Conf. Psycho- tronkRes., JPRS L/5022-1 and 2, Sept. 6, 1974. 1441 D. Mtnnie, "Consumer electronics," IEEE Specrmm, Vol. 12, pp. 34--3S, Mar. 1975. (451 W. P. Zinchenko, A. N. Leontiev, B. M. Lomov, and A. R. Luria, Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 354 . PROCEEDINGS OF THE IEEE, MARCH 1976 "P&,Am~dwd*6rnReted~W'20BOO104it8,~hGWk6-RDPD6~Oe7.8'9R~M~0~80001:-*echinisms and models," M.S.S. vol 9,'PP.-128-136, 1973. Information Corp., New York, 1974. 1461 R. Cavanna, Ed., Proc. Int. Conf, Methodology in PSI Rese4rch. [ 651 B. Julesz, Foundations of 'Cy@lopean Perception. Chicago, IL: New York: Parapsychology Foundation, 1970. Univ. Chicago Press, 197 1. (471 E. D. Dean, "Plethysmograph recordings as ESP responses," Int. 1661 H. Puthoff and R. Targ, in Psychic Exploration _A Challengefor ,r. Neuropsychizrry, Yol. 2, Sept. 1966. Science, J. White, Ed. New York: Putnam, 1974, pp. 522-542. [48] C. Tart, "Physiological correiates of psi cognition," Int. J. Para- [ 671 G. Feinberg, "Precognition-A memory of things future?" in p3ychol., no. 4, 19 6 3. Proc. donf. Quantum Physics and Parapsychology (Geneva, 491 D. H. Lloyd, "Objective events in the brain correlating with Switzerland). New York: Parapsychology Foundation, 1975. psychic phenomena," New Horf:ons, vol. 1, no. 2, Summer 1973. 16S] E. P. Wigner, "The problem of measurement," Anier. J. PhYs., [501 J. Silverman and M. S. Buchsbaum, "Perceptual correlates of con- vol. 3 1, no. 1, p. 6, 1963. sciousness; A conceptual model and its technical implications for [ 691 J. J, Freedmin and J. F. Clauser, "Experimental test of local psi research," in Psi Favorable States of Consciousness, R. Ca. hidden vari abl@ theories," Phys. Rev. Lett., vol. 28, no. 14, p. Yanna, Ed. New York: Parapsychology Foundation, pp. 143- 938, Apr. 3, 1972. 169, 1970. 1701 J. F. Clauser ind M. A. Horne, "Experimental consequences of [511 J. Kamlya, "Comment to Silverman and Buchsbaum," !bid., pp. objective local: theories," Phys. Rev. D, vol. 10, no. 2, p. 526, 158-159. July IS, 1974.@ 52 ] D. Hill and G. Parr, Electroencephalogrophy, A Symposium on 1?11 .B D. Bohm and : . Hiley, "On the intuitive understanding of non. Its VarlousAspects. New York: Macmillan, 1963. locality as iniplied by quantum theory" (Birkbeck College, 5 3) T. D. Duane and T. Behrendt, "Extrasensory electroencephalo- London, England), Feb. 1974, Preprint. graphic induction between identical tAins," Science, vol. 150, 17:1 1 J. S. Bell, "On the problem of hidden variables in quantum p.367,1965. theory," Rev. @fod. Phys., vol. 38, no. 3, p. 447, July 1966. 541 K. Osis, ASPR Newsletter, no. 14, 1972. [?-'1 H. Stapp, "Theory of reality," Lawrence-Berkeley Lab. Rep. (55] R. L. Morris, "An exact method for evaluating preferentially LBL-3837, UrAv. California, Berkeley, Apr. 1975. matched free-response material," J. Arner. Soc. Psychical kes., [?41 A. Einstein, B. Podolsky, and N. Rosen, "Can quantum-mechani. vol. 66, p. 40 1, Oct. 19?2. cal description of physical reality be considered complete?" Phys. 5 6] G. R. Schmeidier, "PK effects upon continuously recorded tem- Rev., vol. 47, p. 777, May 15, 1935. peratures," J. Amer. Soc. Psychical Res., vol. 67, no. 4, Oct. [ 7 51 R. H. Dicke and J. P. Wittke. Introduction to Quantum Mechari- 1973. ics. Reading, MA: Addison-Wesley, 1960, ch. 7. 57 ] W. Scherer, "Spontaneity as a factor in ESP," J. Amer. Soc. (761 E. H. Walker, "Foundations of paraphysical and parapsychologi- Psychical Res., Yol. 12, pp. 12 6-147, 1948. cal phenomenA," in Proc. Conf. Quantum Physics ond Para- 5 8] R. Targ, P. Cole, and H. Puthoff, "Techniques to enhance man/ psychology (6eneva, Switzerland). New York: Parapsychology machine communication," SRI, Menlo Park, CA, Final Rep., Foundation, 1075. NASA Contract NAS7-100, June 1974. [771 0. Costa de Reauregard, "Time symmetry and interpretation of 59 ] R. Ornstein, The Nature of Human Consciousness. San Fran- quantum mechanics," Lecture delivered at Boston Colloquim for cisco, CA: Freeman, 197 3, ch. 7 and 8. Philosophy of: Science (Feb. 1974), Foundations ofPhysics (in 601 R. W. Sperry, "Cerebral organization and behavior," Science, press). vol. 133, pp. 1-@49-1757, 1961. [?E] J. A. Stratton, Electromagnetic Theory. New York: 'McGraw- 61] 0. Bilaniuk and E. C. G. Sudarshan, "Particles beyond the light Hill, 1941. . 15arrier," Phys. Today, vol. 22, May 5, 1969. 1791 A. Sinkov, Elementary Cryptanalysis-A Mathematical Approach. [621 W. Pauli and C. G. Jung, Eds., Tlie Interpretation of Nature and New York: Rahdom House, 1968. the Psyche (Bollingen Set. LI). Princeton, NJ: Princeton Univ. [80] P. Hoel, Zntro@fuction to Mathematical Statistics, 2nd ed. New Press, 1955. York: Wiley, 1954, p. 27. [631 N1. A. Persinger, "ELF waves and ESP," New Horizons Trans. 811 R. Taetzsch, 'Mesign of a psi communications system," Int. J. Toronto Socim for Psychical Research, vol. 1, no. 5, Jan. 1975. Parapsychol., Vol. 4, no. 1, p. 35, Winter 1962. Approved For' Rele@ise 2003/04/18 CIA-RDP96-00789ROO3100030001-4 136 Approved For Release 2003/04/18 clk-W60W-~b~bgPb65)o(Y6~06bil4.2, IIIRt'ARY 1982 The Persistent Paradox of Psychic Phenomena: An Engineering Perspective ROBERT G. JAHN Invited Paper Abstract-Although a variety of so-called psychic phenomena have at- tracted man's attention throughout recorded history, organized scholarly effart to comprehend such effects isjust one century old, and systematic academic research roughly half that age. Over recent years, a sizeable spectrum of evidence has been brought forth from reputable laboratories in several disciplines to suggest that at times human consciousness can acquire information inaccessible by any known physical mechanism (ESP), and can influence the behavior of physical systems or processes (PK), but even the most rigorous and sophisticated of these studies display a characteristic dilemma: The experimental results are rarely rep- licable in the strict scientific sense, but the anomalous yields are well beyond chance expectations and a number of common features thread through the broad range of reported effects. Various attempts at theoret- ical modeling have so far shown little functional value in explicating @rimental results, but have served to stimulate fundamental re-exam- 'expe ination of the role of consciousness in the determination of physical reality. Further careful study of this formidable field seems justified, but only within the context of very well conceived and technically im- peccable experiments of large data-base capability, with disciplined at- tention to the pertinent aesthetic factors, and with more constructive involvement of the critical community. Prologue -The world of psychic phenomena might be likened to a vast, fog-shrouded swamp, wherein are reported to dwell a bewilder- ing array of bizarre phenomenological creatures, all foreign to our normal perceptual and analytical catalogs. Some scholars Who have explored this clouded domain have returned to an- nouhce.-&tegorically that all such life is illusory-mere sunken stumps and swirling subsurface shadows, inviting misperception by the gullible and misrepresentation by the purveyors. But others of comparable conviction have described in minute de- tail their observations of a variety of extraordinary beings of awesome dimensions and capability. Some of these are claimed to appear unexpectedly, erupting from the roily depths to flash m.omentarily i,n the sunlight of human experience, only to disappear again before any systematic calibration of their ch arac teris tics can be taken, Others are reportedly enticed to more replicable and controlled behavior, but only by persons of special talent or extensive training. Much invalid, even fraudulent evidence of such activity has been touted by ex- ploiters of these mysteries, thereby casting deep suspicion on all other testimony. When fully sifted, only a very few legiti- Manuscript received July 15, 1981; revised 6ctober 26, 1981. This work was supported In part by the McDonnell Foundation, Inc., and by the John E. Fetzer Foundation, Inc., The Explorers Club, the Insti- tute of Noetic Sciences, and the Little River Foundation. The author is Dean of the School of.Engineering/Applied Science, Princeton University, Princeton, NJ 08544. mate specimew seem to have been :aptured, by ted:ousl.i _Ye- liberate trolling of the brac.'.ish do'nain, or by mo,e incij;ve . invasion of its turbid interior, and even these have 'Drover, so incomprehensible and so deh,ate to exposure, and the impo:ed criteria for the;-r credibility ?ave been so severe, that ,hey h,.:ve not been fully persuasive. Yet the goal remains alh@ring, .,-nd the search coniinues. INTRODUCTION ITH THIS unlikely bit of a2legorical musing. I veniure to begin the most extraord'I'lary writing tasl ,, I haveyet attempted: to respond to zhe request of the Edi-,ors of this journal for a critical review of the status and prognosis of scientific research into sc-called psychic phenomena. I do so with some trepidation, First because the topic is far from my principal line of schola--ship and my involvement wit]a it has been brief and tightly ci:cumscribed, and second, because of the intensity of reactions any commentary on this subiect tends to call forth from many quarters. For these reasons, it may be well at the outset to specify my perspective on the field and -he purpose that I hope this art"cle lill serve. My formal training is th.at of an engineer a,-,.d applied physicist, and the bulk of rn@- research has concerned a sequence of topics in the broad domain of the aerospace sciences: Fluid mechanics, ionized gases, plasmadynamics, and electric propul- sion. In my present position as Dean of the School of Engineer- ing and Applied Science of Princeton University, I have occa- sion to be involved with an even broader selection of topics selected for undergraduate independent projects, and it was in that context some four years ago that I was requested by one of our very best students to supervise a study of psychic phenomena. More specifically, this young lady proposed to bring her talents and background in electrical engineering and computer science to bear on some experiments in controlled, low-leyel psychokinesis. Although I had no. previous experience, professional or personal, with this subject, for a variety of pedagogical reasons I agreed, and together we mapped a tenta- tive scholarly path, involving a literature search, visits to appro- Priate laboratories and professional meetings, and the design, construction, and operation of simple experiments. My initial oversight role in this project led to a degree of personal involve- ment with it, and that to a growing intellectual bemusement, to the extent that by the time this student graduated, I was persuaded that this was a legitimate field for a high technologist t6 study and that I would enjoy continuing to do so. 0018-9219/82/0200-0136$00.75 (D 1982 IEEE Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 JAHN: PSYCHIC PHENOMENA: ENGINEERING PERSPECTIVE 137 F,%A?IeaPe.2093(0 -RDP96-00789ROO3100030001-4 , 'fl/18 : CIA I have sinAPPFRY19911 pro essiona s a , secured the requisite funding from a few private sources, and undertaken a modest experimental program in selected aspects of the field that could ultimately have some engineering implications. I should emphasize that my fractional involvement with this program remains quite minor in comparison to my other re- sponsibilities, and that the work is still very preliminary and tentative, but it provides the base of cognizance for my broader observations on the field as a whole, The intention of this article is to provide some balanced per- spective on the modern status of this conceptually and logis- tically difficult subject. Certainly no field of scholarly endeavor has proven more frustrating, nor has been more abused and misunderstood, than the study of psychic phenomena. Dealing as it does as much with impressionistic and aesthetic evidence as with analytical substance, and carrying by its nature strongly subjective and numenistic overtones, it has been incessantly prostituted by charlatans, lunatics, and sensationalists, catego- rically rejected by most of the scientific establishment, and widely misunderstood by the public at large, Interspersed with this, and greatly encumbered by it, a pattern of legitimate effort to comprehend and utilize the purported phenomena has evolved to a point where some dispassionate assessment of its accomplishments can be attempted. The questions addressed by this review are whether, once the overburdens of illegitimate activity and irresponsible criticism ar ,e removed, there remains sufficient residue of valid evidence tojustify continued research and, if so. how this research might most effectively be styled, facilitatedI, and evaluated. Before addressing these issues directly, it may be. helpful to review briefly the historical evolution of the field, its contem- porary nomenclature and conceptual organization, and the dimensions of current activity. This can then be followed by a general over-view and critique of the modern research, and that in turn by more detailed description of a few s@pccific efforts, drawn primarily from our own work. Toward the close, we shall attempt to survey several theoretical approaches to modeling of psychic processes and comment briefly on po- tential implications and applications of the phenomena. In all of this, no tone of advocacy is intended, other than for objec- tive asiessment of the evidence in hand. HISTORY In a sense, the study of psychic phenomena is one of the oldest of human endeavors [1]-[71. As far back as can be traced, mortal man has pondered the supernatural in one form or another. Cave drawings at Lascaux and Altamira, circa 20 000 B.C., reflect this preoccupation, and the religious rites of early societies of both the eastern and western worlds were heavily loaded with psychic formalisms. The classic civilizations of Egypt, Greece, and Rome dealt extensively in psychic process. The Delphic Oracle was politically important from the earliest Hellenic times to the age of Alexander the Great, and was consulted on problems as diverse as the proper measures to stop a plague, the constitutions of Greek city-states, and the best locations for new colonies. Even Aristotle, one of the most empirical of the classical philosophers, examined the causal links in prophetic dreams. Virtually every form of organized religion practiced by man has been thoroughly laced with various forms of psychic mech- anism. The Bible, like most other basic theological texts, treats psychic Process as a central i;'Igredient, in a tone so matter-of- fact that one is inclined to believe.that people of those times accepted such events rather Toutiaely, Indeed, the Bible is an excellent catalog of psychic pheromena; virtually every cate- gory of effect identified today is Mustrated there in one form or another. Christian writers and pHosophers, from Augustine to the Reformation, recount many PU7:)orted instances of psychic phenomena, usuaftl. attributed t@ visitations of divine grace or demonic posession. Secular r.-.,:!dieN,al writing also abounds with supernatural and mystical -efeience, and even in the Renaissance period it is still difficult to separate psychic allu- sion from religious @ogma, aithou:h both were then translated into more organized forms in art ind literature. Early in the 16th century the cqlebrated Swiss physician and philosopher Paracelsus wrote ex'tensively on --sychic capabilities and po- tentialities. In his w6rds: The mind of man is the microccimic counteipart of the uni- versal mind ..... .... One man me! commimicate his thoughts to another with wbdmheis in sympk,.hy, zt any distance however great it may be, or he may act upo: the spirit of another pemn in such a manner as to influence his ictions .... [8). Perhaps the first major scientific commentaries on the topic were offered near the.tum of the I -i:h century by Sir Francis Ba- con widely regarded as the origina-.or of the scientific method. I In The Advancement. of Learning '_ne suggested that "supersti- tions and the like" should not te excluded from scientific study, and in his posthumous "7ook, Sylva Sy7varum, he proposed deliberate investigation of telepathic dreams, psychic heating, and the influence of "imqination" on the casting of dice [9]. Some years later, a gj@up of British intellectuals including Henry Moie and Josep1z Glanv],I met regularly to discuss paranormal topics, and in 1681 Glanvill published the substance of the@e studies in I book entitled Saducismus Triumphatus (101. Meanwhile, some four centuries cf public and church hysteria over sorcery and wit .phoraft, as minifested in a sequence of trials, tortures, and executions, hai begun to subside, and by the mid-18th century, the Roman Church authorized Prospero Lambertini, who later became Po-)e Benedict XIV, to carry out a scholarly investigation of rep .orts of psychic events. His conclusions, recorded in De Canon--atione [ 111 , were surpris- ingly uneeclesiastical:. namely, thit 1) psychic experiences were not necessarily divine miracles. but could occur to "fools, idiots, melancholy persons, and biute beasts"; 2) apparitions had little to do with sanctity or demonic entities; 3) prophesy occurs more often in sleep than in waking; 4) it is difficult for a prophet to distinguish his own t@ioughts from extrasensory messages; and 5) predictions frequtntly take symbolic forms. In all of these, he presaged to sorne degree modem thoughts on these topics. At roughly the same time, Anton -Vesmer's discovery of hyp'- nosis opened an alternative route to demonstration and study of unconscious psychi c! process that has continued to this day. Early reports of hypnotized subjects performing telepathic or clairvoyant tasks were common ( 121, (131, and although much of this evidence might now be discounted on the basis of in- adequate experimentalcontrol, interest in hypnosis specifically, and in various altered states of consciousness generally, as facilitators of psychic. experience persists into some of the modem experimentation. Also in this mid-18tii century period, a spiritualist move- ment focused on extraiensory contict With the dead, possibly influenced by the work of Emanud Swedenborg [14), [151, germinated in this country as well as in England, and by the Approved .For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 PROC EEDINGS OF THE IEEE, VOL. 70, NO. 2, FEBRUARY'1982 19th century had reaApordvedrFimiReledse2D9M4dl8AfCWRQB06497~8pFM3slQPQtIW1-4 religion. Symbolic of the popular preoccupation with the topic, Mary Todd Lincoln was reported to have held s6ances in the White House in the early 1860's [ 16 1. A classic two-volume work b, F.W.H. Myers, entitled Human Personalit, and the Survival of Bodily Death [ 171, brought the topic to its acme of sophistication, but eventually the fanaticism the movement attracted and its fraudulent' exploitation created a negative attitudein the scholarly community which prevails yet today. Despite these millennia of human concern with the paranor- mal, orderly and organized scholarly search for verification and understanding of psychic phenomena began only a century ago, with the establishment in London in 1882 of the Society for Psychical Research, in whose Proceedings appeared the first formal publication of controlled experiments in telepathy and clairvoyance [ 131, [181, [19). Three years later the counter- part organization in this country, the American Society for Psychical Research, was founded in Boston by several distin- guished scientists and philosophers. Because of financial difficulty, this shortly merged with the British group, but re- emerged in 1905 as a separate entity with its own professional journA and has continued as such to the present (201. Although the SPR attracted a barrage of criticism from the scientific and intellectual communities, it also attracted significant participation of eminent scholars from established fields. Numbered among its presidents are three Nobet Lau- reates, ten Fellows of the Rbyal Society, one Prime Minister, and a substantial list of physicists and philosophers, iricluding Henry Sidgwick, Frederic W. H. Myers, Lord Rayleigh, Sir J. J. Thomson, William McDougall, Edmund Gurney, Sir William Crookes, Sir William Barrett, Henri Bergson, Arthur, Earl of Balfour, Gardner Murphy, G.N.M. Tyrell, Charles Richet, Gilbert Murray, and one of the most articulate contributors to the evolution of critical thought on this topic in this period, the Harvard psychologist and philosopher, William James. One of the founders of the ASPR, James wrote extensively and eloquently on behalf of objective and disciplined study of psychic phenomena [.211-[251: Any one with a healthy sense for evidence, a sense not methodi- cally blunted by the sectarianism of 'science,' ought now, it seems to me, to feel that exalted sensibilities and memories, veridical phantasms, haunted houses, trances with supernormal faculty, and even experimental thought-transference, are natural kinds of (phenomena) which ought, just like other natural events, to be followed up with scientific curiosity (25 1. Entering the 20th century, anew perspective on psychic phe- nomena was provided by the emergence of psychology as a scholarly discipline, and especially by the early efforts in clinical psychology and psychoanalytic therapy. The patriarch of this evolution, Sigmund Freud, was a member of the SPR and con- tributed, albeit somewhat reluctantly, to its publications (261, [27). His recognition and exploration of the unconscious mind and of the function of dreams prompted Myers to suggest .a possible explication of various psychic effects which is still of theoretical value 117). Freud's interest in parapsychology increased toward the end of his life, and he is reported to have conceded, informally that were he to begin his career anew, he would focus on this topic. Fmud's former proteg6, Carl Jung, who had written his Ph.D. thesis on the psychology of occult phenomena, pursued ex- ploration of the unconscious to deeper dimensions of para- normal experience, publishing Widely on such subjects as telep- athy, mediumship, synchronicity, the collective unconscious, and theoretical models of psychic process (281-[30]. In ... the relationship between doctor and patient, especially when a transference on the part of the patient occurs, or a more or less unconscious identification of doctor and patient, can lead to parapsychological phenoTir.,ena- I have frequently run into this [301. Jung's collaboration with the eminent physicist Wolfgang Pauli on the topic of synchronicity clearly influenced the subsequent evolution of both careers and of fundamental concepts in both disciplines [3 11. Although much of the established psychologi- cal community has since rejected parapsychology as a valid discipline, some interest has be@en retained by a few clinical practitioners, presumably because of the demonstrated con- comitance and similarities of apparent psychic experiences with certain psychological processes (321, [331. It was also early in this centun, that the first organized, academic studies of psychic phenomena were mounted. One of the more visible of these devolved from gifts and a bequest from Thomas W. Stanford, brother of the founder of Stanford University, to endow psychic research at that institution, and to this day the university provides support of a "psychic research fellow" and retains a collection of so-called "apports" indicative of the donor's long personal involvement with the field. Modest research programs were also undertaken at Harvard and a few European universities in the first decades of this century, as evidenced by occasional publications in various established journals. The benchmark academic effort, however, germinated at Duke University in the late 1920's, when William McDougall, who had been James' successor at Harvard, arrived to chair the department of psychology and appointed J. B. Rhine and Louisa Rhine ' "to study the claims to scientific value of the field known as psychical iesearch." Their early tentative efforts in the study of postmortem survival gradually evolved into a laboratory for controlled research in "extrasensory per- ception," as they first termed the process. In.,this laboratory were established many of the basic concepts and protocols of modern psychic research, as well as the first extensive and sys- temati c data bases of several types of psychic experimentation. The professional and personal history of the Rhines and their laboratory is a fascinating saga in its own right, but would take us too far afield here [34]-[377]. A few excerpts from a 1967 address of J. B. Rhine to the American Psychological Associa- tion, in which he attempted to summarize his first two decades of intensive study, give hint of the inherent attractions and frustrations of this field, and ol the man's optimistic vision: The phenomena that were being studied began to show lawful interrelations and even a degree of unity. One by one the major claims, based originally only upon spontaneous human experi- ences, were subjected to laboratory test and experimentally veri- fied... . Certain general characteristics of the psi process became clear during ti-iis period. The most revealing of these is the sub- ject's lack of conscious control over any type of psi ability, a characteristic which accounts for its elusive nature. It was new methodological ground, even for psychology.... Also, we were surprised to find that psi ability is widespread, probably even a specific human capacity rather than a capability possessed by a few rare individuals as had been the popular belief. Evidence that psi is not linked with illness or abnormality was another welcome advance.... By 1951 ... a healthy young science was emerging [381. In 1937, the Rhines began publication of the Journal of Parapsychology, which remains a leading journal in the field today. A professional organization calling itself the Parapsy- chological Association was formed in 1957, and in 1969 was Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 JAHN: PSYCHA'-P"F~jf'699aP6FIkdfLg'"i64"bpt/4"0~tIA-RDP96-00789ROO3.100030001-4 139 1 levitation of objects, up to sorne Association for the macroscopic distortion accepted as an affiliate by the American Advancement of Science. At the present time, there are eight English language publica- tions covering this field [391, supplemented by numerou. less formal magazines and countless books of widely varying quality and relevance. Research activity is reported from some twenty U.S. universities and colleges and at least as many institutions in Western Europe [401, but in most cases it is of very small scale. There are very few academic programs of study, although some fifty M.A. and Ph.D. theses have been accepted on psychic topics at reputable universities over the past forty years [4 1 Some ten research institutes and private corporations in the United States have also authorized publications and reports in the field [42] . The extent of Eastern Bloc and Oriental efforts [43 54 J and of classified research in this country are matters of considerable speculation on which I cannot cornment with authority. Further review of contemporary programs will be attempted in subsequent sections, following an outline of modern noraen- clature and conceptual organization of the topic. In closing this historical overview, we might simply observe that in many re- spects the growth pattern of this field resembles that of the natural sciences in their earliest days, or perhaps even more the incubation of classical psychology, in terms of the ab- sence of replicable basic experiments and useful theoretical models, the low level of financial support and internal profes- sional coordination, and the low credibility in the academic establishment and public sectors. Also like those fields, the survival and early growth of psychic research can largely be attributed to the efforts of a few scholars of sufficient convic- tion, stature, and courage to withstand the rejection of the orthodox communities. NoMENCLATUR!- AND CONCEPTUAL ORGANIZATION or very drastic "poltergo Iist" effects. Psychic healing and man- plant interactions wo@ .ild be two examples of PK in biolog@cal systems. Note that in its major subdivision into ESP and PK, the field confornis to two of i lie main categories of present-day scie;ice and high technology, i,e., that encompassing the extraction, c3n- version, trans mission.. storage, and utilization of informat.iDn, and that comprising tl@c same sequence of processing of ene7gy. Other domains of ps\@cliic research can be identified which do not fit into these major categories of ESP and PK and with wl@_ich we sliall not be further concerned in this article. Exam:@Ies would include resear@h into survival of death, and the fainUy of "out-of-body exp14:riences (OBE)," including astral pro@ec- tion, autoscopy, and bilocation. The following table atlempts to summarize the subdivisions in a concise form. Categories of Psychic Phenomena 1. Extrasensory Perception (ESP) A. Telepathy B. ;Clairvoyance C. *Precognition/Retrocognition D. Animal ESP 11. Psychokinesis (PK) A. :Physical Systems B. Biological Systems III. Survival A. 'Reincarnation B. Apparitions C.:Mediumsliip IV. Out-d-Body Experiences (OBE). Clearly this particular arrangement is neither unique nor orthogonal. Elements of one category frequently appe2r- in the context of another, e.g., precognitive clairvoyance; :'ele- pathic healing, etc., and occasionally an assignment is amlbigu- ous, e.g., a particulak effect may be regarded as precipitated by PK, or simply to. be forecast by precognition. Not,;Ith- standing, the table may aid in keeping the subsequent illustra- tions in some order. Before turning to an assessment of contemporary research, it may be useful to specify some notation and delineation of the field, to an extent consistent with the present limited compyehension of the phenomena. First, let us agree to a global definition of "psychic phenomena" (frequently denoted by "psi" or "@") to include all processes of information and/or energy exchange which involve animate consciousness in a manner not currently explicable in terms of known science. Similarly, let "psychic research" imply any scholarly study of such phenomena employing scientific methodology, as opposed to any dogmatic, ritualistic, or theological approaches. Within these definitions, the field may then be roughly divided into two major categories: extrasensory perception (ESP) and psychokinesis (PK). ESP refers generally to the acquisition of information from sources blocked from ordinary perception. Under this category are included such subdivisions as telepathy, which refers to detection of another person's thoughts; clairvoyance, which refers to contemporary perception of remote physical objects or events; precognition and retrocognition, which refer to perception of future events and eventg'in the past not accessible by normal recollection; and animal ESP, which encompasses a variety of seemingly inexplicable capabilities, such as homing, psi-traffing, collective behavior, communication, etc. PK (occasionally termed tele@inesis, or psychoenergetics) re- fers to a palpable influence of consciousness on a physical or biologica.'I system. The interaction may be deliberate or spon- taneous, and the energy transfer involved may range from microscopic disturbance of atomic-level processes, through PATTERN OF CONTEMPORARY RESEARCH By its nature and heritage, modem psychic research remains rather diffuse and lightly structured, making any attempt to catalog the work by institution or laboratory, or by tracing developments of given lines of effort, rather ineffective and pre- mature. Instead, it may be more useful to comment on the pattern of attention to this field by academic disciplines, noting the variations in emphasis, methodology, representation, and interpretation brought to bear, using specific projects on!y as illustrations with no'implications of hierarchy or attempt at completeness. Even:in this format, no recitation of specific research results or conclusions will be attempted, since these can be misleading or incomprehensible when extracted from the detaiied context: of their experimental arrangements and protocols. In later sections, an effort will be made to follow a' few sample experiments through to their particular results and conclusions. By far the most sustained and broadmt attention 'to this field has been given. by a cadre of scholars with professional backgrounds in classical psychology, comprising a controversial subdiscipline termed:, "parapsychology." This group has tended to approach the field. with the traditional psychological proto- cols and vocabulary,.and to interpret results in the context of Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 their clinical, cognitiveApb+8Wdep6@k6tijo ,&@d conc r dlll:dgW4WP§fQQ7QQfaQB3dl%~031WOlv.*s. Icop with the natural consequence that t cy ave 'N 3 mainly on the ESP category of psychic tasks, although some PK work dots their recent literature. Perhaps the most exten- sive class of parapsychological research has attempted to correlate psychic performance with personality variables. The age, sex, creativity, openness, hostility, extroversion, motiva- tion, and intelligence of the participants as indices of ability to perform ESP tasks have been explored at length, and some sig- nificant correlations, most notably with positive a priori atti- tudes toward the tasks and with outgoing, creative personalities, have been reported from several laboratories. Other studies have searched for connections between psychic performance and dream recall, learning and response strategies, memory, and feedback ( 5 51- (6 11. A more aggressive style of parapsychological research has in- voked a variety of altered states of consciousness in attempts to enhance psychic process. These have included various natu- ral and traditional practices, such as sleep, meditation, and pro- gressive relaxation [621- (671 ; more mechanical sensory in- hibition strategies such as hypnosis, isolation and "ganzfeld" [661, [681-[701 ; and a few controversial efforts with drug- induced states [711 , Physiological correlates have also been sought, using conventional EEG, GSR, and plethysmographic equipment to monitor neurological, cardiovascular, and mus- cular response to psychic effort (35], (721-[751. The diffi- culty of obtaining success'ful replications of previously positive results and an observed common tendency of participant per- formance to deteriorate over time ("decline effect") have led to systematic study of the role of the experimenter in eliciting results, i.e., to consideration of the possible influence of the experimenter's personal attitudes, expectations, and styie of interaction with his subjects, as well as the overall environmental ambience of his laboratory, on the experimental yield [ 761 - [811.. Despite its present recalcitrance toward more systematic study of psychic phenomena, the richly diverse, rapidly matur- ing parent field of psychology continues to offer an expanding array of modern methodologies and models which could be brought to bear on increasingly sophisticated study of this sub- ject. Computer-assisted linguistic analyses-, psych oneurological studies of attention, perception, and concept formation; social learning theory and similar approaches to human interaction&; and the emerging formulations of transpersonal and humanistic aspects of human consciousness, all bear possible relevance to comprehension of various aspects of this ultra-difficult step- chi[d, but at the moment, the low level of financial support, and negative professional peer pressures have discouraged such enterprise. The involvement of physicists in psychic research, while con- siderably less extensive, has been no less dedicated and no less controversial, Since the days of Sir Francis Bacon, a number of noted physicists have made excursions into this field, usually to the berausement and ridicule of their colleagues of the day. One of the most notable of these was Sir William Crookes, discoverer of the element thallium and pioneer in the physics of low pressure discharges, whose broadside pro- fessional and personal battles with the scientific establishment over this issue make entertaining, and possibly enlightening, reading [82). Sir Isaac Newton was intensely involved in the study of alchemy, including some of its more metaphysical aspects (83], and as already mentioned, Lord Rayleigh and J.L Thomson were active members of the S.P.R. In more recent years the attention of physicists has influenced First,theirin- terests have focused more on the PK category of problems, i.e., the interaction of human'consciousness with physical systerw to balance the predominant EF interests of the parapsycholo- gists. Second, more sophisticated experimental equipment than has typically been available to the psychological cora@ munity has been brought to bear on the identification and correlation of very low-level physical effects. Third, the tradi- tional theoretical pl@ysics formalisms have been directed to the proposition of various models of psychic phenomena, from whence has arisen some hope of establishing the traditional dialogue between critical experimentation and theoretical hypothesis essential to any ultimate comprehension and appli- cation of such phenomena. Typical of the modem physicist's specific contributions ro the field have been the development and application of a variery of electronic random event generators (REG) for the purpcGe of identifying and correlating PK abilities in human subjects (841-[931, and similar application of magnetometers (941, torsional pendula [951, lasers [961, interferometers (971, and electronic strain gauges [981, [991 to a variety of other PK tasks. On the theoretical side, a number of applications of qum- turn mechanics, statistical thermodynamics, clectromagne:ic theory, and other formalisms to the representation of psychic process have been proposed [1001-[1131, and aftempts at some philosophical correlation of the phenomena with other previously or presently obscure physical processes have been suggested [1141. Again, despite the open identification of a few distinguished personalities with such efforts, a more broadly held categorical rejection of the field has inhibi-.ed much collaborative or systematic attention to it. Up to this time, the involvement of engineers @%-ith psyc.-iie research has been very recent, very sparse, and very much along the lines of the experimental physicists. Beyond our own program, which will be. outlined in some detail below, I am aware of only a very few engineering laboratories addressing any aspects of the field in any substantial and Lliberate way [ 115 J. These have so far tended to concentrate on applied physics types of experimentation and on aspects of infor.-na- tion processing, rather than on more empirical technolog@cal applications. Another community of scholars to influence the pattern of psychic research comprises the statisticians and other appied mathematicians and logicians who have been concerned ,;ith the proper evaluation and interpretation of the research data. In the absence of any experiments displaying rigid causal replicability, all of the inferences and hypotheses about psy- chic phenomena have necessarily been based on either anec- dotal or statistical evidence. The former defy any systematic representation; the latter are vulnerable to alternative irter- pretations and hence to impressionistic bias and argument. . Early in the emergence of mathematical statistics as an integral discipline, S. S. Wilks found himself involved in a controversy over the validity of the statistical procedures of early psychic researchers, and publisned some recommenda- tions for methods that could be applied to telepathy experi- ments [ 116 1. Since that time, much of the commentary from the critical community has addressed perceived flaws in the statistical methodology underlying the experimental evidence [ 117], [1181, and the advocate community has reacted by paying disproportionate attention to Ws aspect of their logic. Most of the encyclopaedic references in the field contain Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 JAHN: PSYCHIC PHENOMENA: ENGINEERING PERSPECTIVE Approved For Release 2003/04/18 : CIA-RDP96-00789ROO31.00030001-4 substantial components on the statistical methods [1191, and the leading journals regularly display intense dialogues on specific statistical issues [120]-[1241. At least one of these journals routinely refers all articles submitted for publication to a consultant statistician as a part of their review process. A few illustrations of the statistical questions that can arise in psychic experiments appear in the detailed examples presented below. A number of other disciplines have played roles in the evolu- tion of the study of psychic process and continue to contribute, albeit somewhat more peripherally. A succession of philoso- phers, from Aristotle through James and Bergson to C. D. Broad in the present era, have mused on these topics (2l]-[251, [1251-(1291. The intersections of the field with anthropology, theology, and history have been approached from many per- spectives ranging from aesthetic to analytic, and from dogmatic to scholarly 1130)-[134). Its relevance to the study and practice of medicine has been an enduring and intense debate, focusing in the present day on the propriety and efficacy of holistic health strategies and psychic healing. Isolated instances of interest on the part of chemists, biologists, geologists, and archaeologists can be found, and the application of psychic techniques in criminology and law enforcement, while less rigorous than most of the academic efforts, contributes further anecdotal evidence to the overall data base. In the arts and .humanities, the topic continues to provide stimuli for a variety of creative compositions, Still other areas could be cited, but these become progressively more satellite than central to the task of this paper. Rather than pursuing this disciplinary survey further, it seems prefer- able next to focus in greater detail on a few contemporary studies that can serve to illustrate more specifically the bizarre phenomena, the awkward and tedious protocols, and the un- conventional theoretical concepts that arise in this class of research. Although these will provide a better sense of the status of the field than any attempt to summarize results from the diffuse multidisciplinary pattem of effort sketched above, one general assessment may be useful at this point. Namely, throughout 0 of the work just skimmed, and that Parpened som@@what. below, I am aware of no reputable investigator who has claimed, let alone demonstrated, any psychic experiment approaching classical scientific replicability. What have been put forward are a varied assortment of observations, currently inexplicable in terms of established science, which display cer- tain common phenomenological and psychological features, and which could have substantial implications for basic physical theory and ultimate practical applications. The following ex- amples are presented in that spirit. PSYCH6KINESIS The first group of experiments selected as more detailed illus- trations of contemporary psychic research are drawn from the general subdomain of PK. As defined above, this broad cate- gory of purported psychic phenomena encompasses the possible influence of human consciousness on the behavior of physical or biological systems or processes, and corhprises several loosely related classes of effect characterized by different scales of energy, forms of manifestation, replicability, and statistical be- havlor. Confining attention to interactions with physical objects or systems, the most popularly 'publicized class features the de- .formation, levitation, or other macroscopic disturbance of ob- jects, as commonly propounded by professional performers, mediums, and various Eastern practitioners [1351-[1371. A]- 141 though a number of serious efforts have been made to submit such demonstrations to rigorous scientific testing, these have tended to yield only equivocal confirmations, fodder for the critical community, and some embarrassment and frustration to the investigators. Of a yet more bizarre nature are the family of very rare and spectacular "poltergei@t" effects, more technically term. ed "recurrent spontaneous psychokinesis" (RSPK), wherein are reported specific major events of levitation, vibration, tele-por- tation, and breakage of a wide range of objects, a variety of acoustical and electromagnetic phenomena, and various optical aberrations 1138), [09). For years these phenomena were naively attributed to manifes tations of the spirit world, or return of the dead to "haunted" houses, and inspired countless ho.-ror movies and pulp-magazine articles. Recently, some order has been brought to this weird business by systematic surveys of documented poltergeist cases undertaken by A.R.G. Owen, W. G. Roll, J. G. Pratt, and others [138]-[142]. In one of these surveys, 116 cases of reported poltergeist activity, rang- ing back to the year 1612, were re-examined. Of these, 92 were found to be asso6.iated with particular individuals living in the affected dwelling, most of whom were adolescents, and most of whom were affected by some neurological,:emotional ailment, most commonly epilepsy. Often a precipita-drig traumatic event could be identified which seemed to init@ate the activity, after which the general pattern involved a period of relatively mild precursor events, a sequence of major dis- turbances, and a period of "after shocks," extending as much as several weeks beyond:the main events. Controlled experimen- tation. on poltergeist phenomena has proven virtually impossible because of their infrequent and unpredictable occurrence pattern and because of the delicate physiological and emotional situations prevailing in r:nost cases, which have taken precedence over the technical investigations. Nevertheless, these prio'cesses have retained some fascination because of the magnitude of the energy transfer involved and because of their appajent correlation with specific types of individuals and neurological disorders. The most systematic 'and persuasive studies of PK, however, have dealt with much more modest scales of physical distur- bance, in some cases reaching down to the atomic level. This somewhat more viable domain has been addressed by numerous investigators in various Ways, but basically one of two strategies is followed. In one approach, relatively simple physical systems are employed -mechanical, electrical, optical, thermal, etc.,- each of which involves a particular component or process that is ultra-sensitive to disturbance. The experiment is arranged to signify such disturbance.by a Irelatively large change in some dis- play which provides feedback indication to the operator, much in the spirit of a biofeedback instrument, and simultaneously to provide some form of permanent data record, Examples of this class of experiment would in@lude the use of magnetometeTs, torsional pendula, optical interferometers, electronic strain gauges, glow discharges 1143 ), and sensitive thermistors [ 144]. In the second approach, attempt is made to distort the nor- mal statistical patterns of various random physical processes on either a microscopic or niacroscopic scale. In a sense, these ex- periments deal with energy rearrangement within the systems, i.e., with their inf ormatibn content or entropy, rather than with energy transfer-to the system per se. The earliest versions of this class employed dice, or other simple mechanical implements of well-known statistical behavior [371, (921, [1451-[1481, but more recent studies:have tended to employ more sophisti- Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 142 PROCEEDINGS OF THE IEEE; VOL. 70, NO. 2, FEBRUARY 1982 001-4 FABRY PEROT ETALON MOVING _@L AT E EXTENDED LENS LIGHT A r-----T --- IrI REFLECTIVE COATINGS RAMP DRIVE FRINGES ON SCREEN PHOTO @<IPLIER T PIN HOLE @CREEN CHART RE:ORDER Fig. 1. Photograph and schematic of Fabry-Perot interferometer. cated apparatus, such as electronic REG's like those described in detail below. Although the first two categories of PK feature more dramatic effects and thus have enjoyed greater popular attention, the smaller scale phenomena seem more amenable to controlled experimientation and theoretical interpretation, and the re- mainder of this section deals only with such. Perhaps the most efficient means of elaborating on this type of research would be to review briefly the spectrum of such experiments in progress in our own laboratory, and then to display and attempt to interpret data, from one of them. In so doing, we intend no neglect of other work noted in the references, but simply deal from greater familiarity. As examples of the first class of low-level PK experiment mentioned above, we have in operation a Fabry-Perot optical interferometer, a dual-thermistor bridge, and a photoelastic strain detector. In the interferometer experiment, shown in the photograph and schematic of Fig. 1, a Coherent Optics Instrument _#360-370, using a diffuse sodium lamp as source, is adjusted to produce circular fringes on a screen visible to the operator (Fig. 2). Small changes in the separation of the inter- ferometer plates cause the fringes to migrate radially inward or outward. By visual observation of the fringe movement, plate motions of less than 0.1 wavelength can be readily detected. Via a pinhole in the screen, the brightness of the central fringe is recorded by a phot omultiplier/ch art recorder system at an order of magnitude higher sensitivity, thereby preserving quan- titative output data while the operator simultaneously sees an attractNe optical display of his progress for use as feedback. The task of the operator is to elicit significant migration of the fringe pattern in a stated direction relative to the normal baseline drift of the instrument. The protocols involVe rigid control and monitoring of the environment of the instrument and @urrounding laboratory, and the interspersing of baseline responses with active PK efforts obtained under otherwise identical conditions, including the position of the operator and any other personnel relative to the instrument. In pilot studies with this device, a variety of fractional- fringe responses were observed, using several different operators and various initial interferometer settings. A more formal procedure has since been developed which provides more precise conditions for an ongoing series of trials. In this protocol, the central fringe is set initially on a maximum gradient position between a bright and dark fringe, and its progress monitored for subse- quent periods of baseline or PK effort. Encouragingly repli- cable data have been obtained from a number of different operators, in the form of chart recordings of 5-min PK trials with interspersed 5-min baseline drifts of the instrument. Using computerized graphic, regression, and spectral analyses of the data, it is possible to discern characteristics in the hierarchy of trace derivatives and the Fourier spectra which, while not definitive, display certain recurrent features (97]. No physical interpretation has been attempted other than to acknowledge that the observed fringe migrations could also be indicative of slight changes in the index of refraction of the air in the plate gap or in the wavelength of the light source, as well as of a displacement of the plates. The dual-thermistor experiment comprises a much more sen- sitive version of a multiple-thermistor arrangement on which PK influence was originally reported by Schmeidler [1441. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 JAHN: PSYCHIC AHENCIMEN ENV NG 143 e@ERI PECTIVE pprov&@or e ease kfi/64/18 CIA'-RDP96-00789ROO3100030001-4 0 Fig. 2. Interferometer fringe patterns. As shown in Fig. 3, two Omega Engineering thermistors, Model UVA 3254, each with its own electronic bridge and volt- age source, are connected differentially to a Tektronix I A7A oscilloscope preamplifier and to a visual feedback display. With suitable ground planes and cable shielding, sensitivities greater than 0.001 K are obtainable, and by subtracting the two outpul signals the major portion of spurious electrical and mechanica! interfer drice is eliminated. The effects of ambient thermal vari- ations _i@ the laboratory are essentially excluded by enclosing each of the thermistors in identical Pyrex flasks immersed in a large liquid reservoir, in which configuration the undisturbed system exhibits stable baselines over long periods of time. Using the same interspersed baseline protocol as in the inter- ferometer experiment, the task of the operator is to achieve an increase in the reading of one thermistor with respect to,the other or in some more subtle fashion to alter the PIC response relative to the baseline. Some such effects have indeed been ob- served, but little systematic data have so far been accumulated on this experiment. Also in a preliminary stage- is an experiment to monitor in- ternal strain in a solid specimen via photoelastic optical tech- niques. Several studies have been reported on the PK deforma- tion of solids, but most of these have employed conventional engineering strain gauges or microacoustic sensors as detectors (981, [991, [1351, [1491, [1501, both of which require sul@- stantial interface electronics before a feedback signal reaches the operator, leaving unclear the role of the sensor in any pos- sible PIC influence. Although less-sensitive than the electronic 'pethods, photoelastic techniqu'6s have the advantage of relating the operator more directly to the sensitive 'element of the ex- periment via an attractive optical fringe pattern much like that of the interferometer (Fig. 4). This.equipment and technique may.also be applied to a sensitive levitation experiment wherein Approved For Release 2003104/18 : CIA Fig. 3. Dual-thermistor apparatus. ct is suspended on a photoelastic lever arm of suitable the. obje dimensions. Within the second category of low-level PK experiment, we are employing or are now constructing several devices based or. random physical processes, some macroscopic in scale, others deriving from atomic-scale processes. The largest of these in- volves a 6 X 10 ft apparatus, shown in Fig. 5, which drops som- 10 000 @- in spheres through a "quincunx" array of 336 nylon 4 pegs in about 12 min. As a consequence of the multitudinous collisions with the peg§ and with each other, the spheres are dis- persed into a good approximation of a Gaussian distribution as they fall into 19 q 'ollecting compartments at the bottorn. The goal of the operator is to distort the distribution in some prescribed fashion to a significant degree compared to empirical baseline experience. Photodiode counters mounted, in funnels at the entrance to each bin provide real-time digital displays Of aCT_ the bin populations to @up plement them ore qualitative feed b 11 of the growing ball stacks seen by the operator and to provide quantitative data for on-line statisfical analysis. Fig. 5 shows a typical baseline distribution for this device and a distorted distribution obtained in a particular PK effort. Full statistical analysis of the significance of any particular achieved pattern is a challenging problem in its own right, since it must. deal with a combination of 19 bin populations, each of which has its own empirical baseline mean and standard deviation, all constrained by total ball count. A somewhat similar, experiment, not yet refined, employs a device which allows:small metallic or dielectric spheres to bounce on anogticaljy flat, precisely horizontal circular plate of glass, which is oscillated by a vibration coil at frequencies from 10 Hz to 20 006 Hz. In the absence of any external dis- turbance, a sphere started at the center of the plate executes a random walk toward ihe outside edge, arriving with equal prob- ability at any azimuth. Since the sphere may make as many as 105 collisions in the process, it is vulnerable to statistical distor- tion of its trajectory and consequent terminus. The task of the operator is preferent ially to direct the sphere to a prescribed terminal quadrant. . In an attempt to intervene with a random physical process at the atomic level, we .have constructed a large glow-discharge device whose luminous patterns are indicative of the mean free path of the current-(;.arrying electrons against inelastic excita- tion collisions with the background gas. This device, shown in , Fig. 6, presents a 36-in X 2-in diarn cylindrical glow marked by a sequence of bright: and dark zones along its positive column -RDP96-00789ROO3100030001-4 144 Approved For Release 2003/04/1 PROCEEDINGS O@ THE IEEE, VOL. 70, NO. 2, FEBRUARY1982 001-4 Fig. 4. Photoelastic stress pattern. "NORMAL" DISTRIBUTION DISTORTED DISTRIBUTION Fig. 5. Gaussian an3log device and distributions. typical of dc discharges in a given range of gas pressure and terminal @voltage. The number and locations of these striations are sensitive to the electron inelastic mean free path, which in turn depends on the gas type and density, the electron tempera- ture, and the local electric field. Striation position is monitored by photoelectric detectors, and the goal of the operator is to expand or contract the pattern on demand, to a significant extent compared to the normal background jitter and drift. Protocols are much the same, output data take the same gen- eral form, and are analyzed by the same algorithms as in the interferometer and photoelastic experiments. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Fig. 6. Glow discharge experiment. JAHN: PSYCHIC PHENONIENA: ENGINEERING PERSPECTIVE Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 q LOv FILTER DaE . I E,, L-------------------- ELGENGO NOISE SOURCE 0 3602 A 15124 SAMPLE PULSE Fig. 7. Functional diagram of REG. (a) (C) 145 Fig. 8. REG waveforms: (a) Filtered noise. (b) Clipped noise. (c) Sampling pulses. A number of other atomic-scale random system PK experi- ments are under consideration, design, or construction, ihvolv- ing such processes as information storage on a microelectronic chip, the spontaneous decay of phosphorescent surfaces, lam- inar to turbulent transition in a fluid stream, atomic and molec- ular resonators, and resonant acoustical or electrical cavities, but none of these is far enough advanced to merit description here. Rather, we shall concentrate for the remainder of this section on a more detailed presentation of our most serviceable experiment, and the one on which we have the largest data base, the electronic REG. RANDOM EVENT GENERATOR EXPERIMENTS REG's have been the most widely used and most productive facilities for experimentation witl@ low-level PK. Although a broad variety of such devices exist, most involve four conceptu- 'ally and functionally separable components: in electronic noise source; a sampling systpm which examines the noise at pre- Dm 0 , nuu a @fiau Vow- A M@PLE @. A-@ @CLII @A. I @GA scribed intervals and prepares an output pulse train correspond- ing to the samples thus obtained; a system which analyzes the pulse train in accordance with preset instructions and prepares suitable output for a .feedback system; and the feedback dis- play itself, which informs the operator of the results of the sampling process. The particular versi.;on we have employed utilizes a packaged commercial noise soutce module based on a solid-state junction and precision preamplifier (Elgenco Model 3602AI 5124), but modules employing radioactive decay units or glow discharges can be readily substitu ted. This source produces a random noise spectrum up to severg megahertz, which our logic circuit first filters to a flat spectrLim from 50 to 20 000 Hz, then amplifies and clips to the flat-topped profile shown in Figs. 7 and 8. This is'then sampled by a regular train of gate pulses, yielding a cor- responding random sticcemion of positive and negative output pulses indicative of th@ sign of the noise at the time of sampling, and these are then counted. Since the averagp time between Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 146 . I or e Mt R INTEP4AL SAMF SAMF SE I CIRCI E 0 signal Train of ov" W54sourcit N PWI*s +1.0voll +f_@l Gnd__ T I_ T SIG" + I LEVEL METEA LO wit 5 -LNG Int RATE P@:- I a@l n @111 GEKERATOR CRart E.1trnal P.Ite Gen 'clot 1 5 nucto w-iciffil Ponei ,at I 01 CUM ULAT E NV 0.fWI Up, DO. COUNTER AND Y EVENT Cx - j AND LEI @ TRIA_:OL0. .ER Duplical AND @m O(SPLA, Print Cornmand PRik-19 PROCEEDINGS OF THE IEEE, NOL. 70, NO. 2. F-F_aRU.-,RY 1982 SAMPLE - @96 078912!@@Ao%oeqoopA Is r CONTROL E'rtnt CIRCUIT EVENT MODE Sol N SELECTOR SCALE-OF-N COUNTER Slovq DiWoy Unit P,int Gale OFF off Gals ON SCALE - OF - 50 5 T M A COUkTER C TROL C1111 CUIT Mon.o( Start a, Auto@atic Repeat START (50 T-4s) 0 Reset to Printer Trial Counter Reset Sgnol To E,vtni Countvs Fig. 9. Electrical schematic of REG. Fig. 10. REG arrangement, zero crossings of the clipped noise waveform is about 30 Ps, sampling rates to about 15 000/s can be tolerated with statis- tical independence. The full functional array is sketched in Fig. 9, and a photo- graph of the boxed units in Fig. 10. By panel setting the sample.; may be instructed to take "trials" of 100, 200, or 2000 samples, at a frequency of 1, 10, 100, 1000, or 10 0001s. The counting system may be set to count only T.-Siti-ve rulses, only negative pulses, or to alternate positive ant negatiN-. counting on successive samples. The alternating pm:ive:,,-neg!tive mode effectively factors out any systematic bias n the ncise source, and is the mode employed in all the experim-11tS reported here. The counting results are displayed by LED E_-rays tr-,zking both the running count of each trial and the coicurrent mean rela- Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 JAHN: PSYCHIC PHENOMENA: ENGINEERING PERSPECTIVE Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 TABLE I REG 200-SANIPLE DATA SUMMARY Series Instr. No. Trials Mean Std. Dev. t-score P, n.In- REG I B.L. 12000 100.009 6.994 0.144 0.443 5678/5611 PK 4 550 100.264 7.037 2.528 0.006 22330/2056 PK- 3850 99.509 7.063 .@4.313 10-1 1716/1926 A PK 8400 4.890 5 X 10 REG 11 B. L. 2500 100,033 6,875 0.239 0.06 1188/1179 PK+ 1950 100.247 6.849 1.590 0.056 916/919 PK- 1800 99.597 6.775 -2.526 0.006 797/902 A PK 3750 2.920 0.002 REG III B. L. 3500 99.977 7.013 -0.193 0.424 1658/1655 PK+ 2400 100.227 6.821 1.634 0.051 1150/1086 PK- 2600 99.736 7.026 -1-918 0.028 119211270 A PK 5000 2.507 0.006 T, REG I B. L. 18 000 100.006 6.981 0.115 0.454 8524/8445 11 PK 8900 100.250 6.938 3.403 3 X 10-4 4296/4061 III PK- 8250 99.600 6.989 -5.203 10-7 3705/4098 A PK 17 150 6.107 5 X 10-10 REG Ia no B.L. PK+ 2150 100.206 7.091 1.340 0.088 1059/993 PK- 2100 99.945 6.937 -0.365 0.358 954/1019 A PK 4250 1.213 0.113 REG Ila B. L. 5000 100.186 6,974 1.882 0.030 2367/2337 PK+ 2000 100.117 7.041 0.746 0.228 955/950 PK- 1 750 99.941 6.898 -0-360 0.359 803/839 A PK 3 750 0.772 0.220 I REG I B.L. 23000 100.045 6.980 0.978 0.164 10891/10782 Ia 11 PK+ 13050 100.223 6.979 3.644 10-4 6310/6004 Ila III PK- 12 100 99.709 6.968 -4-596 2 X 10-' 5462/5956 A PK 2-5 150 5.828 3 X 10-9 tive to a preset origin and are permanently recorded on a strip printer. For most of the experiments described below, an AIM- @5 microprocessor interface is also utilized to insert the trial- count -data on-line into processing routines supported by a TERAK Model 8510 used as a terminal and PDP 11/45 and VAX 750 employing a UNIX operating system programmed in Clanguage. All of the sampling, counting, and display functions can be simply checked by referring them to an internal or ex- ternal calibrated pulse train generator. The device also has a manual/automatic option, whereby it will either collect its trial samples only when a panel switch or parallel remote switch is pushed, or it will repeat that process for 50 trials automatically once activated Uy the switch. The operator thus has the option of triggering each trial or of initiating a repetitive flow of 50 such trials With no further intervention. The experiments reported here were perfonned by a single operator, seated in front of the device with the remote initia- tion switch in hand and the LED count indicators and TERAK terminal display visible. This operator-attempted, on instruc- tion or volition, to distort the trial counts either toward higher or lower values. The several options of sampling number, sam- pling frequency, +/- polarity, and manual /automatic sequenc- ing were variously determined by'random instruction, operator preference, or experimental practicality, and recorded before the beginning of each trial. Clearl.y, the full-matrix of such pos- sibilities could not be explored, and for our first sequence of experiments only 2QO-sample trials were used, at 100 or 1000 counts/s, all counte ;d in the +,i'- alternating mode. The au-0- matic/manual and Iiigh/low options were more thorou-c'-ly tested, in both the volitional and instructed choice modes. Fifty trials of the 200-sample units comprised a test run, and data from these Were. processed individually and in many' concat- enations via a statistics package in the UNIX system develo;ed specifically for this task. Calculated were the mean, standard deviation, range, kprtosis, skew coefficient, z-score, t-scrre, X goodness-of-fit with both 8 and 16 degrees of freedom, and the corresponding one-tailed probabilities against chance ofthe last four measures., Applied to earlier and ongoing basehne data, this analysis confirmed that in undisturbed operation this REG conforms very' well to a Gaussian approximation to the appropriate full binary statistics. The major portion, of the results listed below comprised three separate experimental series, extending over fifteen months, labeled REG 1, REG II, and REG III, respectively. All other data acquired under slightly less formal conditions of protocol during this period,; included for completeness, are grouped under two other series, labeled REG la and Ila, respectively. Details of these series protocols, calibration tests, and their in- dividual results are. available in the reference [931. All told, 'over 25 000 active'PK trials were obtained, corresponding to more than 5 000 000 binary events. Table I summarizes all of the baseline and PK data acquired during these five experimental series. A total of 23 000 base- Approved For Release 20.03/04/18 : CIA-RDP96-00789ROO3100030001-4 148 _j < 0 PROCEEDINGS OF THE IEEE, VOL. 70, NO. 2, FEBRUARY 1982 Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 6- n - 23000 4 - 2 - 6D 80 106 120 140 COUNT Fig. 11. REG 200-sample baseline data on theory. 0 0 - PK* I- - t_:3050)) : PK 2100 6- 4- 2- 80 10.0 120 14C COU!,T Fig. 12. REG PK' and PK_ 200-sample data on theory. 0 F- COUNT Fig. 13. REG PK' and PK- 200-sample data fitted curves. line trials were taken under a variety of conditions before, during, and after the active PK trials. Their overall mean was 100.045, and their standard deviation 6.980, compared with the values of 100.000 and 7,071 predicted by the theoretical Gaussian approximation to the appropriate binary statistical distribution. As shown in Fig. 11, the fTequency.of count dis- tribution conformed very well with the theoretical curve. The results of the PK trials are also presented in Table I and in Figs. 12 and 13. Briefly, the 13 050 high-instruction trials, denoted PK, yielded a mean of 100.23 and a standard devia- tion of 6.979; the 12 100 low-instruction trials, denoted PK_, yielded a mean of 99.704 and a standard deviation of 6.968. The one-tailed probability of chance occurrence of the former, computed from t-score, is about 10-4; of the latter, about 2 X 10-5. The combined probability of the split, i.e., of this total "direction-of-effort" success, denoted APK, is about 3 X 1.0-9. A number of more elaborate statistical measures have been applied to these data; the resWts are not qualitatively changed thereby.) As is evident from Figs. 12 and 13, and as is verified by the PK' z 2000 -.05 1000 ------ BASELINE 0 ................ W ?! 1000 ----------- --- Pi -.05 2000 O PK_ 3OO L - - _U 2000 4000 6000 8000 10000 12000 14000 NUMBER OF TRIALS Fig. 14. REG cumulative deviations ofPK* and PK_ 200-sample data. > < NUMBER OF P@'dRS OF TRIALS Fig. 15. REG cumulative deviations in direction of effort of all 200- sample data. 0.8 PK' z 0.4 - 0 -T-.05 2t 5 LJ 0 _j ------------ 7 W --------------- < all 0.4 cc 'K W .05@/' 0 2000 4000 6000 8000 10000 12000 14000 NUMBER OF TRIALS Fig. 16. REG cumulative average deviations of PK* and PK- 200-sample data. more detailed statistical tests performed, no significant distor- tion of the fre quency- o f- count distributions other than the shifting of the means has occurred. In other words, the ob- served effect is to shift the total distributions intact, rather than to distort any of their higher moments significantly. This result clearly has felicitous implications for this class of expeni- mentation, since it allows much simpler and faster data collec- tion and analysis than might otherwise have been anticipated. It is illustrative to exhibit the overall data behavior via graphs of the cumulative deviation of the trial score means versus the accumulated data base. Fig. 14 shows such a representation for the total data pool plotting PK', PK_, and baseline data with reference to cumulative 0.05 confidence levels. Fig. 15 uses a similar representation for compounding the PK+ and PK_ data in a "direction of effort" cumulative deviation. (Had REG Ia and Ha been excluded from these data, the overall slopes would have been slightly more severe and uniform.) Alternatively, the cumulative data may be presented in terms of the progressions of the average deviations from the theoreti- Cal mean, as shown in Fig, 16, where the stochastic variations Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 JAHN: PSYCHIC PHENOMENA: ENGINEERING ?F RSPECTIVE 149 of the sin all Apia r9M~Afec%J3*@Rw p299,1494MAde, &@A-RIRPP@@QATA9 terminal values after several thousand trials. The PK+ and PK_ effects also manif-st themselves in terms of the number of trial means recorded above and below the theoretical value. As displayed in Table 1, PK' efforts were generally characterized by an excess of tr.'al means above 100.00 and PK_ efforts by an excess below- The total con- catenations in these terms are significant at a level of 0.003 for PK_' and 3 X 10-5 for PK_. The ensemble of results acquired in these experiments dis- pl2y certain instructive general features: I ) The importance of accumulating very large data bases when dealing with such marginal phenoina @s emphasized by the relative scales of the statistical vagaries and the broader sys-ematic trends in Figs. 14-16. Mthoug'a the trends are established early in the data collection sequ@nce, unambigu- ous departures from the accumulated vagari,-s of chance be- hav'or occur only well into the total 25 000 trial, 5 000 000 bit. sequence. -1) Over this large a data base, there arises @ome quantitative statistical regularity in the PK process, epitom@zed by the mean slopes of the cumulative deviations in Figs. 14 and 15 and by the terminal values of the average deviations ir. Fig. 16. Traced back to the elemental binary samples, these values imply direc- ted inversions from chance behavior of about one or one and a half bits in every one thousand or, alternatively, of 0.2 or 0.3 bits per trial. ;) The differences between the somewhat larger values for the PK_ deviation and the lesser values for PK' are only mar- ginally significant on this database, but prevail father uniformly throughout all the test series. The suspicion -hat these reflect some subtle bias in the REG itself is not supported by the base- line data, which concatenate to a grand mean very slightly above the theoretical value. One of the prima., goals of such -controlled PK studies at this early phase in the understanding of the phenomena is to develop experiments of sufficient yield and replicability that vanous parametric correlations may be systematically explored, thereby hopefully separating the consequential from the incon- -sequential factors. The experiments outlined above hold some promiso' of serviceability for this purpose, but a great deal of data will need be accumulated to establish any such correla- tions. Four classes of parameters could be considered: those associated with the experimental equipment; those associated with the operator's physiological and emotional characteristics; those associated with the operator's technique; and various en- vironmental factors not directly associated with either. So far we have accumulated only small amounts of data from other operators, and given the general indication #1 above regarding the importance of large data bases, we can Make no statement about the generality or peculiarity of our principal operator's performance. Similarly, we have attempted no systematic vari- ation of external environmental factors, and although test times, dates, durations, and laboratory temperature, pressure, and humidity have been routinely recorded, we cannot comment on the importance of this category of parameter. On the matter of operator technique, it should first be re- emphasized that the sole formal difference between the PK+ and PK_ trials is the specified intention of the operator to in- fluence the device to generate numbers in the assigned direction. No other variation in protoco is Permitted, save those subjec- 'tive differences in psychological attitude the operator chooses to invoke. Although no records of such aspects were kept, PAPA PAPOnAl-eliM talent for this or any other psychic task, reported that any conscious variations in psychological strategy, such as focus of visual attention, or intensity of concentration or desire, did not appear to have any evident effect on the yield. Similarly, differences in the laboratory ambience, such as the lighting level, back-ground noise, or peripheral ;presence of other persons, did not seem to influence this op@ Irator's performance. When queried about any impressionistic s.@nse of the interaction process, the opera- tor alluded to a "resonance or identification with the system, leading to a loss of $elf-awareness similar to that experienced in a game, a movie, or some creative occupation." Clearly this class of parameter will be the most difficult to specify and cor- relate, and we are far from any definition of its mechanisms. With respect to experimental options on the equipment pa- rameters, we can make very limited explorations with the acquired data base. Briefly, binary correlations of the data for the 100/1000 counting rate option, for the volitional/instructed direction of effort, and for automatic/manual sampling give little indication of importance of such factors in the overall performance. Each d ategory shows clear and significant separa- tion of the means for the PK' and PK_ efforts, with little to choose between the r-scores for the various categories. Thus, at least for the data base at hand, the process seems insensitive to these particular experimental parameters. We have also atte Mpted correlation in terms of the trial- number sequence. With cognizance of the- ubiquitous "de- cline effect" which i Is reported over a broad range of psychic experimentation, we have prepared an algorithm which cross- concatenates from the data base all scores achieved on the first trials of the exp .erimental run, all achieved on the second trials, etc., up to thei! fiftieth, and arrays those fifty means in a graphical form. The results show little systematic profile of yield versus trial number. A similar exercise has been per- formed to cross-concatenate the data by run-number over the various series to sear@h for a decline effect on that larger scale, but again no significant correlation is found within this data base. The most extensive parametric exploration attempted to date was motivated by the'apparently fundamental question implicit in general conclusion, 2) above, i.e., whether the magnitude of the observed effect correlates with the total number of bits processed, or with the number of trials. To explore this aspect, the same operator has perform .ed a second ensemble of exper-i- mental series totaling 25 000 trials, all consisting of 2000-sam- ple bits rather than @00. As before, various combinations of the automatic/manual and volitional./instructed modes were employed, but to speed data acquisition and reduce the opera- tor's tedium, only the 1000/s counting rate was used. This, coupled with the more elegant data processing capabilities that had evolved over the preceding experiments, allowed this sequence to be completed in less than six months. The results of this effort, as presented in Table 11 and Figs. 17 and 18, are curiously ambivalent. As before, there is clear and significant separation of the means of the PK+ and PK- efforts, and the baseline is well behaved. As could be antici- pated from the larger standard deviation of the 2000-bit data, the cumulative traces display larger statistical fluctuations and require a larger number of trials to settle toward well-defined terminal values. To the quantitative Precision allowed by this data base, these terminal values appear not to endorse any simple bit-level hypothesis in that they fail by a factor of 6 or 7 to achieve the ono ior one and one-half bits per thousand Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 ISO PROCEEDINGS OF THE IEEE, VOL. 70, NO. -2, FEBRUARY 1982 Approved For Release 2003/04/tp iCIA-RDP96-00789ROO3100030001-4 TAB SUMMARY OF REG 2000-SAMPLE DATA No. Std. Instr. Trials M= -Dev. t-score Pt n+/n_ B.L. 12500 1000.016 21.879 0.079 0.468 6IS7/6088 PK + 12200 1000.380 21.906 1.914 0.028 6092/S897 PK- 12800 999@569 22,005 -2.216 O.Q13 621816351 A PK 25 000 - - 2.920 0.002 PK- z 0 3750- 5 > PI W a .. - 7: -- ------------- W 0 > 05 3750 D L) PK- 1 -0 2WO 4000 GOOO 8000 1 12 0 14000 NUMBER OF TRIALS Fig. 17. REG cumulative deviations ofPK' and PK- 2000-sample data inversion accomplished in the 200-bit trials. However, the new values are larger on a per-trial basis by a facter of about 1.7, which is not negligible in this context. Again, much more data of this sort will be required to come to ggrips wi-Lh this class of correlation. In addition to continuing study of this sarnDle-SiZe parameter, our next generation of experiments employs a numler of other operators to explore the variation of yield 1@ith operator type and technique, and a number of alternate noise sour@es, includ- ing pseudo-random SOUTCeS, in an attempt to localize the effect somewhat and thereby to narrow the range of future experi- ments and models. The results outlined above are by no mems the only conse- quential REG data available for contemplat.-lon. 0-1 par-ticular interest are a variety of experiments reported by Schmidt, some employing pseudo-random as well as physically random sources, and others using taped source outputs recorded well in advance of their presentation to the operator [87]-(90]. In another approach, May has recently reported an REG study using electronic gear specifically designed to pretlude very subtle artifacts that might confound the effects of interest, and includes in his paper a thorough search of the modem REG literature [911. In addition, considerable research in the parapsychological community has been performed using REG devices as drivers for various forced-choice video games employed in both the PK and clairvoyance modes [ 151 ). Many of these claim sig- nificant yields, but rarely are the data-bases sufficiently large to present quantitative trends, or to allow much parameteric correlation. Regardless of their particular implementation, any potential vulnerability of random electronic noise sources to incidental or intentional distortion by the means under study here would seem to be of some interest to a number of engineering corn- munities, given the proliferate application of such devices in various functional and computational capacities. REMOTE PERCEPTION As a second example of contemporary psychic research that has dis! layed some substantial yield and,interlaboratory repli- .P z LLJ 2: > _J 3 L) Fig. 18. REG cumulative average deviations of PK' and PK_ 2000- sample data compared to 200-sample data. cability, we select a topic which has come to be called "remote perception" or "remote viewing." The basic concept of this process is far from new; in the early 16th century, Par:celsus stated it unequivocally: Man also possesses a power by which he may see his friends and the circumstances by which they are surrounded, although such persons may be a thousand miles away from him at that time [ 61. In its modern form, the experimental protocol requires a 11 percipient" to describe, by free-response oral or written narrative or drawing, a rernote, unknown target loca-Lion at which is stationed an "agent," with whom there is no normal sensory mode of communication during the course of the experiment. The targets are usually selected by some'Drescribed random process from a previously prepared pool of targets, which is unknown to any of the active participants. The quality of the perception is assessed by various impressionistic or analytical judging methods described below. Historically, this experiment has evolved from several ganera- tions of free-response clairvoyance and telepathy experL-nents, which were found to have certain advantages over the more traditional "forced-choice" ESP tasks, such as the Xener card identifications of the early Rhine laboratory [341-[371, in displaying less tendency for percipient stagnation and "decline- effects" over extended testing, and in maintaining some of the spontaneity of anecdotal clairvoyance experiences. One of the earliest detailed reports ofsuch free-response studies appears as a book by Upton Sinclair entitled Mental Radio, which features an equivocal foreword by Albert Einstein !1521. More modem work of this class was performed at the Mai- monides Medical Center by Ullman and Krippner in the 1960's, and reported in their book Dream Telepathy [641. From this work emerged the so-called "ganzfeld" or sensory inhibition perception studies of Honorton and many others which pro- pounded, the desirability of emotionally stimulating tasks to which the subjects could relate in a personal and spontaneous fashion [661, [671. The contemporary version of the remote perception protocol was introduced in a sequence of publications by Targ and Put- hoff [941, [1531-[1561, which prompted a substantial num- ber of attempted replications [1571-[174], and considerable Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 0 2000 4000 6000 8000 10000 12000 1400Z NUMBER OF TRIALS JAHN: PSYCHIC PHENOMENA: ENGINEERING PERSPECTIVE CIA-RDP96-00789ROO3100030001-4 Fig. 19. Remote perception target: Picnic area, Feathered Pipe Ranch near Helena, MT; 12:00 N MDT, Sept. 5, 1978. Percipient transcript: Princeton, NJ; 8:30 A. EDT, Sept. 5, 1978: "Outdoors. . open landscape-large areas of trees-pines? inter- spersed with open fields. Single road. High overcast, cool, breezy. (Agent) in dark jacket talking to someone near road-possibly turn- out area or picnic area. Assembly of stones-possibly pylon or marker orwall. Large sign somewhere." critical comment. The most extensive of the replications, con- ducted by Dunne and Bisaha in the Chicago area over the period 1976 to 1979, comprised 40 formal trials to which were applied 157 independent transcript judgings, 84 of which assigned first- place rank to the proper targets [ 16 11, [ 16 2). The type of data which can be acquired in such studies is illustrated in the sequence of Figs. 19-22. In each case is shown a photograph of a particular target, selected by some random process, which was visited by an agent on the. date and time indicated. Below each figure is a portion of the cor- responding percipient transcript, with the time and location of the perception effort also noted. The examples shown are drawn from a variety of experimental series conducted under somewhat different protocols, but serve to display some of the characte ris tics which commonly appear in the more successful efforts:: 15 '17he overall ambience of the scene is accurately perceived. 2) Certain details are accurately identified; others are mis- construed or totally ignored.. 3) A feature which is impressive to the agent is not necessarily so to the percipient, and vice versa. 4) The composition of the scene may be distorted by errors in scale, relative positions of key objects, or total right- left inversions. 5) The aesthetic aspects, such as colors, general shapes, degree of activity, noise le 'vel, climate, and other ambient features tend to be more accurately perceived than more analytical details such as number, size, or relative positions. 6) The perception is not necessarily centered on the defined target, and may even provide accurate information on ad- jacent areas external to the target, unnoticed by the agent. 7) The fidelity of the perception seems to be independent of the remoteness of the target, up to distances of several thousand miles. 8) The time of the perception effort need not coincide with the time the agent is at ,the target. Perceptions obtained several hours, or even days, prior to the agent's visit to the target, or even prior to selection of the target, display at least as high a yield as thbse performed in r6al time. 151 Fig. 20. Remoit perception target: Woodrow Wilson School, Princeton, NJ; 2:15 P EDT, Aug. 2 8, 19 80. Percipient transcript: Princeton, NJ; 12:15 P EDT, Aug. 28, 1980: "Some kind of counyard, enclosed by buildings on two sides. Paths or walks arourd periplery, a statue or monument of some kind in the middle surrounded by'grass. Could be a fountain; I have the feeling of water. Trees or t3li hedges on one side. Fairly quiet, but some people v@,alkinz around' Not sure of sound, the idea of a fountain sug- gests sound of water but I'm not sure I really hear it or not." The philosophical apd. practical implications of items 7 and 8 are clearly substantial. If the data are valid, the most parsimo- nious explica-,ions wduld require access of the percipient's con- sciousness to other pprtions of the space-time grid than that in which it is currently immersed, or that it can reach by normal processes of communication or memory. These same items also seriously delimit the potential physical mechanisms for such access. Rigorous evaluation of the data from experiments such as these is confounded.by the psychological components of the process, by The imptessionistic nature of the information in- volved, and by the inevitable subjective biases of all those par- ticipating in -.he experiment. Doubtless the earliest and most primitive asse@ssmenzs were informal a posteriori exchanges of impressions about tfte target between agent and percipient Which, although possibly informative and stimulating to them, lacked any quantitative basis and held little scientific credibility. In a somewhat less vulnerable strategy invoked more recently, the percipient, after, completing his transcript, visited several possible targets dTaA-9 from the pool and attempted to identify the one he perceived, or to rank-order each of them in terms of conformity to his.earlier perception. Statistical arguments could then be applied to these ranks to estimate the likelihood that information about the target had been acquired by means other than chance [ 1751- [ 177). In an attempt to separate the possible ESP functioning of the percipient during the! visitation and ranking process from his original perception eftort, the protocol subsequently evolved to invoke independent judges who were provided copies of the various transcripts and taken to the target sites to perform their preferential rankings.@ Even in this form, the technique has been criticized for possible sensory citing of the judges [ 1781 and has tended in turn to be replaced by a protocol wherein the judges perform their ranking on the basis of photographs of the targets, usually taken by the agent at the time of the trial. In one such version, the judges, who have not been in- volved in any earlier, portion of the experiment, are asked to compare a single percipient transcript with agent-generated descriptions and photographs of a number of alternative targets, Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 152 PROCEEI)INGS Or THE IEEE, VOL. 70, NO. 2, FEBRUARY 1982 Approved F R003100030001-4 Fig. 2 1. Remote perception target: Rockefeller Chapel, Chicago, IL; 2: 15 P CDT, June 10, 1977. Percipient transcript: Mundelein College, Chicago, IL; 1:00 P CDT, June 10, 1977: "I'm seeing a heavy wooden door with a black bolt on It rounded at the top in a dome fashion. I have a feeling of opening the doors and looking in and it's dark inside. My feeling at the moment is that it's a building like a church, And I can see the pews. There is some light but I feel basically a kind of darkness in there and a quietness. I'm seeing little turrets, very elaborate-looking little turrets, a whole series of them like across the entire top of the building and there's a straight line and then up to a triangle. I haye a definite image of an angel-type of statue, marble, flowing robes. I see the door again and I see some stairs. I think it's very high. I'm getting some stained-glass windows that are arch shape and they would took to be dark blue. Whatever the architecture of it the ornamentation on the building is quite elaborate and it looks like there's a section on the top with the turrets and then below that there are some other kinds of designs but more linear designs. "I again have a vision of the doors and then maybe a ledged area or an area of the building that protrudes with some kind of a design and there maybe even be a couple of those before you get to the top part which is either triangular or rounded. There is filigree work, little filigree turrets or something. And within the building there is a sort of a continuation of arches, but possibly they meet columns or something like that, but whatever the decoration is where walls join or separations, it looks like It's arched." Approved For Release 2003/04/18: CIA-RDP96-00789ROO3100030001-4 JAHN: PSYCHIC PHENOMENA: ENGINEERING PERSPECTIVE 153 Release 2003/04/18 CIA-RDP96-00789RO03 100030001-4 d F or Approve tations which a human judge might capture, this method does provide a rudimentary framework foi- evaluation of signal-to- noise ratio in tae information transfer, and an assessment standardization less dependent on subjective interpretation. In essence, the stategy is to replace impressionistic assessment of the quality of a perception by the identification of specified elements of information therein, after which a mechanical scoring and ranking procedure takes over. In the hope of con- veying a bit more substantive flavor of the data acquired in A thereol. remote percepticn eXperiments and the processing . . permit us to describe this analysis in a little detail. The heart of the method is the establishment of a code, or alphabet, of simple descriptive queries which may be addressed to all targets and all perceptions, responses to which serve to distinguish them and to permit quantification of the informa- tion acquired in the Perception process. In one version, these "descriptors," thi7ty in number, are posed in binary form and -um from quite factual discriminations, e.g., range over a specir whether the scene is indoors or outdoors, whether trees a-7,- er there are automobiles, to much more im- present, or whethe pressionistic aspects, such as whether the ambience is noisy or quiet, confined or expansive, hectic or -tranquil. The particula-@ Zil._t ensemble of desciptors has evolved in part through personal experience and imuition, and in part through trial-and-error application to vanous pilot data. The goal has been a balanced alphabet whose elements are a) relatively unambiguous; b) com- Fig. 22. Remote perception target: Danube River, Bratislava, Czecho- monly perceived ty a broad selection of percipients; c) individ-u- slovakia; 3:00 P European Standard Time, Aug. 24, 1976. - ally instructive in defining the scene; d) complementary to one Percipient transcript: Minoqua, WI; 8:30 A CDT, Aug. 23, 1976: another; and e) sufficient in number to permit reasonable syn- "I have the feeling that the agent is somewhere near water. I seem to thesis of the scene, but not so numerous to burden the data have the sensation of a very large expanse of water. There might be collection or computation excessively. boats. Several vertical lines, sort of like poles. They're narrow, not heavy. Maybe lamp posts or flag poles. Some kind of circular shape. Given this deschlitor. alphabet, each target in the pool is then It's round on its side, like a disc, it's like a round thing flat on the represented in terms of 30 binary bits, corresponding to the al@- ground, but it seems to have height as well. Maybe with poles. Could propriate YES/No responses to the queries. This encoding is possibly come to a point on top. Seeing vertical lines again. Seems to be a strong impression, these vertical lines. Predominant colors seem normally performed by the agent at the time of target visita- to be blue and green. Water again. Some very quick impression of a tion, although reference may be made to the target selector's fence, L low fence. The steps seem to go up to some kind of fence. It's a dark fence and it's along like a walk sort of at the top of the judgment or to photographs of the target for verification. steps. The steps sort of lead up to like a path or walkway. Like a Each perception is similarly rendered into a corresponding boardA-31k. And there's a fence along it. There's people walking ' sequence of binary digits, but only after the percipient has been Along it and there s vertical lines along that walkway." allowed to form a free-Tesponse impression of the target. Vari- ous scoring recipes are then invoked for quantitative comparison including the proper target, and to rate or rank-order them by of the perceptions with the targets, using for computation. the some prescribed criteria. Again, statisticaf assessment of the UNIX operating system of a PDP 11/45 or VAX 750. significance of the rankings follows. The simplest recipe merely counts the number of correct The bulk of the remote perception data reported in the litera- responses to the 30 descriptors, Le., the positive correlations ture has been evaluated by some form of these independent between the target and descriptor matrices. This does not nor- judging processes, and displays sufficiently high yield to encour- mally provide a particularly accurate index of the quality of the age further refinement of the protocol and analysis before individual perceptions, since the a priori probabilities of the attempting categorical judgment on the validity and viability various descriptors are widely different. For example, a given of the phenomena. Beyond minor tightening of the target pool may have more outdoor than indoor targets, and hence selection, agent maneuver, and perception acquisition and re- a correct identification: of an indoor context should be given cording phases of the experiment, the major potential improve- higher credit than identification of an outdoor context. To ments would still seem to pertain to the judging process, which facilitate such weighting, a step is included in the computational remains potentially vulnerable to subtle cues in the transcripts, program to provide the: a priori probabilities of all descriptors to vagaries in the judges' capability, to their subjective biases in the prevailing target pool, on the basis of which more elabo- toward the individual experiments and to the topic as a whole, rate scoring recipes may. be invoked. and to possible psychic input of their own [ 1611, 1621. Since the various targets have substantially different charac- In an effort to improve thejudging process further, our labora- teristics and hence different capacities for achieved scores, a tory has explored the applications of various information theo- variety of normalization procedures also have been developed, retic methods to the quantification- of the data, and an analytic using as denominators the total number of descriptors, the per- technique has been devised w@ich is based on a limited binary fect score, i.e., the scoTe that would be achieved for a given .alphabet of target/perception desrrip;ors [j791. While less target if all descriptors were identified correctly, and various sensitive to-the Gestalt impressionism and symbolic represen- "chance" scores for the target, defined by some random or Approved For Re'lease 2003/04/18 CIA-RDP96-00789ROO3100030001-4 154 PROCEEDINGS OF THE IEEE, VOL. 70, NO. 2, FEBRUARY 1982 Approved For Release 2003/(14M !ICIA-RDP96-00789R603100030001-4 " PRECOGNITIVE REMOTE PERCEPTION 2 4 X 24 "CHI CAGO SERIES PROPER TARGET RANKS Scoring Method Human Judges Perception No. A B C D E Avg. Avg.124 Mean Rank b 1 3/4 7 7 9 2 5.9 0.25 2.7/8 = 0.34 2 1/2 2 4 1 1 1.9 0.08 1.018 = 0.13 3 1 1 1 1 1 1.0 0.04 1.5/8 = 0.19 4 2/8 7/2 2/2 4/2 2/2 4.5 0.19 2.7/8 = 0.34 5 7/3 9 11 4 8 8.0 0.33 1.7/8 0.21 6 9/5 12/2 7/2 16/2 7/2 11.0 0.46 3.5/8 0.44 7 13/2 11 14 10 13 12.3 0.51 2.3/8 0.29 8 20/3 22 20 14 19 19.2 0.80 1.8/8 0.23 9 4/2 4 8 1 5 4.5 0.19 2.6/7 0.37 10 10/7 13 13 9 5 10.6 0.44 1.4/7 = 0.20 11 9/4 9 11 6 12 9.7 0.40 3.6/7 = 0.51 12 1/3 2 2 2 5 2.6 0.11 1.8/7 = 0.26 13 1/3 1 3 1 3 2.0 0.08 2.2/7 = 0.31 14 1 2 2 1 1 1.4 0.06 1.4/7 = 0.20 15 1/2 2 2 1 1 1.5 0.06 1.0/7 = 0.14 16 2 2 2 1 2 1.8 0.08 1.0/10 = 0.10 17 1/4 9 1 9 1 4.5 0.19 1.015 = 0.2C 18 2/2 10 5 1 2 4.1 0.17 1.015 = 0.20 19 14/3 14 14 19 17 15.8 0.66 unjudged 20 7/6 11 11 8 10 9.9 0.41 5.0/6 = 0.83 21 1/2 2 3 4 1 2.3 0.10 2.0/6 = 0.33 22 5/6 7 12 2 7 7.1 0.30 3.0/6 = 0.50 23 16/4 23 11 is 9 15.1 0.63 3.0/6 = 0.50 24 3 3 4 3 4 3.4 0.14 2.0/6 = 0.33 M ean 6.73 7.75 7.13 5.96 5.79 6.67 0.28 0.31 'Computed raAk/number of ties for that rank bAssizmed rank;number of possible ranks. arbitrar-y process of descriptor response. A "selective" scoring/ normalization process has also been applied which effectively allows the percipient to reject any descriptor on which he feels unqualifled to comment. and thence to be scored only on the reduced descriptor set. The statistical significances of these various normalized per- ception scores are assessed by a collective ranking process reminiscent of the traditional human judging techniques, but having the advantages that the ranking proceeds on a much more standardized and analytical basis, and that many more alternative targets can be ranked by the machine than by a human judge. Specifically, the program scores each transcript not onl, against its proper target, but against every other tar- Y get in the pool, and then ranks these targets in order of descend- ing score and specifies the rank of the match with the proper target. This process is repeated for every scoring method, and the results displayed in corresponding matrix arrays. Table III displ3yS typical results of these analytical Tanking procedures as applied to a group of 24 perceptions of 24 targets in the (.-lucago area. Tabulated are the ranks of the proper tar- gets compared Aith all other targets for each of the perception efforts, as computed by five of the scoring methods we have found to be most instructive, namely, A) number of correct descriptors/total number of descriptors; B) weighted full de- scriptor score/perfect score; C) weighted full descriptor score/ number of descriptors; D) weighted selective descriptor score/ perfect score -, and E) weighted selective descriptor score/chance score. Mso included in the table are the mean ranks assigned by independent human judges subjectively comparing these per- ception.s with a much smaller number of alternative targets. Although the bases of comparison are quite different, it appears that in the majority of these cases the analytical and impressionistic evaluations concur at least roughly in their estimate of the quality of the perceptions, particularly for those which consistently obtain low rank assignments. If the analyt- ical computation is carried throuih using as target pool only those alternative targets available to the human judges, the agreement in mean ranks is found to be somewhat closer, per- haps fortuitously so, given the categorically different bases of assessment implicit in the two methods. To this analytically scored and ranked data it is possible to apply a variety of statistical assessments of widely ranging sophistication and complexity. Consistent with the rather broad mesh of the descriptor code and the elementary scoring recipes invoked in this version of the concept, we confine our- selves to correspondingly simple statistical measures which provide at least semi-quantitative indication of the yield beyond chance. Specifically, we address only the distribution ofproper target ranks achieved in the series of perceptions, such as sum- inarized in columns 2-6 of Table 111. Using the common z- score method for a discrete distribution, the probability of achieving the mean rank of any of these columns by chance may be directly computed. Table IV displays the results of such calculation for the same 24 X 24 "Chicago" series. Note that, whereas all of the methods suggest significant departures of the computed mean ranks from chance, there is relatively little disparity among them, indicating that the specific method of scoring and normalization is not a sensitive element in the overall evaluation of the perception series. The departure of the shape of the proper target rank distribu- tion from chance is also displayed in Table IV in terms of the number of perceptions achieving first-place ranks, the number Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 TA13LE IV PRECOGNITIVE REMOTE PERCEPTION 24 X 24 "CHICAGO" SERIES Approved For Release 2003/04/18: MA47DIRM00789RO03100030001-4 Mean No. Ist (2nd) No. Ranks Method' Rank Ranks Below Mean Z P2 X2 (4) PX Chance expectation 12.5 1.0 (1-0) 12 A 6.73 4.4 (4.0) 19 B 7.75 2.0 (6.0) 20 C 2.0 (4.5) 20 D 5.96 8.0 (2.0) 20 E 5.79 6.0 (3.5) 21 aAs described in text: A) Number of correct descriptors/total number of descriptors. B) Weighted full descriptor score/perfect score. Q Weighted full descriptor score/number of descriptors. D) Weighted selective descriptor score/perfect score. E) Weighted selective descriptor score/chance score. -4.08 2 X 10-' 18.5 <0.005 -3.36 4 X 10-4 11.8 <0.01 -3.80 7 X 10-5 19.6 <0.001 -4.63 2 X 10-6 24.0 <0.001 1 X 10-6 27.6 <0.001 -4.75 TABLE V PRECOGNITIVE REMOTE PERCEPTION 6 X 24 "OCEAN" SERIES SCORE SUMMARY Mean No. 1st (2nd) No. Ranks Methoda Rank Ranks Below Mean Z P, Chance expectation 12.5 0.25 (0.25) 3 A 4.5 1.0 (1. 1) 6 -2.83 0.002 B 4.2 2.0 (0.0) 6 -2.95 0.002 C 5.0 1.0 (1.0) 6 -2.65 0,004 D 5.2 1.0 (0.5) 6 -2.59 0.005 E 5.3 1.0 (1.0) 6 -2.54 0.006 aAs described in text: A) Number of correct descriptors/total number of descriptors. B) Weighted full descriptor score/perfect score. Q Weighted full descriptor scorelnumber of descriptors. D) Weighted selective descriptor score/perfect score. E) Weighted selective'descriptor scorelchance score. ranked better than the chance mean, and a simple X test with its associated probability. The method is not restricted to square arrays, i.e,, to equal numbers of targets and perceptions in one-to-one correspon- dence, Table V displays the results of a recent "Ocean" series in which six perceptions were ranked against 24 targets in the pool from which the six actual targets were drawn. (The per- cipient for this series was at the time sailing alone across the North Atlantic Ocean.) Because of its particularly severe protocol, we also include as illustration 'the results of a "European" series judged by this method. In this series, the agent was traveling in eastern Europe, and on five successive days., at 3:00 P.M. local time, utilized whatever location he happened to occupy as the target. The perceptions of these targets were recorded by a single percipient in northern Wisconsin at approximately 8:30 A.M. local time of the preceding day, i.e., each perception was roughly 24 h precognitive. (Fig. 22 depicts a target and percep- tion from this series') Table V1 displays the analytically judged results of this series, and compares them with the results of previous human judging. Although the data base is small, the consistency of yield is striking. To obviate the possibility that this method of analysis may somehow process even random inputs to apparently significant scores, artificial target data matrices and artificial perception data matrices have been constructed from the output of REG, and the computational schemes applied to various co binations of these with each other and with true data. The pattern of results conforms to chance expectations. An alternative form: of control is provided by application of f_ method to discernibly unsuccessful test series, which yie: appropriately insignificant results. The project outlined in the last few paragraphs is de-scril-I in greater detail, with many other experimental examp' :, both successful and unsuccessful, in reference C 179 ]. We h., e since developed the capability of employing ternary rathei than binary responses to the descriptors, in order to con-_y more shaded information about the aspects queried. A gi r feature may thereby be specified as a) present and dominant b) present but .secondary; or c) absent. Or alternatively, th4 feature may be 'described as a) definitely present; b), ambigu__E or unspecified; .or c) definitely absent. While these approat - clearly provide: more specific target and perception data, th, scoring thereoftecomes more complex, especially in the def-m, i tion of certain of the normalization denominators. We have also'been exploring a modification of this-analy-_E judging procedure which would bypass the ranking steps altc gether and move directly to compute individual statistical sc7c for each transcript. The key to this variation is the defirti and utilization. of a generalized target pool composed of ovt Approved For-Re4ease 2003104/18: CIA-RDP96~00789ROO310003OO(Yl~;4 FEBRUARY t9&2 156 Approved For Release 2003/04/19P~LIA~PbA?YobuAffibdTib-do'~'odo'i-4 TABLE VI PRECOGNITIVE REMOTE PERCEPTION 5 X 5 -EUROPEAN" SERIES SCORE SUMMARY Mean No. Ist (2nd) No. Ranks Method3 Rank Ranks Below Mean Z P, Chance expectation 3.0 1.0 (1.0) 3 A 1.2 4.3 (0.3) 5 -2.85 0.002 B 1.4 4.0 (0.0) 5 -2.53 0.006 C 1.4 3.0 (2.0) 5 -2.53 0.005 D 1.8 4.0 (0.0) 4 -1.90 0.029 E 1.6 3.0 (1.0) 5 -2.21 0.013 Seven human judges (avg.) 1.9 2.1 (1.6) 4.6 aAs described in text: A) Number of correct descriptors/ total number of descriptors. B) Weighted full descriptor score/perfect score. C) Weighted full descriptor score/number of descriptors. D) Weighted selective descriptor score/perfect score. E) Weighted selective descriptor score/chance score. 200 targets, local, national, and international, assembled for a broad range of earlier and ongoing experimental series, from which correspondingly generalized a priori descriptor proba- bilities may be calculated. Evaluation of the transcript scores on the basis of these gendralized probabilities, rather than those calculated for the specific series target pools, has been found to alter only slightly the relative ranks of the perceptions determined by any of the five methods used above and provides the desired common basis for evaluating the individual statis- tical significance of those scores. To pursue this, the program next assembles a set of empirical chance distribution functions, one for each scoring method, by concatenating all mismatched perception scores assigned by that method, i.e., all offidiagonal elements of that perception- target matrix. With reference to Gaussian fits to these empiri- cal chance distributions, the proper perception-target scores can then be assigned z-values and corresponding probabilities azainst "chance." As one example of this approach, Fig. 23 shows the empirical distribution of scores compounded from some 1400 mismatched targets and transcripts by method E, on which are superimposed the scores achieved by the 24 proper target perceptions of the "Chicago" series. The corresponding z-values and probabilities against chance are listed in Table VII. These individual significance values can subsequently be com- pounded into an overall significance level for the entire series by various standard procedures [ 180 ]. The latter result should agree with that derived from the original ranking method, to the order of approximation implicit in each form of this anal- ysis. In this case, the series value is about 10-8, compared to 10-6 for the ranking method. The specific results shown in Tables III-VII represent some of the most successful data we have acquired; many less successful examples could also be displayed. To summarize our total ex- perience with over two hundred remote perception efforts, all performed with volunteer percipients claiming no special abili- ties, we might note that the data tend to fall into one of four categories, in roughly comparable quantity: I ) the target is accurately represented in detail and composi- tion; 2) particular features of the target are accurately perceived, but the context is incorrect; 3) the ambience of the target is perceived, but the details are inaccurate; MISMATC@+ED o o 4- TARGETS @,@,_,@FITTCD GAUSSIAN o MEAN - 1.0016 U @3 o -0.1163 Z 3- S.D. LU o Z) , 'o a LU 2- MOPER - TARGETS 0.6 0.8 1.0 1.2 1@4 NORMALIZED SCORES Fig. 23. Remote perception empirical chance distribution (menod E). 4) the perception seems unrelated to the target in :-.)ntext or detail. A@ Survey of the pertinent literature indicates a compara@ie pat- tern of yield across the experience of others involved ir. 3irri2ar experiments. To date there has been little progress in correlating tht dep-ee of success of such efforts with the prevailing expe!7.mental conditions or with the personality traits or attitudm, of the participants. A certain body of lore has compounded f:,om the testimony of the more successful perciplents, such as the desir- ability of a, personal 'rapport between the agent and pe,,:ipient, the value of a lighthearted attitude, the importance of !xclud- ing any associative or constructive logic, etc., but muc_- of this is still too vague and inconsistent to provide any bash, for ex- perimental refinement or theoretical modeling. At :@resent, the only fair statement would be that empirical evi&nce for this class of phenomena continues to accumulate, lat Aith frustrating irregularity and little basic comprehensio-- Not- withstanding, the present and potential applications of the Process in a variety of arenas, combined with the rtiati-vely simple and inexpensive nature of the experiment, ktep this type of study active. THEORETICAL CONCEPTS No remotely satisfactory physical theory of psychic,,rfienom- ena yet exists. Indeed, next to the evasiveness of tht effects under controlled experimentation, the second greatesi frustra- Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 JAHN: PSYCHIC PHENOMENA: ENGINEERING PERSPECTIVE Approved For Release 2003/04/18TAq~i-~PP96-00789ROO3100030001-4 PRECOGNITIVE REMOTE PERCEPTION INDIVIDUAL STATISTICAL SCORtS (METHOD E) 24 X 24 LICH ICAGO" SERIFS Perception No. Z-Score PZ 1 1.210 .0.113 2 0.089 9.465 3 2.737 0.003 4 1.700 0.045 5 -0.553 P.709 6 1.533 0.063 7 -0.315 .0.623 8 -0.315 .0.623 9 0.451 .0.326 10 -0.197 0.578 11 0.191 @0.4 24 12 2.037 Q.021 13 2.948 0.002 14 2.244 0.012 15 2.661 0.004 16 2.478 0.007 17 2.180 P.017 18 1.904 0.028 19 0.083 0,467 20 -0.152 0.560 21 2.679 b.004 22 0.701 0.241 23 1.057 0.145 24 0.113 0.455 tion in the study of psychic processes has been the absence of viable theoretical models with which to begin the traditional dialogue between theory and experiment on whichall scientific progress eventually depends. This may, of course, be indicative of an illegitimacy of the phenomena themselves, or at least of an evanescence that fundamentally precludes any analytical representation. On the other hand, it may be an indication that modem physical theory, elaborate and sophisticated as it is, has not yet evolved to a stage where it can properly acknowl- edge and deal with the role of consciousness in the physical world [ 1811, and that this should be one of its new frontiers. Before pouncing toward either extreme, it may be worthwhile to.play through some of the more canonical attempts that have been -maile to deal with this domain, both from a formalistic and philosophical point of view. Efforts toward theoretical explication of psychic phenomena over the past century have proceeded from various levels of presumption as to the fundamentality of the effects observed. Some have insisted that the effects are totally illusory, i.e., arti- facts of.poor experimentation or data processing, or that they are the chance results of random processes. Others have assigned the effects to known physical and physiological mechanisms associated with, but not deliberately precipitated by the partic- ipants, e.g., electromagnetic radiation from brain circuitry or intercardial potentials, or heat transfer, vibration, or aerochem- ical changes in the experimental environment caused by human presence or exertion. More ambitious efforts have contended that no such conservative models can suffice, and that it will be necessary to identify new forms of energy or information transfer to retain the established physical formalisms, or pos- sibly even necessary to expand the physical laws themselves, as was required in the generalization from classical mechanics to quantum mechanics or to special.and general relativity, with the present forms becoming subsumed under more compre- hensive statements. Still others have concluded that the scien- tific paradigm in general Is inadequite and that basic revision in the representation of, the process of conscious observati:n of physical events will be!!. required. Beyond the uncertain@-ty as to the fundamentality of tli-2 model required, these modeling efforts have labored under t'--t confusion of whether the purported phenomena are most b2.5-@ cally psychological, physiological, physical, or some inextiica1L,'-_ combination thereof, and hence which class of concepts shou: J-1 dominate the model and .which should be secondary. VirtuaLy all permutations have been explored to some degree, uncti labels of "psych o-physi ological, " "biophysical," "psych:- physical," etc., but none' of these can claim more than suggt@5- tive analogies or philosophical stimulation.. This author @s unqualified to assess any models based in the psychological or physiological domains, other than to note that those must frequently discussed tend to acknowledge the role of randcm processes, information oirdering, and statistical, rather thin directly causal mechanisms [1821-[187), all of which have their counterparts in several of the more physically oriented modes which have been proposed. Confining ourselves to su:h physical theories, the history of credible analytic effort is cco- veniently short and, in my view, more instructive in its phil-)- sophical than in its functional characteristics. i Electromagnetic and Geo 'physical Models For perhaps naive reasons, the earliest physical models tendtd to presume wavelike propagation of psychic effects, usually in the electromagnetic mo@es [ 188]. Doubtless the concurrent emergence of radio technology with its revolutionary wireless capabilities influenced the concepts and nomenclature of these versions, and frequent reference to psychic "transmitten-," "receivers," 11 antennas .... .. tuning," and "static" are found in them. Upton Sinclair's b :ook, Mental Radio, mentioned earlitT, is one example of such an analogy [ 152). More modem models of this genre, appearing predominandy in the Russian literature t431, [501, [1001, [1891, [1901, have focused on very low frequency bands, of the order of 10 Hz, Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 158 Approved For Release 2003/04/18 PROCEEDINGS OF THE IEEE, VOL. 70, NO. 2, FEBRUARY 1982 : CIA-RDID96-0078914003100030001-4 PK effects are reported to'be achieved just at th6 first effort or immediately after the effort is terminated. Such "beginner's luck'? evidence might be construed to indicate that psychic processes are invalidated b@ prolonged and careful examina- tion; alternatively, it may be a hint that they are inherently "inductive," in the electromagnetic sense, Le, that they are in- trinsically unsteady pjienomena wherein the.time derivatives influence the magnitudes of the effects. Yet another characteristic that psychic effects share with certain electromagnetic processes is their tendency to be in- direct, tangential, or peripheral: direct effort on one PK task fails, but a secondary effect is noted on another component or device; central elements of a remote perception target are ignored by the percipient, but minor or peripheral aspects are identified with precision; the pattern of physical disturbance in a poltergeist event is reported to be vortical rather than radial, all of which call to mind cross-product and vector curl effects in the electromagnetic domain. The point in suggesting such analogies is not to endorse direct physical correspondence between electromagnetic and psychic prociesses, but rather to speculate whether the human mind may tend to perceive and assess phenomena in the two domains in certain similar fundamental ways. characteristic of various physiological frequencies, especially the brain wave spectra, Some variations of these have invoked modulation of the earth's magnetic field or of the electrostatic fields of the atmosphere. Wave models involving other thaft clectroirriagnetic environmental media have also been proposed, such as infrasonic atmospheric waves, geoseismic waves, and barometric fluctuations, possibly stimulated by contemporary attention to the inexplicable homing capabilities of birds, fish, and animals, and the hive or swarm behavior of certain insects [1011,(1021,(1321. As suggested by the pre-occupation with screen-rooms in the early days of the electromagnetic concept, and by more recent long-distance remote perception experiments, some of the ob- vious validation/disqualification tests examining attenuation, diffraction, interference, and polarization effects have indeed been attempted, but the very large dimensions involved for these wavelengths leave the studies less than conclusive. Some con- tempo),ary workers retain support for this category of model, claiming that the necessary electromagnetic signal levels re- quired for influence on the brain circuitry are so small that no coarse-grained tests can properly discriminate against them. More Problematic, in my own view, are the absence of any demonstrated velocity of propagation of psychic effects and the large body of precognition evidence which, if accepted, cannot be accommodated by any reasonable advanced wave characteristics. However, setting aside for the moment the quantitative func- tional difficulties with the electromagnetic approach to psychic pherioniena, certain broader philosophical analogies in the conceptual and experimental aspects of the two topics may be worth rioting. In electromagnetism, beyond the bald empiricism of the definition of the fields themselves as representations of "action-at-a-distance," there are many features which to a naive or primitive observer would, and indeed historically did, appear as "paranormal," or at least anomalous: the noncolinear aspects of Ampere's and Biot-Savart's laws and of the Faraday effect; induction effects and switching transients; the Maxwell displacement current; the propagation of waves in a vacuum, with a specified finite velocity-each of these in some sense digressed from contemporary "normal" experience, was con- ceptu,dly difficult in its time, and required certain leaps of empiricism to get on with the formalism. To the extent that we just now hold a similarly naive and primitive view of psychic phenomena, it may be necessary to tolerate similarly empirical representations until a more com- prehensive model can knit itself into a more fundamental representation. For this purpose, there may be some merit in looking to just such electromagnetic effects for analogies. As one example, the pervasive "decline effect" in p§ychic experi- mentation, wherein the performance on psychic tasks is widely reported to be highest immediately after initiation, to decrease over protracted testing periods, and then to improve again just before termination, is somewhat reminiscent of the induced signatures of certain electromagnetic processes. The decline effect has been commonly ascribed to a psychological boredom or reduction in the emotional intensity of the operator per- forming the task; just possibly it may be a more fundamental characteristic of the phenomenological domain. In a similar vein, many psychic effects are reported to be in- trinsically transient, e.g., the "fleeting impression"; the "sudden vision'%the "unexpected effect." One of the favored techniques of some remote perception percipients is to "sweep through" their image of the target repeatedly until it is clarified. Many Entropy and the Random Process A second, more recently opened class of psychic model addresses the intera@tion of consciousness with natural random processes [ 951, t 1871, [ 1911. A common aspect of the estab- lished physical formalisms of kinetic theory, thermodynamics, statistical mechanics, and information theory is the role of randomicity as the reference plane for information and energy exchange. By whatever representation, the second law of ther- modynamics, expressing the tendency of isolated physical systems to drive irreversibly toward configurations of mini- mum order and information co 'ntent, stands starkly asymmetric in the time coordinate, thereb'y' raising profound philosophical issues in virtually every domain from biophysics to cosmology.`- Some of the most controlled and replicable experiments in PK, such as the REG studies outlined earlier, could be construed to challenge the second law, or at least to suggest modifications of the concept of an isolated physical system. Namely, i1nder the circumstances of those experiments, human consciousness could be postulated to be inserting order, albeit to a small degree, into a random physical process., This possibility can be extended conceptually to the anoma- lous acquisition of information in remote perception experi- ments, to psychic healing, and to animal and plant PK, but to my knowledge, no attempts at formulation have yet specified any details of the ordering capability, e.g., its physical or phys- iological source, its propagation modes, or its manner of inter- action. Pending these, one can again only proceed with high empiricism to attempt to represent the observed correlations in a useful fashion, a strategy which has sustained many other observational fields in their primitive phases. Notwithstanding, this class of psychic model poses a profound question: The long-accepted essence of consciousness is its ability to extract information from its environment; may the reverse also be possible? May consciousness have the ability to insert infor- mation into its environment? One extreme variation of this model escalates the question even one step further, to ask whether it is possibly an indige- nous property of extremely elaborate and complex systems that they may embody inherent functional consciousness of Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 JAHN: PSYCHIC PHENOMENA: ENGINEERING PERSPECTIVE 159 their own -AoprowdoFvocruReibgs*2oft3le4l~i8)atpI"DBPMQ7A~g#MlqgglWA-4,e have come to perceive it, interlockingg of their systems, they derive not only abilities but rather as frequency and amplitude information, and that to learn, to reproduce themselves, to adapt to their environ- the human consciousness essentially performs "Fourier trans- ment, but also to exert an entropy reversing form of "Con- forms" on this to oTderthat information into the more familiar sciousness" on themselves [ 1921, [193 1. form. In the sens6 that the space and time coordinates are thereby downgraded from the fundamental coordinates of Hyperspace Representations experience to useful ordering parameters, one could interpret A few attempts have been made to represent paranormal such models to imply that consciousness may, by this mecha- effects by re-casting the basic laws of physics in more than the nism, access any portion of space and time to acquire informa- four coordinates of normal human experience and applying the tion, and then interpret it in some characteristic form. consequent new terms to the representation of paranormal The physicist David Bohm, in conjunction with the psychol- effects [ 1041, 1105), 11941. For example, one such approach ogist Karl Pribram, has elaborated such concepts to a consider- adds an imaginary component to each of the spatial and tempo- able degree of generality, proposing a so-called "implicate ral coordinates, i.e., invokes complex space and time, somewhat order" or "enfolded order" of fundamental reality from which in the spirit of ac circuit theory or exponential wave mechanics the more familiar "jexplicate order," i.e., the commonly mani- 1195). The imaginary components and their "cross-talk" with fest perceptions are assembled in accordance with the prevail- the traditional real ones thereby permit representation of anom- ing circumstances of their observation [ 2011, [ 2021 . alous effects within the framework of established physical laws. To date, no convincing fundamental definition of these new Quantum Mechanical Models dimensions has been offered, thereby reducing such models to Probably the most exercised category of contemporary model an heuristic or empirical level. In this sense, the approach bears attempts to apply the concz;pts and formalisms of quanrum me- some similarity to the "hidden variable" aspect of quantum chanics to represent some of the paranormal effects presented mechanics, mentioned below. in the psychic domain. Of all the forms of physical analysis, One may speculate that if such models are to address the in. quantum mechanics. invokes the largest array of empirical pos- teraction of consciouness with physical process, the requisite tulates that are at variance with conventional rationality, and new coordinates or components must relate to, or in some way yields in their implementation a corresponding array oll results define or localize, the processes of consciousness. In other which contradict common impressions of reality. The quan- words, to the normal "hard" coordinates, whereby events are tization process it4lf, which limits measurable properties to ; conventionally specified in the physical world, it may be neces- discrete values; the. representation of particulate systems by sary to add certain "soft" coordinates to specify the conscious wave functions; the: role of observation in collapsing the wave processes by which those events are perceived and possibly in- functions to a single state vector; the uncertainty principle; fluenced. Clearly, thecoordinate frame in which one chooses to the exclusion principle; the indistinguishability principle; and observe and represent any physical process is intimately linked most drastic of all the commitment to totally probabilistic to the perception of that process, Two common examples mechanical behayiors-aD, in some sense concede. a degree of would be the perceptions of kinematic and dynamic effects in paradox in human perception of physical processes. The famil- a rotating frame of reference, such as a merry-go-round or an iar conundrums of "Schr6dinger's Cat," "Wigner's Friend," or.,, orbiting spacecraft, or the even more bizarre appearances of the "Einstein-Pod olsky- Rosen Paradox," all suggest that the physical processes in rapidly accelerating frames, A la general laws of quantum mechanics are not so much statements of relativity. fundamental physical reality, as of our ability to acquire in- To compound this interdependence of perception and refer- formation about that reality. Quantum mechanics, in other enC67- frame by including "consciousness coordinates" in the words, does not so.much describe the state of a physical sys- specification of the latter is an awesome proposition, but an tern as it describes our knowledge of the state of that system. intriguing one. And the consequences need not be restricted it is somewhat in. this spirit that a number of authors have to the mechanical behavior of physical systems, but may also aspired to model psyphic process in quantum mechanical terms. influence their perceived substance. The noted British astron- Some have attemptod to invoke the so-called "hidden*' or un- omer Sir Arthur Eddington, some fifty years ago, presaged this used variables of the formalism to involve conscious process concept most boldly and heretically [ 1961, [ 1971, [ 203 ] - more explicitly in the behavior of physical systems [1071, .The whole of those laws of nature ... have their origin, not in [2041-.[2081. Others have endeavored to draw analogies any special mechanisms of nature, but in the workings of the between the synaptic processes in the brain and quantum mind.. mechanical "tunnelibg" [209]. All through the physical world runs that unknown content Given the primitive state of the phenomenological data base, which must surely be the stuff of our consciousness.... much of this effo"rt may be prematurely elaborate and complex. Where science has progressed the farthest, the mind has regained Our own approach to quantum mechanical modeling has been from nature that which the mind has put into nature.... far more superficial and generic, attempting only to explore We may look forward with undiminished enthusiasm to learning possible analogies between the paradoxical consequences of the in the coming years what lies, in the atomic nucleus-even though formalism, and the paranormal evidence of certain psychic ex- we suspect that it is hidden there by ourselves.... The stuff of the world in mind-stuff. periments. Again with the indulgence of the reader, we might sketch a bit of this: argument, for the purpose of illustrating Transform Models this class of approac@ with the example closest to hand. Another rather extreme approa@h proposed recently has come One conventional interpretation of the application of quan- to be referred to as the "holographic" or "transform" model tum mechanical forrihalism to the observable behavior of phys- [198]-[200). Essentially .@rhat is suggested here, as I under- ical systems is to ass@pciate appropriate mathematical operators stand it, is that the information of the universe is arrayed, not with a corresponding measurement process. When applied to Approved Foe Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 160 Approved For Release 2003/04/18PReiA~Ofy~g-lo6lbW6o~ibddsddol,-4EBRUARY 1982 the prevailing wave function of the system, these operators call out the observable values of the property in question as eigen- values ofan equation of the form Moi = Miol where IvIr is the measurement operator, ini are the observable values of the ineasured property, and Oi the corresponding eigenstates of the wave function. Our approach is to general- ize this representation to include conscious systems as well as conventional physical ones, and to allow the measurement operator concept to include specification of psychological as well as physical properties. Thus, we denote a particular individual consciousness by a "state function" Oi, and represent a situation to which it is exposed 'by an operator S. Application of the situation opera- tor to the consciousness wave function then yields the possible psychological responses, si, as eigenvalues: S Oi = si @i- We then invoke certain aspects of quantum mechanical in- teraction theory to develop the capacity for "paranormal" be- havior of both physical and conscious system' s. For example, in the traditional theory of the covalent chemical bond between two hydrogen atoms, one constructs from the separate atomic functions, @' and 0 b, using arguments of symmetry and in- distinguishability, a composite molecular wave function 0,b, which yields expectation values for the molecular energy levels substantially different from simple linear supbrposition of the atomic energy eigenvalues, i.e.. P P, = P P, IF P = P1 i.e., the percipient would perceive nothing about the target that was not accessible to his normal perceptual modes, and the agent would react to the target under no influence from the percipient. However, if the percipient and agent are strongly enough interacting to require a new "molecular" wave function, Opa, "pa.ranormal" terms will appear in their response patterns: pa = P I I Pip + Pq +AD Fa + Ap@ I I where in Appla we may accommodate the anomalous acquisi- tion of information about the target, and in Ap'iP the com- monly reported experience of the agent of having his attention attracted to specific details he would "normally" have ignored. Application of this formalism to a PK experiment proceeds in a similar fashion. Here we might represent the experimental operator (person) by @0, and the experimental device by 0 d. Again the experimental protocol is represented by a mathe- matical operator, K. In the absence of major interaction, the device behaves "normally": K@@ = kiq @ql and the operator has the "normal" psychological experiences: K 09 = k 9 9 But if the operator and the device are in some state of reso- n 0 do, ance, each behaves somewhaz differently, i.e., @d k1o = kq + k? + Ar.1yo + Aki . e qb = eg + d + Aeqb where 4 and eJP are the energy eigenvalues of the atomic systems, and Leqb embodies an "exchange energy" term which is classically inexplicable, but devolves formally from the postulate that the electrons are indistinguishable in the bonded configuration. Stated more bluntly, surrendering in- formation about the identity of the atomic electrons in the molecular configuration leads directly to a significant and ob- servable component of the binding energy, thus posing an equivalence between information and energy far more stark than that: implicit in the second law of thermodynamics. Using similar formalism, we may represent the state function of two interacting individuals. or of an interacting individual and physical system, by a composite state function 0,b whose behavior characteristics also differ significantly from those of the separated systems, i.e., sqb = S, + Sb + ASqb + ASb, where i I' denotes the "normal" response of the first individual to the prevailing situation, and sib that of the second individual or of the physical system, and As"b and AsPa denote modifica- tions of those behaviors arising becaus 'e the two systems are strongly interacting during the observed situation. As a specific example, to apply this approach to remote per- ception experiments we could denote the percipient by OP, the agent by 04, and the experimental protocol by the mathemati- ca. I operator P. In the absence of interaction between the per- cipient and agent, each would have certain "normal" reactions to the experimental situation, p P, p F, derived from the eigen- value relations: The anomalous modification in :he behavior of the system Akdo is termed PK; Ak?d accomodates any paranormal. psy- chological reactions of the operato.-. Development ', of further illustrations of this general method here would be inappropriate and would require more detailed specification of the nature of the consciousness wave functions, their functional form, their proper "soft" coordinates, and the interpretation of their quantum numbers. Some of this has been attempted, along with vario-us other applications to the psychic domain, and is available in a reference [210). The point of exposition here is largely a philosophical one: namely, the "paranormal" effects emerge as a Gonsequence of the com- parison of the behavior of an interacting system with that of its separated components. Quantum mechanics may have quite another analytical precedent to contribute to the representation of psychic phenomenology. Clearly, much of the psychic research data will continue to be acquired and processed in statistical form, using established statistical methodss. Yet, all statistical models ultimately trace, back to certain fundamental probability rules for the elemental systems involved. For the statistical models to be viable, these probability rules must a) exist; b) be known; and c) be analytically tractable. At present, virtually all pro- cessing of psychic research data presumes the applicability of classical statistics, yet the basic probability rules for the ele- mental processes are, in point of fact, unknown [2111, [2121. ' It may fortuitously be the case that much of psychic process can be adequately treated as marginal deviation from classical chance behavior. In some cases, however, it may be necessary to invoke categorically different statistics, tracing back to funda- mentally different probability rules, to deal with the effects. The quantum mechanical precedents of relevance, of course, are Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 JAHN: PSYCHIC PHENOMENA: ENGINEERING PERSPECTIVE 161 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 the two systems of quantum statistics, i.e., the Fermi-Dirac acteristics underlying: those formalisms may be relevant to .and Bose-Einstein systems, which are based upon the phase- ultimate representation of such processes. Specifically, they space population rules for half-integer and integer spin particles suggest that the following rather general hypotheses may he as imposed by the Pauli principle, i.e., the wave function sym- worthy of more detailed examination: metry requirements. For most common physical systems, these 1) The phenomena :play be inherently statistical, rather than quantum statistics conveniently degenerate into the classical directly causal, and we: may be observing them "on the margin." form, but for certain special situations, e.g., the specific heats That is, the observed phenomena may represent marginal of metals and certain radiation properties, their full applica- changes from 110MIal. 'behavior on a very grand scale and with tion is essential. fluctuation times which tax human observational capability. It By analogy, one could postulate that physical reality, as per- also may be necessary' to deal with more complex statistical ceived and influenced by human consciousness, actually plays mechanics, appropriate to more involved basic probabilitN by more elaborate probability rules than commonly attributed rules, to represen! theimost drastic effects. to it, and hence strictly requires a more complex statistical 2) Just as human consciousness has the ability to extract in- mechanics. Thus, in this view, the processes commonly regarded formation from an external system, e.g., by observing it, tha-. as "normal" would be those for which the "classical" approxi- consciousness may also have the ability to project information mation to this more complex system is adequate; small-scale into it, e.g., by ordering random processes. psychic effects would then comprise those displays of minor 3) Quantum mechanics may be more than a system of phys-;- deviations of the complex statistics from the "classical" limit; cal mechanics; it rnay be a more fundamental representation of the more drastic phenomena -poltergeists, levitations, metal human consciousness and perception processes, and the empir.- bending, etc.-would presumably become explicable only in cal pillars of this formalism, such as the uncertainty principle. terms of the full statistical formulation. the exclusion principle, the indistinguishability principle, and the wave/particle dualities may be as much laws of consciow- Holistic Models ness as laws of physics. Yet more extreme in conceptual difficulty is a body of con- 4) Psychic processes may be inherently holistic, and thus tention that psychic processes are inseparably holistic, and that the ultimate model may need to integrate both the scientif.@ : no model rooted in any sector of established science can ad- pcts in order to identify the sources c" and the aesthetic asp equately represent them [2131. In particular, the suggestion the phenomena. That is, psychic processes may be manifesta. is offered that psychic processes are manifestations of the inter- tions of the intersection of the analytical, scientific wor'@J diffusion of the analytical, scientific world with the creative with the creative, aesthetic world, and thus, to represent then, aesthetic world, and thus to represent them effectively it will effectively, it may be necessary to integrate both perspectives be necessary to combine the philosophical perspectives and without sacrificing the integrity of either. techniques of both domains. To resort to metaphor for illus- Clearly, any of these intuitions will have to be developed L-i tration once again, one could ponder such analogies from the far more philosophical and analytical detail before a trenchant common physical world as the interface regions between the theoretical model can emerge, but at this primitive stage it is sea and land, where the diffuse patterns of the ocean wave probably stimulating to consider a few such radical possibilities, structure meet the solid promontories and sloping beaches of along with more prosaic explications. Changes of tlifs magn;_ the coast to produce the crashing breakers and hissing foam of tude in representation of human perceptual reality inevitably, seashore phenomena, or the overrunning of a warm, moist and properly, would'be attended by much philosophical 7-.- atmosphere by a cold-front of drier, cooler air to initiate the calcitrance and agony, but the broader personal and collective sfrikifig* electrical and acoustical phenomena of the summer insights that could dorive from legitimate efforts to bridge the thunderstorm. analytical /aesthetic interface could be of at least corresponding By whatever analogy it may be illustrated, theoretical repre- benefit. sentation of the interpenetration of causal physical mechanics with creative conscious process must be a formidable under- THE NEGATIVE SIDE taking, yet not totally without precedent or allied effort. For Contemporary criticism of psychic research and re ection of i examples, interest is just now growing in the humanistic psy- the phenomena it purports to demonstrate tend to focus on a chology community in the analytical study of human creativity number of specific objections each of which has some degree [ 214 ] - [ 216 ), and, on the other side of the in@terface, a few , of validity and merits some thought in any balanced assess- physicists are beginning to muse openly- about the role of ment of the topic [1181 1 -[224). The most [1781 [2221 aesthetics in submiclear and cosmological physical behavior , , commonly cited concerns include: [2171-[2201. In yet a different arena, certain futurists are now examining the interplay of aesthetic and functional hu- I ) demonstrable fraud; man needs and values in the evolution of social and political 2) naivete of technique, including inadequate controls, structure[221]. To be sure, none of these has produced much faulty equipment,,,sensory cuing of participants, other in the way of analytical formalism, but the peculiarity, magni- experimenter biases, selective treatment of data, im- tude, and significance of the interpenetration effects are being proper statistical methods, and general experimental acknowledged. and theoretical incompetence; Any summary assessment of the status of physical models of 3) little improvement in comprehension over many years psychic phenomena should properly begin with reiteration of of study; the opening statement: none of the.approaches outlined above 4) absence of adequate theoretical models; has yielded anything approaching a functional theoretical basis. 5) suppression of negative results; Yet, theensemble of empbrical experimental experience seems 6) poor experimental replicability; to suggest that certain of the con6eptual and perceptual char- 7) elusiveness of effects under close scrutiny; Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 162 PROCEEDINGS OF THE IEEE, VOL. 1.0, NO. -2, FEBRUARY 1982 Approved For Release 2003/04/18: CIA-RE).P9 -007 8) sensitivity of results to participants, attitudes, and labora- ratio of positive lo negaMiRe 9911WP@PAQ?Aof experiment, tory ambience; thereby. contributing to a broader sense of the grana statistics 9) tendency for many results to be only marginally signif- of the phenomena. In addition, documentation of the specific icant compared to chance expectation; conditions prevailing in unsuccessful experiments may be help- 10) inconsistency with prevailing "scientific world view;" ful in excluding irrelevant parameters from further considera- 11) contradictory to personal psychology, philosophy, tion, and in identifying and reducing counterproductive influ- theology, or "common sense." ences. Perhaps most pointedly, however, the body of negative and equivocal data emphasize that psychic phenomena, if real, Obviously this list runs a gamut from rather technical and pro- cedural objections, through phenomenological inconsistencies, to rather categorical and subjective rejections, and only a few of these issues can be constructively addressed here. Unfortunately, but undeniably, one or the other of the first two judgments may legitimately be applied to a large body of the propounded results. By its nature, the field is immensely vulnerable to fraudulent exploitation and naive gullibility, and such have indeed occurred to a distressing degree. It is also true thzt: the topic has attracted a disproportionate share of less than fully competent researchers, and that it presents extra- ordinary pitfalls for even the most disciplined scholars. Yet despite the substantial validity of these claims, and the suspicion they inevitably cast on all other results, it does not seem that they should predicate categorical rejection of the entire field. Rather, the vulnerable cases should be patiently ferreted out using obvious scholarly criteria, and only those efforts surviving such scrutiny used for scientific insight and judgment. The lack of definitive progress toward comprehension of the phenomena and the absence of viable theoretical models have already been acknowledged in the foregoing text, and although these assessments should perhaps be qualified by the relatively minute integrated investment of resources made in this field in comparison to many of the more favored areas of science, they nonetheless constitute legitimate concerns about the ultimate tractability of the field. It should perhaps also be noted that, despite the prolonged effort, it is only very recently that more sensitive and powerful experimental equipment and data pro- cessing techniques have been brought to bear, and equally re- cently that more sophisticated physical formalisms have been invoked. and that these have had a much briefer opportunity to render the phenomena into comprehensible terms. The subsequent five objections, 5)-9), are more specific and substantive, and merit examination from two orthogonal Points of view. Namely, to what extent do such characteristics indeed invalidate the results, and conversely, to what extent might they illuminate the basic nature of the phenomena? With respect to supprezsion of unfavorable results, there has undoubtedly been some tendency in this field, as in most others, to advance POsi- tive or definitive findings more enthusiastically than negative or equivocal ones. Indeed, this paper has been guilty of the same bias. In an effort to provide concise representation, of the style and substance of psychic research and of the nature of the effects it can produce, we have tended to invoke as illus- trations some of the more successful and familiar pieces of work without balancing the presentation with comparable ex- amples of the negative or equivocal results that are regularly acquired in these efforts. To the credit of the psychic research community, it has offi- cially encouraged thorough and objective reporting of negative data, and much of these indeed appear regularly in its established journals [2251- [ 228), with a number of consequent benefits. First, beyond adding credibility to the body of positive results, such data compound to provide some quantitative index of the Approved For Release 2003/04/18 are highly irregular and sensitive to intangible influences well beyond current scientific control, and, if their study is to be pursued, this caveat must be accepted ab initio, at least for the present. A similar interpretation also applies to the irreplicability complaint, and to the three following it. Without question, the dominant experimental frustration in this field is the in- ability to replicate on demand previously observed paranormal effects, not only at other laboratories, but even in the original facility, using the original participants, under apparently identi- cal experimental circumstances [1721, [173), [2291-[2311. This ubiquitous characteristic has precipitated major philosoph- ical excursions which are well beyond our capacity to review here. Only briefly, four possible categorical interpretations have been advocated: a) The phenomena are illusory. b) The phenomena are rare and bizarre chance occurrences, beyond any hope of regularization. c) The phenomena are precipitated, at least in part, by psy- chological and/or physiological factors which are presently beyond experimental control, but which if fully compre- hended would conform to established scientific paradigm. d) The phenomena are inherently statistical, and possibly quanturn mechanical, on a macroscopic scale, thus mani- festing themselves with finite but tractional probability on any given occasion. -, The latter pair of options are not necessarily mutually exclu- sive, particularly if one takes a rather generous doubly statistical point of view, namely that the human population embodies a range of capability for engenderingsuch effects and that beyond that, any individual may display a variable range of personal ca- pacity, depending on a variety of environmental, physiological, and psychological factors prevailing at the time. . The evasiveness of the phenomena under carefully controlled and observed study may be the most damning criticism of all, or it may also constitute a valid and illuminating phenomeno- logical characteristic. The tendency of a given preliminary or anecdotal effect to disappear or diminish when the experiment is tightened, or when it is displayed to a skeptical jury of ob- servers, obviously casts major doubt on the scientific integrity of the process. Yet it also brings to mind at least two other pro- ces@es which, while superficially dissimilar, may not be totally irrelevant, namely artistic creativity, and quantum mechanical measurement as limited by the uncertainty principle. In the former, there should be little quaff el that the creative processes of artistic, musical, or literary composition, or of lofty philo- sophical thought in general, are not usually facilitated by rigid constraints or by the presence of a body of unsympathetic observers. The importance of favorable ambience and mood for such efforts is intuitively and demonstrably clear, and little creative achievement is likely to occur in overly sterile or hostile environments, a truth Richard Wagner vividly conveyed to his own critics by his portrayal of the fate of young Walther at- CIA-RDP96-00789ROO3100030001-4 JAHN: PSYCHIC PHENOMENA: ENGINEERING PERSPECTIVE tempting his aNapEemadfiN R*a%%f2QO&041lcBnbO)k-RE)P96-00789ROO3.100030001-4 Meistersingers. Virtually every creative artist preserves some form of retreat or sanctuary, and even the most rigorous of scientists will concede the role of unstructured mental imagery in enhancing their technical insights. The analogy of the quantum mechanical measurement process is somewhat more strained, in that it requires generalization of the concept to the macroscopic level of information or energy exchan.ge between two persons or between a person and a physical systern. The point will not be developed here, other than to note that if there is any validity to the application of quantum mechanical logic to this class and scale of intellectual/ intuitive process, as discussed earlier and in the references, some form of "uncertainty principle" could predicate a limit to the precision with which psychic effects could be observed. More specifically, if the "hard" and "soft" coordinates of repre- sentation are canonically conjugate, some form of Aq Ap - h rule may apply, so that attempts at excessive precision in speci- fication of a psychic effect could dissipate its cause, and vice- versa [2 101. The final two reservations regarding inconsistency of the phe- nomena with established scientific and personal views, while. constituting powerful professional and personal discriminators and properly predicating great caution and discipline in ventur- ing into any anomalous field, also cannot be allowed total veto authority if new domains of conceptual experience are ever to who -@,be challenged. In responding to a critic of an earlier paper stood on these points I wrote, perhaps too floridly, Authoritarianism such as this encourages established knowledge to sit smugly on its duff and categorically reject all new evidence that does not support or fill in its contemporary "world-view compatability criterion" -whatever that is. Worst of a it stifles the most precious attribute of human consciousness, the yearning for ever new, ever higher wisdom that has driven the mind and spirit of man to evolve upward, rather than merely to replicate [2321. More persuasive to this issue, however, would be a simple histor- ical count of the number of leaps of scientific insight, from Aristotle to this day, which would have been, and in most Oases )yere, for a time, rejected on the basis of these criteria. @@iouilk, it has often been those giants of science who with soaring insight and courage of conviction violated such tenets to lead their fields to new plateaus of understanding, who also, in a later day, led the recalcitrance of the establishment against comparably sacrilegious visions of their successors, while still endorsing in general terms the importance of visionary thought. Galileo, early champion of scientific methodology and revolu- tionary concepts in terrestrial and celestial mechanics against vicious dogmatic opposition, rejected Kepler's elliptical orbits as "occult fantasy"; Thomas Young, whose brilliant interference experiments finally established the wave character of light, contended with Fresnel's theoretical formulations of the same processes; Ernst Mach disputed relativity and atomic theory; Rutherford, who showed the world the nuclear atom, dismissed any practical significance for nuclear energy; Lavoisier and Ostwald disputed atomic theories of chemistry; D'Alembert opposed probability theory; Edison discounted alternating current; Lindberg despaired of Goddard's rocketry; and Albert Einstein retained an enduring uneasiness about quantum theory despite his many contributions to its evolution [ 23 3 ] - [ 23 5 ]. Incidentally, the same Albert. Einstein who would invoke the establishment criteria against Upton Sinclair's clairvoyance ... the results of t@e telepathic experiments carefulb- and plainly set forth in tl-ds book stand surely far beyond those which a nature investigator holds to be. think-able [ 15 2 ] . could in quite another: tone testify eloquently to the impor- tance of the aesthetic dimension in creative science: The most beautiful'and most profound emotion we can ex- perience is the sensatOn of the mystical. It is the sower of' all true science. He to whom this emotion is a stranger, who can no longer wonder and sta: d rapt in awe, is as good as dead. To know that what is impenetrable to us really exists, manifesting iis-elf as the highest wisdom ald the most radiant beauty which our dull faculties can comprehend only in their most primitive forms- this knowledge, this feeLing is at the center of true religiousness. The cosmic religious experience is the strongest and noblest mainspring of scientific research [ 236 J. Individually and collectively, these critical challenge; to ps-,- chic research raise valid concerns which merit deliberate atten- tion and predicate grea1T caution, and can also help to fflurninate some of the subtle phenomenological features. However, the', can perform these furctions well only if they themselves are informed, reasoned, an; d fair. Regrettably, from my reading o-1 the critical literature, this has not invariably been the case, and instances of naivet6, selective representation of data and pro- tocol, and excessive generalization also appear therein. The role of the criti, in psychic research is a most essenti-1.1 one, perhaps more so than in any field ol scholarship @__L broached. When the riticism is based in fact and experience and is objective and fair, it can instill healthy disciplille in the study of this or any :other difficult field and ensure that the fundamental requisite$ of scientific methodology, e.2., dispas- sionate rigor, humilit@. in the face of observations, limitation on extrapolation Of Te '@sults, and openness of mind Nk@l preval in the search. But it is equally essential that the process of criticism play by thesp: same rules. If it violates any of them, if it lapses.into categorical rejection, guilt by association, or sloppy logic, it can become as suspect as the object of '-s complaint, and thus fail in its proper role [ 23 7 IMPLICAT.IONS Despite their compounded length, the foregoing historical outline, survey of contemporary activity, and selected ex- amples of ongoing research and theoretical efforts are still far from adequate to con :vey the full essence of this complex and contradictory field. Fortunately an extensive body of refer- ence literature existsi including a number of comprehensive general volumes, whereby an interested reader may flesh out this sketch and exterid it to many aspects not broached here [61, (441, [451, [$61, [571, [921, [1301, [2381-[242). Hopefully, such more' thorough study would tend to confirm the general impressions of the status and prospects of the field conveyed above. To restate these in summary, it appears that once the illegitimate p@search and invalid criticism have been set aside, the remaining accumulated evidence of psychic phenom- ena comprises an array of experimental observations, obtained under reasonable proiocols in a variety of scholarly disciplines, which compound to a philosophical dilemma. On the one hand, effects inexplicable in terms of established scientific theory, yet having numerous common characteristics, are frequently and widely observed; on the other hand, these effects have so far proven qualitatively and quantitatively irreplicable, in the strict scientific sense, and appear to be sensitive to a variety of psycho- logical and environm0tal factors that are difficult to specify, Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 164 Approved For Release 2003/04/ft?ldff-'Kb'F%-'ff7W6ftOv6~16bo~bdoi~4111" 1982 let alone control. Under these circumstances, critical experi- mentation has been tedious and frustrating at best, and theoret- ical modeling still searches for vocabulary and concepts, well short of any useful formalisms. Given these difficulties, what then are the motivations, if any, to proceed'? As for most speculative topics, three potential generic benefits could be considered: a) acquisition of fundamental knowledge, b) practical applications, c) humanistic benefits. In this particular field, basic knowledge might accrue in two ways-the attainment of new scientific information in the usual sense, and the broadening of scientific methodology to deal more effectively with irregular phenomena of this type. In other words, study of this topic not only might provide certain phe- nomenological answers, but also might serve to broaden the context in which science can formulate its questions. New mechanisms for transfer of information or energy might be identified, or broader understanding of those properties, and how they are perceived and measured might emerge. The latter half of the opportunity clearly is a majorchallenge to science, but hardl, a new one. William James posed it rather bluntly Y some eighty-five years ago: The spirit and principles of science are mere affairs of method; there is nothing in them that need hinder science from dealing successfully with a world in which personal forces are the starting point of new effects. The only form of thing that we concretely have is our own personal life. The only completed category of our thinking, our professors of philosophy tell us, is the category of personality, every other category being one of the abstract elements of that. And this systematic denial on science's part of personality as a condition of events, this rigorous belief that in its own essential and innermost nature our world is a strictly im- personal world, may, conceivably, as the whirligig of time goes round, prove to be the very defect that our descendants will be most surprised at in our boasted science, the omission that to their eyes will most tend to make it look perspectiveless and short [24]. The potential applications of psychic process are best con- sidered with conservatism and restraint, especially given the tendency of certain elements of the public media and private exploiters to extrapolate the possibilities far beyond any dem- onstrated accomplishments. Clearly, the process of remote perception described 'earlier, along with other forms of clair- voyance, could hold some potential interest for intelligence agencies, law enforcement units, and any other activity relying on surveillance, as well as for archaeological searches, natural resource prospecting, and the like, and such operations have indeed engaged in empirical efforts to evaluate the efficacy of such strategies in their particular domains. From a strictly engineering standpoint, however, the potential efficiency and precision of such tactics are unclear, given their apparent ten- dency to trade more effectively in impressionistic generalities than in analytical detail. Low-level PK effects, such as the REG distortions indicated above, could have more pervasive implications for high technol- ogy. If, -for example, the basic functions of microelectronic elements could be even slightly disturbed by intentional or in- advertent: intervention of human consciousness, it would seem important to obtain some assessment of the potential magni- tude of such effects, and of the factors favoring or inhibiting such interference, before much more elaborate integrated cir- cuit arra, s, graphic display systems, and other sensitive man/ @y machine interfaces are committed to delicate or critidal opera- tions. To focus our -assessment of such possibilities, we are now examining PK disturbance of the memory function of a single microelectronic chip [ 243 ] . If tfi@ indications that psychologi- cal and environmental factors bear on the precipitation of such effects aie sustained, it may be necessary to expand considera- tion of such parameters beyond the usual scope of human factors engineering, especially in situations involving high psy- chological stress. The potential humanistic benefits of better comprehension of psychic phenomena could be addressed on either a personal or social level, but to do so in any.detail would far exceed the purview:of this article. Again, extensive references on various facets of the issue abound [2441-[252). Ultimately, most of these philosophical excursions arrive at the same monumental question, namely whether convincing demonstration of the c.apability of human consciouness to influence its reality to a significant degree would substantially alter individual and col- lective perception of the human state, its value system, and its behavior pattern, and thereby facilitate its evolution to a higher life form. Such projections have been offered from a variety of perspectives. En ginee r/ futurist Willis Harman fore- casts an "inner experience" paradigm: Just as conventional science depends upon a prior cons6n@is' on how knowledge of the sense-perceived world shall be publicly tested.and validated, so the complementary paradigm will have to include consensus on how knowledge relating to the world of inner experience shall be publicly tested and validated. Its essential characteristic would be that consciouness and its contents axe primary data, rather than being secondary and de- rivative as in the conventional paradigm. Where the conventional paradigm involves redu ctionistic models the complementary para- digm would add holistic models; where the first employs deter- ministic (or stochastic) explanations of events the second would add teleological, purpose-recognizing explanations; where the first is little involved with matters of values and meaning, the second find§I'.these of central concern; where the first is dominated by technology-focused values of prediction and control, the second would tend to value understanding relating to human well-being, development and evolution [ 221 Biologist/immunologist Jonas Salk phrases it more in terms of a resonance of human volition with natural processes: Man has come to the threshold of a state of consciousness, re- garding his nature and his relationship to the Cosmos, in terms that reflect 'reality.' By using the processes of Nature as metaphor, to describe the forces by which it operates upon and within Man, we come as close to describing 'reality' as we can within the limits of our comprehension. Men will be very uneven in their capacity for such understanding, which, naturally, differs for different ages and cultures, and develops and changes in the course of time. For these! reasons it will always be necessary to use metaphor and myth to provide 'comprehensible' guides to living. In this way, Man's imagination and intellect play vital roles in his survival and evolution [249]. And philosopher/paleontologist Teilhard de Chardin states his hope in terms of a collective consciousness of the human race: Thus we find ourselves in the presence, in actual possession, of the super-organism we have been seeking, of whose existence we were intuitively aware. The collective mankind which the sociol- ogists needed for the furtherance of their speculations and formu- lations now appears scientifically defined, manifesting itself in its proper time and place, like an object entirely new and yet awaited in the sky of life. It remains for us to observe the world by the light it sheds, which throws into astonishing relief the great en- semble of everyday phenomena with which we have always lived, without. perceiving their reality, their immediacy or their vastness [253). Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 On a somewhat less lofty, but possibly f previously stored data by any . *t P 49 a recent hWvvWd@FPVRW*0*q12 M1110 UUM104 t XMilMattiM :ten parametric indices, e.g., sampling rate, t & tives Science and Technology Committee, in a section encourag- manual/automatic:, volitional /instructed, operator characteris- ing serious assessment of further research on "the physics of ties, etc. For any: such parametric explorations to be indicative, consciousness," stated that recent experiments: the pertinent data, subset must be adequately large to display suggest that there exists an 'inter connectiven ess' of the human any systematic dev:iation from the baseline distribution beyond mind with other minds and with matter; . . . that the human mind the statistical noise background. This we have found to re- may be able to obtain the information independent of geography quire at least several thousand trials; the implication for the and time total data base is 6ear. and later concluded that: Thk vp.T-v 1,9rue, ilAtn cannhilitv innvibblv nredinateq q ,zonhk- a general recognition of the degree of interconnectiveness of minds could have far-reaching social and political implications for this nation and the world [254]. The details and tones of these visions clearly are matters of indiyidual heritage, experience, and intuition, but the messages share a common theme: the next stage inhuman evolution may involve expansion and interconnection of human consciousness, features clearly central to the psychic concept. Beyond the difficulty of the phenomena and potential bene- fits of the knowledge, a More Prosaic factor to be considered in contemplating further psychic research is the requisite cost of the effort. To date, such experimentation has been extra- ordinarily inexpensive by usual scientific standards. The primi- tive level of comprehension, the. lack of organized interest within the established scientific communities, and the Bohemian status of many of the investigators have predicated projects of very low budgets compounding to a total annual national investment of the order of one million dollars or less [2551. Clearly, if more incisive progress is to be attempted, some increase in the sophistication and interaction of the principal programs must be funded, but it is most unlikely, and for the time probably undesirable, that this will consume any magnitude of resources comparable %ith, or distractive from, the better established re- search domains. A comparable statement could be made with respect to commitments of the requisite scholarly personnel. If on the basis of such cost/risk/benefit considerations a modest ongoing program of research seems justified, it re- mains to consider the selection of topics, experimental styles, and evaluation criteria which would optimize the effort at this stage. In addition to the obvious desiderata that the specific experiments be clearly posed, conceptually simple, lend them- s@lv6s- logistically to rigorous, tightly controlled protocols, and 00' ;on the more tractable and potentially applicable effects, u f three more specific recommendations could be offered, First, given the irregularity of the phenomena, their possible dependence upon a broad range of physical, psychological, physiological, and environmental parameters, and their ten- dency to display effects as marginal deviations from some "normal" distributions, some premium should probably be placed on the capacity of the experiments for large data base accumulation and processing. While less prolific studies may continue to provide interesting anecdotal effects and suggest procedures for more detailed programs, it is unlikely that much correlation of those effects with pertinent prevailing parameters can be achieved without large quantities of data. More specifi- cally, the favored experiments should 1) deal with processes found to have relatively high intrinsic yields; 2) employ equip- ment and protocols which permit data acquisition at rapid rates; and 3) have access to computational equipment and software which allow storage of large data arrays and procming of that data in many selective cross-c6ncatenations. As an example, the latest refinement of the REG experiment described earlier allows acquisition, storage, and primary processing of several "hundred, experimental trials (-10' - 106 bits) per hour, and tication in equiptlient and software which, beyond the initial and operating expenses, may introduce some undesirable effects of complicating the phenomenological processes and clouding the experimental aMbiences. Specifically, one needs to consider whether any observed effects, in either the PK or ESP categores, still trace unequivocally to the primary physical processes, or whether there now might arise confounding interactions v-ith elements of the data collec tion and processing equipment or techniques. Related to this concern are possible uncertainties of the participants in defining and focusing on the primary tasks. Reservations of this sort lead directly to a second general Tee- ommendation for effective psychic experimentation. Namelyif the phenomena derive to any significant degree from conscious or subconscious processes of the human mind, it is important that such not be inhibited or excessively complicated by :he design and operati 0@; n of the experiments. For this and numer- ous other reasons, it is probably essential in planning and L-n- plementing the experimental programs to include the insighats, interpretations, and intuitions of the human operators, espe- cially those who have demonstrated some success in the gen- eration of the phenomena. It is quite possible that the differ- ence between a sterile experiment and an effective one of equal rigor lies as much in' the impressidnistic aspects of its ambience and feedback as in the elegance of its instrumentation, and the former need to be viell-tuned to the participants who are asked to function as components of the experimen0l system. On oc- casion, there seems io have been some tendency in this field to treat the experimen :@al participqnts in rather perfunctory fash- ion, discounting any insights they might offer on the tasks at hand. If one subscribes at all to the concept of the phenomena emerging from some interpenetration of analytic and intuitive processes as suggested by the holistic models, there would seem no better place to combine perceptions and insights from these two domains than in the design, opmtion, and interpretation of the experiments addressed to illumination of the interface. Finally, it seems most evident that given the intrinsic trans- disciplinary nature of the business, research on this topic in any established sector should become much more communica- tive and interactive across traditional scholarly boundaries if it is to have any hope of rendering the phenomena into compre- hensible and serviceable terms. This cross-talk cannot be limited to naturally contiguous fields, a la the usual exchange between physicist and engineer, or between psychologist and sociologist, As its lehgthy heritage WUStTates, in this domain the interests and insights of the theologian, philosopher, statis- tician, technologist, hard scientist, and creative holistic thinker are all potentially valid, and need to be melded in scholarly symbiosis and common respect. No insular approach is likely to prevail. This requisite has implica 'tions for the staffing of particular 'projects, for the institutional environment in which they oper- ate, and for the professional societies and publications which choose to attend to this topic. Individual laboratory personnel e 2003104118: CIA-RDP966-00789R003100030001-4 Re leas ... ... . ... 166 PROCEEDINGS OF THE IEEE, VOL. 70, NO. -2, FEBRUARY 1982 Approved For Release 2003104118 CIA-RDP96-00789ROO3100030001-4 groups !should comprise a broader range of experience and in- sight than the conventional hierarchy of technical specialists, and their cognizance of other contemporary work should be broader. The institution housing that laboratory needs to dis- play considerable tolerance and support for the unusual tone and special requirements of the research and not force its con- formity to established scientific subdivisions and research styles. Likewism-, the professional community at large cannot at this time profitably ask for total adherence to its own reductionistic supers tructure, but can only inquire dispassionately regarding the respective implications of this conglomerate field for the traditional areas of endeavor. In this last regard, I should like to express my personal re- spect for this particular Society, and for this particular Journal, for the openness and generosity of spirit with which they have solicited and presented the results of legitimate scholarly effort in this difficult field. Their attitude could well stand as a model for other institutions and organs in dealing with this topic or with any other present or future projective area of human inquiry. ACKNOWLEDGMENT na author would like to express his indebtedness to his re- search colleague, Ms. B. J. Dunne, who has provided immense logistical help in the preparation of this manuscript and who, as laboratory manager of the Princeton Engineering Anomalies Research program, has played a primary role in the generation and interpretation of much of the data reported herein. Dr. R. D. Nelson also contributed heavily to the experimental pro- gram and to the refinement of the manuscript. Prof. W. H. Surber designed much of the REG circuitry and software, and wrote the original programs for the remote perception analytical judg:ng procedure. The program has also benefitted immensely from the interests and elforts of several present and past students and staff, most notably C. K. Curry (Dunham), E. G. Jahn, T. A. Curtis, and 1. A. Cook. REFERENCES [11 12) 131 [41 [51 161 [71 [81 (91 1101 1111 1121 (131 (141, M. Ebon, "Ahistoryof parapsychology," in Psychic Exploration, E. D. Mitchell et al., J. White, Ed. * New York; Putnam, 1974. A Gould, The Founders of Psychical Research. New York: Schocken, 1968. J. B. Rhine and associates, Parapsychology from Duke roFRNM. Durham, NC: Parapsychology Press, 1965. H. Carrington, TheStoryofPsychic Science (Psychical Research). New York: Ives Washburn, 193 1. R. C. LeClair, Ed., The Letters of William James and Thbodore Flournoy. (Foreword by Gardner Murphy.) Madison, WI: Uni@ versity of Wisconsin Press, 1966. B. B. Wolman, Ed., Handbook of Parapsychology. New York: Van Nostrand Reinhold, 1977. D. S. Rogo, Parapsychology: A Century of Inquiry. New York: Taplinger, 1975. F. Hartmann, Paracelsus: Life and Prophecies. Blauvelt, NY: Rudolf Steiner Publications, 1973, pp. 103-131. M. Bell, "Francis Bacon: Pioneer in parapsychology," Int. J. Parapsychol., VI, no. 2, pp. 199-208, Spring, 1964. J. Glanvill, Saducismus Triumphatus, Introduction by C. 0. Par- sons, Facsimiles edit. Gainesville, FL: Scholars' Facsimiles and Reprints, 1966. (Originally published in 1689.) R. Haynes, Philosopher King-The Humanist Pope Benedict XI V. London, England: Weidenfeld & Nicholson, 1970. C. Richet, "La suggestion mentale et le calcul des probabilit@s," Rev. Philosoph., 18, pp. 608-674, 1884. C. Richet, "Further experiments in hypnotic lucidity or clair- voyance," In Proc. Soc. for Psychical Research, 6, pp. 66- 83, 1889. E. Swedenborg, Divine Love and Wisdom. New York: Sweden- borg Foundation, Inc., 1979. [151 G. Trobridge, Swedenborg, -Lift and Teaching.' New York: Swedenborg Foundation, Inc., 1976. (10] 1. Ross, The President's Wife, Mary Todd Lincoln. New York: Putnam, 1973. . I 'S 1171 F.W. H. Myers, Human Personality and the Survival ofBodi7y Death, (Abridged and edited by S. Smith.) Secaucus, NJ: Uni- versity Books, 1961, (Originally published in 1903, 2 vols.) [181 W. F. Barrett, "Appendix to the report on thou ght-rea ding," in Proc. Soc. for Psychical Research, 1, pp. 47-64, 1882. [191 0. J. Lodge, "An -account of some experiments in thought-trans- . ference, " in Proc. Soc. for Psychical Research, 2, pp. 189-200, 1884. 1201. The American Society for Psychical Research, Inc., 5 W. 73 St., New York, NY 10023. 121) W. James, Human Immortality: Two Supposed Objections to the Doctrine. Boston, MA: Houghton Mifflin, 1898. 221 -, "Report of the Committee on Mediumistic Phenomena," in Proc. American Soc. for Psychical Research, 1, pp. 102-106, 1896. 1231 G. Murphy and R. 0. Ballou, William Jameson Psychical Re- search. New York: Viking Press, 1969. 124) W. James, "Psychical research," In The Will to Believe and Other Essays in Popular Philosophy. New York: Longmans, Green, and Co., 1897. - 1251 -, "Review of human personality and its survival of bodily death by F.W. H. Myers," in Proc. Soc. for Psychical Research, 18, p. 23, 1903. 126] S. Freud, "A note on the unconscious in psycho -analysis," in Proc, Soc. for Psychical Research, 26, pp. 312-318, 1912. [271 -, Studies in Parapsychology. New York; Collier Books, 1963 1281 C. G. Jung, Synchronicity. Princeton, NJ: PrincetorrUhiiersity Press (Bollingen Series #20), 19?3. 129) -, On the Psychology and Pathology ofSo-Called OccultPhe- nomen@. Princeton, NJ: Princeton University Press (Bollingen Series # 20), 1957. (Originally published in 1902, doctoral dis- sertation, University of Basel.) [301 -, Memories, Dreams, Reflections. (Recorded and edited by A. Jaffe, translated by Richard and Clara Winston.) New York: Vintage Books, 1965, p. 137. (31] W. Paul! and C. G. Jung, Eds., The Interpretation ofNature and the Psyche. Princeton, NJ: Princeton University Press (Bollingen Series #51), 1955. 1321 J. Ehrenwald, "Psi, psychotherapy, and psychoanalysis," In Handbook of Parapsychology, B. B. Wolman, Ed. New York: Van Nostrand Reinhold, 1977, pp. 529-540. [331 M. Ullman, "Psychopathology and psi phenomena,"in Handbook of Parapsychology, B. B. Wolman, Ed. NewYork: Van Nostrand Reinhold, 1977, pp. 557-574. (341 J. G. Pratt, 1. B. Rhine, C. E. Stuart, and B. M. Smith, Extra-,, Sensory Perception AfterSixty Years. New York: Holt, 1940. 1351 J. B. Rhine, New World of the Mind. New York: William Mor- row and Co., 1953. [361 J. B. Rhine, Ed., Progress in Parapsychology. Durham, NC: Para- . psych ol ogyPress, 1971. [37] L. E. Rhine, Mind Over Matter: Psychokinesis. New York: Mac- millan, 1970. 1381 J. B. Rhine, "Psi and psychology. Conflict and solution," J. Para- psychol., 32, pp. 101-128, 1968. 1341 Eur.'J. Parapsychol., Int. J. Parapsychol. (defunct), J. Parapsy- chol., J. Amer. Soc. Psychical Research, J. Soc. forPsychicalRe- search, Parapsychol. Rev., Proc. Soc. f.or Psychical Research. [401 US Universities /Colleges: Chicago; Colorado; Columbia; CUNY; Delaware; Denver; Drexel; Duke; Finch; Harvard; Illinois; John F. Kennedy; Kent State; Michigan; Missouri; Mundelein; Newark College of Engineering; New School for Social Research; NYU; North Carolina; Oakland; Pittsburgh; Princeton; St. John's; St. Joseph's; South West Minnesota; Syracuse; UC/Berkeley; UC/ Davis; UC/Irvine; UC/LA; UC/Santa Barbara; Virginia; Wesleyan; West Florida; West Georgia; Wisconsin; Yale. Foreign: Amster- dam; Andhra; Argentina, Benares; Bonn; Cambridge; Chile-, Copenhagen; Edinburgh; Freiburg; Ghana; Halifax; Iceland; Jung Inst.; Leningrad; Leyden; London; Loyola; Lund; McGill; Mu- nich; Mysore; New South Wales; Oxford; Paris; Peking; Rhodes; South Africa; Tasmania; Tel Aviv; Utkal; Utrecht; Warsaw; Waterloo; Zurich. [41] List available upon request. [421 Research Institutions: Amer. Soc. for Psychical Research, New York; Ballistic Research Labs., Aberdeen, MD; Foundation for Research on the Nature of Man, Durham, NC; Inst. for Noetic Sciences, San Francisco, CA; Menninger Foundation, Topeka, KS; Midwest Psi Research Inst., Chicago, IL; Mind Science Foun- dation, San Antonio, TX; Parapsychology Foundation, NY; Psychical Research Foundation, Durham, NC; Psychophysical Research Labs, Princeton, NJ; Science Unlimited Research Foundation, San Antonio, TX; SRI International, Menlo Park, Approved For Release 2003/04/18 CIA-RDP96-00789ROO3100030001-4 JAHN: PSYCHIC PHENOMENA: ENGINEERING PERSPECTIVE 16- Approved For Release 2003/04/18 : CIA-RDP96-00789RO03 100030001-4 CA; Society for Psychical Research, London; Forschungsinstit 1721 J. Eccles, "The human person in its two-way relationship to the fOr Psychotronik, West Berlin; Matsui Hospital, Kasahara, Japan; brain," Invited address to the Para psychological As*ociation Con- Inst. of Pharmacology, Polish Academy of Sciences, Krakow; vention, State Uni@.ersity of Utrecht, The Netherlands, Aug. :0. Shanghai Municipal Human Exception Functions Research Soci- 1976. Research in'Parapsychology 1976. Metuchen, NJ: Scare- etY. Corporations: Airesearch Company of California; Boeing crow Press, 1977, op. 251-262. Scientific Research Laboratories. 731 J. Beloff, "ESP: The search for a physiological index." J. A trier. 1431 1. G. Pratt, "Soviet research in parapsychology," in Handbook Soo, Psychical Research, 47, Sept. 1974. ofParapsychology, B. B. Wolman, Ed. New York: Van Nostrand (741 1. Ehrenwald, "Psi phenomena and brain research." in Handboo@ Reinhold, 1977, pp. 883-903. of Pa rapsych o fogy, B. B. Wolman, Ed. New York: Van No!stranJ J 44 E. K. Naumov and L. V. Vilenskaya, Bibliographies on Parapsy- Reinhold, 1977, pp . 716-729. chology (Psych oen erge tics) and Related Subjects. Arlington, 1751 D. Dean, "The plethysmograph as an indicator of ESP," J. S,v. VA: Joint Publications Research Service, 1972. (Originally Ps)-chicalResearcii.,41,pp.351-353, 1962. published in Moscow, 1971.) J 761 J. C. Carpenter, "Intrasubject and subject-agent effects in ESP 14 5 E. K. Naumov and L. V. Vilenskaya, Bibliography ofParapsy- experiments," in Handbook ofParapsychology, B, B. Wolman, chology (Psychorronies, Psychoenerge tics, Psych obiophysics) Ed. New York: Van Nostrand Reinhold, 1977, pp. 202-272. and Related Problems. (Literature published in the USSR), 1771 R. A. White, "The influence of experimenter motivation, atti- 1979 (in manuscript). tudes, and methods of handling subjects on psi test resu!ts," m 461 E. C. Wortz, A. J. Bauer, R. F. Blackwelder, J. W. Eerkins, and Handbook of Parapsychology, B. B. Wolman, Ed. New York: A. J. Sour, "An investigation of Soviet psychical research," in Van Nostrand Reinhold, 1977, pp. 273-301. Proc. Special Session: The State of the Art in Psychic Research; 1781 C. Honorton, M. Ramsey, and C. Cabibbo, "Experimenter effects Electro 7 7 Professional Program. New York: Apr. 19-21, 1977. in extrasensory perception," J. Amer. Soo. Psychical Re3earck. 1471 A. J. Lewis, A Report. Document EW-76-01 1, Airesearch Manu- 69, pp. 135-149, 1975. facturing Co. of California, January 1976. (A supplement to 791 G. Murphy, "Psychical research and human personality," Pro-.. Wortz et al., #46 above.) Soo. Psychical Res 'garch, 48, pp. 1-15, 1949. 148) M. Johnson (Chairman) et al., "Symposium: Current directions [801 J. B. Rhine, "Conditions favoring success in psi tests," J. Par.:- in European parapsychology," in Research in Parapsychology psychol., 12, pp. 58-75, 1948. 1978. Metuchen, NJ: Scarecrow Press, 1979, pp. 1-10. 811 G. R. Schmeidler.and M. Maher, "The non-verbal communicl- 491 S. Ostrander and L. Schroeder, Psychic Discoveries Behind the tions of psi-conducive and psi-inhibitory experimenters," 1. Iron Curtain. Englewood Cliffs, NJ@ Prentice-Hall, 1970. Amer. Soo. Psychical Research, 75, pp. 241-257. 1981. 501 L. L. Vasiliev, Experiments in Distant Influence (Edited with 82 ] W. Crookes, Crookes and the Spirit World, collected by R. G. Introduction by A. Gregory). New York: Dutton, 1976. Medhurst. New York: Taplinger, 1972. 1511 T. Moss, "Psychic research in the Soviet Union," in Psychic Ex- (831 B. J. T. Dobbs, The Foundations of Newton's - Ichemy. Car2- ploration, E. D. Mitchell et al., J. White, Ed. New York: Putnam, bridge, MA: Cambridge University Press, 1975, pp. 7-8. 1974,pp.469-486. 1841 C. Honorton, "Has science developed the competence to confront 1521 R. A. McConnell, "Parapsychology in the USSR," J. Parapsychol., claims of the para6ormal? " inResearch in Paraps:, chology. 19 39, pp. 129-134, 1975. Metuchen, NJ: SCiTecrow Press, 1976, pp. 199-: 23. t531 V. P. Zinchenko, A. N. Leont'yev, B. F. Lomov, and A. R. Luria, (851 B. Musicante, Experiments in Psychokinesis: Crtique and Pr "Parapsychology: Fiction or reality?" Voprosy Filosoflif (Ques- posed Study, Senior Thesis submitted to the Department .f tions of Philosophy), 9, pp. 128-136, 1973. Psychology, Princeton University, May 1979. [541 Zhang Feng, "New advances made in study ofexceptional human [861 R. Jungerman and, J. Jungerman, "Computer controlled random functions," Shanghai Ziran Zazki (Nature Journal) in Chinese, number generator PK tests," in Research in Parapsychology 197. no. 8, p. 606, Aug. 1980. Metuchen NJ: Scarecrow Press, 1978, pp. 157-162. [551 titudes and personality traits in experimental ESP J. Palmer, "At [871 H. Schmidt, "A PK test vAth electronic equipment," J. Parapsy- ' research," in Handbook of Paraspsychology, B. B. Wolman, chol., 34, pp. 175-181, 1970. Ed. New York: Van Nostrand Reinhold, 1977, pp. 175-201. - [881 H. Schmidt and L. Pantas, "Psi tests with internally different [56] S. Krippner, Ed., Advances in Parapsychological Research: 11, machines," J. Parapsychol., 36, pp. 222-232, 19-2. Extrasensory Perception. New York: Plenum Press, 1978. [891 H. Schmidt, "PK effect on pro-recorded targets." J. Amer. Sc-.-. [571 K. R. Rao, Experimental Parapsychology: A Review and Inter- Psychical Research, 70, pp. 267-291, 1976. , pretation. Springfield, IL: Charles C. Thomas, 1966. [901 -, "PK tests with pre-recorded and pre-ins@ected seed num- [581 G. Schmeidler, "The psychic personality," in Psychic Explora- bers," in Researqh in Parapsychology 1980. Metuchen, NJ: tion, C. Mitchell et al., J. White, Ed. New York: Putnam, 19 74. Scarecrow Press, 1981, pp. 47-50. [591 H. J. Eysenck, "Personality and extrasensory perception,"J. Soo. 1911 E. C. May, B. S. Humphrey, and G. S. Hubbard, "Electronic sys- Psychical Research, 44, pp. 55-71, 1967. tem perturbation techniques," SRI Int., Final Rep., Sept. 30, 01 16 -G, R. Schmeidler and R. A. McConnell, ESP and Personality 1980. , , . Westport, CT: Greenwood Press, 1973 (originally [921 S. Krippner, Ed. Advances in Parapsychological Research, Vol. 1, published 1958). Psychokinesis. New York: Plenum Press, 19 7 7. [611 J. Palmer, "Scoring in ESP tests as a function of belief in ESP. [93) B. Dunne, R. G. Jahn, and R. Nelson, "An REG experimentAith Part 1. The sheep-goat effect,"J. Amer. Soo. Psychical Research, large data-base caliability," in Research in Parapsycholog-y 19-A I. 65, pp. 373-408, 1971. Metuchen, NJ: Scarecrow Press (in press). [621 A Parker, States of Mind: ESP and A Itered States of Conscious- 194] H. Puthoff and P. Targ, Mind-Reach. New York: Delacorte, ness. New York- Taplinger, 1975. 1977. 101 R. L. Van de Castle, "Sleep and dreams," in Handbook ofPara- 1951 -, "Physics, entropy, and psychokinesis," in Quantum Physics psychology, B. B. Wolman, Ed. New York: Van Nostrand Rein- and Parapsychology, L. Oteri, Ed. New York: Part psychological "hold, 1977, pp. 473-499. Foundation, 1975. (641 M. Ullman, S. Krippner, and A. Vaughan, Dream Telepathy. 1961 W. W. Eidson, D. L. Faust, K. Getsla, J. McClay, and B. Conoway, New York. Macmillan, 1973. "An investigation. of possible anomalous interaction between a ' [651 1. Strauch, "Dreams and psi in the laboratory," in Psi Favorable human being and a polarized laser beam," preliminary rep. on States of Consciousness, R. Cavanna, Ed. New York: Parapsy- investigations Juno-Sept. 1977, Drexel University, Physics Dep. chology Foundation, 1970, pp. 46-54. (unpublished.) . (661 C. Honorton, "Psi and internal attention states," in Handbook [971 R. D. Nelson, B. T.. Dunne, and R. G. Jahn, "Psychokinesis studies of Parapsychology, B. B. Wol man, Ed. New York @ Van Nostrand with a Fabry-Per4t interferometer, "in Research In Parapsychol. Reinhold, 1977, pp. 435-472. ogy 1981. Metuqhen, NJ: Scarecrow Press (in press). 1671 W..Braud, "Psi conducive conditions: Explorations and interpre- [981 J. B. Hasted, "Physical aspects of paranormal metal bending," tations," in Psi and States of Awareness: Proc. Int. Conf. held J. Soc. Psychical kesearch, 49, pp. 583-607, 1977. In Paris, France, Aug. 24-26, 1977, B. Shapin and L. Coly, Eds. 1991 -, The Metal-Berders. London, Boston and Henley: Routledze New York: Parapsychology Foundation, 1978. and Kegan Paul, 1981. .[681 E. Dingwall, Ed., Abnormal Hypnotic Phenomena (4 Vols.). 11001 M. RyzI, "A model for parapsychological communication," J. New York: Barnes and Noble, 1968. Parapsychol., 30, pp. 18-31, 1966. 1691 C. Honorton and S. Kripprier, "Hypnosis and ESP performance: 11011 M. A- Persinger, ';'Geophysical models for parapsychological ex- A review of the experimental literature," in Surveys In Parapsy- periences, " Psychbenergetic Syst., 1, pp. 63-74, 1975. chology, R. White, Ed. Metuchen, NJ: Scarecrow Press, 1976, [1021 G. E. Wolstenholme and E. C. P. Millar, Eds., Extrasensory Per- pp. 227-270. , . k - . ception: A CIBA:Foundation Symposium. New York. Citadel, 1701 C. Sargent, "Hypnosis as a psi conducive state: A controlled repli- 1966. cation study," J. Parapsychol., 4, pp. 257-275, 1978. [1031 0. C. de Beauregard, "Time symmetry and interpretation ofquln- (71 1 C.' Tart, '!Drug induced states -of consciousness," in Handbook turn mechanics," Foundations Phys,, 6, pp. 539-559, 1976. ofpdrapsychology, B. B. Wolman, Ed. NewYork: Van Nostrand [1041 G. Feinberg, "Possibility of faster-than light particles," Phys. Reinhold, 1977, pp. 500-525. Rev.. 159, p. 108@, 1967. Approved F& Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 168 (1051 (106 ] 1107) [1081 [1091 (1101 fill] [1121 11131 (1141 11151 11161 11171 [1181 [119) [1201 11211 (1221 (1231 11241 11251 1126) 11271 11281 1129) (1301 (1311 (1321 (133 1 11341 1135 1 FffA_W6W6eb6%&E6VOL. 70 NO 2 FEBRUARY 1982 Approved For Release 2003/04/18 8 Ro 310000061L4 W. A. Tiller, "A lattice model of space and its relationship to multidimensional systems," in Proc. A.R.E. Medical Symp., Phoenix, AZ, Jan. 1977. L.Oteri, Ed., Quan rum Physics and Parapsychology: Proceedings of an international Conference held in Geneva, Switzerland, A,ugust26-27,1974. New York: Parapsychological Foundation, 1975. J.H.M. Whiteman, "Quantum theory and parapsychology," J. ,,imer. Soo. for Psychical Research, 67, pp. 341-360, 1973. F. H. Walker, "Consciousness and quantum theory," in Psychic Exploration, E. D. Mitchell el W., J. White, Ed. New York: Nitnam, 1974, pp. 544-568. It. D. Mattuck, "A quantum mechanical theory of the interaction between consciousness and matter," Colloque International "Science et Conscience, " Cordoba, Spain, Oct. 1-5, 1979. 1). Bohm, "The enfolding-unfol ding universe and consciousness," Colloque International "Science et Conscience, "Cordoba, Spain, Oct. 1-5, 1979. 1-1. Margenau, "ESP in the framework of modern science," J. A mer. Soo. Psychical Research, 60, pp. 214-228, 1966. J1. H. M. Whiteman, "Parapsychology and physics," in Handbook ofParapsychology, B. B. Wolman, Ed. New York: Van Nostrand Reinhold, 1977, pp. 730-756. A. Puharich, Ed., The Iceland Papers: Select Papers on Expert- rriental and TheorerfcalResearch on the Physics ofConsciousness. Foreword by B. D. Josephson. Amherst, WI: Essentia Research Amociates, 1979. A. H. Klopf, Mental and Physical Phenomena: Toward a Unified Theory. July 197?. (Unpublished manuscript.) Duke University, Dep. of Electrical Engineering-, Princeton Uni- versity, School of Engineering/Applied Science; Syracuse Uni- versity, School of Computer and Information Science. S. S, Wilks, "Statistical aspects of experiments in telepathy," H.C.L.R. Lieber, Ed. -A lecture by Samuel S. Wilks given at the Galois Institute of Mathematics, Long Island University, Mimeographed, 1938, 18 pp. 11. Diaconis, "Statistical problems in ESP research," Science, 201, pp. 131-136, July 14, 1978. C.E.M. Hansel, ESPand Parapsychology: A Critical R eevaluarion._ IBuffalo, NY: Prometheus Books, 1980. D. S. Burdick and E. F. Kelly, "Statistical Methods in Parapsy- chological Research," in Handbook of Parapsychology, B. B. Wolman, Ed. New York: Van Nostrand Reinhold, pp. 81-130, 1977. 1. A. Greenwood and C. E. Stuart, "Mathematical techniques used in ESP research," J. Parapsychol., 1, pp. 206-225, 1937. W. W. Carington, "Experiments on the paranormal cognition of drawings," J. Parapsychol., 4, pp. 1-134, 1940. L. L. Gatlin, "A new measure of bias in finite sequences with applications to ESP data," J, Amer, Soc, Psychical Research, 1, pp. 29-43, 1979. C. T. Tart, "Randomicity, predictability, and mathematical in- ference strategies in ESP feedback experiments" (Discussion of Dr. Gatlin's paper). J Amer. Soo. Psychical Research, 1, pp. 44- 60, 1979. J. G. Pratt, "Is ESP only a misnomer for response sequences chosen to match inferred target sequences?" (Discussion of Dr. Gatlin's paper.) J. Amer. Soo, Psychical Research, 1, pp. 60-66, 1979. D. S. Rogo, Parapsychology: A Century ofInquiry. New York: Taplinger, 1975, pp. 29-30. H. Bergson, "Presidential address to the Society for Psychical Research (1913)," Proc. Soo. Psychical Research, 27, pp. 157- 175, 1914. -,Mind Energy. New York: Henry Holt, 1920. C. D. Broad, Religion, Philosophy and Psychical Research. New York: Humanities Press, 1969. J. Lud"ig, Ed., Philosophy and Parapsychology. Buffalo, NY: Prometheus Books, 1978. G. R. Schmeidler, Ed., Parapsychology: Its Relation to Physics, Biology, Psychology, andAsychiatry. Metuchen, NJ; Scarecrow Press, 1976. R. L. Van de Castle, "Parapsychology and anthropology," in Handbook of Parapsychology, B. B. Wolman, Ed. New York: Van Nostrand Reinhold, 1977, pp. 667-6a6. R. L. Morris, "Parapsychology, biology, and anpsi," in Hand- book of Parapsychology, B. B. Wolman, Ed. New York; Van Nostrand Reinhold, 1977, pp. 687-715. W. H. Clark, "Parapsychology and religion," in Handbook of Parapsychology, B. B. Wolman, Ed. New York: Van Nostrand Reinhold, 197?, pp. ?69-780. J. M. Backus, "Parapsychology and Literature," in Handbook ofParapsychology, B. B. Wolman, Ed. New York: Van Nostrand Reinhold, 1977, pp. 781- 802. C. Panati, Ed., The Getter Papers: Scientific Observations on the Paranormal Powers of Uri Geller. Boston, MA: Houghton Mif- flin, 1976. [136 ] C. Crussard and 1. Bouvaist, "Study of certain seemingly abnor.- mal deformations and transformations of metals," Memoires Scientiflilues Revue Metallu)yle, Feb. 1978. (English transla- tion H & P de Malgret, Nov. 1978). 11371 J. G. Pratt, ESP Research Today: A Study of Developments in Parapsychology since 1960. Metuchen, NJ: Scarecrow Press, 1973. 113a] J. G. Pratt, "The Pearisburg poltergeist," in Research in Parapsy- chology 1977. Metuchen, NJ: Scarecrow Press, 1978, pp. 174- 182. [1391 J. Krieger, D, DalCorso, and G.'Solfvin, "RSPK in Philadelphia," in Research In Parapsychology 1978. Metuchen, NJ: Scare- crow Press, 19 7 9, pp. 5 9 - 6 0. [1.401 A.R. G. Owen, Can We Explain the Poltergeist? New York: Garrett Helix, 1964, [1411 W. G. Roll, The Poltergeist. Metuchen, NJ: Scarecrow Press, 1976. (1421 W. G. Roll, "Poltergeists," in Handbook of Parapsychol6gy, B. B. Wolman, Ed. New York- Van Nostrand Reinhold, 1977, pp. 382-413. (1431 W. Tiller, "A technical report on some psychoenergetic devices," A.R.R. J., 7, p. 81, 1972. [144] G. Schmeidler, "PK effects upon continuously recorded temper- ature,"J. Amer. Soo. Psychical Research, 67, p. 32S, 1973. 1145 J R. G. Stanford,. "Experimental psychokinesis: A review from diverse perspectives," in Handbook of Parapsychology, B. 13@ Wolman, Ed. New York: Van Nostrand Reinhold, 1977, pp. 324-381. 1146]! H. Schmidt, "Psychokinesis," in Psychic Exploration, E. D. Mitchell er al., J. White, Ed. New York: Putnam, 1974, pp. 179-193. 11471 W. Cox, "The effect of PK on the placement of falling objects," J. Parapsychol., 15, pp. 4 0 - 4 8, 195 1. 1148) -, "The effect of PK on electromechanical systems," J. Para- psychol., 29, pp. 165-175, 1965. (1491 J. Hasted, The Metal Benders. London, Boston and Henley: Routledge and Kegan Paul, 1981, p. 131. [1501 M. G. Shafer, "PK metal bending in a semi-formal small group," . in Research in Parapsychology 1960. Metuchen, NJ: Scarecrow . Press, 1981, pp. 32-35. (151.1 C. Honorton and L. Tremmel, "Psitrek: A preliminary effort toward development of psi-conducive computer soft-ware," Research in Parapsychology 1979. Metuchen, NJ: Scarecrow Press, 1980, pp. 159-160. (1521 U. Sinclair, Mental Radio (Rev. 2nd printing, Introduction by W. McDougall, Preface by A. Einstein, with a report by W. F. Prince). Springfield, IL: Charles C. Thomas, 1962. (153] H. E. Puthoff and R. Targ, "A perceptual channel for informa- tion transfer over kilometer distances: Historical perspective and, . recent research," in Proc. IEEE, vol. 64, pp. 329-354, 1976. 11541 R. Targ and H. E. Puthoff, "Information, transmission under conditions of sensory shielding," Nature, 252, pp, 602-607, 1974. (1551 C. T. Tart, H. t Pu thoff, and R. Targ, Eds., Mind at Large: IEEE Symposia on the Nature ofExtrasemoryPerception. NewYork: Praeger Special Studies, 1979. (1561 H. E. Puthoff, R. Targ, and E. C. May, "Experimental psi re- search: Implication for physics," in The Role of Consciousness In the Physical World, R. G. Jahn, Ed. Boulder, CO: Westview Press, 1981, pp. 37-86. [157] B. J. Dunne and J. P. Bisahs, "Precognitive remote viewing in the Chicago area," J. Parapsychol., 43, pp. 17-30, 1979. [1581 -, "Multiple subject precognitive remote vie'Aing,"in Research . in Parapsychology 1977. Metuchen, NJ: Scarecrow Press, 1978, pp. 146-151. (1591 -, "Long distance precognitive remote viewing," in Research in Parapsychology 1978. Metuchen, NJ: Scarecrow Press, 1979, pp. 68-70. 11601 J. P. Bisaha and B. J. Dunne, "Long distance and multiple sub- ject precognitive remote viewing of geographical locations," in Mind at Large: IEEE Symposia on the Nature of Extrasensory Perception, C. T. Tart, H. E. Puthoff, and R. Targ, Eds. New York: Praeger Special Studies, 1979, pp, 107-124. 11611 B. J. Dunne, Precognitive Remote Perception. M.A. thesis, Com- mittee on Human Development, University of Chicago, 1979. 11621 B. J. Dunne and J, P. Bisaha, "Precognitive remote perception: A critical overview of the experimental program," in Research in Parapsychology 1979. Metuchen, NJ: Scarecrow Press, 1980, pp. 117-120. [1631 R. G. Jahn, "Psychic process, energy transfer, and things that go bump in the night," Princeton Alumni Weekly, Dec. 4, 1978, S-1-12, (1641 E. A. Rauscher, G. Weissmann, J. Sarfatti, and S.-P. Sirag, "Re- mote perception of natural scenes, shielded against ordinary perception," in Research In Parapsychology 197-5. Metuchen, Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 JAHN: PSYCHIC PHENOMENA: ENGINEERING PERSPECTIVE 169 A -4 Prov9d For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001 " @ P , , NJ: Sc e row ess, pp. 41-45, 1976. 11941 Trecognition-A memory qf things future, in G. Feinberg, [1651 S. Allen, P. Green, K. Rucker, R. Cohen, C. Goolsby, and R. L. Quantum Phy$:ics and Parapsychology, L. Oteri, Ed. New York: Morris, "A remote viewing study using a modified version of the Parapsychology Foundation, 1975, pp. 54-64. SRI procedure," in Research in Parapsychology 19 75. Metuchen, [1951 E. A. Rausch&, "Some physical mod6is potentially applicable NJ: Scarecrow Press, 1976, pp. 46-48. to remote perception," in The Iceland Papers, A. Puharich, Ed, (1661 A. C. Hastings and D. B. Hurt, "A confirmatory remote viewing Amherst, WI: Essentia Research Associates, 1979, pp. 49-83. experiment in a group setting," Proc. IEEE, vol. 64, pp, 1544- [1961 A. S. Eddington, Space, Time and Gravitation: An Outline of 1545, 1976. the General Theory ofRelativity. Cambridge, MA: Cambridge [1671 J. Vallee, A. C. Hastings, and G. Askevold, "Remote viewing University Press, 1920, pp. 198-201. through computer conferencing," Proc. IEEE, voi. 64, p. 1551, 11971 A.S. Eddington, The Nature of the Physical World. Cambridge, 1976. MA: Cambridge University Press, 1928, p. 276. 11681 T. W. Whitson, D. N. Bogart, J. Palmer, and C. T. Tart, "Prelim- (1981 K. Pribram, Languages of the Brain. Englewood Cliffs, NJ: inary experiments in group remote viewing,"Proc. IEEE,vol. 64, Prentice-Hall, 1971. p. 1550, 1976. [199) K. Pribram, "Toward a holonomic theory of perception," in 169) H. Chotas, "Remote viewing in the Durham area," J. Parapsy- Gestalttheorie kodernen Psychologie, S. Ertel, Ed. Darmst3dt: chol., 1, pp. 61-62, 1978. Steinkopff, 1975. 1170) M. Schlitz and S. Deacon, "Remote viewing: A conceptual re- 12001 K. Pribram, "Problems concerning the structure of conscious- plication of Targ and Puthoff," in Research in Parapsychology ness, "in Co.nsciousne@s and the Brain, G. Globus, Ed. NewYork: 1979. Metuchen, NJ,. Scarecrow Press, 1980, pp. 124-126. Plenum Press, f976, 11711 E. Karnes and E. Sussman, "Remote viewing: A response bias [2011 D. Bohm, "Quantum theory as an indication of a new order in interpretation," Psycholog. Rep., 44, pp. 471-479, 1979. physics. Part A, The development of new orders as shown through 11721 E. Karnes, J. Ballou, E. Sussman, and F. Swaroff, "Remote view- the history of: physics," Foundations Phys., 1, pp. 359-381, ing: Failures to replicate with control comparisons," Psycholog. 1971. Rep., 45, pp. 963-973, 1979. 12021 D. Bohm, "Quantum theory as an indication of a new order in (173) E. W. Karnes, E. P. Sussman, P. Klusman, and L. Turcotte, "Fail- physics. Part 13. Implicate and explicate order in physical law," ures to replicate remote viewing using psychic subjects," Zeteric Foundations Phys., 1, pp. 139-168, 197 1. Scholar, 6, pp. 66-76, 1980. [ 2031 A. Eddington,,Relativity Theory ofProtons andElectrons. New (1741 M. Schlitz and E. Gruber, "Transcontinental remote viewing," York: Macmillan, 1936, p. 329, J. Parapsychol., 4, pp. 305-317, 1980. [2041 0. C. de Beauregard, "The expanding paradigm of the Einstein 11751 C. Scott, "On the evaluation ofverbal material in parapsychology: paradox," in The Iceland Papers, A. Puharicb, Ed. Amherst,M: A discussion of Dr. Pratt's monograph," J. Soc. Psychical Re- Essentia Associates, 1979, pp. 162-191. search, 46, pp. 79-90, 1972. [ 2051 E. P. Wigner, "On hidden variables and quantum mechanical 1176 1 R. L. Morris, "An exact method for evaluating preferentially probabilities, "Amer. J. Phys., 39, p. .1005, 1970. matched free-response material," J. Amer. Soc. Psychical Re- [ 2061 J. S. Bell, "On the problem of hidden variables in quantum search, 66, p. 401, 1972. m-nhanics," Rev. Modern Phys., 38, p. 447, 1966. J 1771 G. SoIfyin, E. Kelly, and D. Burdick, "Some new methods of (2071 D. dohni and J. Bub, "A proposed solution of the measurement analysis for preferential ranking data," J. Amer. Soc. Psychical problem in quantum mechanics by a hidden variable theory," Research, 72, p. 93, 19?8. Rev. Modern Phys., 38t p. 453, 1966. 1178) D. Marks and R. Karnmann, The Psychology of the Psychic. [ 2081 J. Mehra, "Quantum mechanics and the explanation of life," Buffalo, NY: Prometheus Books, 1980. Amer. Scientist 161, pp. 72@!-728, 1973. (1791 R. G. Jahn, B. J. Dunne, E. G. Jahn, "Analytical judging pro- [2091 E. H. Walker, "Foundations of paraphysical and parapsycholog- cedure for remote perception experiments," J. Parapsychol., 3, ical phenomt.,&," in Quanturn Physics and Parapsychology, pp. 207-231, 1980. L. Oteri, Ed. New York: Parapsychology Foundation, 1975, [180] B. J. Winer, Statistical Principles in Experimental Design. New pp. 1-44. York: McGraw-Hill, 1962, pp. 43-45. [2101 R. G. Jahn and. B. J. Dunne, "A quantum mechanical approach [181) R. G. Jahn, Ed., The Role ofConsciousness in thePhysical World. to the representation of paranormal phenomena," 1982 (to be Boulder, CO,. Westview Press, 1981. published). 1182 ] R. G. Stanford, "An experimentally testable model for sponta- [2111 T.N.E. Greville: "Are psi events random?" J. Amer. Soc. Psy- neous psi events. 1. Extrasensory events,"J. Amer. Soc. Psychical chical Research, 74, pp. 223-226, 1980. Research, 68, pp. 34-57, 1974. "Il. Psychokineticevents,"Ibid, 1212] J. A. Greenwood and T.N.E. Greville, "On requirements for pp.321-356. using statistical analysis in psi experiments," J. Parapsychol., 43, [183] C. T. K. Chari, "Som.e generalized theories and models of psi: A pp. 315-321, 1079. critical evaluation," in Handbook ofParapsychology, B. B. Wol- 213 ] R. Jahn and B. Dunne, "Phy"sical models of psychic process," man, Ed. New York: Van Nostrand Reinhold, 1977, pp. 803- in Concepts and Theories of Parapsychology, B. Shapin and L. .822. IZ41 -,;R. G. Stanford "Conceptual frameworks of contemporary psi Coly, Eds. New York: Parapsychology Foundation, Inc., 1981, , research," in Handbook of Parapsychology, B. B. Wolman, Ed. 1214) pp. 69-79, R. May, "Creati v'ity and encounter " in The Creative Imagination New York: Van Nostrand Reinhold, 1977, pp. 823-858. " , , H. M. Ruitenb k Ed. Chicago 1965 11: Quadrangle Books Mind and body: A contribution to a theory of 1185) B. B. Wolman, " , , , , Op. 283-291. in Handbook ofParapsychol. parapsychological phenomena, ogy, B. B. Wolman, Ed. New York: V Ian Nostrand Reinhold, (2151 F. Barron, "Tho needs for order and disorder ag motives in crea- 197?.pp.861-879. tive activity," in the second (195 7) University of Utah research (1861 A. M. Young, The Geometry ofMeaning. New York. Delacorte conference on the identification of creative scientific talent, Press/Seymour Lawrence, 1976. C. W. Taylor, Ed. Salt Lake City: University of Utah Press, [187) B. Shapin and L. Coly, Eds., Concepts and Theories ofParapsy- 1958, pp. 119-i28. chology: Proceedings of an International Conference held in 12161 D. McClelland, "The calculated risk: An aspect of scientific per- New York, NY, Dec. 6, 1`980. New York: Parapsychology formance," in Scientific Creativity, itsRecognition andDevelop- Foundation, Inc., 1981. ment, C. W. Tjylor and F. Barron, Eds. New York; Wiley, [1881 F. Cazzamalli, "Phinomenes t6l6psychiq'ues ot radiations cere. 1963, pp. 184-192. brales," Rev. Mitapsych., 4, p. 3, 1925, 12171 S. D. Drell, "When is a particle?" Phys. Today, 3 1, pp. 23-32, [1891 1. M. Kogan, "Is telepathy possible? "Radio Eng., 21, p. 75, 1966. 1978. "Telepathy, hypotheses and observations," Radio Eng., 22, [218] W. Heisenberg, '@The nature of elementary pwicles, " Phys. To- p. 141, 1967. day, 29. pp. 32-39, 1976. "Information theory analysis of telepathic communication [2191 F. Capra, Paper LBL-796, Lawrence Berkeley Laboratories, Uni- experiments," Radio Eng., 23, p. 122, 1968. versity of California, Berkeley, CA, 1978. [190) M. A. Persinger, "ELF field mediation in spontaneous psi events: (2201 J. A. Wheeler, "The universe as home for man," Amer. Scientist, Direct information transfer or conditional elicitation?" in Mind 62, pp. 683-69t, 19?4. at Large, C. T. Tart, H. E. Puthoff, and R. Targ, Eds. NewYork: (2211 W. W. Harman, "Broader implications of recent findings in psy- Praeger Special Studies, 1979, pp. 189-204. chological and p@ychic research," in The Role ofConsclousnem (1911 L. L. Gatl 'in, Information Theory and the Living System. New in the Physical World, R. G. Jahn, Ed. Boulder, CO: Westview York; Columbia University Press, 19@2. Press, 1981, pp. 113-131. 11921 W. von Lucadou and K. Kornwachs, "Development of the system- [2221 A. Angoff and B, Shapin, Eds., A Century ofPsychical Research: theoretic approach to prchokinesis," paper presented at the ' The Continuing Doubts and Affly7nations. New York: Parapsy- 1979. Parascience Cohf., Londo, , chology Foundation, 1971. -1193) K. Kornwachs and W. von Lucadou, "Psychokinesis and the con- 12231 C. A. Murchison, Ed. The Case For and Against Psychical Belief. cept of complexity," Psychoenergetic Syst., 3, pp. 327-342, New York: ArnoiPress, 1975. (Originally pubUshed in 1927.) 1979. 1224] W. F. Prince, TheiEnchanted Boundary. New York: Arno Pre", Approved For-Release 2QO3/04/18 : CIA-RDP96-00789ROO3100030001-4 170 Approved For Release 2003/04/18 P Of FX AfflOhMondivi 24 FEBRUARY 198-2 j975. (Originally published in 1930.) Scarecrow Press, 1976.. (22s] 1K. M. Goldney and S. G. Soal, "Report on a series ofexperiments 241] R. A. White and L. A. Dale, Parapsychology: Sources ofInf6r- with Mrs. Eileen Garrett," Proc. Soc. Psychical Research, 45, "tation. Metuchen, NJ: Scarecrow Press, 1973. (Rev. edit, by pp. 43 - 87, 193 8. R. A. White in preparation.) N 12261 1. Beloff, "The 'sweethearts' experiment," J. Soc. Psychical Re- (2421 N. Bowles and F. Hynds with J. Maxwell, Psi Search. San Fran- search, 45, pp. 1-7, 1969. cisco, CA: Harper and Row, 1978. (2271 M. Maher, D. ?eratsakis, and G. R. Schmeidler, "Cerebral locali- (2431 R. G. Jahn, B. J. Dunne, R. Nelson, and H. Mertz, "Anomalous zation effects in ESP processing: An attempted replication," operator influence on elementary microelectronic devices," J'Amer. Soc. for Psychical Research, 73, pp. 167-177, 1979. Tech. note, School of Engineering/Applied Science, Princeton 12281 S. Blackmore, "Correlation between ESP and memory," Eur. J. University, 1981: Parapsychol., 3, pp. 127-147, 1980. 2441 J. Bronowski, Science and Human Values. New York: Harper [2291 J. EiserLbud, L. Hassel, H. Keely, and W. Sawrey, "A further and Brothers, 1956. study of teacher-pupil attitudes and results on clairvoyance tests 2451 M. Polanyi, Personal Knowledge. Chicago, IL: University of in the fifth and sixth grades," J. Amer. Soc. Psychical Research, Chicago Press, 1962. 54, pp. 72-80, 1960. [2461 A. KOTzybski, Science and Sanity: An Introduction to Non- 12301 F. Belvedere and D. Foulkes, "Telepathy and dreams: A failure Aristotelian Systems and General Semantics. Clinton, MA: The to replicate," PerceptualMotor Skills, 33, pp. 783-789, 1971. Colonial Press, 1933. (231 ] J. Beloff and D. Bate, "An attempt to replicate the Schmidt [ 2471 W. Penfield, The Afystery of the.Vind, Princeton, NJ: Princeton findings,"J. Soc. Psychical Research, 46, pp. 21-31, 1971, University Press, 1975. [2321 R. G. Jahn, reply to Bradley Dowden, Zetetic Scholar, 7, p. 127, [248] M. Ferguson, The Aquarian Conspiracy. Los Angeles, CA: J. P. 1980, Tarcher, 1980. (2331 C. C. Gillisple, Ed., Dictionary of Scientific Biography. New (2491 J. Salk, Survival of the Wisest. New York: Harper and Row, York: Scribner's, 1970-1980. 1973. 2341 R. L. Weber, A Random Walk in Science, E, Mendoza, Ed., (2501 A. M. Young, The Reflexive Universe. -New York: Delacorte Foreword by W. Cooper. New Yorki Crane, Russak, 1973, Press, 1976. pp. 66-67. [ 251 E. Jantseb, The Self- Organizing Universe: Scientific and Human 2351 R. S. de Ropp, The jVew Prometheans. New York: Delacorte, Implications of the ErnergingParadigmsofEvolutlon. Elmsford, 1972, p. 2-28. NY: Pergamon Press, 1980. [236 ] L. Barnett, The Universe and Dr. Einstein (Rev. edit.). New 252 D. Dean et al., Executive ESP. Englewood Cliffs, NJ: Prentice- York: The New American Library, 1952, p. 117, Hall, 1974. [2371 R. G. Jahn, "Psychic research: New dimensions or old delusions?" 253 P. T. de Chardin, The Future of.Van. New York: Hailieri and Zetetic Scholar, 6, pp, 5-16, 1980, . Row, 1964 (pp. 166-167 in Harper Torchbooks edit.). 238) J. Beloff, Ed., Arew Directions in Parapsychology. Metuchen, [2541 Report to the House of Representatives, Committee on Science NJ: Scarecrow Press, 1975. and Technology, June 1981. [2391 1. B. Rhine and J. G. Pr.att, Parapsychology: Frontier Science 1-2551 C. T. Tart, "A survey of expert opinion on potentially negative of the Mind (Rev. edit.). Springfield, 11: Charles C. Thomas, uses of psi, United States government interest in psi, and the level 19 6 2. of research funding of the field, " in Research in Parapsychology 2401 R. A. White, Ed., Surveys in Parapsychology. Metuchen, NJ: 1978. Metuchen, NJ: Scarecrow Press, 1979, pp. S4-55. Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 12 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Psychology and Anomalous Observations The Question of ESP in Dreams Irvin L. Child Yale University ABSTRACT Books by psychologists purporting to of- fer criIn.cal reviews of research in parapsychology do not use the scientific standards of discourse pre-,-alent in psychology. Experimms at Maimonides Medical .Ceraer on possible extrasensory perception (ESP) in dreams are used to illustraie this point. The experi- menis have received little or no mention in some re- views to which they are clearly pert 'inew. In others, they have been so severely distorted as to give an en- tirel'v erroneous impression of how they were con- ducted. Insofar as psychologists are guided by these re-vie-ws, they a're preventedfrom gaining accurate in- fon-ruvion about research that. as surveys show, would be of wide iraerest to psychologists as well as to others. In r=nt ycars cvidcnce has bc@cn zccumulating for the occurrence of such anomalies as telepathy and psychokincsis, but the evidence is not totally con- vincing. The evidenct has come largely from cxpcri- mcrits by psychologisis who haye dcvoted their carcers mainly to studying these anomalies, but members of othcr disciplines, including engineering and physics, have also Laken part. Some ps@-_holcgists not primarily concerned ,vith pa-rapsychology have taken timc:out from other profc:ssional concerns to explore such anomali@s for themsclvm Of these, some have joined in the cxpc-rimcntation (e.g., Crandall & Hite, 1983; LxAvry, 1981; Radin, 1982). Some have cTitically re- viewed portions of the evidence (e.g., Akm, 1984; Hyman, 1985). Some, doubting that the phenomena could b-- real, have explored nonrational processes that might encourage belicf in their reality (e.g_ Ay- croff & Abelson, 1976). Still others, considering the evidenoc substantial enough to justify a oonstructivc thoorctica.] effort, have struggled to relate thc apparent anomalies to bcticTestabLished knowledge in a way that will render them less anomaJous (e.g., IrArin, 1979) or not anomalous at all (e.g., Blackmore, 19.84). These psycholoosts differ widely in their surraise about whcthcr the apparent anomalies in questionwill eventually be judged real or illusoff, but they appear to agree that the cvidemce to date wan-ants serious consideration. Serious considcration or apparent anomalics seems an csscritial pan of the procedures of science, November 1985 - Amcrican Psychologist COPYri& IM b7 tk A. hd,.kv;CJ AA.0-b--k 1W- 000J%4"X/351=.75 rcgardlm of whclhcr it Icads to an undcrstandii:ig of nc'w discoycrics or to an undcrsi2nding of how p-c-r- guasive illusions all-st. Apparcnt anomalics-jusi like the more numc:rous obscrvadons that arc not anom- adous---can rcccivc appropriatc attention only as they bccomc accuratcly known to the scientists to whos-c work th-cy are relevant- Much parapsychological rc- search is barred from bcing scTiously considered be_- cause it is cithcr nc7glcc1cd or misreprtscritcd in ings by some Mcbologists-among them, some who h-avc placc)d themselves in a primc position to medizte intcraction bttwcen parapsychological rcscwcb and the gcncral body of psychologicaJ knowledge. In this article, I illustrate this important general point viLh a par-Licular case, that of cxperimental rascwch on possible ESP in drcam 's. It is a case of espm a] I y great intertsi but is not un'represcDtatiYe of how psycho- logical publications have trcatcd sirailar anomalies. The Maimonides'Research 7be cxpcTiment2d evidence suggesting that dreams may actually be influencod by ESP comes almosi en- tircly from a resewch program carrie;d out at the Maimonides Medical Center in Brooklyn, New York. Among s6entists active in para psych o) ogy, this pro- gnam is vidcly known and grtatly respected. It ha-s had a major indirect influence on the recent cours-c- of parapsychological research, although the great ex- pensc of dream -laboratory work has prevented it from being a direct model. None of the Maimonides rescarchwas published in the journals that are the conventional mtdia for psychology. (The only possible exception is that a summary of one study [Honorton, Krippncr, & Ull- man, 1972) appearod in convention proceedings of the Amcrican Psychological Association.) Much of it was published in the specialized journals of parapsy- chology. 71,-- rest was published in psychiatric or other medical journals, where it would not be noticed by many psychologists. Most of it was summariztd in popularizcd form in a book (LJUman, K-rippncr, & Vaughan, 1973) in which two of the rescwchcrs wcTc joined by a popular writer whose own writings are clearly not in the. scientific tradition, and the book departs frorn the pattern of scientific reporting that characterizes the original research reports. 1219 Approved Far Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 How', Lhcn, would this research come to the at- tcn6on of psychologists, so that its findings or its crrm might in time be cv-aluatcd for their significaricc to the body orsys1cmalic observations upon which psy- chology has bccn and will be buflt? T_hc cxpcTimcnts at Mairno-uicics were publ_ished bctweeri about 1966 and 1972. In the years sincc-Dow over a decade- five book-,, have bcr-n pubLishod by acadcmic psy- chologists @hat purporl to offer a schohu@y rrvicw and cvaluaLion OCPaTapsychological research. 7-hey vary in the extent to which they sccm addressed to psy- chologists themselves or to their students, but they seem to bc the principal route by which either present or future psychologists, unless they hayc an already establisho@ interest strong enough to lead thcm to scarch out the origin2J publications, might become acquaintei@ with the experiments on ESP in dreams. I proposc !o review how these five bomoLs have prc- sentc@d knowledge about tl@c experiments. Fim, how- ever, I must offer a summary of the experiments; without that, my review would make scrise only to rcadcrs already well acquainted w@ith thera. The cxpcr-imcnt_s at Maimoaides gre-w out of Montague UUma-n's obsen,-ations, in his psych)'atric practice, o@`apparcnt telepathy unde:rly@ing the content of some d@eams reported by his patieriLs-obser-a- tions para.111d to those reported by many other psy- chi-atrists..Hc sought to determine whether this ap- parent phenomenon would appew in a sleep labora- tory under controlled conditions that would seem to exclude intcr-pretations other than that of ESP. He was joined in this research by psychologist Stanley K.irippner, now at the S@aybrook Institute in San Fran- cisco, and F firde later by Charles Honorton, now head of the Psychophysical Research Laboratories in Princeton, New Jersey. Encouragcd by early fiDdings but seeking to improve experimental controls and identify optimal conditions, thcsc rcscaicheTs, assistc:d by numerous helpers and consultants, tried out v@Lr- iOus modifications of procedure. No one simple de- scription ofprocedurc, therefore, can be accurate for afl of the expt-riments. But the brief citscTiption that follows is not, I believe, misleading as an account of what was generally done. 7We Experimental Procedure A subject would come to the laboratory to spend the night there as would-be percipient in a study of pos- sible telepathic influence on drcam@. He or she met and talkcd with the person who was going to serve as agent (Lhat is, the person who would try to send a telepathic mess-age), as'wed as with the two experi- menters taking pan that night, and procedures were Rrqucsts for rcprinLs should bc scni to Irvin L. C10d at the Csc- panmcnt of Psychology. Y2Jc Univu@iy. P.O. Box I I& Ncw Hxvrn. Connccticut 06520-7447. 1220 cxpuined in dctail unless the pcTcipicDt mas a rrpcater for whom illat acp P.-is not necessary. Virlico ready- to go to bed, ffic pCTMpIcnt was wired Up in Lhc usuaJ way for MOTIitOrii3g Of brain waves and eye move- mc-nL,, and: he or she h@d no further contact A-,'Lh the agent or a&cnt's cxpcrimr_i3tcT until after the %@:ssioi:i- was compl:Lc& The cxp--_r-imcI3TC-r in the next room monitored Lhc pCTCipiCDt Is sleep and at the bc-@nmng of each 'od orrapid eye movements (REM), when PCTI it was rC2.so.nably certain the 5tlccpcr would k &cam- ing, no6ficd the agent by pressing a buLzc-r. The atcrit was in a remote room in the builiding, pro-@idcd with a target picture (and sometimes a cce-s- - sory material echoing the theme of the picture) ran-- domly chosen from a pool of potential ta-r-geLs 2-s Lhe message to'; be concentrated on. The procedu, c for random chQice of a target from the pool wis dcsigned@ to prevent anyone clst [Tom knowing the Ideni;ty of the Lirgct. the agent (lid not open the packti con- taJning Lhc:targct until Lsolated for the ruighT (-,-xccpt., for Lhc one-way buzzer commualcatioo). Wh.-ncver@1 signaJed'Lha:t the percipient had entered a RI'll pc- riod, the agent "-as to conctntrate on th-e Largc, "Irlith the a-im of communicating it Ic1cpathically to L@@ per--- cipient and:thus influencing the drnm the pcri'pient was hav@ing_.Thc percipient was oriented towa-rd L-Ying to receive this message. But of cour&e if clairyoyanct- and telepathy artboth possible, the PCTLipir-lit Might have used the formcT-that is, might have been pick- ing up information directly from the tar-get picture, without the mod_iation of the agent's thoughts or ef-- forts. For this reason, the term general extrasensory perception (GESP) would be uscd today, thoug-h the rese-archm.more often used the term telepath.V. Toward Lhe approximate end of each RE@@i pc--' riod, the percipient was awakened (by intercom) by the monitoring exp@Hmcntcr and dcsc:Tibe@d any dream just txpericnced (with prodAing and qu--6on-- ing, if nemssary, though the percipient ofcoursc knew in advance -what to do on each awakening). At the end of the night's sleep, the percipient was intefNic-ed and was asked for impressions about what the target- might have been. ffhe inter-view was of course. double- blind; neither percipient nor interviewer knc-,y the identity of 6c target.) T'he cLrr-am descriptions and morning impressions and associations were recorded abd later Lranscribed. T"hc original research reports and the popular book both present a number of very striking sIrndar-: i6cs between passages in the dream Lranscripts and the pic-ture'Lhat happened to bc the night's La-rgct. Thesc similair-itics merit attention, yet they should in themselves yield no sense of conviction. @crh2ps 2ny@ transcript of a night's dreaming contains passages of striking similarity to any picture to which they might be comparo. The Maimonidcs research, howcvcrc. consisted or carefully planned experimeriLs dc-5igned- Novcmbcr 1985 American Psychologist Approved Foe Release 2003/04/18 CIA-RDP96-00789ROO31,00030001-4 A RrovMe F?r po§e"? gMq4(slgrrQWRDP96-q%~§ERQOi2~LOoio3-oolall-4cfcr-s to one scgmcni to M it ua I sirnil-arity, and I -must now tum to that asp=. or the research, and scgmcmLs for the mosl part a--, Re sults laklod as they wtre iri the table of Ull man cl al. () 973, - To cN-aluate the chance hypothesis, the rtsczrchm pp. 275-277). Scgmcnts that rollowcd the gcncT :J procc-durc I dc:scribod-2JI-night sessions, Aith an obtained judgments of similarity bctwmn the dream ' ' agent concc:nlra6ng on the target during c:ach of t-c content and the actual target for the night, and al the : n percipient's R-EM pcn'od,&-a-rc gathered together sarric time obtained judgmcnLs of similarity bcrwmn the Erst eight Encs, A through H (in fivc ofthc-,-- sclz- the dream c@onlcnt and cach of the other potential mmts, aD but A, C, and H, a single percipient co:,,- ta-rgcLs in the pool from whkh the target had been tinucd throughout a saim, and in four of these t@c selected al random. The pcrson judging, of course, percipient was a psychologist). MCT tYPcs ofscgmc-,@@s had no information 91>out which picture had been arc presented in the rcsl of the table. Lines 1, 3, a:-.d randomly selected as taMct; the entire pool (in du- K summarize prcoogniti .vc sessions; here the tajTfi plicate) A-as pi-c-scritc;d logcthcr, with no cluc a-s to was noisclcctod untB &RcT the dreaming andinlcPricl-v which picture had been the target and which ones had had been compIctod. The target consis-tcd of a set not. That is, in the experimental condition a picture Sumuli to b,-- prcsc-mcd direct-ly to the percipient a_@i-r was randomly sclccicd from a pc>ol and conccnLraicd it had b-ccri sclcctc@d in the morning. Lines L and "I on byLhc agent, and in the control condition a picture reprcscnt GESP sessions in which the PcTciplcn@'s was left khind in the pool. Any consisicht diiTti-crice dreams were monitorod and recorded throughout i`c betwc-cri target and nontarget in sim-@arity to dream night, but the agent was attempting to t=smil or*@' contcat, exceeding what could reasonably bc.ascribcd before the percipient went to sl=p or just after, )r to chance, ww considcrod an apparent anom2Jy. sporadically. Unc N refers to a few clairvoyance &_-s- Tbc data available for the largm numbcr of ses- sions; thcsc wcTc @kc the Ytandard GF-SP sessions c.;- sions came from judgments made by judg 'es; wbo had cept that tbcre %-a-s no agent (no one knew the idcn@--v no contact with 1-he experiment except to r=ivr- (by of the target). Finally, Linc. 0 reports on some GES? mail, generally) the materia) necessary for J;udging ions in which each dream was considered (tanscTipts of dreams and niervicw and a @@ of rately; these formed a single experiment with fc-:r the target pool). For many sessions, judgments were pcrcipicnt .s-, comparing ri@ights involving a differc--it also avai@lable frorn the dreamer, he or she, 0.f coum, target for each REM period vith nights involving .-z- madc judgments only afieTcompicting parLicipation pealed use- of a single target, in the cx periment as dreamc-r (except in some series Regardless of the type ofsmion (considering -L-.e where a separate target pool v,-as used for each night ' five types I have dcscTibed), each session fcU in to 0:11 e and the dreamer s judgments could be mad-, at the I t a of two categories: (a) pilot sessions, in wbich ciih - end oftlic session). For many.sessions, judgments were new dreamer or a new procedure was being t:He@d 0--t; made for the dream transcripts 2Jonc and for the total th= appear in bncs H, K, amd N, or (b) sessions @n Lranscript including the moming intcrvick-, for con- an expenmental sc:lics, plannod in adva-ricc. Ls onc or sisicncy I have used the latter, because it involved more. smions for each of two or more subjects, or as judges who had more nearly the same information as a number of sessions veilli thesamc @lreamf_T throuL]7i- the subjccu. out. Most of the researchers' publications were cc- The only form in which the data are available voted to the results obtainc4 in the experimental for all series of scssions is a count of hits and misses. Ties, but the results of the pilot smions have also b---n If the actual target was ranked in the upj,@_ThaN or briefly ircportod. the urget pool, for similarity to the dreams and in- A glance at the score columns forjudgcs and tcT-view, the outcome was considcrod a hit. If the actual subjects is sufficient to indicate a strong tendency -Or target was ranked in the lovtr half of the pool, the an excess of hits oyc:r misses. If we avcrage the oulcom C outcome was considered a miss. The hit-or-miss score for judges and for subjects, we find that hits Cxc=d is presented separately in Table I for judges and for miss,-s on every one of the 15 independent lines on subjects in the fim two data columns. Where infor- which outcome for hits and misses differs. (On L-ne mation is not supplied for 'one or tht other@ the reason F_ hits and misses occur with equal frequency.) B-,@ a is gener-2illy that it was impossible for the researchers simple sign-tcst, this outoDme would be significznt to obWn it, and. fbTa similar reason thenurnbcr of beyond the 0.000 1 level. I would not stress the c;,@ct cases sometimeS varics.1 value here, for scyaral reasons, There was no adv2@ce 'Of course, usable judZTri@ents could not be obWned fircom the subject in precognitive Pessions@ because al the time orjodging he judgments. In a few or the pilot sessions (Uncs H, Y@ ;LrA N) cnh' or she would aIrcady kmow wh2t tbc urgci had be=. Forl_iDcF, the subj='sjudpD=t w-as sought, and in sorr@c scssions onl\ _@al tbc singk subject was unable to rive the extra time required for ofooc or moitjwirs; in a fi7w the rne2njudgrs' rati ng was n6--hcT jtkWnt, &M for IJDc 0 one or ihc four PA)joas failed to Mike a hit nor a miss but cxAcdy at tbc midd1c. November 1.985'- American Psychologist Approved For Release 2003/04/18 : CIA-RDP9 6-00789ROO3100030001-4 A IA n -lQrip-QA-QQ789RO03 100030001-4 App[VVCU I WI lx@- Table 1 Summary of Maimonides Rasufts on Tender@cy for L>rG&m.3 lo pe Judg,&d More Oke Target Than Like Nontargers in Target Pool A49-, x or I r""v ftrgm pdv- See*" &-6?oC-* GESP: Dreams morwored and r&cord&-- uwDug6ow night; &;lerrt -Uansn4ttk-,g- durlr@g each REM penod A. I st. screening 7 5 10 2 @Ib Z - 0. Z - 1. 33b LMrrvan, Krippoer, ! F&lclstain (19,66) B. I st Erwin 5 2 6 1 z - 2.5 3"' Z - 1. 9c@ Ullman et al. (1966' C . 2nd screening 4 8 9 3 Z - -.25b Z - 1. 770 U1.1man (1969) D. Posin 6 2 6 2 Z - 1.05' Z - 1.05' U hman (19,6 9) E. Grayeb 3 5 5 3 Z - -.63C Z - 0. 63, Ullraan, Krippoef, & Vaughan(1973) F. 2nd Erwin a 0 1 - A.9a' UUn-Lan & Krippoer (1969) G. Van 6e Castle 6 2 8 0 f - 2.81' t 2.74' Krippf-@-- & Ullman (1970) H. Pilot sessions 53 14 42 2-2 z = 4.26rb z 2.2 Ib UlIrran et al. (1973' Precognition: Dre-arn-s moni-tored &nd recorded throughout night; target expefky-poe next day 1. 1st Bessent 7 1 t = 2.81.' Krippr,-r, Ullrivan, Ko@ on (19 7 1) J. 2nd Bessent 7 1 1 = 2.27' Krippn,-. Fkx)ortc-.. I Ullman (1972) K. Pilot sessions 2 0 Z - 0.67c Unman et al. (1973,@ GESP: Dreams monitored and recorded tblv oughocrt night; agent active oNy at b&ginNN- or sporadically L. Seftsory bombard- a 0 4 4 z = 3.110 Z =..O. 00' Kripprw, Honorton, ment Ullman, Mastars. Flo@iston (197 1) M. GratefLA D-ead 7 5 6 4 Z = 0.6t, Z 0.81C Krippr@,-_ Horxxton, UHman (1973) Clairvoyance: Dreams moni-tored and m@ tJvoughmm (1977) argued bliat a fur,@c-_,t mprotb1cm with both the diram tclcpa(hy research Ot viewing Icsis c is that the reports suffer from wh t called "shoc-fit6ng- Language; she cited a study in whi h i:% scndcr veas insulled in a room draped in white fabric and "d ice cubes poured down his back. A rocciycr who cp,@-ricd "white" was im- mediatclyjudged to have made -@a- by an independent panel. Yet, as she observed, WO ds such as -mL@crablc", 'vo a ..V"c or. icy w ould have been I ncr hits. . Again, the obvious need is for a control grou Why art they not usc-d? ou , Jn,, (p. 163) What Romm described as ce- fitting" (misintcr- @fons) is an important preti ng even ts I o fi t on c's cx kind ofcrror that is repeatedly rn-lde in inicrpretaEion y of everyday occurrcnccs by plc who believe theN are psychic. But the dream tel pzthy rescarch at Mal:_ p@ monides was well protected a inst Lhis kind of error by the painstaking controls i al Alcock seemed not to have noticed. _@L@rclvjom must be rcfcrrijP_g to some other and very sloppy d czm research? 1226 November 1985 @rncrican Psychologist Approved For Release 2003/04/18 CIA-RDP96-00789RO031'00030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 Not.ai 0. The dctiik in this paragraph, arA even m6rc'iri Romm's anicic, point unmi5takably, though inaccurately, to the fifth night of the firm PTC- cognitive series at Maimonidcs. The actual details of target and rcs-ponsc would alone deprive it of much ofits value as an example orshoc fit6ng. As rc-poricd by Krippncr, Ullman, & Honor-ton (197 1 ), the tayEn v,w a morning cxpcr-icnec that included being in a room that was draped Aith whitc shcets. The subject's first dream report had includcd the statement, -1 Nk--s just standing in a room, surrounded by white. Evcrv imaginable thing in that room was white" (p. 201). There is more similarity hcre than Romm and AJoc>ck acknow1cdgod in mentioning from this passage only the single word "white." Marc impor=t, howcN-_-r, is the fact that the ex- pcrimcnx they .xtrc refer-ring to provided -no oppor- tunity for shc-c fit6ng. T-he proout the cxpcrimcnL,; othm _S 1-m- - P-T @_ neglect them. I believe it is fa-ir to say that _n - __ ... I o-n6 of tbcsp books has correctly idcritified 2zy ckfw in the Mai onidcs expeTiments other than ones rel- evant only -o the hypothesis of fraud or on iaappro- priate sta i ical reasoning (easily remedied by new calculations from the pubhshe@d data). 1 do not mean To that the Mai onides experiments are models of design and executi n. I have already called attcridon to a design flaw prevents sensitive analysis of some of the experi'm ts; and the control proctduTts vetre vi- t ,@n olated at an session, *as Akers ( 19 8 4) poi n te:@' out on the basis of full information 'supplied in the orig- inal report- (Neither of thesse genuine ckf@_@.s was 2 mentioned il any of the five books I have rrviev,,ed here, an indication of their authors' gencra! lack of corred info tion about the Maimonides experi- ments.) Rea who doubt that the falsification is as ex- treme as I pictured it noed only consult the sources I ha referred to. Their doubt migbi also be reduced by miliarity *th some of James Bradley's research 0 9 1, 1984). In his 1984 article, be reported simiLar misr tations of fact on a topic, robust- of pr ures of stadstical inference, ou which .would not be thought to have nearly the strength of nception that many are known to hm a iow much more likely, then, fahi- - ficatio=onl otionally laden a topic as ESP is for man In the earlier article, Bradley (198 preset) 1981) ted experimental evidence (for college 5tudexits, in case, not psycbologis@) that oonfi- ibcr 1985 American Psychologist Approved For Release 2003/04/18 CIA-RDP96-00789RO03@100030001-4 Approved For Release 2003/04/18 : CIA-RDP96-00789ROO3100030001-4 &rice in the cor-rcciricss orones own erroneous opin- E,@i&-ncc' fcpr impropcriy focused psi? Journal Ofthe American ions is positively corrclatodvith the dcUcc ofcxpcr6-,c Soc-Lrryfar Ps@chical Research. 77. 209-228. one bclievcs oncsclfto have in the field of knowledge Foulkm D.. BeJvn3crt, E., Mxstcm R. E- L., Housion. J.. Krippncr, " ,Ai thin which the moncous opinion falls. This finding &, K:,norton, C., & Ullnun, M. (1972). Lofig-dinkncc 3cnsory- boa@:,ardmcnt- ESP in dreams: A ft@urc to rcp@iattc- Prrccpual may hclp in undemanding why the authors ofsomc ar&.Voeor Skilli. 35. 73 1-734. of thcsc books did not find it necessary to consider G lobu& G , Kram P- Ski n ncr, J., & Hc2k-Y, 3. ( I % 8). A n appr-aj sM critically their own erroneous sialcmcrits. octecp2thic commun;cx6on in drcxn@ Py>chop@ysiology 4. R5. A very considerable proportion of psychologists H-aJ 1. C- ( 196 7). E i pcri mcn tc z u r te k-pa th L%chcn Bcci n 0 um u ng Yo n Triu-nc-n. (F-xpcrin-,cnLs ou t6cp,30jc innumcc on drcarns), have a potential inicresi in the quc:stion of ESP. In a Zeia-hrififii@ Paraps)@@e Wid Grralt'rbicle der Ps)cAdogie, rcccni survey (Wagner & Monrict, 1979) ofunivmity /0. 11-47. profe:ssors in v2Hous fields, 34% ofpsycholoj@sts wcrc Hamc[@ C. E, M. (1990). ESP and parapsychology: .4 critical re. found to consider ESP rithcr aj L 2 r 9 rval-ion. Buffalo, NY: FTorncthcus. . . . Gkcly possibility, exactly the saMt PrOPWiQAkj C1W, HonorLm, C., KripprK:r, S., & Ullaixn. M. ( 1972). Tclc-paLh3c M- ccp6-n of an, prinu urKSa t-o oondidons. Procefdings ofihr sidc:rcd it an impossib ility In dLis survey, psychologists 80th Annual Comv-wion ofihe American Psychological Asioci. r -- k ss Ttquently expi-c-,scd a positive opinion than did aiioe-. 7. 319-320. rT)cmbcrs of other disciplines, a finding that May be Hyrnan, F@- (1985). Tlx &ax&cld psi cxparimcnt: A ai6c@d xppra.L%al. attributable to psychologists' b-ttter understanding of Jovrr.:] ofParops),cho(ogy. 49, 3--49. Irvein, F.. J. (1979). Psi and the mind.-An inform--vion proczising sourots ofmor in human judgment. There s-,cms to opprrcch. Mcwd-,c-n, NJ: ScarccTow. b@ no Nually sound reason for the curious fact that Jxhn, R G. (1982). The pcrsis7tcn1 par-2dox ofpsychic p@@norncna: psychologis-Ls diffcrcd overwhelmingly from others in An mv@n=ing pcr.,cp@. Proceedings ofthe lrztitwe ofElm- their tendency io consider ESP an impossibility. Of Incal znzi Elearceiics Enginerrs. 70. 136-) 70. natural scientists only 3% checked that opinion; of Krippnc@ S., Hooonoo, E-, & Ubm-az, M. (1972). A x@cond prc- i-;, J l d d i h M B l , ec Tcam stu a cs&cni. Journa t co m cogn y w of (he 11)c 166 professors in other sc>cial s6cri=, not 2 single Anurram Socieryfor Psychical Re-search, 66. 269-279. one did. KrippDr-- S., Honortoo, C., & Ullman, M. (1973). An cxpcrimcni .- Both of these groups of psychologists have been' in drcarn tclcpat-hy v/iLh -Tbc Gratcful Dcad.- Journal of the ill served by the apparently scholarly books that seem Ankrr@:ari Society ofPs@chosornajic D,-ivisiry and Medicine, 20, 9-17. to convey information about the dream experiments. The same may be said about some otheT lines ofpara- psychological'rcseuch. Interested readers might well consult the original sourccs and form their ow-n judg- rnents. REFERENCIES Akcrs, C. (19184). Mcthodo;ogic2J cri6cisms of parapsycWogy. to S. X-rippncr (Ed.). Advances in parapsychological research (Vol. 4, pp. 112-16-4). Jcfferson, NC: McFarl2M- Alcock. J. E. (198 1). Parapsychology. science or rr@c? A psycho- logical perspecripe. New York: Pagamon Prcss. Alcock. J. E. (1983). Bringing &nom3dks back into psychotogy. Contemporary PsycAology 28. 351-352. Aycroff. F- & Abelson, R. P. (1976). ESPand ESB: BeLicrin pa-son,21 succcss at mcnW telepathy. Jow7ud of Personality and Social Psychology 34. 240-24 7. Mvodcrr- I-, & Fvalk@ D. (197 1). Tek-pathy amd drca.Tns: A Uw-c to rcpljatc. Perceptual and)Wo(or SWIs. 33, 783-789. Blackmort, S. J. (1984). A psycho;*cal tboory of the out-or-body cx Pcricrcc. Journal ofFarapsychology. 48. 201-218, Bradky. J. V. (1981). Ovacoo6dcnoe in ignorant experts. Bulletin ofih,- Psychonmic society. 17. s2-4@4. Bradky. J. V. (19M). Anti Donrobust nc= A case study in the so- 6ology of %CK*=. Bulletin afthe P@ydsorsornic Society. 22, 4463- 466. Braud. W. (1977). Long-distan;oe drcam amd p,,,Iccp tek-pathy. In J. D. Morris, W. G. Roll, & R- L Morr-is (Eds.), )Zesearrh in parapsycholop- 1976 (pp. 154-155). Mctucbm, NJ: Sc2fccro-- Caild, I. L, KAn4=aaj-j, H., -A Swmncy, V. M. (1977@ A sirnpli6cd exPcrimmt in drtarn tek-pathy. In 3. D. Morr-is@ W. G. PoU. & P- L Morris (Eds.), pesearch in paraps)@- 1976 (pp. 9 1 - 93). Mciuchen. NJ: Sauvcraw. Cxxndall. J. E., & Hite. D. D. (1993). Psi-missint and dLqAa=mcnt: ...Krippnc-, S.. Hononon, C., Ubrnan, M., Mzstcrs, R., & Houslon, 1. (19-1). A Iong-di=s=cr "scnsory-bcrrnbardrncnC' study ofESP in dra-ras. Jow7ialqf1&-Anu--@ SocicryjorPs)o@imlpksearch. 65, 4@3-4 7 5. Krippom S.. & UlJm2n, M. (1970). Tckpathy and drcarns: A cion- trolke experiment with ck=ocrccphLLDgram-c~cmo-ooAog-am rnor6ixing. Journal afhlervous and Merual Disease, 151, 394- 403. KrippDr- LTIJ man, M., & Hooonon, C. (197 1). A prrcognitiNt dr= study %%;,th it singk Journad ofthexnu@@can Society for Ps@rhical Res,-axch, 65, 192-203. Lo-ry, K (198 1). Apparemi -PK cffoci on corn puicr-pmcraicd ran- dorn dgil s=i= Journal ofrhe American Socieqfor Psychical Resea,ch. 75. 209-2.20. Marks, D., & Karnminti, R. (1990). The psycWogy ofthe psy-hic. Buffa3c, NY. Pronxtheus. Mostcl]ia- F. & Busli, R- P- (195-4). Sckmcd quantit2tivc tcchr@crucs. In G. L@in