Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4'5-"(--~ &t4 (k,11, SPE"CIAL ACCE,(L3":3 P"ROE-31RAP-1 4ww""Mo GFIECIT'AL FR F:, P%, -1- SUN. STI [-.':.'VAL-Uo1'1'01,q ........ . ...... . ....... .... ..... ................... . . .. ...... ......... . ......... ........ WARNING NOTIC'X':.'-. THIS MATEF~'IAL IS ff ~1:1 ~11I - ri. -1D.... ......IC. T-1 1C)SE WITH VERITIUM ' TO ':':)t..JN LEVI'L 0316-: S 2 NOVEMBER 1.989 CLASSIT"IED BY-. DIAVST DECLASSIFY BY. OADR SPT.:.C I AL ACC,EE-)S PROGRAM Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 U 1-A C L. A!311-3 I F- 1. 1'::. 1.) CON]"EN'TE3 BACKGROUND I I V A L U A T 1: C) N 1.1.1 F''INDING11:3) ..-.N D I" X IPPI' A I k I F-~ 0 J E C'T .11 DET'AILED j. I A. -12 RE:CORD DETAILS) I NG-1-13UCTI DNS TO ANALY.13TS/DA"I"A REVIEWERS UNCLASS I FT E.D Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 UNCLASSI F'I F-D .T.1 "ON C PU I """' F" 0 S I B A CF~ 0 U N D 11 EVALLAI-ION . ..... . a S e L"-* V a 1. u a t i cj nT ei h 1-1 j. ~'b e 5 1"'v,~:kI i..i at i on R e s ul t s I 1\1 1) 1 N G S-------------------------------------------3 APFENDI X RECORD N.."I"AILS II. D L-.' 1- A I L E 1) INSI-RUCTIONS TO ANALYSTS/DA-l"A REVIEWERS UNCLASSIFIED Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 U N C, L A F.-) S I 1 1'. BI- FIC.3UIREIS' 2. Fl (3UFI~'E -~,4 P 1: GUIRE, 5 F'I G U RE 6. I::-- 16 U R E7. NL.Irfibei- of 1:~,rojects--i as a F'unction of type Data Clategciri(-;~s Surnrnary D,--A t aE-.valW-Ati.oni :Shec--t---E.,.,,afflple Ov'er a 1. 1. Dat'a Ccw r el. a J. ons Overall Averages -for all Project Types Expected Results -For- Approximately .70% Data Acc.L~racy--lSel ected Personnel- ONCLASS 1'. F" 1'. ED All Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 Approved For Release 2000/08/08 : CIA-RDP96- 00789ROO0700260003-4 SUN ...... . ............... ......... ..... ........... .......... I., .. ..................... ... ............ o+ this rePOr-t Is to provide an evaluation o+ the GUN ope a]. Projects condkACted since 19136. dwmiiww" Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 Approved For Release 2000/08/08: CIA-RDP96-00789ROO0700260003-4 B()CJ:.'.(:3ROUND (U) S F, / S S - 13UN is, an ij-i-hOLASIi~ Dlr~i project f or (J v (-.--%:L a pi n g a n o p e r a ti o I,-, al p S. y c h a e n e r g e ti c s ( i e r e mito t e v i e w i n q capability for the C01MMLAI-lity. T w e 1. v (.;_1 G e n e i-z-% 1. De-Fense Intc...A.ligence Program (GDTP) billets were ZkUthorized for DIA in V?86 for ti-iis -Activity. F'ersonnel from the AIRMY TIqSCOM CEN"I"EIR LANE F'roject who ha(-.1 bc,(,~n c:-~xz:kfnining similar phenomenon were to DIA to form t1--te SUN 13TREAV.' C01"C4 group. D 1A had (1.985) received control from 1-10 DA for Arri'iy Unit. (S/NF/SS-2) ln 1.985, thc::.-~ D1A 1,3UN STREAV" Program Manager an Action P'lan th~.--tt- (1) detailed -the steps necessary to t-.Vie CENT'E' IR LANE' unit to Dlv'-),- (2) identified SUN STIREAk". s -L a 4: +i I ig and SLAIDPOrt ri(,ieds-, ~--xnd set forth key programmatic r e.:.~ (:,.I LA j. r E~. m e i +c1r- SUN S'T*IR'1V'.`.**AI-:.. activity. The Action Fllan -tha:~: t-i:rr,:eJ7pmqL(:L r(Rd fo.r ach-leV1419 -.A,. ~prototype J. o n a c ap~a b J* :N'p opera -~y-`Wbuld -bLA. p rts_T a~td' -3. ~S/NF/6S-2) Key aspects of this Action Plan, along with additioi-ial procedur,~O. information, were sent to congressional committt.i~es ii-i 1986. The 1C staff was also briefed at this time on the Action P.lan_and on. -,Fth-ticipated. SUN STREN' operational developmert t i i pr C ' Ur. 5 K -And data, evalua or 0 e.C I e (G/I\IF/SS-2)- -Prog.rzxfnm,-..Atic and operational. requirements identified i-n the Action Plan were -to." (.1) gain..,Special. Access Program (SAP) statLAS (accomplished in March 1985); (2) gain human -A -ited, n. March, 1!1~05) (3) Set up a Senior LA L A ~apj::)roval (graf Ovei-~..-iight and zi Coardii'-ii--itting C,'oM1T1J[ttbe' 1'.'. h C) U g 1-1 1 ot ac-tivated) , (4) establish -tight project al.c.3rig With al"'I EiLAtC.-1(Y11Ettc.:-.,d datat)jAs-,e mav-iagemient and records s Y!::;,t e (T., c (:) ml I.:) "I. i s t--, e d arid to an R/D. link for d(--.)vel(:)I:)(vient (accomplished vi a S'GFU) + LAI"'] di n ~:j ;,.Arid a DO-ARPA MIPR) T'1-1(-..-~ R/D via SIR1. Tnternational, has yield(..-:-~d data eVa1LA0ttiQI-i prOCedUres, has idC.-F-1ti+ied techniqLAL-S, and Vias contributed to methods that -are currently in project use. Thc~:, z:tctivity at SIRl hai--i recei.ved and continues to receive t *:- I 11;;ave by a 9---meml-Der peer review panel to il'ISLAre -that in all theii- activities. 'I-I--1f,-.-? L)as, employed by GUN GTIREAI-::.' -J.c- appro toward devcz..l(:.-)I:)ii ig a opev-,rktional 1:~emote Viewing (RV) (-:ar.:)a1.:)ility is -to with potential RV capability Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 Approved For Release 2000/08/08: CIA-RDP96-00789ROO0700260003-4 anc:1 to devel.q_-) ti-w-,ise al-jilitieiiin, via appr-opf-iate, tr,--.kining/ (Jevel ol:iment ("'Ince satisfactor-y pr-ogr-e5s is noted on tr-airiirig individUals ar-C., 1-1 v- ci---., s e n t e d advanced and c;per-a-Lional sitm..O.ation tar-gets. ~.:;imUlatiCA"I tav--gets Usually Us 1-flilitary or- US sc-J.('-.~~rlti+ic tar.-getS gV-0U1'_1CJ tV-LAtl--i is totally known or can be r-eadily LG~.-.ktis+actory l-:)eF--+c.)F"mance on thes(-:_~v t a s I,.- s c-.luz--.-dAfy i.:.Ari individUcAl foF- al-.1jer-atiorial projects of inter-est to the i n t e 11. J. g ei ice C"C:)M(T1t_kF'11ty. Ir., mainy of the ol:)erational projects, hiowever-, gF-CAMIC] truth il usually not known. (or- is only partially le.r.au-Jr-1) . coml..-;lete evaluation of the viewer-'s data c-annot be ffiad(-~ until a later- -time when grOUCICI tr-uth does become a v a i 1. 1. E., In t1--ie inter-ifri, Y-easor-tz-tble estimates of the overall validity of the viewer-'s data can be made -for many of the oper-ational f.:-w-(::)jects bzu-iied on what is generally known or st-ti::4jected tar-get. interii-yi evaluation results are whenevc,)r- nc..1/4 ti-k-tth is received. The opbrational prpjects ptkrsued-by..BUW.STREAK a r E., a p It) r- -.are --:Ln d" p af4 o'ved by 'Vie - agrarn ahager~-' an' the Pr-ogr-a(Y-1 tlanager's +amilar-ity with Intelligence CommUnitV (IC) needs and on solicitati(-.)r) +r-om other-s Within the 1C who have been br-iefed J.n-Lo the: SUN ST'l-'%'EP)V:: pr-ogr-am. (S/NF/SS-2) The Evaluations performed -For'thi's report cover all :the oper-ati(:inal and operati-onal si-mulation projects, 200) t1--i,~A have been wor-ked by SUN STREAk..". per-sonnel 1986. A few of tl--ie special (.-)pbrations i n c u called upon consultants fr-ofyi the SRI: talent pool.~ These. results are also ini over-al]. (`~?Va'lLkation. Approved For Release 2000/08/08 CIA-RDP96-00789ROO0700260003-4 Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 I T. IEVA I U A -r ci ri (U) 1. DATA (U) ......... . ........................................ ,., ) Th e 6 U l',1 ",,r i"', i::* (.) i.::' p r- o j e c: t m at i n t a i I-, s a n ex (S/NF/Gs 0 _~ , t er-i s i ve r-ec(.-)r-d Of all pr-c)ject activity. Deti:-ki Is include project tif-iiing, people involv(,~d (i.e., vi .ewer-s, inter-viewer-t-ii, and possibly obs(--...--)r-vers), and a var-iety of Other data considerHd essential for- good ancl for- c-.-~v~-xluating project results. T 1-1 i s data, along with pF-oject summaricas, a-F-r-.~ i-iiaintained in an a U t. o fnz-t t e cl da t a b a s e f c) r- c o n v e ri i e I-) t F-- e t r- i e v a 1. . Copies of project summar-ies ar-P also sent to the F'F-ugF--am Manager for his review. 1r-, adclition, all raw data (i.e., sk:etcl-iles, viewer's notes) are riia j. n t a i Iec-1 ini z-t separzAte file that is available for r-eview and a ri,7.--k I y s j. 15, (AclclJ, L i anal pi,-oj ect r ecar- d detai. 1. s ar- e j. n append I EV;~~l oa i ant c ~,d for.. Ah MF ~a comp l.etc-.-.., ir-c.?--examination a+ the er-,tir~e SUN STRIEAF-". 'operational data b ' ase. Many Of the (,%?arlier pr-ojects had only been partially eVEk I Uat ecl , (:)I,- not evaluated at all, clue to lack of Suitable gr'OUnd trUth at the time they were completed. These projects were r-e-evalu,~tted at this -time to ad.just for. new intelligence data that has since b.ecome available. (S/Nl:1:'/S5,-2) For- this evaluation, the data base was subdivided,. i.nto 6. m,ai,n project types: . (I Sc.i ent i f i c and U.3/71-) (2) Counter-terrorist (CT) - (3) COLAI cs (CN) (4) Cot.,tr-it(:.~ri r-it.el' gericLa. (C'), DO L ei t, Mc)c: Cont) ~ ai A ~6) pr-e-clicti..'Ve (pred Tbt al 'pv-oj6ct Is w(.-jr-k:ecl for- these c,~:tteg(:.)r-J.(-?s ar-e shown in Figure 1. .1 - - -a--ojec.-tc IS /I 1\1 1- / ~3 S - OT 2CM) r. S (:)r-,e--hal+ c-arir-iot be evalua-ted since gr-OUr-,d truth is not suTffic-Aently knowri at this time. For approxifriately one- Of gl"C-ILAnd tF-Uth is known (or- highly cc.-m-t.ain), and +or- the ot.1-ic.-w- ot i e - + c) u 'r-t["t, gr-OLAnd tF-Uth i.S:~ Only 1.1-1-1cmn but suTH-icient 4:cjF- making a F-easonable ((_::)/NF/S~:3---2) E)oime~ of these pr--c)Je(-.::t categories can overlap. (.Ja-tE.--t is, Fkl~.:io an asp(--!~Nct Of Most of -the CN, c)-f ti-le C.T a few Of the S/T l::)r-c.)Je(:::ts,. The prediction (.-a-teg(.)F--y i1-1 J. l..)r-J.(-,iaF--JAy -to pr-(,_~clicticins of a r-tatt-ki-e. ]'I-) 4:LttLkr'-e aj talysis, py-edictive data will W:- (,?valuated as a S(:~J_),5ff-ate aiiiipect Of -the these project C a t g (::) r- i e aii . Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 Approved For Release 2000/08/08 CIA-RDP96-00789ROO0700260003-4 CD co ui Ij 0 cc CL 0 CF) 00 a) (D OD Lij ~3 0 cc CL z 0 z z -j0 z 4 z :,~:P 0 cc z < 0 ~e a.z 11~"; ,4 Ci) C> C> C) C) CO C> C> C> t- CJD U*) "t CIO C4 SIDWOUcl z Ir 0 LL 0 z Approved For Release 2000/08/08 CIA-RDP96-00789ROO0700260003-4 Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 2. 1". VAI LATION (U) S / N I::-- / G E) 2) TecI'-ini(:.1LkeS.~ Ui'xied +or evalUatil-Ig the 13UPA STR'EAI---" and (::.A.mulated operationa'.1. database depend on the ri,Fd'_-t..kre o+ the task and type (::)-f project. S/T' projects are the to This di+-ficuity arii_i~es -from the general comir.:0.exity existing at most S/T' target sites, -F-rom possible affil--i-iguous aspects o-f known ground truth about -the -target E-i-ii-te, -Frofn the 1 latUrCi? c-.)-f the ii--i-Formation desii-ed, and in a few cases, poi:.,sibly -fromi thie. viewing (RV) targeting method e fil p 1. C) Yd . It is easier to data on G/T targets i+ only a sing1c.--a iGSLte, such as presence or absence (o-f a particular- systc,m-,, +or example) is desired, then it is to evalu,~.-tte how well a viewers' detailed but possibly +ragmentary description corre-ilatei..; with asr.:)c-~c-ts of a coiy-iplex site. In this case a ammlllt C.14: Subjectivity can be i nvol ved in evaluating t 1-1 e (;:I gr e e + d a t a j: a r g e t c o r r e I a t i o n . fC") ~'assj:.S I d Li'c!`*.'L'ng-. o'v 61W' a. -su j6C.'%~l-, "'t GG - n r e VI Y a evaluating c(::)mplex S/T targets, the viewers' data is examined and compar( d to ground truth with several data categories in mind. IT-iese categories are shown in Figure 2, and include geographic descriptioris, large and small scale objects, 'Large and small c data. Not all s -ale +Lkllctions, p6rsonality data, and predictive these categories may be relevant to a speci-fic project, and. in.. ~:X)me may even be. part o+ the RV targeting pirocedure used (e.g., when a photo o+ target building is used as an RV targeting re+erence +or accessing its unknown contents). ate data US/Nr- SS-'42) 0 + t e J.. d e n t i + yi n g h e :'. zkpp:ropri category, the i ic.-.nt step is to exaf-fdrie the viewers' raw (or summiariz(.7-ml) data -for- compariscin to or 2Stim,~-tted (-round 3 trUtI--t ~.-trid t..(D fy-ij.-tke J Ltd gerf-,c--.-..j it art what approximate degree O-f d a t. a C" C) F" F" LDI. At J. CA actt.xally C-.m."ists. F-igUre .3 defir,(.:~s the scale ratii-igs uined ~.klui-ig with their ~.-%ppro.ximate degree oT' data r r 1. z--k t i (_-) n A 1: F.) e n (.Ji x .1 containi: detailed instructions +or ji-t i ialyi::~t when reviewii ig ti ie d~_-ita) Fi n a 1. ' evalt_lati(:Jn~5 al-IC-1 !'-_;Lkmmaries are prepared by the Progrart-i 11ai-iager ai ic.:I/or his project representative (who is not part o+ SUN s-rREM:-`. sta++) in conjunction with the responsible or ar comrf-lul lity J::)C.Ant-o4:- r, -L..z.-L c t I ateas.t.... data i,'-.krid reports on that target ~--J.te alcA::) i-eviewfDd CJLU"-i1_)~.J t[Iiii.i In some cases, area Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 DATA CATEGORIES Q DESC, MPTIONS e.g., TERRAIN FEATURES; WATER, MOUNTAINS C~ 1 R G;: E) - - e.g., AIRFIELDS, FACILITY- 0 LAYOUT, DOCKS, SIL S -C*`_S",_-iAALL SCALE) -,.e.g., LASERS, TANKS- ANTENNA S IL.ARG~".. SCIALE) e.g., R/D, WEAPON TESTIN G PRODUCTION K) NJ S"SNPIALL SCALE) e.g., LASER TESTING, CBW STORAGE, MISSILE FIRING i;~]- Y 0 Ae g STATE- OF -HEALTH, PHYSICAL'-.., FEATURES, BACKGROUND .12! U I i~41 e.g., PLANS, INTENTIONS, FUTURE LOCATIONS FUTURE ACTIONS/ ACTIVITY SPS353-13 EVALUATION SCALES APPROX. DEGREE OF NUMERICAL DATA CORRELATION RATING PERCENT DESCRIPTION 0-10 LITTLE OR NO CORRELATION 1 10 - 3 0 MIXTURE OF RELEVAINIf'..'. AND INCORRECT DATA 1 + 3 0 - 5 0 WITH MAJORITY INCORRECT 2 50-70 ( MIXTURE OF RELEVANT.'.. AND INCORRECT DATA WITH MAJORITY CORRECT 2 + 70-90 ANDUNAMBIGUOUS 9 0 - 10 0: VERY HIGH TARGET CORRELATION WITH ESSENTIALLY NO AMBIGUOUS DATA SP5353.11 Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 SG1A SG1A SG1A SG1C S,G 1 B. i t t c -1 a r 'a i , I s zrtn,~.Alysts and -the :1c provide wr -i pp La to assist i.r., the +ina:1. evatuation pr'oces,'_.-~. evaluations are rez.cordecl C.)CII SUUI)m~..~ry -f-cw-fris and are friaintairied iri the F-'rcjgrzmTI files, C1.3 / N F, /11:3 G 1.1 ex ~..x ni C) + a n 13 T t a r- g e t e va 1. LL ~.A t i 0 1-1 i S j. UI 4. :In -this case -the -tar-get site was M SG1A m The project Was COMPIEFted IFI JU 37 ,~%.incj involved Tcm..tr viewers (2 prc)ven aric.1 2.' 1 ovi ce.-O . 1'. n t hi j. s eXaMple, -the brZkCJ%et. ( ) indici.-Ateiiii a best (.-.;!stiTja-t-.e of data validity was for tl--iztt clata c:&.-ttegor-y since ground truth is not yet totally known +c)r~~~ A " d as h mc?ans -that data categqry was j jot in the data. One of the viewers (#101) attecripted -to desc-ribe thc.~? site 6 fifionths in the the UtUre 3'ri t I-i . i i:.i Some the A - ca-.~~,gprj es,.,. o c dat-L .66~t u ...... . . ts~ edgr.ap h r.es,j ar-~g.q.,.46al.e,' a r (,~ iciipcirtar-itan intelligence data v i ew p l c'e n c) -from o i n S t iey t' are kri6wn +romi conventiorl'al c-allection ts. However , the asse vic--)wee-s'data in these ci..ktc.~gories are ded in -the valuation inclu e -For-- projec-A sinc(they tend to provide confidence that other this (as yet unknown) details , correct. in the data ffi~ay. be (E-)/114F/Sq--2) A 6-iore'~ sophisticzzktv.~d analysis methodology has been c_I(--WelQpeC'I by GRI -for use in evaluating complex, projects. This method's main 'advantage is that it allows qL.(~.--tnt-.-L+io._d estimates -to be made for each and every data element generated 1::)y the viewer- with rcic.,pect.. to both actuz.0. target ~existeric(--~ and ify-qiortaricc:~. ''417his tc.-~L-hnique iis currently being e,,.,z-.--;mj.r,ed use in the F)UN STREAk' program and has been applied zt fc.--...-w projects. However, D *:., a labor intenc-s-ive te-achnique be used (:ji-',I.y high it iterest pro-jects t. Ita t w j. 1. 1 in +UtUre. US / IA I.--- / (.3 S -- 2 111c)~.ft of the other- SUN STRI':.-_'AI< V:)r-o-ie(:::-t types do iic.)-t recju'ire a cc.-Jfrlplc~~x mc.-A-_ hod o'.1. cq Y. For example, due to the nature of what t,~I::) e of data ici (and availability of collectic.-m assets th~.rkt can be cuec.1), most of -the CT, CN, or predictivc-.~ projects, ground trUtI I is I...nown can be evaluated in a "t-ilack. or whitF:.," manner. The viewc.:-!r-s' data, ever-, if not ac:-A-.-.ec:I t..q:)c-jn, ei.t1--ic.~r- with the ~-:iubc5eq[.AC:?I_)t locatiori of the fugitive or ship, c_-)r J.-L dcjc~~iF, not. T'hic_-.~ event pr-c_-)d:LctecJ did, or C:)VeI"-Zk.I..1, for- iyiEiiny of thiese a ~~t criat.-t-c-,r c)-f c--cn-triting hlitE.-~ and Hi t raticm,--i or- of iForm the basis of overall. data G madc.:~ iri this rej::)cjr-t for- types of p r. cj jc... f --Vj- x Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 SUMMARY DATA EVALUATION SHEt 0-3 -SCALE DATA CATEGORIES JN.F jE d PPOJECT COMPLETED SESSIONS SOURCES b. C. d. e. f. a 9. OBJ. OBJ.FN. FN. GEO F PERS PR N. . . . L.S. S.S.L.S. S.S. . L.S. S.S 8701 11 JUNE 87 -12 003 2 2 1 1+ 1 ols 0 1 0 021 1 2 (1) - - 101 1 2+ 2+ 1+ 1 v 6MONTHPR.:-` 7"~]'!~ Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 A(J(::Iit-.i.(::)nal data ii-.inalysiis is, of to IIOW ( '-IOSC`~ t.C.) (..JF-(,)L.kl'i(:j trt..tti-i data iactuai'ly was. Th i s m ay be of value in hC)W to COFICIUCt -FLAtUr"e 1:)roiects of this ri j.-t t. u r e . 3. E'VALLATION RIE'C"]3ULTS (U) ( S/ N 1*':'/ F:) S -:3 Over-all data correlations for all SUN STIRE01<'. operational and operational SifflUlatiOCII pr(--ijects performed since 1.986 are shown iri Figure 5. results were obtained by averaging the data entered on the Summary Data Evaluation sheets for eac-1--i V.)r'-(::)jec-L primlarily for two data categories (i.e. large scale and small scale object and function). 'The top lines on -the bar- chartsi reflect data zi-tverages obtained from the 1::)roverI or e-q-)(~-_%rienc.ed F-or c5(:.jme projects, especially some of the CN arid -CI 1--it tI i(::~ dis.-A-Arictiori between large scale and small sc,~.~Ie is not Clei-.ArCL(t,' +Urthermare., t,htis differentiation m a y no.t. . ..' t .(n s f projects, data correl,'-_Ations W(._~Fe based mainly on a hit/miss C" a 1. C LA 1. a t i o n S / N F / S S FigUre.-! 5 indicates that, on the average, data f rom ~proven SUN STREAK viewers for- S/T projects will tend to have a 20 percent correlatJon with gr'ound truth for small scale -targets, and a 3'0 percent to 50 percent correlation wit.h grOLAI-11CI truth for large scale target features. Likewise, 4,or ur or.- CN. projects, about 20 percent to 50 percent of SUN STREAK data W'ould L')e e,--'pected to cor-'rel'ate, with q .r C) Lk rid tr'L (th Cautio n muf-.5-t: be exerC'ISed.in i rig this dalta, however, since the d a t a 1::) z~.-t s w i lz. 1-1 k r, W r) - g F- C) Ut i d t r u t h i S' 'tJ LAI t C-2! .1. 0 W i 1_~ t 1-1 e S 6' c a L. F..-! g C:) r s For CT and 1)(::)(:::LtrAc.?nt Reading projects, sukl s-rREAV* i _lvel with (J t a 1i(:)ws a A.C) it to 60 1( gF"0UI'I(.J '[~.'.F_Uth, I-iowever (::)-f: t.his datj--kL)ase is also quite. 1. (::) w . IT'ie DCDCA.kment F%'eadirig were, however, carefully isolated and in a 1.::nown or- location. Pr-(-_:~dictive. data o 4: t. I evei it type (e. g. itical /mi 1. itary EiiitLkatiOnS) sl I (-) w ci a v e r y I (:-.)w d a t a C.. Or r' (.-a .1 a ti ri i r-.e 1. i a bi 'I i t y) o f ab 0 LA t .10 p e t--.(:::erit or (S/NF/SS-3) If all SUN STREAV', averaged projects are as show n in FigUre 6, ata (:::orrelationwould range from d t. -ev"(:::erit 4:(::)F" !HMU-11:1.i..i~:1-e(:::ts LTO 1--)er.cent 1::~ (::) k..t to about for- 20 1.) I a r- g e scale Wl-iilc..~ aver-agi rig " SUCI-11(yii-.-.--ty data indicate results i l I tl-,(...-~ aver-agirig t ei it::k:; to 101")g FLAY'), WaShOUt t I results have Gi.-II-IgUlar'high su(:Ji as I- Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 OVERALL DATA CORRELATIONS...,. 80 z 60 0 U. i= 04 z 40 LU cc p 0 CC ui 20 0 S/T CT LARGE SCALE SMALL SCALE (TEST FACILITY, AIRFIELD) (LASER TEST, ANTENNA DISH) 10FORN SP5353.02 OVERALL AVERAGES FOR ALL PROJECT PERCENT OF DATA CORRELATION 60 50 40 30 20 10 0 LZ:.-.:-. NOFORN ~TYPES ALL PROJECTS SP5353.01 Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 identi+i(:::atic)n ~;evc-il-al mofiths-ii in, advanc:e of a s.;pecj.+j.c-- area in I A .1. - - e , SUN -L 1-i e LJ G w h e, r c.~ a 4: u g i t i v e w a s I a t ca r + o (..t n c .. 11.1 thiS CZAS STIRE.1011< (Jz,.tt,..A wal:ri I-KA aCted LAID011, +0r'tUI-),FAtely, the 4:ugitivc-_A was f-i ever t h, cA, a1'.')C.'Ik_kC...tec:I in this arc_-?a -to the alertne;_-.;s; o+ a loc:al lzXV-4 ei i4c.)i-cement of-fici,7.0.. G) / PA F- /,1:3 S - '.3) Another way o+ (-_--onsi deri fig overa'll SUN STREW::: d~.-,.kta correlzixtion isr, to cons;ider only the I:iroven viewers. ri-ii,z data is; si-iown in [-igLAre 7, +or timcas wifien thes~e experienced viewers.; received a 2' or-- 2+ in the nuviierical ratings assigned to their data c-_-c.)rre'Latic)ns~. Only two types~ o+ data -Are presented ., 3 /1", 'AF- iUS CT hr e - Ec and Personality data as., obtained +rom the vz CN and CI projc.~~cts. For s/"r projects, proven viewerS would be e.xpec-ted to receive a high (i.e. "/(-') percent) data c-orrelatic-)n. rati.rig on about 20 1DE,.,rcent of -the S/T projects ii.-.-t t t e m 1:-.) t e c-I F-ov` Pc.,v--s,,oi i;.-t 1. i t y 1-3 r a i ec t s ( j. . e. , I-Dac- 1.:: g r OUn d , v;t at e-o+ - t h e-1-i ea,'I. I: h ,sj:)L,-~cific ac--tivities) , around 50 percen't -to 6C) percent 0+ the projects woul,d yi el d hi gh resul ts. - Essenti al 1. Y, tI Ii S f a r.t;" 'rf-4imn. - s-tre 7' -th p es:!int. SUN I g~hq_. Wea messes P 'TREAk: sit-+f and sugge-AiF~ jects~ n -foreign or C 1. 3 cl - that morc.? p r o 0 F targe It persona]. i -ties are warranted. ~3 Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 EXPECTED RESULTS FOR APPROXIMATELY 70% DATA ACCURACY SELECTED PERSONNEL 80 70 cn 60 w n 0 50 cr (L u- 40 0 z 30 w 0 cc w 20 IL 10 0 Z.-. Z T X 0 F 0 R N S/T PERSONALITY DATA PROJECT TYPE SP5353.03 Approved For Release 2000/08/08 : CIA-RDP96-00789ROO0700260003-4 .1.1.1 F'INDINC3S (LD (S / 1`4 F* / S'._i --:') Althot..tgr, ti-ie over--all data cor-relatioi F, p r-- o vi d (J iri the previC_')Ui-S section hiave, in instances, have a '.Low over-zill aver**~,-;ge, th(.-_~ r-eSUltS are UlliqUe enOUO-h to WaF-rallt attention and conitinued GTREA4--:: activity. Even in thie lowest I-elial::)ility case (i.e., pr~edic-tive), identificatiori of even one .mj::)(_-)r-tant +Utur-('_~ event out of ten could in fact be high' i I y s:i.onj.-TrJ.(-.:ant for-- cost (:,)r- life, saving. These " aver- ages " do not do justice to tl'ie single ur-aclue cases that cost little -to act upon, as in the case of ti-ie fugitive locatiari cited in section 3. (,S/NF/S',3--3) Sp(:x.',cJ.+ic findings that F-CDSUI-ted -From in-depth r-eview (--)4: eritif-e (*.:')UN G)TREA1,.' data j.1 IC 1. Ude 0 1 ndi Vi dual s' per-F or'fflan,ce corr-el ateii~ wi th - proi ec t t _P Y W-77b . .... . ad S s i t,t"e'd A. O-ttpr mai. 'c": 1"13. n g , -and' 6 A. I c ~~'at ons wo V a -data Yorve" 63 d be expe ct ea .0 impr-ove in the future. 0 SLIN ST1RE-A1-:'.' has a distinct poteritial -for- diF--ect . . .. . ................. . ........ ....I............. . ....... . ....................................................,..", . ........... contr-ibution. to cer-tain Cl: CN ai id...C.T cases, as burn out by speci+ic instances.. over- fl-i'd past.two year-s. 0 Fredictive dat ;IjIg tnder certain....conditions . ...... . ..... sUc~_a-s__n,_,_e_ar- ter-m events or situations that do not involve c-omplex 'interactions. t -IOU -i~ -i r- o `c taiifl