AD-A250 786I llIWIlIl 11111111 11 1 ll! 1111llI~l

TOWARD ADVANCED HUMAN RELIABILITY PROGRAMS

STRUCTURAL DEVELOPMENT CONSIDERATIONSAND OPTIONS FOR

EXTREME RISK ENVIRONMENTS

DTICS LECTESM\Y 27 1932A L

Office of Emergency OperationsWhite House Military Office

May 1992

This document has been approved

distibution is unlimited.

92-13629

92oll

-1 . ~ U U I w OS a~g ma mwv

1. AGOCYWE OhIY .Up) L REOR DATE 3. FWT TYPEAND DATE@ 01WR

IMay 1992 Basic research: 10/90-6/91

4. T1TLEA0OWITfLE & 5. N N30hSUMERTOWARD ADVANCED HUMAN RELIABILITY PROGRAMSStructural Development Considerations and Options for ExtremeRisk Environments

&AJ1WORM

Richard C. Nelson

7. PERMW00OWBM~TiW PM)AD0DOWM L. PEFOM W ORGMA"TONOffice of Emergency Operations RPR AE

white House military Office - '-

Suite 600, 5113_ Leesburg PikeFalls Church, VA 22041-3230

REORT MUER

11. affftweed" NOTES

Approved for public release; distribution is unlimited

susai susata oseytmstb bet b eiduont opee hi iso

Ioperational populations exposed to extreme risk

environments (ERz-s) ight expect to

or missions regardless. Existing human (personnel) reliability programs are inadequateto assure that personnel capable of meeting both the necessary security and operationalrequirements are available for response to such conditions. This study explores anumber of issues to consider in building a robust human reliability program (HRP) struc-ture capable of supporting single to multiple operational populations, scenarios, andmissions, using any of several program structure formats. The HRP structure format maybe used within a single agency or government-wide.

K411 WTIUN reliability, human reliability, personnel *reliability,14extreme risk environments, reliability programs, psychological ia plGEC0reliability

17g60CWvTE ANUMM * 1a6NMTVLAWWnffcr~ 1pUTOD.GAWIWT 2L LN fATCN ~W TOOP= OF @PW AAE OPANTMCT

tJNCTASSIFIED I UNCLASSIFIED I UNCLASSIFIED unlimited

NS70I MO SS~mdU-m W MD4

1

Table of Contents

Executive Summary .......................................... 5

Table of Exhibits .......................................... 7

Section I. Introduction ................................. I-1A. Acknowledgement ................................. I-1B. Tasking ......................................... 1-2C. Concept ......................................... 1-2

Section II. Methodology ................................ II-1A. Baselines ...................................... II-1B. Study Design a, Execution ...................... 11-2C. Regulations .................................... 11-3D. Research Team .................................. 11-3E. Interviews ..................................... 11-4F. Abstract Services .............................. 11-5G. Information Sources ............................ 11-5

Section III. Stress and Human Reliability ............. III-1A. Introduction .................................. III-1B. Background Issues: Stress and Reliability

Conceptualizations ........................... III-1C. Human Stress in Extreme Risk Environments ..... 111-3D. Decision Making in Extreme Risk Environments..IIT-4E. Team Decision Related Issues and Stress ....... II-8F. Cohesion and Infrastructure .................... 111-8G. Coping in Extreme Risk Environments ........... III-9H. Interpersonal Conflict ....................... III-10I. Personnel Selection and Monitoring ........... IiI-11

1. Behavioral Issues ....................... III-112. Performance Issues ...................... 111-123. Physiological Issues ..................... 111-12'. Psychological Issues ..................... 111-125. Monitoring Issues ....................... 111-14

J. Implications of Extreme Risk EnvironmentsOn Operational Personnel ..................... 111-14

Section IV. Behavioral Reliability Issues InExtreme Risk Environment Contexts ................... IV-1A. Introduction ................................... IV-1B. Current Position ............................... IV-1C. Future Considerations .......................... IV-2

Section V. Performance Reliability Issues InExtreme Risk Environment Contexts .................... V-IA. Introduction .................................... V-iB. Current Position ................................ V-iC. Future Considerations ........................... V-2

2

Section VI. Physiological Reliability Issues InExtreme Risk Environment Contexts................... VI-1A. Introduction ..................................... VI-1B. Current Position................................. VI-1C. Future Considerations............................ VI-2

Section VII. Psychological Reliability Issues InExtreme Risk Environment Contexts.................. VII-1A. introduction.................................... VII-1B. Current Position................................ VII-.C. Future Considerations........................... VII-2

Section VIII. Monitoring Issues InExtreme Risk Environment Contexts................. VIII-1A. Introduction................................... VIII-.B. Current Position............................... VIII-.C. Future Considerations.......................... VIII-2

Section IX. Summary of Interview Results................ IX-.A. Introduction..................................... IX-.B. Areas of Agreement............................... IX-.

Section X. Models for Multiple Population, Agency,Level, and/or Scenario Extreme Risk EnvironmentHuman Reliability Program Structures................. X-1A. Introduction...................................... X-1B. Selection Strategies.............................. X-1.C. Task Analysis..................................... X-2

1. Orientation.................................. X-2D. Support Personnel................................. X-3E. Testing............................................ X-4

1. Requirements................................. X-42. Validation Processes......................... X-43. Testing Process.............................. X-7

F. Models--Front End Selection...................... X-10O1. Introduction................................ X-102. Sliding Scale............................... X-103. Layered Model............................... X-1.24. Modular Model............................... X-1.35. Mixed Model................................. X-146. Nondeployment Model......................... X-1.5

G. Monitoring Models................................ X-1.61. Introduction................................ X-1.62. Continuous Monitoring....................... X-1.63. Periodic Monitoring......................... X-1.74. Aperiodic Monitoring........................ X-1.85. Mixed Variable Monitoring.................. X-1.96. Monitors and Process........................ X-1.97. Retesting................................... X-208. Intervention................................ X-21

H. Recommendations.................................. X-221. Strategies.................................. X-222. Application................................. X-233. Evaluation.................................. X-23

3

4. Models--Front End Selection ............... X-245. Monitoring Models ......................... X-246. Intervention .............................. X-247. oversight ................................. X-25

Section XI. Other Issues and Recommendations ........... XI-1A. Introduction ................................... XI-1B. Stress Training ................................ XI-IC. Crosstraining .................................. XI-2D. Motivation and Responsibility .................. XI-3

1. Contracts ................................. XI-32. Grandfathering ............................ XI-43. Ergonomics ................................ XI-5

E. Phased Response ................................ XI-6F. Nonobtrusive Evaluation Techniques:

Voice Stress Analysis ......................... XI-7G. Personal Relationship Issues ................... XI-8H. Alternate Scenarios ............................ XI-9I. Community Coordination for Fit ................ XI-10

Section XII. Summary and Recommendations .............. XII-i

1. Recommendations Summary .................. XII-i

Appendix A. Interview Formats ........................... A-i

Appendix B. Terminology ................................. A-7

Appendix C. Experts Interviewed ......................... A-9

References ............................................... R-1

Accesion For tNTIS CRA&IDTIC TA3-

Jilsofication,

B y _ _....... ... ... .... ......... .......... .

y.................

i , J'a a 2 or

Dis

I1-~ I

4

This page intentionally left blank.

5

Executive Sulmary

This study was undertaken at the direction of MajorGeneral Robert M. Marquette, Jr., USAF, Director of theOffice of Emergency Operations (OEO), White House MilitaryOffice. The study explores the feasibility of establishinga personnel reliability program for use in extreme riskenvironments (EREs) and identifies issues pertaining toprogram implementation. Upon review of background materialand existing military and intelligence programs, it becameobvious that:

1. Existing programs are inadequate to provide thenecessary degree of assurance that they, and theirpopulations, would be able to accomplish their purposes inthe event of exposure to EREs.

2. Unclassified human reliability programs are notrequired to be functional during an extreme risk environmentsituation.

3. In the absence of specific scenario or threat, itwould be necessary to examine application of a programstructure or framework from the standpoint of usability in asingle on up to multiple threat/population/agency approach.

A multifaceted approach was taken to find or developsolutions to the problems noted above in order toeffectively conduct the study. During the course of thisnine month, part time effort, a multi-agency team wascreated to identify professional sources of information.Two dozen psychologists, psychiatrists, and otherprofessional specialists from a number of Government andnon-Government organizations were consulted both to obtainthe most pertinent information in highly condensed form andto identify academic sources of information. Six nationalscale databases were thoroughly examined for relevantinformation and reports. Over one hundred scientificpapers, six books/compendia reports, and nine FederalGovernment regulations were obtained for backgroundresearch. Several key conclusions emerged from thisprocess:

1. Without a reliability program designed to deliverpersonnel specifically selected for operational service inparticular ERE scenarios, there were minimal chances thatenough personnel would be available, or able to function, toprovide an effective response to the ERE.

2. Scientific research has not been conducted in thearea of human reliability in virtually any EREs.

6

3. Measurement tools or instrumentation forapplication in EREs has not been developed, but can be.

4. No existing reliability program integratesoperational functionality and human performance issues inEREs with suitability issues in a context applicable to morethan the narrowest and most limited of mission applications.

5. There are a number of subtle factors affectinghuman reliability that must be examined in detail andintegrated into a program as a whole in order to provide thegreatest assurance of success in EREs. These includedecision making in extreme risk environments, team decisionrelated issues, cohesion and infrastructure, copingstrategies, interpersonal conflict potentials, and others.

6. There are also a number of subtle factors that havegreat validity in some or most ERE scenarios that shouldalso be considered in detail and integrated into a programas applicable in order to provide even greater likelihood ofsuccess. These include stress training, crosstraining,motivation and responsibility factors, phased deployment,personal relationship issues, and others.

With all these factors manifest and bearing in mind thevariety of potential candidate populations and theirspecific needs, a number of highly versatile architectureswere developed to construct a program from. These includefive front end selection and four monitoring models, anddiffering validation and intervention mechanisms. Thestructures and models will all accommodate at least somereconfiguration as well. The format is designed to befilled in with specific tools to address the specific needsof particular populations and scenarios. Sixteenrecommendations are made in the text and are summarized atthe end of it. A seventeenth would be appropriate to placehere, as it is the most important of all:

A human reliability program should be developed toensure that adequate numbers of personnel capable of meetingboth the necessary security and operational concerns areavailable for response to extreme risk environments.

7

Table of Exhibits

EXHIBIT PAGE

1.Model of Stress and Performance..................... 111-6

2. Precursors to Personal Injury....................... 111-7

3. Sliding Scale Model.................................. X-11

4. Layered Model......................................... X-12

5. Modular Model........................................ X-13

6. Mixed Model........................................... X-14

8

This page intentionally left blank.

I-1

Section I. Introduction

The following study was prepared for the Director ofthe Office of Emergency Operations (OEO), Major GeneralRobert M. Marquette, Jr., USAF. Its primary objective is toprovide a vehicle for lay decision makers regarding theutility, applicability, development, and coordination ofhighly advanced, extremely flexible, human reliabilityprogram (HRP) structural formats designed to factor for asignificant number of pertinent variables of both securitland operational significance. It does however also seek toaddress issues, configurations, and coordinationrequirements of interest to the scientific community aswell.

The complexity of tne task undertaken in this volumewas increased by the requirement to simultaneously addressthe disparate needs of a variety of candidate populationsexposed to particularly hostile conditions where substantialcasualties might be anticipated, but in which personnel wereneeded who could meet both significant security requirementsand also perform under conditions they and/or those close tothem might not survive--what are defined herein as extremerisk environments--in a single HRP. Therefore, althoughgeneral areas of concern that apply to all potential extremerisk environment (ERE) populations are addressed in thefollowing sections, models for this study are designed toaddress multiple population needs and formats in a single,overarching HRP format.

It also became apparent as research on this studyprogressed that conventional personnel (human) reliabilityprograms do not explicitly address consideration ofoperational performance factors in the security screeningand monitoring process. These factors have been addressed inthis study as part of an integrated program.

A. ACKNOWLEDGEMENT

This report differs in approach from most others on thetopic in that it relies heavily on the practical experienceof a number of highly talented people, many of whomcurrently operate in or upon human reliability programs.Additionally, the capable support of United States ArmyMilitary Intelligence personnel was available to identifypersons to interview for this study and to assist insearching out reference material. Persons who participatedin this effort are included below: all have heartfeltthanks for their support, and in many cases, enthusiasticencouragement. Thank you all so much:

S/A Lee Anklin, USA

1-2

S/A Donald Brenno, USAMr. Mark CentraCAPT Chandler, USNMAJ Ronald Chapman, USAFS/A Ann Clawson, USADr. Kent CrawfordMs. Lorna DodtMr. Pat DowdDr. Douglas Eddy, NTI Inc.Dr. Joe Fagan, USADr. Edna FiedlerDr. Edwin Gerwell, USAFDr. Harold GinzburgDr. Fred Glogower, USNMr. Dewey GoffDr. Allan GreenwaldDr. Martin KurkeMr. Richard McMurrayMr. Robert OppenheimerS/A Michael Parker, USADr. John Patterson, USAFDr. John Plewis, USAMr. John ReardonDr. Richard ReesMr. William RobbinsDr. Walter Sipes, USAFMAJ Gail Stark, USADr. David Strome, SRL, Inc.Dr. Michael WigglesworthDr. Martin Wiskoff, BDM Inc.

B. TASKING

This project was assigned on October 11, 1990 toexamine feasibility of development and implementation ofHRP's for multiple candidate populations. Additionaltasking elements supporting, amplifying, or clarifying themain themes were added by the Director, OEO at In ProgressReviews (IPR's) and have been listed as appropriate in thefollowing material.

C. CONCEPT

The operational concepts behind the examination of ahuman reliability program for use in EREs differ in a numberof ways from standard department of Defense (DOD) PersonnelReliability Programs (PRP's), although there are somegeneral objectives in common (the selection and retention ofreliable personnel).

One major point of variance between the OEO HRP and DODPRP concepts is that DOD's is specifically stated in DOD5210.42 to be a peacetime program; in time of conflict, it

1-3

may be set aside with additional manning/reinforcements notnecessarily having to meet the same strict standard as theoriginal staff. The concept presented for this study, whilenot necessarily intended for use in a combat environment,holds that due in part to much smaller numbers of personnelinvolved and the magnified damage a single individual mightthus cause, in combination with their use in environmentsthat may be fully as hazardous as direct action tacticalcombat, original personnel may have to persevere withouthope of reinforcement or replacement, and any reinforcementor replacement personnel would have to meet the sameselection criteria as original personnel. In short,functionality in EREs is as essential as mere survival.Unlike the military, this may mean a short/medium termrequirement to shift existing assets rather than acquiringnew ones.

There are other issues to consider. Military PRP's aregoverned ultimately by compliance with DOD 5210.42 ofDecember 6, 1985, Nuclear Weapons Personnel ReliabilityProgram. This policy document mandates in a fairly generalway the requirements for selection and retention ofpersonnel in military PRP's. Standards for selectioninclude both qualifying and disqualifying factors.Qualifying factors stress physical, attitudinal, andbehavioral characteristics, all to be decided on a purelyjudgemental individual basis. There is neitherclarification nor explanation of how or what to assess, forexample "evidence of.. .ability to exercise sound judgementin meeting adverse or emergency conditions" (DOD 5210.42 ofDecember 6, 1985, Enclosure (4)). What type of conditions,and under what circumstances? The leeway is presumablyallowed both for the quoted factor and the others in thereliability standards in order to allow certifying officialsto factor in additional variables to assess a multitude ofsituations. Similar latitude is granted in medical andpsychological selection, which may well be no more than arecords review for approval. There are a variety of otherspecific weaknesses in the selection process, such as therequirement for National Agency Check (NAC) as theinvestigative tool for Nuclear Controlled Positions (saidNAC's being adequate only to determine existence of criminalor certain other records, and not provide indication of anyother aspect of a person's suitability). In sum, however,the problem with DOD type PRP's is that they attempt tocover all foreseeable variables with a uniform programapplication and thus are inadequate for many specificsituations or population requirements. Additionally, othercritical factors impact the capability of such systems todeal with a large variety of adverse factors. First, PRPselection using the DOD screening proces is primarily donein peacetime and is based upon past indications ("the bestindication of reliability is past performance" per AFR 35-99(Cl), May 5, 1988, page 9) and current observations:stressors or failings are essentially everyday ones and

1-4

though these may be sufficient to "select out" someone withsignificant problems, they may not be adequate to addressthe additional extremities involved in threat to life or thelives of loved ones. Testing in support of DOD effortsconcentrates on predictability of groups in currentconditions, not in periods of extreme stress. Further, themajor components of the selection process, adjudication andmedical/psychological, are geared toward "selectingout"--accepting all personnel not contraindicated forapproval by specific interpretation of the generalstandards. Thus the certifying official is presented with avariety of indicators and expert opinions. Only he or shehas the "big picture" from all sources and thus thecapability for a comprehensive selection judgement, but itis questionable how often such an official (who is alsoseldom trained in the selection areas he/she has expertsfor) might disagree with lower level recommendations basedprimarily on approval of all whom meet individual baselinerequirements. Secondly, the DOD type selection processconcentrates on front-end loading, with by far the mostextensive review during selection. Monitoring relies onpersonal observation by designated personnel (all members ofthe PRP, in the case of the United States Navy, for example)whose training ranges across the full spectrum fromuntrained to expert, and whose rate of observation varies.Monitoring also relies on the self-reporting of participantmembers who may or may not be willing to do so. Thirdly,although some research has been undertaken regarding DOD'spersonnel security continuing evaluation programs (Crawford,Abbott) similar work was not found specifically for PRP's.

Although the general validity of the above remarksconcerning DOD's programs are relatively well known, and itsrare failures sometimes spectacularly so, in raw terms offailures in comparison to successes, the program must beconsidered overall to be quite successful.

No studies of the effects of life-threateningsituations upon DOD PRP populations was found in the limitedtime available for this study, but a nagging suspicionremains that substantial losses might be expected under suchconditions. Representatives of certain Federal agencieswith PRP-like selection/retention processes of their own alladmitted similar concerns for their own programs, andexpected major reliability problems in the event of asignificant life threat to their personnel.

For all the reasons above, it was felt necessary tofocus on what additional procedures might be added to orused in lieu of existing in-place operational methodologiesin order to acquire the types of human reliability selectionand retention tools demanded by varying populations andmissions operating in (an) extreme risk environment(s). Thebaseline concept was to take the best of existing humanreliability system procedures for the good that might begleaned from them and, knowing their limitations, seek meansto enhance those areas that needed it to better support ERE

1-5

requirements. Also, since the first steps in establishingan HRP must be acquiring the background necessary to makeeffective recomnendations and decisions about establishing aprogram and the structure or direction it might take,discussion of pertinent background areas would be necessaryas a means of establishing a more level and integrated basisfor the recommendations and decisions. As it became obviousin the course of this study how little existinginstrumentation was suitable for direct application underthe specified conditions, new mechanisms to develop theprogram-specific tools needed became a necessity. At thesame time, flexible formatting was considered essential tomeet the requirements of differing populations and agenciesor departments, and these were accordingly developed.

1-6

This page intentionally left blank.

II-I

Section II. Nethodology

A. BASELINES

The initial impression of existing HRP/PRP's gained inthis study was that they were geared primarily to abehaviorally and physiologically oriented front end loadprocess that served as an adjunct to enhance personnelsecurity adjudication. This perception was concurred in byall persons asked during interviews conducted for thiseffort, and all personnel who commented agreed that suchsystems were inadequate to assess reliability underlife-threatening conditions or stress. All commenting feltthat far more work was needed in the area of post-selectionmonitoring as well, and that current instrumentation wasinadequate for the task. Most of them did agree thatsuitable instrumentation could be developed, however.

The implication initially drawn from a juxtaposing ofthe initial assessment above and the requirements for bothdependability and functionality in variable EREs was that ifan overarching system could be devised that assessed theserequirements, it would also cover the necessary suitabilityaspects of current systems while giving assurance thatpersonnel selected would be able to maintain stability andcapability in EREs. Although no interviewee disagreedopenly with this analysis, it is now considered to be toosimplistic and oriented to evaluation of differentmeasurable criteria (performance vice suitability), based ona review of academic information (Driskell et al, 1990).Thus though performance does not necessarily includesuitability or vice versa, they remain separate keyvariables for assurance of the degree of reliability neededfor an assignment to an ERE.

No baseline parameters were established for this studyby OEO management, allowing maximum freedom to seekeffective responses. To focus the effort, however, thefollowing five parameters were laid down by the leadresearcher:

1. Determine whether available research had been doneinto extreme risk issues of dependability vice suitabilityaspects.

2. Determine feasibility of development of a flexiblescale for personnel selection using standard testing andevaluation mechanisms.

3. Assume the HRP structure developed would have toserve selection against multiple extreme risk scenariossimultaneously, and provide assurance of continuingsuitability and capability in each or all.

11-2

4. Determine feasibility of development andapplication of standardized testing method3 and flexibleevaluation criteria to a broad population and multiplecultures in a single PRP-type program.

5. Collect and integrate existing suitability PRPdata.

To these parameters the Director, OEO, added thefollowing during an IPR on March 18, 1991:

Determine how to: Implement a modular HRP;Implement with other organizationsoutside OEO;

Develop models;Develop selection criteria.

Determine what: Criteria to use with differentpopulations;

Type of development personnel areneeded;Categories (modules) would be used bypopulations;We're trying specifically to protect.

B. STUDY DESIGN AND EXECUTION

This study was intended to support development ofpractical applications for findings as soon as possibleshould a decision be made to proceed. The means chosen toeffect this purpose were through use of a multi-facetedapproach that would allow maximum data retrieval in alimited time. The information would then be subjected to anintegrational analysis and means and models developed tomeet baseline requirements. All findings would have to passthe tests of applicability, capability, and necessity foruse in multiscenario EREs.

The accomplishment of the task as set forth above ismade no easier by an examination of various classifiedthreat documents, as they are often far better at definingwhat is expected to be lost than what might survive.Additionally, the Murphy Law of Comb3t that no OPLAN(operations plan) survives first contact can be expected inany ERE to be absolute, whether or not actual combat isinvolved.

The areas of information to be concentrated upon inthis effort were those aspects of reliability selectiondealing with behavior, psychology, physiology, performance,and monitoring. The reason for focusing on these areas isthat they are the least standardized across the FederalGovernment, and because most other parts of the review andselection process are clearly defined, validated in

11-3

extensive and continuing testing, and mandated by Nationalauthority: these constitute investigative requirements andthe investigative and adjudicative processes. Further, someagencies currently have reliability selection and retentionprocedures in place in their organizations, although theirpractitioners freely admit that such procedures are unsuitedto selection to meet EREs.

Having defined areas of information to focus collectionefforts upon, it was determined by the lead researcher toproceed simultaneously along several tracks:

1. Collect current PRP regulations for backgroundmaterial.

2. Form a research team to develop information andhelp analyze it.

3. Conduct extensive interviews with personnelcurrently running or working upon design aspects of HRP's.

4. Conduct full-scale national database checks toidentify relevant research material.

5. Obtain and review relevant academic information.

C. REGULATIONS

As a first step, DOD and each of the services wereasked for copies of their PRP regulations. All respondedexcept the Navy, who advised that their manual was currentlyin revision. It has remained so throughout the course ofthis study, due to negotiations between DOD and Navyregarding what are seen as policy or perceptual issues,depending on agency orientation. The Army and Air Forceregulations are of course service-specific implementationpolicies of the DOD policy.

D. RESEARCH TEAM

A research support team was formed to identify andobtain human and record sources of information, and provideprofessional technical assistance. The latter capacity wasfilled initially by Dr. Michael Wigglesworth of the NationalSecurity Agency (NSA), who volunteered the assistance.Later on, Dr. Richard Rees of the Central IntelligenceAcency (CIA) was essentially co-opted for a similar purposeon an infrequent basis. Their involvement was in providingadvice regarding human sources, directions and relatedissues to examine, and in responding to specific questionsdeveloped from this research. U.S. Army MilitaryIntelligence was enlisted to help identify human and recordsources, and they eventually identified the bulk of people

11-4

interviewed during this study. In addition to leads theyassigned field agents like S/A Brenno, S/A Parker and S/AClawson attempted to develop sources of record and humaninformation on their own. S/A Lee Anklin assisted withanalysis of collection efforts at one point. The teamleader/lead researcher assigned from the Office of EmergencyOperations was Richard Nelson, who also prepared this study.

E. INTERVIEWS

Despite a significant academic effort made for thisstudy, the core of it was intentionally a series ofinterviews with professionals involved with PRP's or similarprograms, or with the study of aspects of human reliability.This course was chosen to maximize professional input, giventhe limited time available for this study. A standardizedinterview format was developed to assure all applicableareas were covered; this was later modified twice toprovide guides for certain special purpose interviews.Interview formats are contained in Appendix B (attached).Virtually all interviews were conducted by Richard Nelson.In practice, most persons interviewed had proposeddiscussion formats or presentations (sometimes withmaterial) that they wished to follow initially, and Mr.Nelson willingly accepted this approach as it gave fullerperspectives of the interviewees' understanding andpositions. Any issue areas in the formatted interview notcovered incidentally earlier were pursued after theinterviewees completed their presentations when such aprocess was used. Interviewees often covered half or moreof the formatted items unwittingly in the earlier portion ofthe interview. Other interviewees followed the formattedquestions, which allowed focused responses in minimal time.An extensive amount of valuable--and even more importantly,practical--information was thus gathered in this manner fromprofessionals in a variety of communities including clinical(behavioral) and industrial/organizational (I/O) psychology,psychiatry, from medical service officers, and otherauthorities.

The interviewees made a number of general and specificsuggestions and recommendations, which will be cited atappropriate points. All expressing an opinion were highlysupportive of undertaking research in the areas identifiedfor this study, though a number expressed dismay at beingtold an unclassified version might not be done, or that noaction might be taken on the classified report if any.

R E CO MN E N DAT ION

Compile an unclassified version of this study whetheror not a classified version is done.

The Director, OEO, verbally authorized the developmentof an unclassified version of this report on July 16, 1991.

TI-5

F. ABSTRACT SERVICES

A variety of National-scale databases were used in thisstudy to help identify pertinent material in related areas.These databases operate by use of key word/phraseidentification association. Interviewees also made a numberof valuable suggestions that assisted in developing relevantinformation, to include categories of information to pursuein the abstract services, areas that would be unproductive,and locations and accessibility of abstract services.Preliminary foci for research into pertinent, more closelyrelated information were research on the Israeli DefenseForces (IDF) psychological literature (suggested by Ginzburgand Rees), nuclear related psychological issues (Reardon andGinzburg), police/forensic psychology (Rees and Kurke), andaerospace medicine (Fiedler, Patterson, and Sipes). Thelead researcher conducted extensive research in the NEXIS,National Technical Information Service (NTIS), and Medlinedatabases himself, and was capably supported in searchingand acquiring information from the Defense TechnicalInformation Center (DTIC) database by Reference LibrariansDewey Goff of Strughold Aeromedical Library, USAF School ofAerospace Medicine, and Lorna Dodt of the Pentagon Library.Ms. Dodt also had an extensive check made of thePsychological Database through the DIALOG service for thelead researcher.

A more general but unifying area became predominant asthe abstract services were checked: the criticalrelationship of various aspects of stress to humansuitability and performance reliability. This in turnbecame one of the underlying factors studied in this report.

G. INFORMATION SOURCES

In addition to expert interviews and abstract servicesa number of libraries were checked to obtain previouslyidentified and new material: all provided at least someinformation valuable to this study. They included theNational Library of Medicine (NLM) and Pentagon libraries inWashington, D.C.; Strughold Aeronautical and HumanResources Libraries in San Antonio, Texas; and the U.S.Naval Hospital Library in San Diego, California.Regretfully, time pressures made it impossible to checkother libraries that appear likely to contain pertinentinformation, such as the Air War College Library inMontgomery, Alabama.

A selection of other valuable information often notidentifiable through abstracts or available throughlibraries was provided by the interviewees and others:

Dr. Ginzburg - copies of classified and coauthoredunclassified research papers (including one not yet releasedat the time);

11-6

Dr. Rees - results of classified internal CIA research;

Federal Bureau of Investigation - PsychologicalServices for Law Enforcement Symposium compilation;

Dr. Kurke - original and modified models of humanreliability in an I/O context;

Dr. Fiedler - Air Force Military Evaluation Test(AFMET) program and validation testing, co-authored researchpapers;

Pat Dowd - copy of William Archer Jones, Jr.'s thesison Evaluation of Voice Stress Analysis Techniques in aSimulated AWACS Environment;

Dr. Sipes - copies of the U.S. Army's GTA 21-3-4, 5,and 6 dealing with battle fatigue issues at the troop, troopleader, and company commander levels;

Dr. Crawford - copies of numerous unclassifiedco-authored PERSEREC studies, a compilation review for theDefense Nuclear Agency, and a Personnel Decision ResearchInstitute study of investigative interviews;

Dr. Wiskoff - coauthored research papers and a copy ofan incomplete personnel security databased interview formaton floppy disk;

Dr. Chapman - copy of Osterkeamp and Press' bookStress? Find Your Balance;

Robert Oppenheimer - background material and historyconcerning an earlier classified effort on a subject relatedto this study undertaken by another organization;

Dr. Gerwell - copies of pertinent textual materials;

A number of classified and unclassified manuals,regulations, and other guidance from a variety of Federalmilitary and civilian sources was also consulted during thisstudy. They are listed in the references.

llI-I

Section 1II. Stress and Human Reliability

A. INTRODUCTION

Given the wide range of experience, training,education, and viewpoints of those who will read thisreport, a general background of stress-related issues andtheir relationship to EREs has been provided in this sectionin order to offer both a somewhat more level field forcommentary and an identification of some of the conceptualdifficulties underlying genuine operational needs. It isnot intended to be all inclusive.

B. BACKGROUND ISSUES: STRESS AND RELIABILITYCONCEPTUALIZATION

For a topic that has given employment and investigativepurpose to entire disciplines for ninety years, stressremains a curiously elusive subject. A large body ofliterature exists concerning various aspects of stress andresponses to it, but as Barge, Hough, and Dunnette (1984)point out, "the concept appears to be defined in whateverway happens to fit the particular investigator's frameworkrather than by reference to any ongoing explication of theconcept."

Despite conceptual murkiness however, stress has longbeen linked to reliability--and not always unfavorably. Asdemonstrated by the Yerkes-Dodson Law (1908), to a pointstress is necessary to obtain optimal performance, afterwhich stress degrades performance. This law can beconceptualized by the familiar bell shaped curve, withvertical stress and horizontal reliability axes. Thingsbecome more problematic beyond the position of the curve ina given individual at a given place and time with a givenstressor, however.

In the absence of a common definition for stress, Bargeet al (1984) choose to identify indices currently used instress research. The indices fall into three categories:behavioral, psychological, and physiological, for which theylist characteristic behaviors or symptoms of stress. Theseare cataloged below:

Behavioral Indices of Stress

performance inefficiency psychopathyerrors/accidents self-destructive behaviortarget (non)detection sleeping disturbanco-sreduced productivity suicidereduced job involvement deterioration of relationshipsunreliability drug use

111-2

absenteeism alcohol useattrition hypervigilancewithdrawal impaired decision-makingsabotage

Psychological Indices of Stress

job tension job-related threat to well-beingboredom job (dis)satisfactiondepression apathyirritability fatigueanxiety frustrationresentment guiltneuroticism absence(lack of)self-confidence temper(lack of)self-esteem moodinessburnout loneliness

Physiological Indices of Stress

headaches coronary heart diseasedigestive difficulties brain waveshyperuricemia serum lipidsgout free fatty acidsrheumatic disease serum cholesteroldermatological symptoms adrenalinediabetes noradrenalineulcers serum glucoserespiratory ailments cortisolskin conductance amino acidsrespiratory rate corticosteroidsrespiratory rate prolactinhypertension thyrotropinblood pressure growth hormoneheart rate myocardial infarctioneliminative function difficulties

(after Barge, et al)

Thus they define the concept and problem of stress byciting its symptoms as an interaction between the person andthe environment. Their list is not, nor is it intended tobe, all inclusive. For example, one source adds trembling,restriction and narrowing of the perceptive field, longerreaction time to peripheral use and decreased vigilance,performance rigidity, and lowered immunity to disease(Driskell et al, 1990). Yet another source (Stephens, 1987)takes a somewhat different approach by defining stress as astate of imbalance, reflecting the unrelieved dominance ofeither arousal or inhibition, which is applied tobehavioral, psychological, and physiological reactions tothe environment. Stephens quotes Holmes and Rahe (1967)that "we caused tissue damage by just talking about a

111-3

mother-in-law coming to visit" to demonstrate potentialsensitivity in furthering his argument that the unrelievedarousal must be perceived as distressing. A physiologicaldefinition of the concept of stress might vary in a rumberof important aspects however, e.g., its definition as thephysiological consequences of a threat (Herd, 1991).Indeed, Benson et al (1977), citing the predictability ofhuman reaction to a stressor by involuntary activation ofthe fight-or-flight response concur that excessive orinappropriate arousal of the fight-or-flight response mayincrease pathogen or health disorders.

The literature also abounds with models of aspects orperceptions of the person/environment stress relationship,all appearing substantially valid for their purposes, butoften with little or no resemblance to each other. Theythus serve as additional clarification of the difficulty ofharnessing the concept.

Despite divergences of opinion or perception, anincreasingly unifying trend in the literature is to viewstress as the product of an interaction between the(individual) and the environment (Keinan, 1986). Thistendency was also reflected in the great majority of expertopinions obtained for this report.

C. HUMAN STRESS IN EXTREME RISK ENVIRONMENTS

In the narrower field of human reliability under thestress of EREs offering a distinct possibility of death orserious injury, there is somewhat more substantial unanimityregarding disabling stress impact upon participants incombat environments (Rahe, 1988), although Barge et al(1934) found conflicting evidence regarding increased stressor negative effect in dangerous situations based on a reviewof studies of dangerous military and civilian environments.It appears however that Barge et al formed their opinionbased upon review of literature focused more upon review ofpredominantly physiological aspects and self-reporting ofpsychological aspects. The two combat examples they presentof multiple loading/non-firing of rifles at the Battle ofGettysburg (from Waller & Burkhardt, 1965) and weaponsfirings by only the same fifteen to thirty percent of troopsengaged in World War II units (from Marshall, 1944; 1947)are simultaneously consistent with perceptions of increasedstress in combat and more ambiguous due to unansweredquestions concerning organizational and environmentalaspects. That there is also very little data availableregarding reliability in emergency situations due to theirbeing low probability events is agreed upon both byliterature (AGARD, 1989) and the great majority of expertsinterviewed for this study. There is a significant body ofinformation available in the areas of reaction to combatenvironments and coping with those environments. In fact,as stated by Novaco, Cook, and Sarasan (1986), "the study of

111-4

human stress has not a better context for investigation thanin military environments." They base this opinion upon twocited 2actors: the harsh physical circumstances that affecttissue needs, and the threatening psychological ambience ofcombat.

Stress casualty rates in combat are exceptionally highin comparison to other types of losses. During World WarII, one example cites stress losses at 50+ men per day (10%of effectives per month) for U.S. Army Divisions (Driskellet al, 1990), another at 18-48% of all casualties sustained(Mandelbaum, 1954). During the 1967 Arab-Israeli War, theEgyptians were said to have suffered stress casualties threetimes higher than wounded in action, and the Israelis 900 ofthe first 1,600 during the 1973 war (Labuc, 1984). Over thecourse of the entire 1973 war, the Israelis were reported tohave suffered psychiatrically-related casualties amountingto 23% of all non-fatal casualties (Belenky, Tyner, andSodetz, 1983). Currently, U.S. experts predict stress lossrates at 25% of all casualties for conventional war and 50%or higher for nuclear or chemical warfare (Driskell et al,1990), or at one in every three to five wounded in combat(GTA 21-3-6, 1986).

Such losses can extend beyond combat itself, too: onlyone percent of German held POW's died before release inWorld War II, while 33% of Japanese held POW's did. Basedon an analysis of similar factors, it has been concludedthat the difference was the greater stress of Japanesecaptivity (Wolf, 1950).

Categorization of Post Traumatic Stress Disorder (PTSD)in DSM-III-R cites overwhelming natural, manmade, orpersonal disasters such as tornados, combat, or rape--experiences well outside the normal human experience--asbeing likely producers of severe stress reaction. There isa substantial and relatively homogeneous literatureconcerning military aspects available, and a literatureregarding natural disasters is being built. Currently thelatter is having to deal with a high level of controversy(Steinglass and Gerrity, 1990). Some very broad themesappear to be relatively consistent between the combat anddisaster related literature, though: stress rates arecomparatively high, many suffer some or most of thesymptoms, substantial attenuation can occur with increasingtime after threat termination, and most persons exposeddon't suffer combat stress reaction (CSR) or PTSD symptomsat levels sufficiently high to warrant treatment (Rahe,1988; Steinglass and Gerrity, 1990; GTA 21-3-6, 1986).

D. DECISION MAKING IN EXTREMF RISK ENVIRONMENTS

Decision making is a complex process under anycircumstances. It has been defined in a number of ways, forexample as analysis surrounding a psychological moment ofchoice (Pitz and Sachs, 1984; Berkeley and Humphries,

111-5

1982); as a complex task depending on a number of separatecognitive abilities that are themselves multifaceted(Wickens et al, AGARD-CP-2/58, 1988); or as a continuumbetween recognizable and analytical decision strategies(Klein and Thordsen, 1989).

An ERE that offers genuine possibility of death orserious injury presents the greatest challenge to effectivedecision making due to the consequences of an error. Oftenon such occasions, however, time and circumstances conspireto provide the least effective opportunity to make accuratedecisions. This does not necessarily mean that decisionmaking quality is degraded, though. A common thread inWickens et al and Klein and Thordsen is that using what thelatter call a reccanizable strategy for decision makingallows experienced decision makers to arrive at effectivechoices with minimal analysis. This holds true both fortime effects (Klein and Thordsen, 1989) and circumstances(Wickens et al, 1988). Both note the need for higherdegrees of analytical reasoning on the part of lessexperienced persons, who are thus more vulnerable to stresseffects, and both note the degrading effects on decisionmaking from increases in spatial requirements. Wickens etal also warn however that experienced decision makers tendmore towards overconfidence, resulting in additional riskand less reliance on memory. Keinan (1986) also relatesconfidence to experience level, but feels it did not deriveexclusively from direct experience with risk, and noted theimpact of time stress on peripheral tasks (a factor alsocited by multiple interviewees during this study).

For experienced decision makers, cognitive failure dueto overarousal or underarousal is cited by Wickens et al(1988) as being the most consistent immediate cause ofaccidents. Cognitive skills are also more prone todegradation than labor intensive ones (Driskell et al,1990). Factors affecting task performance are shown inExhibit 1 (Driskell et al, 1990) and leading tocasualty-producing accidents in Exhibit 2 (Kurke, 1991).

111-6

Lm

' I-

i..

= ..

L -D S

a..

111-7

0 1

iii

JILInaU

II" •i "3

ASS

1"a8 -'.1

U'. ,, Ii i'!*

• e - -. .|

111-8

E. TEAM DECISION RELATED ISSUES AND STRESS

Stress leads to more reliance on recognizable decisionmaking strategies versus analytical ones in teams as well asindividuals (Klein and Thordsen, 1989). Perhapssurprisingly, the presence of other team members tends todegrade rather than enhance individual performance (Driskellet al, 1990): even a two man team experiences a smalleffect, three a moderate to large effect, and larger teams avery large effect. Not as surprising, team leadership canalso exercise a significant impact, with negative leadershipaffecting both team efficiency and error rate (Chidester andFoushee, 1988). Another source has identified eight otherpotential problem areas for team decision making as well:1) distorted perception; 2) difficulties with situationassessment handoff; 3) difficulties in the formulation ortransmission of intention; 4) directed attention; 5)missing expectancies; 6) restricted improvisation; 7)synchronization; and 8) metacognition (Klein and Thordsen,1989). Additionally, Dr. Fiedler when interviewed by thelead researcher for this study cited the importance in atraining environment of quickly isolating problem personnelto prevent additional losses due to emotional contagion.

F. COHESION AND INFRASTRUCTURE

Despite adverse effects on group and individualdecision making, there is often no other way to effectivelyaccomplish a task than by use of a team approach, andindividual performances can be enhanced by teamparticipation, as can be observed from team sports. Why isless certain. One study (Zakay et al, 1986) that testedcognitive and motor tasks under heat stress suggests thatstress reduction observed was due to co-actor participationin the task they all, by experience, knew what theindividual members were experiencing. This is borne outanecdotally by Dr. Fiedler and John Reardon. Reardon warnshowever of the importance of maintaining a functionallyoperative infrastructure. He cites as examples ofrelatively effective group performances in a civil disasterthe responses of civilian emergency and local militaryassets to the San Francisco earthquake of October, 1989.Steinglass and Gerrity (1990) suggest in their study thatthe town's perception of a flood in Parsons, West Virginiain 1985 as a community-wide disaster led to more effectivestress coping as well despite weak infrastructure, butReardon also cites the 1972 Willsboro, Pennsylvania flood asone in which the effects of the disaster were magnified byabdication of surviving local leadership. Steinglass andGerrity (1990) suspect that good infrastructure wassufficient to prevent a relatively higher proportion of PTSDcasualties as compared to Parsons, West Virginia in a

111-9

community devastated by tornado (Albion, Pennsylvania),which they suggest was due to a perceptual division on theparts of survivors into victim and non-victim categories.

Military literature is more uniform in stressing thatcohesion of military units in combat is crucial to effectiveor continued performance and casualty minimization. This istrue in nuclear (Mickley, 1989) and other types of combat(Rahe, 1988; GTA 21-3-4, 1986; GTA 21-3-5, 1986; GTA21-3-6, 1986) as well as for prisoners of war (Rahe, 1988).

G. COPING IN EXTREME RISK ENVIRONMENTS

Perhaps not surprisingly given the controversy over theconcept of stress itself, coping responses appear to be anissue of little less controversy than their precipitators.A wide variety of perceptions on the parts of bothresearchers and subjects affects this, but there is alsosome general agreement on theoretical approaches to theissue of coping strategies: psychodynamic,person-environment/relational, and individual differenceviews. The psychodynamic theory essentially postulates thatcoping is a defense-oriented tension reduction processresponding to intrapsychic conflict; theperson-environment/relational that coping aims at cognitiveand behavioral strategies to alter a stressful transaction;and individual differences on fluid, multi-dimensionalvariables surrounding locus of control aspects. Foci ofcoping processes are seen as being either emotion (internalcontrol) or problem (external adaptive) oriented (Barge etal, 1984).

Appropriate coping strategies are obviously critical inan ERE. Dr. Fiedler reported that in highly stressedcommunities she dealt with, stressors were the same forthose who succeeded and those who didn't--only the copingand social net capabilities varied: Dr. Chapman, referringto a similar population, agreed. Both Dr. Fiedler and Dr.Gerwell (1990) feel that how individuals are prepared forstressors is as important as individual adaptiveness. Theyuse short term intervention--group sessions--to enhancecoping skills. Elsewhere, (1990) they cite a high stresscommunity whose members required more significantintervention perceived themselves to have many fewerpersonal resources; for this effort the Occupational StressInventory (OSI) was used to measure stress factors andlevels. Dr. Glogower in interview also noted the value inhigh risk environment personnel selection of measuringcoping skills to determine an individual's use ofconstructive or destructive coping strategies.

Coping in EREs obviously must be both effective andconstructive in order to better assure both immediatesurvival and continued functional performance of anindividual. Training is only a partially effective means toboth teach personnel constructive coping techniques and

III-10

inure them to stressors that require use of the techniques,since in many cases it is impossible to train realisticallyenough to cope with real-life challenges or to cover allspecific situations that may need to be dealt with bycoping. Some EREs might be sufficiently alien to directexperience that new coping mechanisms might need to bedeveloped and utilized. Finally, the particularcircumstances of a specific environment might require use ofeither new or previously used coping mechanisms in waysdifferent from those learned or utilized before in order torespond to situations radically different from thoseexperienced before, with varying lead times betweenemployment of a particular coping response and itsconsequences. Intensity can also vary. For example, as GTA21-3-4 (1986) states bluntly, war is more stressful than anytraining can be, but as Mickley (1989) adds, nucleardetonations can produce such devastation as to create aperceived unreality of the aftermath. And underlying thegreat majority of research into coping is the fact that itis generally of pathogenic orientation (Antonovsky andBernstein, 1986).

Despite the novelty and extremity of circumstances,significant numbers of personnel do appear to adapt more orless successfully to EREs. Steinglass and Gerrity (1990)noted the striking resilience of most people in their studyof tornado and flood victims in two communities andpercentages of personnel not lost to stress in combatenvironments can be identified from statistics cited earlierin this study. Substantial research has been done intomeasures to prevent PTSD and enhance coping, the mostsignificant measure being reported as development of a senseof pace (Rahe, 1988); this is also implicit in Milgram's(1986) comparison of successful coping techniques byprisoners (Shachak and Weir) of different Middle Easterninterests: both apparently structured and compartmentedtheir activities, then dealt with each compartmentseparately. This is rather similar to Dr. Kurke's adviceregarding the approach to the interrelated core issues ofthis study. Similarly, though Mickley (1989) warns ofevidence of direct influence on psychological changes byionizing radiation (leading to more compliant, subduedbehavior, depressed motivation, and attentional focusing),he also notes that most victims of the Hiroshima andNagasaki bombings "behaved in a manner compatible withestablished social norms" (p. 160), nor was there increasedincidence of psychosis.

H. INTERPERSONAL CONFLICT

One of the most serious threats to security andoperations within EREs might be interpersonal stress amongthe members of the group or team. Research into isolatedmicrosocieties (Warren, 1988; Helmreich, 1987; Evans et

III-11

al, 1987; Emurian and Brady, 1981; Gunderson, 1966)strongly suggests that even in highly professional confinedpopulations, there is a serious potential for significantinterpersonal conflict to develop. Indeed, Warren (1988)cites disruptive effects in all of four separate tests usingprofessionals isolated for only four to seven days in acomparatively luxurious habitat complete with recreationroom featuring stereo, television, and gymnasium; both heand Helmreich (1987) cite similar problems in both theUnited States and Soviet space programs amongastronauts/cosmonauts, including (Helmreich, 1987) a"mutiny" on board the U.S. Skylab. Warren (1988) and Dr.Glogower in interview also cited interpersonal conflictresulting from the isolation of wintering-over in theAntarctic. The Federal Emergency Management Agency (FEMA)recognizes the interpersonal conflict potential in CivilDefense shelters as well (CGP 2-21 of May 1988).

Interpersonal conflict arises from a very large numberof specific potential sources, but in most general termsfrom stress produced by the nature of the environment and/orhabitat time (Warren, 1988; Gunderson, 1966), andindividual unsuitability for adaptation to the conditionsexperienced (Evans et al, 1987; Helmreich, 1987;Gunderson, 1966). Warren (1988) also notes the strongpossibility of greater discord in mixed gender populations,though stressing the need for more research in this area.

Factors mitigating interpersonal conflict includeleadership (Warren, 1988; Gunderson, 1966);compatibility/cooperation (Evans et al, 1987; Emurian andBrady, 1981; Gunderson, 1966); emotional control (Warren,1988); work role/performance (Warren, 1988; Helmreich,1987; Emurian and Brady, 1981; Gunderson, 1966); andlikability (Warren, 1988). The chief means of avoidinginterpersonal conflict other than modification of the livingand work environment is by personnel selection (Warren,1988).

I. PERSONNEL SELECTION AND MONITORING

For purposes of this study, personnel selection forEREs will touch upon (1) behavioral, (2) performance, (3)physiological, and (4) psychological issues. These are seenas primary fields of interest in such environments.

1. Behavioral Issues

Behavioral aspects of personnel selection have beenexhaustively researched, yet remain a significant area ofcontroversy. As has been stated, the axiom that pastbehavior is the best predictor or future behavior is socommonly observed in everyday life that it's difficult toconceive of a situation in which predictive behavior had noprior analogue. Perhaps the most comprehensive approach to

111-12

the issue is taken by the same source in suggesting aselection, monitoring, and interaction approach based onpersonality characteristics, vocational interests, andbiodata in an interaction with environmentalcharacteristics, life events, stress, coping, and targetbehaviors (Barge et al, 1984).

2. Performance Issues

It is a precept of this work that personnel should notmerely survive but be able to function in an ERE. Yet asmight be expected, stress degraded performance is mostlikely when least tolerable (Driskell et al, 1990).Stress-induced performance degradation even in an ERE is notuniform, however: one study observes that knowledge skillsare not degraded by stress (Wickens et al, 1989), anotherthat cognitive skills are degraded by stress more thanlabor-intensive ones (Driskell et al, 1990). Another factormay be critical as well in some EREs: as Driskell et al(1990) have observed in a military context, theperson-machine system is the fundamental military unit, butin combat systems, so much has been done to improvereliability and effectiveness that the most unpredictableelements are human operators (Trejo et al, 1987).

Regardless of context though, it appears that a highconfidence expectancy engendered by experience can mitigatepsychological stress sufficiently to allow effectiveperformance even in the face of extreme risk (Keinan, 1986).Ursin (1989) essentially agrees, noting also that such astate need not cause high levels of endocrine or autonomicreaction if the individual expects a positive outcome. Dr.Sipes, in interview, also held that performance should focuson stress resistance enhancement vice merely coping.

3. Physiological Issues

The requirement for a person to be physically capableof performing necessary duties, particularly in an ERE, isobvious. Physical fitness, especially endurance related,may be helpful in enduring psychological stress (Rahe,1988), but it is not necessarily a factor: Herd (1991)cites numerous studies showing cardiovascular reactivity tostress had no consistent relation to exercise conditioningor physical working capacity. Indeed, he cites familyhistory as the most significant modifier of psychologicalstressors.

4. Psychological Issues

Psychological factors are of critical significance inEREs not only because of issues cited earlier in this studyimpacting on reliability, but because of the cumulativeeffects stress can produce in such an environment (Novaco etal, 1983; Mickley, 1984). The basic element of

111-13

psychological consideration is personality, which ischaracterized by traits (Anderson, 1989; Barge et al,1984). This study deals less with traits than might beexpected however because although traits can be an importantfactor in selection of personnel for certain types of highrisk environments or EREs, and though traits can and havebeen measured effectively both quantitatively andqualitatively (Anderson, 1989), selection methods involvingtraits must be very carefully constructed, as perception ofneeds can vary somewhat from results obtained. For example,the U.S. Army looks for resourcefulness, ingenuity,pragmatism, patience, self discipline, dependability, andemotional stability in its Special Forces personnel; intesting, they found the most successful candidates were lowanxiety, high persistence, high vigor, risk takers (Plebanet al, 1989). An experienced combat pilot identified traitsof an ideal combat pilot as being leadership,masculine/adventuresome, intelligence, stability,extroverted, archly humorous, and patriotic (Spiller, 1989);a number of psychologists and psychiatrists at the sameconference (AGARD, 1988) cited Type A behavior patterns(Damos, 1988), self-confidence (Barnes and Lurie, 1988;Siem, 1988), and a complex defense mechanism (D. R. Jones,(1982), cited by Barnes and Lurie). The samecharacteristics and mechanisms that made them successfulalso were cited in cases of high or extreme risk to havepotential to cause illness or emotional problems (Barnes andLurie, 1988), or overconfidence (Wickens et al, 1988), whileChappelow (1988) argued that under-arousal was a verysignificant accident factor. The variations in concept andfocus at the AGARD Conference led one observer (Billings) tonote the contradictions presented, those othercontradictions presented by military and civil air transportpilots (opposite characteristics in the same person neededin different jobs), and the inability to select accuratelybased on traits. Earlier he (Billings) had noted lowcorrelations in trait measurements as the reason they wereno longer used. Another observer/participant (Ursin) inresponse stated that it was important to be very accurate inwhat was to be predicted if traits were to be used. Itcould be noted that since Desert Storm, there would probablybe some reaction to the pilot's requirement for masculinetraits, too.

The points are that traits are extremely difficult touse in a selection process and that those necessary may varydepending upon the type of ERE: for example, research intoisolated, confined environments (ICE's) suggest thatcompatibility and stability of groups in such environmentsincreases with homogeneity in lack of need for prominence(Gunderson, 1972; Nelson, 1964; Nelson, 1964). Barge etal (1984) also note that "personality measures seem to havetheir greatest predictive power when applied to adjustmentcriteria, those same criteria that are most important forthe reliability of behavior." (p. 10)

111-14

5. Monitoring Issues

The problem with a selection process without continuousmonitoring is that it provides only a "snapshot" of anindividual, as Dr. Glogower expressed in interview.Monitoring programs focus on detection of unsuitable orunreliable behavior or indicators of them vice prediction ofsuch behaviors, as done by selection. Thus monitoring cancompensate for two failings of selection: inaccurateprediction and behavioral change (Barge et al, 1984). Thevalues of monitoring are generally agreed, but there are anumber of alternative means to accomplish such programs.For example, Bosshardt et al (1991) and Barge et al (1984)tend to focus on commander/supervisor responsibilities; theU.S. Navy stresses that all personnel have responsibilitiesto report unusual behavior by any others in the PRP; Dr.Rees suggests use of the informal net of psychological andpsychiatric personnel as trained observers, and the U.S.Army requires a surety board, officers, and coordinators (AR50-5, 1989). Scope of monitoring effort, specificprocedures utilized in monitoring, training level ofreporters, and frequency of review are subject to similarvariation.

J. IMPLICATIONS OF EXTREME STRESS ENVIRONMENTS ONOPERATIONAL PERSONNEL

With very little doubt, stressors of all kinds would beimposed upon operational personnel in the event of an ERE ofany scale, whether the personally devastating losses orthreat of losses that might be experienced by an individualor the massive cataclysmic devastation by natural or humanmeans of a large area or the whole nation, that should inmany cases be expected to exceed anything in their directexperience. Some types of devastation, such as wide scalegeological, biological, nuclear, atmospheric, chemical, orother type of invasive disruption may in fact exceedeffective conception by participants that might be requiredto respond effectively to the effects experienced, both on apersonal and a much wider scale. It is likely thatpersonnel would have to deal simultaneously with a varietyof threats at multiple levels in an extreme stressenvironment. As can be seen from preceding materialregarding extreme stress and human reliability, however,although the challenge is daunting, if personnel can survivethe event with means to carry out their duties at some levelof effectiveness, they can function to mitigate effects ofthe event. As it is known that personnel can survive andfunction, in extreme risk events, it is possible toconstruct mechanisms for identifying best-risk andworst-risk candidates for assignment to deal with suchevents. It also appears that it would be possible toestablish specific criteria for the particular requirements

111-15

of teams, organizations, and activities that directly orindirectly support operations in such environments. Unlessa reliability program is established to select personnel whocan meet suitability (often security-oriented) requirementsand function, even if isolated, in EREs however, thepreceding information in this section also strongly suggeststhat elements deployed to deal with such events will sufferstress casualties in sufficient numbers to render them atbest ineffective and at worst eliminated. In many cases,deployed elements aren't the only personnel who should beincluded in an HRP, however.

It would be most desirable to establish an HRP for allpersonnel directly and (at least as far as direct support todirectly involved elements is concerned) indirectly involvedin operations in EREs. The selection process thus would beenhanced by providing a mechanism beyond current means formost such populations that would assure a higher degree ofreliability in personnel at selection, throughout theirassociation, and to some degree after separation. It wouldalso assure that acceptable replacements were more readilyavailable to fill in for anticipated losses. The scope ofeffort that such a system might require could be substantialif multiple planning, response, and/or support activitiesare to be covered. For reasons of cost, though, populationsindirectly involved in responding to EREs might notnecessarily be screened in an HRP as rigorously as thosedirectly involved, but again for cost effectiveness, thescreening mechanism(s) used should be directly and rapidlyexpandable to full scale. It would also appear that atleast low-level HRP criteria could be developed forapplication across multiple populations.

Use of an HRP could also ensure selection and retentionof emotionally hardier staff personnel who also have bettermorale and unit cohesion (these desirable features may bebuilt into the selection procedure). Focus on performanceunder stress could significantly reduce stress losses.

An added consideration is that in particular extremerisk scenarios it may be necessary for directly and/orindirectly involved personnel to present an outwardappearance of normalcy to minimize the disruption and panicthat might otherwise ensue. Such conditions may wellrequire striking a balance between providing too much or toolittle information to others at a given time where eithercould result in the same magnified undesirable effects. AnHRP for EREs must be sensitive to this potentialrequirement.

An HRP can also provide more general, non-emergencyenhancement for activities dealing with EREs by increasingselection and retention capability for people with betterabilities to deal with performance and security relatedissues before they become crises. Hardening of suchactivities would be provided operationally or during periodsof quiescence through increased reliability andfunctionality of personnel under stress. This translatesdirectly into increased mission capability.

111-16

It is recognized that it may be necessary to considertrade-offs in selection between quality of one individual'sperformance in peace or quiescent times and that of anotherwho has better performance under stress. In such events, itis critical to examine the specific scope and nature ofscenarios a population selected to deal with them might facein order to effectively analyze the extent of trade-off thatwould not compromise mission accomplishment. A number ofother factors may impact as well: for example, securitystandards may or may not be able to be relaxed in order tomore readily permit mission accomplishment while stillmaintaining necessary protection for the mission, and ifthey may be relaxed, they may be able to only at some stagesand not others. Some performance degradation should beexpected from the best personnel in an actual ERE, butfunctionality must be retained. Overall, suitability toaccomplish the mission is the baseline requirement.

IV-I

Section IV. Behavioral ReliabilityIssues in Extreme Risk Environment Contexts

A. INTRODUCTION

As could be seen from the previous section, the issueof behavioral reliability is a complex one now generallyseen to involve an interaction of the person with his/herenvironment. The issue basically devolves though to theindividual's suitability for assignment in a particularposition and enviroament. Suitability determination in theFederal defense, intelligence, and security communities is amatter of personnel security adjudication based upon amandated scope investigation and at the most sensitivelevels utilizing selection requirements in Director ofCentral Intelligence Directive (DCID) 1/14.

B. CURRENT POSITION

Certain agencies in the Federal Government such as theCentral Intelligence Agency (CIA) and the National SecurityAgency (NSA) employ psychological services and polygraph,whose primary purpose is, as psychological review is in DODPRP's, to reinforce the suitability selection process forthe agency as a whole. According to interviewees, bothagencies cited above employ at least in part currentlyavailable commercial psychological testing instruments.Instrument selection and application at CIA and NSA is madeto enhance the general selection process in use at the usingagency and is different for each. Interviewees at bothagencies however noted that their processes are aimed atother criteria than those identified in this text, and wouldbe unsuited to application to populations in EREs.

The issue of polygraph use is a highly divisive one.Defenders such as Dr. Rees of CIA feel that althoughpolygraph is known to fail to obtain all available data,information from lifestyle polygraphs is more faithful tothe whole person approach. The issue is not just a failureof intentional self-reporting: as cited by the NSApsychological staff, honesty with self is essential topsychological testing, and people often aren't honest evenwith themselves (or have an unrealistic self-view). The NSApsychological staff also points out that self-reporting istotally inadequate for identifying problems (as might beexpected). Both of these honesty factors, deliberate andavoidant perspectives, are of course susceptible topolygraph examination due to that medium's focus onconscience. Having said these things, however, it must alsobe pointed out that research by Jayne (1988) cited by

IV-2

Bosshardt et al (1988) seeks to demonstrate that thepolygraph is less successful at obtaining confessions thanintegrity-type personal interviews or even computerquestionnaire interviews: this being used by Bosshardt etal in support of their contention that an effective initialprocedure is a suitable alternative. The point made by Dr.Rees (who strongly disagrees with Jayne and Bosshardt) andNSA is that an integrated process involving interviews,psychological reviews, background investigations, andpolygraphs (not necessarily in that order), followed bymonitoring and intervention, were all necessary for the mostcomprehensive behavioral reliability review. Sincebackground investigation and adjudication requirements arethe most standardized, modifications to them for purposes ofuse in EREs would be far more difficult to arrange than theother selection process components available. As theselatter vary, they are seen as offering the best prospectsfor development for the more specialized and criticalrequirements postulated in this study. They might thus bedesigned to enhance behavioral reliability selection in acomprehensive program yet remain sensitive to the needs ofmultiple populations. This it is felt can be done, andspecific processes for doing so will be presented further inthis text.

C. FUTURE CONSIDERATIONS

Assurance that behavioral reliability requirements willmeet mission objectives requires identification andvalidation of mission-oriented criteria. Existing selectionstandards such as DCID 1/14 have served well to providesuitable candidates in times of peace or tranquility. Theymay be less valid in an ERE: for instance, credit solvencyis a reliable indicator of an individual's vulnerability inan environment where classified information must beprotected and the individual is not at great risk, but it isboth explicitly of less significance in an ERE where theindividual's life may well be at risk and implicitly ofunproven value as an indicator of an individual'sreliability in such an environment. It is understood thatin multi-mission, multi-population environments,accomplishment must be geared to specific population needsand resource availability. Further, any behavioralreliability modifications recommended should not affectexisting means of reliability selection for other purposesor pose unrealistic burdens. At the same time,modifications to existing behavioral reliability processesare inevitable if personnel are to be suitable for andfunctional in EREs.

V-i

Section V. Performance Reliabilityin Extreme Risk Environnent Contexts

A. INTRODUCTION

The man/machine interface is the definitive basicelement of effective response to EREs, whether the interfaceis direct or indirect, operational or referential,regardless of the specific scenario presented. Whetherhis/her tool is a computer or a weapon, a bulldozer or aspacecraft, a telephone or a soldering iron, on virtuallyany scale, people use tools to alter a less than idealenvironment. Today, however, even in high technology fieldssuch as aerospace the human operators are acknowledged asbeing the limiting factors in the man/machine interface(Trejo et al, 1987; Driskell et al, 1990), yet this aspectoften receives the least systematic, solution-orientedreview. Even where it does (perhaps especially in hightechnology fields), the solution is likely to be perceivedas the elimination of the human factor in the equation. InEREs, though, machinery itself will be stressed, quitepossibly well beyond its design limitations. The result ofcomplete reliance on machines where it might be avoided isat that point obvious: sooner or later, the machines willfail.

B. CURRENT POSITION

Machinery can be thoroughly tested in actual orsimulated EREs. In many cases, people cannot, at leastrealistically enough to achieve the performance levelsarising from confidence expectancy cited earlier by Keinan.Artificialities growing out of simulated EREs effectivelypreclude the ability to truly "train as you fight", to usethe U.S. Air Force phrase, but few persons would willinglyface a true extreme risk situation - -re death or seriousinjury was a significant possibilit. Those that might,especially on a repetitive basis, may well not meetbehavioral or psychological suitability screeningrequirements. Another important aspect to consider too isthat the more senior members of groups or populationsresponding to or functioning in EREs are likely to be themost advanced in age as well, and the physiological factorbecomes increasingly vital.

Planning for operations in EREs must include detailedexamination of interpersonal and group dynamics. In anisolated operational environment (one in which there is notime or opportunity for relief of personnel), the success ofonsite personnel in dealing with these issues will define

V-2

the success of the mission. As mission length increases, sodoes potential for sleep disturbance, anxiety, depression,irritability, and disruption of group cooperation andaccomplishment (Gunderson, 1972). These problems are causedby both physical and social stressors relating to crowding,artificiality of sterile environments combined with likelyseparation from (and natural worries about) loved ones, lackof privacy, and forced interaction (Evans et al, 1987). Asthese issues are exacerbated over time, short termartificialized training in and of itself cannot deal withthem. Again, selection appears to hold the key,particularly if combined with training.

Many mission requirements address operational issuesfrom a perspective of system network and equipmentefficiency; positions for operational personnel are thendeveloped with an eye to perceived mission needs for theinformation or service generated by or from a type orcategory of source. In such environments, the enrichmentcaused by crosstraining in a variety of work is beneficialto stress resistance (Gunderson, 1972; Evans et al, 1987).Dr. Ginzburg in interview though noted that technical andmission operational staffs are likely to more fully adjustto decision making than decision makers to technical ormission operations. If candidate selection for particularoperational requirements is politically driven or (even morelikely) availability driven, analysis of candidate selectionfor those requirements is adversely impacted in directproportion to the influence of the driver. Without suchanalysis, however, there will be insufficient assurance thatcandidates, minus additional screening aimed at identifyingthose among them who are able to cope with the requiredstressors and environment, would be capable of performing asnecessary to accomplish the mission. Thus, the selectionprocess may be placed between two difficult issues: cansufficient candidates be found if a more comprehensivescreening is done versus will they be able to perform ifmore comprehensive screening is not done? The first part ofthe question offers both more promise and more hopefulresponses than the latter.

C. FUTURE CONSIDERATIONS

It is essential to assuring that performance-basedoperational reliability requirements will meet missionobjectives that unavoidable training limitations,interpersonal/environmental factors, and lack oftype-oriented selection criteria do not mitigate againstselection and retention of capable personnel for actualmission status. Obviously, use of new selection mechanismsof any type will cause a higher rate of personnel fallout ifadditional screening of any type is implemented, but withoutit the only assurance that personnel will perform asrequired in an ERE is their not having self-selected

out...and this may well change when the chips are down.

V-3

It is possible though to enhance motivation,resporqibility, and stress tolerance both in currentlyavail ole and future candidates through a variety of means,and possible to develop programs that do so. Some havealready been touched upon, and more will be addressed inSections X and XI of this study.

V-4

This page intentionally left blank.

VI-I

Section VI. Physioloqical Reliabilityin Extrene Risk Environment Contexts

A. INTRODUCTION

Extensive study of physiological reactions to stressimposed by a very wide variety of circumstances andenvironments including those involving personal risk orisolation and confinement has been made for generations.Cannon published his findings on what he called theemergency reaction and has more commonly become known as thefight or flight reaction in 1914 (Cannon, 1914).Substantial advances in physiological research into theeffects of stress caused by the impact of harmful physicalstimuli was done by Selye in the 1930's (Selye, 1936). Inmore recent years, much research has been done regarding thecatecholamine and other chemical and chemically-relatedinfluences. The empirical nature of the subject has allowedfor a more thorough understanding of both individualresponses and the interactions between responses to bearrived at than in less readily accessible fields. As Herd(1991) has noted:

"Because psychological stressors are difficult toapply in a controlled and quantitative manner, thephysiological responses observed become an index of thestressor." (p. 314)

B. CURRENT POSITION

Despite the large body of validated research intocatecholamine (e.g. norepinephrine, epinephrine, anddopamine) and corticosteroid (e.g. cortisol) relatedresponses to cite only a very few of the pertinent ones, atranslation of the results into extant PRP and PRP-likeprograms has not been effectively accomplished. We canascertain to a meaningful degree the impact of various typesof stress (e.g. heat and work load--Zakay et al, 1986;Carter and Cammeyer, 1989) on various critical functioningcapabilities and perceptions--the cardiovascular system, forexample (Herd, 1991; Henderson et al, 1990), and even lessdirect relationships such as motivation and aversion (e.g.Lovallo et al, 1990), relaxation (e.g. Benson et al, 1977)or emotions (Simonov and Frolov, 1984). Yet the commonphysiological review technique extant in PRP's and PRP-likeprograms is a medical records review, or at most a medicalrecords review and comprehensive conventional physicalexamination aimed at verification of current ability toparticipate in a program at current stress levels. Although

VI-2

medical reviewers do take into account anticipated stressand research results in certain highly specific isolated,confined environments such as Antarctic over-winters orsubmarine service, less attention appears to be paid to therigors of some even more stressful environments, forexample, nuclear combat. The reasons for this are becausestandards (DOD 5210.42 in the nuclear arena) aresufficiently vague to allow less screening to beaccomplished, and the reasons for the vagueness in publishedscreening is generally a reflection of the practicalconsiderations of volumes of personnel requiring generalreview, unpredictability of specific nuclear engagement orother non-contained scenarios, general availability ofmassive redundancy in replacements for broad scenarios, anda need for lowest-common-denominator screening capabilityfor scenarios requiring massive populations.

Nor are these the only problems. Some populations thatmight be expected to respond to ERE scenarios are allowed toparticipate at risk, i.e., with the individual participanthaving made a cognitive decision to participate with atleast partial knowledge of the fact that to do so placesthat individual at greater life risk than others due tospecific health problems. In some cases, comprehensiveconventional physical examinations might be offered, but ona voluntary basis, for responding populations. In thelatter case, experience has shown that substantial numbersof participants fail to avail themselves of the opportunityeither because they feel no pressure to do so on anexpeditious basis, because they don't really wish to findout (or have others find out) their actual current healthstatus, and/or because they have unilaterally chosen todeploy at risk regardless. The consequences of all theabove cited situations are potentially serious shortfallsduring exposure to EREs.

C. FUTURE CONSIDERATIONS

Obviously, a continuation of the current state ofaffairs in the physiological review arena does not provide agreat deal of assurance that populations deploying to meetERE scenarios will in fact be adequate to the task unlessredundancy of manpower is so massive and crosstraining soextensive and effective as to amply mitigate the effects ofstress casualties; in an environment of drawdown of Federalresources to meet all currently anticipated contingencies,it is increasingly questionable whether these conditions canbe met. Now more than ever, quality must be stressed overquantity. This is even more particularly true whenconsidering smaller populations or those that must be reliedupon in the greatest extremities. The field of physiologyis broad enough and well enough understood to be able toplay a far more significant role in the selection andretention of personnel whose physical constitution can be

VI-3

best anticipated to respond appropriately when confrontedwith life-threatening stress. Physiology (and itssubfielCs, particularly neurophysiology, psychophysiology,biophysiology, and endocrinology) is also capable ofprescribing effective means to retain that appropriatestatus, and means of enhancing individual abilities tosuppress the emergency reaction and permit continuedsuitable response to EREs, though it must also be admittedthat much remains that can and should be done. Without athorough review of existing capabilities and resources tomeet specific ERE scenarios and/or the development ofadditional means to do so, however, the operationalapplication remains the only test of efficacy. In an ERE,there may very well be no capability for recovery from (oreven awareness of) physiological deficiencies that spawnedfailure of the necessary response.

VI-4

This page intentionally left blank.

VII-I

Section VII. Psycholoqical Reliabilityin Extreme Risk Environment Contexts

A. INTRODUCTION

If compatibility and adjustment are primarydeterminants for stable group structure in a highly stressedenvironment, there is an implicit requirement for a selectin process to obtain the most desirable candidates for EREs.Such a requirement does not always fit political andoperational realities, and is inadequate by itself to meetsecurity needs. These facts place additional demands uponprocesses used for stress tolerance enhancement.

B. CURRENT POSITION

No Federal agency was identified during this study thatcurrently has in place a psychological reliability processdesigned to factor in issues central to EREs, which in sumrelate to stress hardening to cope with such environments.Institutionalized programs in existence are geared tocurrent, nonemergency or peacetime needs.

The authorities interviewed for this study whocommented upon ability to provide psychological stresshardening were unanimous that it could be done. With thenext breath, they generally agreed that they were unaware ofany existing research and mechanisms to provide a systematicprocess for it at the desired level, or a selectioncapability for it. All were aware of psychological stressenhancement techniques for less severe environments,techniques such as envisioning and mentally rehearsing anact, fed by actual opposition tactics, and envisioningsuccessfully completing the act; this technique has provensuccessful with athletes providing the activity and successenvisioned are reasonable substitutes for field performance(Dr. Rees). Coping-oriented techniques might includerelaxation through a meditative (Benson et al, 1977) orself-awareness/self-regulation (Stephens, 1987) approach.

Another approach as explicated by Driskell et al (1990)is stress training utilizing indoctrination, skillstraining, and confidence drill. Indoctrination informsindividuals about the stress environment. This is commonlydone in hazardous professions, but descriptions and evenvideos cannot adequately prepare personnel for the reality(or surreality) of actual EREs where they themselves arelikely to be at life risk.

Skills training exposes personnel to the stressenvironment and causes them to experience stress based skilldegradation adaptation to the stress environment so

VII-2

performance returns to higher levels. Training or exerciseartificialities will weaken the effectiveness of measuredresponse, however, particularly if the participants arevolunteers operating in a thoroughly disagreeable (andtherefore more meaningfully simulated) environment. Evenwith nonvolunteers, such artificialities will distortperception of reliability if not of ultimate results.

The purpose of confidence drills are to give personnelthe opportunity to operate in a stressed environment usingsolutions found and adaptation to increase their confidencein their abilities to perform regardless. Even if tests orexercises are successfully conducted from an operational andequipment standpoint, though, if the simulations do notaccurately confront personnel with the emotional andfunctional stressors they will encounter in reality, therewill remain serious doubts about the ability of personnel todeploy and function under actual operating conditions. Thesubject of stress training is explored in additional detailin Section XI of this study.

C. FUTURE CONSIDERATIONS

In the absence of ability to effectively use trainingto increase stress hardiness in personnel to the levelnecessary, such training as can be done in combination withother mechanisms offers the promise of substantially betterresults than might otherwise be expected. These mechanismsmay include procedural, organizational, operational, andphilosophical alterations of existing arrangements, andintrospectively-focused self awareness techniques. In EREs,chemical means for response enhancement do not appear to bean adequately acceptable alternative due to the likelihoodof lack of medical personnel to diagnose and administerthem, and dosage, duration, and specific medicationavailability concerns, though some of the newer psychoactivedrugs have been shown to be very effective in enablingenhanced performance from individuals who might otherwise beincapable of performing at all.

VIII-I

Section VIII. Monitoring Issues inExtreme Risk Environment Contexts

A. INTRODUCTION

Even though existing PRP's focus on a front-endselection process, all systems known to the lead researcherrequire a monitoring process of some kind afterward toassist in giving assurance that selected personnel continueto meet requirements for retention in the programs. As hasbeen shown, the nonemergency focus of such programs and theoften inadequate nature of established monitoring processesdo not meet requirements for use in extreme riskenvironments. Some cultural or current popular impressionsalso suggest that monitoring implies a lack of faith inpersonnel integrit-, unwarranted invasion of privacy, or"Big Brother" concerns, instead of a simple realization thatpeople and personal circumstances change over time, andpersonnel directly involved are not always in a position torespond objectively to the changes. In many EREs, there arealso genuine requirements to assure protection of classifiedinformation to the fullest extent possible.

B. CURRENT POSITION

Screening systems lacking a reliability program relyprimarily on reinvestigation (if any), self-reporting ofpotentially adverse information, and voluntary observationand reporting by other personnel (who have very seldom beentrained to monitor others). In the lead researcher'sextensive experience in the area of Personnel Security(having among other things adjudicated over sixty thousandinvestigations for special access), there are seriousdeficiencies in the process, particularly in reliance uponself-reporting or voluntary reporting by others. Despitewarnings as to the extreme sensitivity of highly classifiedinformation, clearly identified issues on whichindoctrinated personnel are required to report, and sternwarnings about the potential consequences of failure toreport, admissions obtained are often literally years lateif obtained at all, and observations by others are oftenperceived by one and all as accusations by untrainedobservers which have potentially serious consequences foraccused and accuser. Many clearly identified areas arerationalized away by nonreporters through logical gymnasticsthat the pe-sons themselves know are indefensible. Thesingle biggest reason for not reporting is obviously fearover the repercussions of the act of reporting. Of theprofessionals interviewed in conjunction with this study,

VIII-2

NSA and CIA representatives felt self-reporting did not work(except to self-select out of undesirable duty, per NSA'srepresentatives), Drs. Gerwell and Fiedler felt it worked ona situation-dependent basis, and Drs. Patterson and Sipesfelt self-reporting generally did work, although admissionsmight come only after personnel had accomplished their ownobjectives.

C. FUTURE CONSIDERATIONS

As Dr. Glogower expressed (1991), any evaluation nomatter how good is static. Situations are fluid andtherefore need close review. Additionally, it may be seenfrom the material above that self-reporting and voluntaryreporting by others is inadequate to assure timely, reliablereporting of events potentially affecting personnel. Itthus appears that for programs in any, environment wherecontinuing reliability is important that effectivemonitoring is necessary.

Current monitoring programs for military personnel, aspart of peacetime-use PRP's, are also geared to anonemergency, peacetime environment. Monitoring personnelwould be expected to report unreliability in a wartimeenvironment to be sure, but this is largely implicit, andthere appears to be little appreciation in existing PRP'sfor the change of monitoring focus in the transition from apeacetime to a wartime environment.

In a nonemergency monitoring environment, primary focusis on personnel suitability and physical issues. Inoperational EREs, the primary suitability focus would be onpsychological readiness, and in the physical area on abilityto perform a certain task vice good physical health. Asreported in Radiological Factors Affecting Decision Makingin a Nuclear Attack (1979) and Mickley (1989), it may benecessary to see some personnel sacrificed to accomplishcritical mission tasks--much narrower, but much moreimportant requirements. Personnel conducting monitoringactivities must be able to shift focus and be aware of theincreased need for vigilance in the different evaluationareas at a time when operational concerns are increasinglypressing. It is vital that they must be trained for bothenvironments and the review shift process. Additionally,those who will monitor others must be thoroughly trained todo so, and neither the accused nor accuser faulted in theevent of an understandable though inaccurate perception onthe part of the monitor--this represents instead a trainingdeficiency.

IX-I

Section IX. Suumary ofInterview Results

A. INTRODUCTION

The interviews of authorities in the field of humanreliability selection and monitoring conducted for thisstudy constitute the broadest effort known to have been madein this area. Many authorities in a variety of closelyrelated human reliability assessment fields were consultedin order to more quickly and effectively obtain a broadercontext cumulatively to evaluate more specialized experienceand/or knowledge of individual interviewees against. Theinterviewees were advised that this study was a Director,OEO initiative to conduct basic research to determine thefeasibility of development for possible implementation of aPRP for use in extreme risk environments, with actualscenario(s) unspecified. This, in combination with the openuse of OEO/White House Military Office affiliation toencourage responsiveness, both intrigued and excited many ofthe interviewees, who were pleased that such an importantbut little-studied area would receive such high levclinterest. Many provided copies of related information theyhad authored and other research as cited earlier in thiswork, along with recommendations on additional sources andmuch good advice.

B. AREAS OF AGREEMENT

On a number of areas the interviewees were in generalagreement. These included the following weaknesses incurrent research/application:

1. Information in general about human reliability inemergency events is lacking (due to low rate of occurrence).

2. There is a lack of specific information regardingreliability when events involve life risk to an individual,or to an individual and those emotionally close to him/her.

3. There are difficulties in accurately quantifyingreliability break points.

4. Difficulties exist in quantifying the universallyrecognized difference in reliability where family or closeassociates are at life risk versus only the individual.

5. There is a lack of specific research into thesuitability and performance areas sought vice suitabilityselection alone.

IX-2

6. For many EREs, there are no measurement tools andvery little factually-based information that addressspecific issues of suitability and performance dependabilityneeded for personnel to operate in a particular environment.

The interviewees were in general agreement that thefollowing could be accomplished:

1. Instruments could be developed to assesssuitability and dependability of an individual in order forhim/her to respond to life risk situations.

2. Instruments could be developed to assessfunctionality of individuals in life risk situations.

3. Personnel facing life risk situations could beinured to them by a variety of physical, physiological, andpsychological means.

4. Reliable personnel selection could be made if theywere chosen against the operational requirements of aspecific scenario.

5. Physical fitness significantly enhances mental andemotional toughness.

The authorities interviewed were less sure orconsistent upon a number of other issues, and tended to leantoward their own familiar areas on these:

1. Type of approach to the problem (behavioral,performance, industrial/organizational, or integrational).

2. Ability to assess variables in theinterrelationship between life risk selection andresponsiveness when family and/or close associates are atlife risk versus when only the individual is.

3. Requirements for monitoring, or how to accomplishthem. All in all, there were more strong opinions aboutvalue, frequency, purpose, and scope of monitoring than anyother area, though no one actually opposed doing it.

4. Viability of reliance on self-reporting accuracy.

5. Type of validation process to be used for testingan HRP or its components.

6. Which currently available instruments would besuitable for integration into an HRP.

7. What model a program should follow.

IX-3

In some areas only a few interviewees had comments, butthese were relatively consistent:

1. Reardon, Ginzburg, Gerwell, and Fiedler agreed thatcohesion was a significant factorable.

2. Rees, Kurke, and Fiedler emphasized the importanceof not relying on a single medical/professional authority tobuild a program with due to potential essentially for ablindered, tangential outlook.

3. Reardon, Ginzburg, and Hibler (who was interviewedby Donald Brenno) emphasized the critical value-addedbenefit of cross-training.

4. Dowd, Strome, and Eddy all felt that voice stressanalysis had reached the point of viability for inclusion inoperational reliability screening.

Observations and analysis regarding the above will becommented upon in succeeding sections of this study.

IX-4

This page intentionally left blank.

X-1

Section X. Models for NultiplePopulation, Agency, Level, and/or Scenario

Extreme Risk EnvironmentHuman Reliability Program Structures

A. INTRODUCTION

One of the larger challenges presented for this study wasto find or devise a human reliability program suitable for avariety of different populations with widely divergentmissions, agency cultures, regulations, and approaches in asingle structure that both accommodated their needs andeffectively addressed their vulnerabilities in order toselect and retain candidates of the best possible qualityfrom both security and functionality aspects. Additionallyand most significantly, the screening process would have toproduce these results against and during multiple extremerisk scenarios. By use of a variety of methods, means willbe proposed in this section to deal with the operationallimitations that might be faced, mechanisms to developprograms, instrumentation, and validation that can betailored to fit both mission requirements and humanreliability concerns, and approaches to implement them. Itshould thus be possible both to develop a workable umbrellaprogram and better fit the needs of the various populationswithin it as well as conform to political realities. At thesame time, there may well be similarities that may beexploited: for example, there may be both operational andsupport populations, and these populations may fulfillsimilar functions and face similar challenges to performanceand suitability from a particular extreme risk environment,regardless of agency affiliation. These, then, are the fociof proposed HRP development and implementation.

B. SELECTION STRATEGIES

There are basically two strategies used in anypersonnel selection process. As stated by Benner (1986),

"The 'select in' strategy assumes that the desired andappropriate characteristics are known and further, away to identify them exists. The 'screen out' processis less ambitious and infers the ability to identifyundesirable characteristics and the elimination ofcandidates who exhibit them." (p. 15)

After dismissing the "select in" strategy as prematuredue to lack of consensus and specificity, Benner goes on tonote in the same source that both stability and suitabilityare issues within the "screen out" strategy, which he feelsto be more possible and practical.

X-2

Benner isn't the only one who feels a screen outprocess is more feasible: all personnel security processescurrently in use by the U.S. Government rely to some degree,often near completely, on a "screen out" process.Operationally oriented selection processes on the other handuse just the opposite to select in appropriate candidatesbased on their background, expertise, and perceived "fit"within the organization. As Dr. Wiskoff observed ininterview, selecting in is more of a capability assessment,screening out more of a behavioral assessment. InGovernment and contractor organizations where securityclearances are required, a select in/screen out(Personnel/Personnel Security) process is used.

The opposite process also has clear validity in certainsituations. For example, Captain Chandler, USN, who sat inon part of the interview of Dr. Glogower noted that NewZealand uses a combination of screen out to get rid ofobvious or substantial problem personnel, then select induring a practical field test for the remainder in some oftheir (admittedly small) sensitive communities. In fact,this method is particularly well suited to team building andis not infrequently used for the purpose by manyorganizations.

The challenge in constructing a valid selection processfor suitability and reliability is to avoid the tendency ofpsychologists to build what they personally think is valid,but rely too heavily on their own expert judgement tocreate, with the risk of establishing an arbitrariness thatcandidates can reclama (Benner, 1986). Dr. Fiedler, Dr.Kurke, and Dr. Rees among the interviewees all concurred aswell on this aspect. Dr. Fiedler in particular expressedconcern over the potential for tangential development ifreliance was placed on a single dominant psychologicalauthority, and Dr. Rees in response suggested a group ofrecognized experts for the purpose. Inwald (1986)recommends a number of specific guidelines to establish areliability selection process within various checks andcross validations to avoid the problem as well. Options andrecommendations regarding aspects of selection will bediscussed further in reference to testing procedures laterin this text.

C. TASK ANALYSIS

1. Orientation

At its most basic, there are only two types ofpopulation in any extreme risk ervironment that must betested in order for a mission to be accomplished:operational populations (including direct mission supportduring operations) and populations indirectly supportingoperational populations. This holds true regardless ofother factors influencing either selection, retention, or

X-3

events such as mission, operational expectations of specificpopulations, and degree of sensitivity. For operationalpersonnel, there are requirements for reliability insuitability and performance areas pre-, trans-, andpost-emergency; for indirect support personnel, forsuitability and performance reliability primarilypre-emergency. The operational populations must necessarilydeal with much more serious and possibly extended stressthan supporting, non-operational personnel, but the latteralso have a vital role to play: without their pre-emergencyreliability, operational elements may be unable to performtheir own missions.

From this basic stance, other aspects may be factoredthat demonstrate operational similarities, then those thatmake specific populations operationally unique, and tools ortypes of tools identified to create an effective and ongoingassessment process for each individual in each population.

D. SUPPORT PERSONNEL

Personnel performing support duties to operationalpersonnel fall into two basic categories: direct andindirect. Depending on particular scenario, direct supportpersonnel may or may not be at life risk equal to orsignificant in comparison to operational personnel, and mustbe evaluated accordingly in an HRP. Indirect supportpersonnel in EREs are less likely in many scenarios to befaced with as broad or continuing a life risk. The supportcommunities may thus need evaluation and monitoring focusedmore to a pre-, trans-, or post-emergency scenario insteadof the full scale, multi-level evaluation and/or monitoringprocess operational personnel are more likely to require.Use of even a minimal front-end screening can achieve anumber of benefits in a more narrowly focused HRP for EREs,however:

1. It could fulfill the traditional PRP role of providingadditional, otherwise less adequately covered, personnelselection support.

2. It could provide an initial screen for use in anoperational or more directly supporting population.

3. It could provide an early warning mechanism for theindividual and the organization that help was or would beneeded (with a clear focus on treatment vice separation).

4. It could be used as a tool for identification of areasto focus stress tolerance training upon.

5. It could be used to support a request for waiver of someinvestigative requirements in an emergency.

6. All of the above.

X-4

RE CO N END A TI ON

Establish a limited HRP screen even for indirectsupport personnel for EREs (which could also be used as aninitial screen/identifier for operational selection and/ormonitoring test instrument).

E. TESTING

1. Requirements

Should implementation of an ERE HRP be undertaken, thespecific components of the HRP and the models for it must bedeveloped and tested to demonstrate psychometriccharacteristics of validation, reliability/repeatability,and sensitivity. There are a variety of ways in which aprogram can be undertaken, and a number of them (withoptions) will be discussed in the following pages. In caseswhere multiple satisfactory options are feasible, the coursebest calculated to deliver the most effective product mostquickly will be recommended.

2. Validation Processes

There are two basic validation formats that will bebriefly examined in conjunction with this study: criterionand content and construct. The method chosen will dictatethe approach taken to development and testing. Each hasadvantages and disadvantages.

a. Criterion ValidationThis process essentially involves the creation of a

measurement tool by (a) researcher(s) from individual expertjudgement. The tool is then applied to an unscreenedpopulation of sufficient size to be able to mathematicallydocument some measure of statistically significantsensitivity (often low in single digits). The tool mustthen be retested a sufficient number of times to verifyrepeatability of the results. Then, if necessary, themeasurement parameters are reevaluated and different oradditional variables or more specific language applied inorder to increase sensitivity of the tool by narrowing thewindow of sought-after response.

The advantages in using such a format are that HRP'svalidated through such a process would meet scientificrequirements for accuracy and repeatability and wouldprovide the capability for a solid legal defense ofdecisions based on use of the validated parameters. Thedownside of using this method is that the smaller number ofcleared psychological researchers increases potential tofail to identify alternatives to specific parameters andcourses of pursuit, and the trial-and-error process requiresa substantial time (one to five years) to produce refined

X-5

results. Additionally, there would have to be access to asubstantial body of unscreened personnel for use as a testpopulation, which is likely to be difficult or impossible inthe case of ERE testing, and which would require substantialcoordination if such a population was to be made available.Finally, facilities would be needed at least in the vicinityof the test population, with appropriately cleared andcertified spaces or containers if classified research wasundertaken.

b. Content and Construct ValidationThis process was suggested by Dr. Rees as a means to

link scores obtained in testing to specific evaluativefactors. This method would involve identification andselection of a panel of nationally-recognized expertsincluding from outside the intelligence, defense, orsecurity communities. Criteria regarding what thepopulations the program would be applied to needed toaccomplish and what they needed to be free from would haveto be identified to the panel. These would then becomebaselines for the panel to build parameters to select upon.The panel would rely upon a consensus of their expertopinions to develop windows for selection against thebaseline criteria, then establish windows for initialtesting. These windows could then be tested against theactual populations.

Despite a shortened fine-tuning process compared tocriterion testing and validation, there would still be aneed to adjust and narrow criteria. Therefore, the use ofstandard Boolean configurations for fuzzy set application todevelop mathematical models for computerized simulations ofaspects agreed upon by such a panel is suggested, whichshould allow substantial refinement of parameters in alphatesting (without field or beta testing). Once narrowed bythis process, the resulting product could be validated byfield/beta testing against the actual populations.

There are a number of advantages that could be realizedby utilizing the content and construct methodology citedabove in conjunction with the mathematical processdescribed. It would produce valid selection criteria thatwould be acceptable to the scientific community and thus belegally defensible, and would do so in far less time thancriterion validation--months to years less. It would allowbetter concurrent testing of all aspects of humanreliability, not just behavioral reliability for example.Panel size (five to seven voting members) would assure thatlikelihood of examining insufficient options or errors dueto inadequate knowledge were minimized. One member of thepanel would be responsible for moving the professionalparameter identification and establishment process along,presenting majority and divergent opinions to the projectmanager, and getting closure on issues: he/she would nothowever be an arbitrator. Use of simulations testing wouldallow work to proceed even when operational factors wereunavailable.

X-6

This method could also be accomplished in a manner thatnurtures further and broader study, in that the panelmembers could be chosen from multiple facilities if asuitable representation of facilities were solicited andwere to nominate candidates for the positions sought andhave one or more of them selected for the panel. Much orall individual or even panel evaluation tool and mechanismdevelopment could be made available to the medical andscientific communities while in progress, and panel membersshould be encouraged to publish. It is anticipated thatthis would produce an effective cross-pollination effect, aspanel and panel member work inspired others not directlyassociated with the effort to produce research of their ownin the area, which may well collaterally resolve issues thepanel might have otherwise had to undertake itself, whilealso suggesting alternatives that might prove useful. Thusthe benefits of a single research project could bedisseminated rapidly in a manner that both advanced thestate of the art and engendered further interest andresearch that could benefit all, and do so more broadly,thoroughly, and quickly than a criterion-based effort.Additionally, the effort could be accomplished with panelmembers working in that role part time, such as two days amonth, which would impact their home facilities onlyminimally, reduce research costs, and spur interest outsidethe panel in its research efforts. Although it may beargued that the operational method proposed above for aprofessional panel could be better suited to measurementtools of the pen-and-paper variety, both simulations testingand actual field testing are also proposed for the testingof instrumentation of a fieldable system. It isacknowledged that there are disadvantages to the content andconstruct methodology as a whole and for the proposedversion in particular. These include on a broader scale theneed for iden"ifying anC obtaining the services of aslightly larger professional group than criterion validationmight (5-9 persons as opposed to 1-3), and getting a largergroup appropriate security clearances as necessary. Itwould require use of some of the finest talent in theirfields to assure that panel decisions were not readilyassailable scientifically or legally, though themathematical methodology proposed would mitigate at leastsome of this. It would also likely require higher travelexpenses to bring remotely-located experts to a commonlocation on a regular basis for panel meetings, andpersonnel turnover could also be a factor due to panel size.With regard to the specific content and construct panelformat proposed above, it would also mean that during betatesting and possibly some simulations testing, it might benecessary to have the services of the panel members for morethan the suggested two days a month (plus travel), but anumber of factors involving training cycles and specifictesting will mitigate this. In that the specific contentand construct methodology proposed above offers both

X-7

substantial offsets to the disadvantages and can beaccomplished for far less expense (essentially travel andexpenses associated with panel meetings, beta testing, andpurchases of samples of existing instrumentation forevaluation vice maintaining facilities and staff or hiringsomeone else who does), the format appears to offer distinctadvantages worth pursuing.

3. Testing Process

a. OversightRegardless of validation approach selected, experts

selected to develop and operate the testing and validationprocess are likely to have their particular areas ofspecialization as their primary perspective on HRPdevelopment. Coordination and management oftest/evaluation/implementation/operation of an HRP requiresa far broader perspective, but with a primary securityorientation. A Program Manager from Security could alsothus serve as the focus for logistical and coordinationissues pertaining to the development, testing, andimplementation effort, while retaining an overall focus ofeffective performance within suitability parameters before,during, and after operations in EREs.

RECOMMENDATION

Implement development, testing, and fielding of an HRPunder Security aegis.

b. Professional staffAs noted earlier, development and testing personnel

requirements reflect the type of validation strategy chosen.Projected minimum requirements for an HRP of the sortenvisioned in this study would include the followingpersonnel.

CRITERION--Two clinical psychologists with extensiveexperience screening in an I/0 environment, or one clinicalpsychologist and one I/O psychologist throughout. Apsychometrician would probably be necessary periodically toassist in assessment of results. A secretary/techniciancapable of assisting with testing administration andperforming secretarial/clerical duties would be neededthroughout. During at least part of the development andtesting process, some or all will need to be co-located withthe test population. Operational personnel would have to bemore sensitive to any security nuances, however, as thismethod does present more inherent security risk than thecontent and construct method. Materials p-oduced thoughwould in many cases not need to be classified.

X-8

CONTENT AND CONSTRUCT--A panel of at least five andpreferably seven leading professionals would be needed todevelop and guide the testing and validation effort. Thougha primary purpose of the group size is to assure greaterlikelihood of more effective results by requiring consensus,it also presents the opportunity to increase breadth ofexpertise available: it is suggested that a five memberpanel be composed of an I/O psychologist, two clinicalpsychologists with experience in I/O environments, apsychiatrist, and a physiologist as voting members, one ofthe panel additionally serving as Professional Director. Aseven member panel would utilize the same specialists plus aneuropsychologist or neuropsychiatrist with I/O environmentexperience, and a medical administrator to serve asProfessional Director in leading the panel. Non-votingmembers should include a psychometrician and a PersonnelSecurity Specialist.

BOTH STRATEGIES--As deemed necessary or appropriate,occasional specialized governmental support as follows wouldbe needed: a doctor with medical policy expertise to assistwith policy integration issues (an excellent choice forProfessional Director if experienced in medicaladministration), a personnelist, a lawyer familiar withgovernmental scientific research liability case law, atraining officer, a mathematician with psychometric orstatistical experience, and a programmer for the operatingsystem employed. All would require sufficient familiaritywith ERE scenarios screened against to provide effectiveinput, which may result in additional requirements forsecurity screening for such panel members.

c. Population Requirements and OperationalCoordination

As earlier stated, there would be substantialdifferences in test populations depending upon validationprocess chosen.

CRITERION--This process requires a large population(several hundred to preferably thousands) of entirelyunscreened personnel. This unfortunately eliminates theeight thousand member U.S. Air Force Security Policecommunity, whose field level training elements had expressedinterest in being tested for such an effort. Preliminaryinquiries have indicated that it might be possible toutilize unscreened Air Force trainees, presumably differentphases of testing being done on different groups to minimizeimpact on training schedules. An effort based on this testpopulation would of course require establishment of apresence on or near Lackland Air Force Base in San Antonio,Texas.

It may however be more appropriate (and accurate) touse test populations more analogous to the actualpopulations to be screened for EREs, should it be possible

X-9

to identify them: patterns of thought, relationships, andresponsibilities may vary substantially between basictrainees and many ERE populations, particularly operationalones, and might throw off higher levels of sensitivitymeasurement to some degree.

CONTENT AND CONSTRUCT--Testing could be conducted onactual populations on a continuing basis, with testing inthe field under simulated conditions. Although most testingcould be done in more quiescent periods, such testing underfield conditions has obvious advantages, but it is suggestedthat non-obtrusive methods, if used, would be a moredesirable approach under such circumstances. This approachhas the advantage of time saving due to availability ofpopulations if development and implementation are desired,accuracy for the specific purpose (to the degree possible inan artificial environment), substantially greater securityassurance, and lack of need for facilities on an extendedbasis. Time lost for additional panel selectionrequirements in a content and construct process should beoffset by a less complex, shorter set-up time as opposed toa criterion-based strategy.

d. InstrumentationA variety of instruments will need to be devised for

ERE's, including highly specific tools for use in particularthreat scenarios due to the current lack of availabiletools. Although some are anticipated to be pen-and-paperinstruments, these are not seen as adequate of and bythemselves, at least for deploying/operational personnel:the problem is that people intentionally (due tounwillingness to self-report for whatever reason) orunintentionally (through misunderstanding, or more often,inaccurate self-perception) can skew results of a singleinstrument. The uniqueness of extreme risk operationalenvironments would be especially likely situations forunintentional error due to lack of experience withcircumstances of the types postulated, or in some scenarios,even an analogue. Exercises, due to the inability tointroduce sufficient realism into the training experienceare not likely to be analogues (and if training were capableof producing the necessary realism for EREs, it would soonalso produce a dearth of participants). Althoughinstruments can be devised to address behavioral,performance, psychological, physiological, and monitoringareas, specific tools and mechanisms of and by themselvesare similarly limited: life style polygraph, favored by Dr.Rees for its ability to force accuracy in self-perception,is a behavioral and physiological measuring tool only; itcannot address psychological or performance aspects.Physiological screening processes, while covering thataspect well and possibly some psychDlogical and/orperformance aspects (depending on configuration), can becumbersome. But an effective, validated combination ofinstruments can provide the necessary coverage and accuracy.

X-10

The best approach to the issue of specificinstrumentation for an HRP needed for missions in EREs wouldappear to be a multi-faceted one incorporating both existinginvestigative means (background investigations withscreening and investigative interviews) and tools speciallydesigned to thoroughly cover particular reliability aspectsbased on specific mission requirements and status(operational or nonoperational). The specific reliabilityinstrumentation must be developed by experts. They shouldbe provided with necessary baseline material to formulnitethe tools and testing process needed. In addition, morespecific criteria regarding system structures, application,and assessment requirements within scenarios can beaddressed to fit the mechanisms into an HRP.

F. MODELS--FRONT END SELECTION

1. Introduction

A number of specific and distinctive models of HRPsystem structures will be proposed on the following pages.With one exception for indirect support elements only, allare designed for use with operational, support, or bothcategories of personnel. They are designed as schematicstructures only to accomplish the overall purpose ofreliability assessment, but do not include yet-to-bedesigned specific instrumentation or existing measurementtools (e.g. MMPI II) that require professional justificationor others (e.g. polygraph) that require a policy decisionbefore incorporation into an ERE screening process. Themodels will provide adequate detail to allow productivedecisions regarding desired system structure. Havingselected a model structure and decided upon preferredvalidation strategy (either of the two proposed earlier canbe applied efficiently to any of the models), a projectmanager and experts can be obtained to flesh to structuresout.

It should be stressed that for all models suggested,the purpose of evaluation is for diagnostic vicedetermination purposes. Early detection in screening ormonitoring must be followed by intervention. After athorough review of the issue(s) surfaced, the diagnosticianshould make a report to the appropriate Personnel Securityofficials to pursue from there.

2. Sliding Scale

This model is designed to provide a full-range programfor multiple populations, with each population beingevaluated separately against each of the assessmentcriteria:

X-11

Figur4 3

WeightingP1 ------------------------------------------------ Level I

P2------------------------------------------------ Level II

P3------------------------------------------------ Level IIIAa Ab Ac Ad Ae

where P1 through P3 represent different populations to beevaluated, levels I-III represent minimum score levels foracceptability within a particular population, and Aa throughAe represent different assessment criteria. Thus populationP1, though evaluated using the same tools as population P3,may be considered to require a higher degree of capabilityto perform certain activities at a higher stress level dueto its specific mission requirements. Population P3 may beable to perform its mission perfectly well at the lesssensitive Level III assessment category. Only the weightingof the criteria varies. Factors outside the variabilityscale, such as a requirement for a background investigationmeeting DCID 1/14 investigative and adjudicative criteriawould continue to be applied as currently directed with thisas with all models proposed.

There are a number of advantages to the Sliding Scalemodel:

1. All personnel are assessed by all instruments,giving the broadest overview.

2. Tailoring for population variables is possible, dueto variable weighting.

3. It is relatively easy to administer.

4. It allows an overall system to be in place fairlyquickly.

5. Different assessment tools can be used in acentralized or decentralized locations.

Disadvantages of the Sliding Scale are as follows:

1. It needs to be fully developed before effectiveuse.

2. Facilities and personnel may not be available toprovide all testing required for an effective and efficientoperation.

X-12

3. Layered Model

Whereas the Sliding Scale model is designed to test allpopulations to some degree on all assessment criteria, thelayered model tests all populations equally on criteria usedbut not against all criteria. All would be tested againstsome criteria, and some against all, thus:

Figure 4

P1 I--------------I I

P2 I-------------------------I I I

P3 I-----------------------------------I I I II I I I

Aa Ab Ac Ad

where P1 through P3 are evaluated populations and Aa-Ad areassessment criteria. In the model, populations P1 throughP3 are all tested equally in criteria Aa and Ab, but PI isnot tested further.

Advantages to the Layered model include the following:

1. The same instruments and weighting are usedthroughout.

2. It's highly cost effective: testing beyondspecific population needs is not done.

3. It places less demand on testing and assessmentfacilities and personnel compared to full-range systems.

4. It's easily administered.

5. It allows easy shift in personnel status due toaccretional nature as opposed to modular or mixed systems.

Disadvantages include:

1. It has the least flexibility to politicalsensitivities.

2. There is structural inflexibility in formatarrangements.

3. It is easiest to target from a security perspective(though still very difficult, especially at more extensivetesting levels).

4. It needs to be fully developed before effectiveuse.

X-13

4. Modular Model

This model varies from the others in that it tests allpopulations against at least some assessment criteria, withthe same weighting applied to all criteria used.Diagrammed, an example of it would appear as follows:

Figure 5

( ) ( ) ( ) ( )Aa )-. ..( Ab )-....( Ac )-. ..( Ad( _ _ _ ) ( _ _ _ _ ) ( _ _ _ _ ) ( _ _ _ _ )

PI P1P2 P2 P2

P3 P3

Any number of configurations can be developed for testing aparticular population,limited only by the number ofassessment criteria times the number of populations to betested, yet since each assessment criterion (instrument)remains the same (or works the same way), manageability isretained.

Advantages to be found in the Modular model include thefollowing:

1. It provides a highly flexible structure whichallows ready tailoring for specific population needs.

2. It has great sensitivity to policy issues.

3. It's very cost effective: only testing perceivedas required is done.

4. Administration of testing instruments remainsconstant.

5. Tt can be implemented as testing instruments,facilities, and personnel become available.

There are also disadvantages to be found in this model:

1. Program administration is more complex due todifferent arrangements for different populations.

2. It's more difficult to assess continuing evaluationsupport requirements.

X-14

5. Mixed Model

In the Mixed model, all personnel would be testedagainst all or some of the assessment criteria at the sameor different weighting. A sample schematic might appear asfollows:

Figure 6

I I--- PI P2 P1,P2Weighting I

II I--- P2, P3 P1,P2 P3Level I

III I--- P3 P3 P3I

Aa Ab Ac Ad Ae

As such, this model is designed to incorporate the greatestdegree of personnel selection variability while at the sametime providing maximum structural flexibility and capabilityfor system adjustment.

Advantages of the Mixed model include:

1. It has maximum population assessment flexibility.

2. It demonstrates maximum structural and adjustmentflexibility.

3. It presents a very tough target to defeat from asecurity perspective.

4. It allows the best individual fit for the broadestnumber of different populations and missions.

5. It could be implemented as testing instruments,personnel, and facilities become available.

There are a number of disadvantages, however:

1. It's highly complex to administer and operate.

2. It's not very cost-effective unless truly largenumbers of candidates are to be processed.

3. It requires substantially more effort to developand evaluate.

4. It's likely to promote confusion and frustration.

5. It requires more centralized testing.

X-15

6. Nondeployment Model

Earlier in this study, a testing instrumentation wasproposed for use by both operational and support elements.Admittedly, the scope of effort required to administer suchinstrumentation could be substantial in large populations,depending on configuration. Alternatively, such tool(s)could be used either for nondeploying (indirect support)personnel, as a first screen for deploying (operational anddirect operational support) ones, and/or for monitoringpurposes.

Despite their individual screening limitations,pen-and-paper instruments offer the best option from alogistical standpoint to enhance personnel security byallowing improved intake screening and early detection ofpotential difficulties for effective intervention andtreatment. In order to achieve all these objectives, such amodel should contain instruments that meet the followingcriteria:

1. Testing instruments should be readilyadministerable by local security personnel with minimaltraining.

2. Testing instruments should be sensitive enough tomeet general program needs while not presenting undueburdens to tested personnel (i.e., no MMPI I clone).

3. Testing instruments should not be sostraightforward that desired answers are obvious, yet not socomplex in language as to confuse the average intellect.

4. If multiple question formats are used in apen-and-paper instrument, multiple formats should be usedfor monitoring purposes where regular observation isimpractical, impossible, or where a Monitor's (a person)effectiveness is being checked.

5. Screening of test instruments should be centralizedand automated if possible.

6. Testing instruments should be able to provideclarification of personal observations obtained duringmonitoring, and should be used for that purpose.

7. Format of the tool(s) should be suitable for useboth in intake screening and monitoring.

As noted previously, pen-and-paper instruments areimperfect screening/monitoring tools, and will not achieveas high a success rate evaluating personnel as a morecomplex, integrated process will. They should be able to beformatted to mitigate to a statistically significantpercentage either intentional or unintentional skewing,

X-16

however, and the obvious advantages of such easilyadministered, flexible, cost effective, and securityenhancing tools far outweighs the weaknesses presented.Since a tool(s) should be used solely as a diagnostic not asa determination, there should be no fears either thatincorrect responses mean that the person being tested is tobe separated: should significant pathology or investigativecriteria be suggested by the device(s), a thorough followuppsychological or psychiatric screening or limitedinvestigation of the criteria information should beconducted, as appropriate.

G. MONITORING MODELS

1. Introduction

A thorough but practical monitoring process isessential for the maintenance and operation of an effectiveHRP, but as noted previously, existing programs generallyshare a weakness in this area. This is however one area inwhich extant programs share a common general purpose withthe one proposed in this text: continuing assurance thatpersonnel in particularly sensitive populations can continueto be relied upon in life-threatening emergency situations.The DOD programs are designed to provide continuousmonitoring, but have some serious weaknesses in training andapplication. Additionally, it may not be possible toprovide continuous monitoring for all populations that mustbe relied upon to respond to particular emergency scenarios.Therefore, it may be necessary to look beyond thestraightforward approach advocated by the DOD programs inorder to achieve more practical results in such scenarios:for this reason, multiple monitoring configurations areexamined in this section.

2. Continuous Monitoring

As it implies, this is a relatively continuous,ongoing, essentially nonobtrusive review process during workhours and possibly social occasions by supervisors and/orpeers. Self-reporting of criterion information is arequirement. Intervention by the appropriate agent upon thesurfacing of criterion information by any means leads alsoto suitability rereview if deemed to meet regulatoryrequirement. This model was supported by the largest numberof interviewees, including the NSA staff, Dr. Rees, and Dr.Glogower.

A number of assumptions must be made for a continuousmonitoring system to operate efficiently:

1. All monitors must be trained to detect criterionactivity or information, and how to act upon it.

X-17

2. Populations to be monitored must be readilyobservable on a daily or near-daily basis, and for extendedperiods.

3. Intervention support must be readily available on a

continuous basis.

4. Monitors themselves must be monitored.

5. Professional retesting on a periodic or aperiodicbasis is necessary to maintain the validity of ongoingnonprofessional observation.

Although a continuous monitoring system as described iscapable of detecting and intervening on criterion issueseven where self-reporting fails to function adequatelythrough either deliberate (concealment) or nondeliberate(ignorance) means, there are serious weaknesses in applyingsuch a format in scenarios where key personnel or evenmajorities of particular populations who would be expectedto deploy for or support operations in an ERE either traveltoo frequently, work at offsite locations too much of thetime, work in relatively isolated areas, or work in areaswhere they might be the only individual required to respondto a particular scenario. Since these features are likelyto embrace many personnel and populations required in ERE's,the continuous monitoring model may well lack sufficientpracticality for application in ERE's.

3. Periodic Monitoring

Periodic monitoring need not be simply the same ascontinuous monitoring on a less frequent basis. It canreadily differ in a number of operational assumptions:

1. All monitors still must be trained to detectcriterion activity or information and how to act upon it,but may be fewer in number and/or more professional inconcentration.

2. Populations monitored must be available atscheduled times.

3. Intervention support must be accessible as needed.

4. Monitors themselves must be periodically monitored.

5. Professional retesting on a regularly scheduledbasis is still needed as a check on validity of monitoredobservations.

The periodic monitoring system is obviously easier toadminister, but its many faults are almost as obvious:

X-18

I. Personnel with problems can and will dodgemonitors, or be able to conceal their problems for thelimited time monitors are available.

2. The "snapshot" nature of the process (to use Dr.Glogower's word) makes it very difficult for monitors todevelop a thorough understanding of the individual, and thuseffective decisions regarding him or her.

3. If monitors are a separate or semi-separate group,by whom and how are they monitored? Most likely, they wouldbe monitored periodically by each other.

4. Aperiodic Monitoring

The aperiodic monitoring model has close parallels inthe DOD random urinalysis or polygraph programs. As such,populations know only that they will be monitored and/ortested on a random basis. This method is supported by Dr.Kurke for activities for whom continuous monitoring wouldnot be feasible. It does offer a variety of advantages to adispersed operational environment:

1. It requires trained monitors, but as with theperiodic model, these may be fewer in number and specializeto some degree.

2. It's very well suited to variable, highly mobilepopulations.

3. Intervention support is required only on anas-needed basis.

4. It allows discreet monitoring and variation inmonitoring patterns.

5. It allows monitors to be readily monitored.

6. It can accommodate professional retesting on acontinuing basis to verify validity of monitor observations.

There are of course some tradeoffs involved in using anaperiodic monitoring model:

I. As with periodic monitoring, it presents only a"snapshot".

2. Personnel with problems meeting interventioncriteria may be missed, at least for a time.

3. Retesting is more disruptive to tight schedules orcertain types of assignments, though the nonobtrusiveobservation process does not affect schedules.

X-19

5. Mixed Variable Monitoring

If a continuous monitoring model is impractical, and aperiodic or aperiodic one is felt to leave too many holes,perhaps a combination of the three can resolve some of thefailings of all. The combination proposed in this modelwould utilize supervisory and peer observation of allpersonnel on an "as available" basis wherever they might be,with random retesting.

The Mixed Variable Monitoring model is designed toincorporate as many advantages of the other models aspossible. These include:

1. It accommodates maximum flexibility in availabilityof monitored personnel.

2. It minimizes deviancy crediting by a singlemonitor.

3. It eliminates need to separately monitor themonitors in a smaller, narrower system.

4. It backs up observations with a random retestingmechanism, whose overall frequency and format is adjustable.

5. Though not continuous, it provides a far more

comprehensive view than "snapshot" models.

As with other monitoring models, there are tradeoffs:

1. It requires a far broader and more continuoustraining effort than models using fewer monitors.

2. Intervention support must be continuouslyavailable.

3. It is more costly and complex to administer than"snapshot" systems.

6. Monitors and Process

The types of personnel considered especially suitablefor selection as monitors has been discussed earlier in thisreport. Regardless of who or what type of personnel arechosen as monitors, however, certain requirements remainconsistant:

1. They must receive effective initial training withperiodic followup training. Simply telling people tobasically "look out for and report anything 'hinkey'"without their having a clear understanding of criteria toreact to will be disastrous.

X-20

2. Monitors must be aware of and trained in the shiftin criteria that will take place transitioning from anonemergency to an emergency environment.

3. Monitor observations must be backed up on a regularbasis by professional retesting.

4. For objectivity, monitoring and intervention mustbe completely separate.

5. For effective performance, monitoring must be anonobtrusive process. Creation of a "big brother" or"Gestapo" atmosphere would be destructive of bothperformance and the evaluation process.

6. Monitors must have sufficient access to personnelto be able to reliably detect variances in behavior,performance, and health.

7. Monitors themselves must be monitored by others atsome point, if not peers at the operational level, then atsupervisory or organizational level.

Suggestions that emerge from the above criteria arethat personnel selected as monitors should have sufficienttime available to observe individual activity in enoughdetail as to be able to make informed decisions. They mustbe alert to the issue of deviancy credits, yet at the sametime adequately trained (with reinforcement) not tooverreact. And they must be constantly alert to the need toremain objective.

Another issue that suggests itself upon review of theabove is that monitoring will also be needed in anoperational environment: in fact, it will be vital.Monitors in the operational environment may very well not bethe same ones as previous, nor will the reactive criteria bethe same. Additionally, it is obvious from earlier remarksthat the most effective monitoring arrangements will not beable to contend with the situation presented by the mostmobile population members. For them at least alternativemeans of continuing evaluation will be necessary.

7. Retesting

Some form of retesting process is necessary in anymonitoring program as a verification tool to providecontinued assurance that the rest of the program remains ontrack, and to provide a regular professional review insupport of any nonprofessional components of the process.Depending on the frequency of retesting determined necessaryby the experts the process itself might range from a toolwith characteristics similar to the one proposed earlier fornondeploying personnel, to that or another tool and an

X-21

interview with a psychologist, to a more comprehensiveprocess probably incorporating review of physiologicalstatus (in fact, the NSA staff suggested that morephysiological than psychological change might take placeover a five year period for a population as a whole) andpossibly proactive cognitive testing (also suggested by theNSA staff). Separate instrumentation specifically designedfor the monitoring program might also be developed but isfelt to be redundant and possibly productive of confusingresults since monitoring retesting should be remeasuringpreviously assessed criteria.

8. Intervention

An intervention capability and mechanism is anessential part of a monitoring program, although of courseit may also be used for entry evaluation as well. Thepurpose of having intervention capabilities is as previouslystated, to provide professional review of potential issuesdetected, and to afford retesting and revaluation ofmonitored populations. As a diagnostic vice a determinativeprocess, it should provide a professional opinion concerningcontinuation or separation to the existing personnelsecurity process for a final decision. Since this must beprofessionally staffed, there are only two options foracquiring the capability: create one or gain access to anexisting one. The NSA staff recommended the former courseas best suited to providing a capability geared to specificorganizational needs. In addition, there are otheradvantages to such a system:

1. It offers better assurance of quality controlthrough concentration on a specific capability.

2. It accommodates centralized administration andgreater responsiveness to change in operational or securityrequirements.

3. Allowance is made for clear control of ongoingtesting and refinement of new instrumentation andmethodology.

There are a number of obvious signif4 r'ant disadvantagesto such an approach, too, including:

1. It requires creating a new organizationalcapability, with attendant staffing, logistic, andadministrative requirements.

2. It's very expensive.

3. It creates a bureaucracy that will be less flexibleto organizational/administrative fluctuations.

X-22

By comparison, acquiring the necessary capability fromoutside offers other possibilities, both good and bad, andcan be sought either from a government or contractedorganization. These are summarized below as:

GOVERNMENT ORGANIZATION--

Pluses: Uses existing structural and professionalcapability, so is faster online, and may well be the mostcost effective option.

Negatives: Unless special arrangements are made,in-house needs will likely be met first.

Host agency fluctuation in evaluated populations willaffect external support capability.

Host agency policies may affect ERE HRP efforts,"orphaning" the program or affecting perceived supportrequirements.

Such program capabilities as are in place currently aregeared to administering fully developed instrumentation andnot originating improvements or new tools.

CONTRACTORS--

Pluses: Probably the most flexible from security,operational, administrative, and organizational standpoints.

Very capable of modifying or creating instrumentationto meet changing needs, and can be available as necessary.

Negatives: Probably the most expensive option.

It places a very critical personnel security evaluativeresponsibility outside the government.

H. RECOMMENDATIONS

1. Strategies

The model structures set forth earlier in this sectionare capable of being used either for selecting in, screeningout, or any combination of the two. The decision as towhich or what order or combination should be chosen may beeither political or expert opinion.

X-23

RECOMMENDATION

Continue with the current select in (Personnel)/screenout (Personnel Security) process with informal select in forteams/staffs not requiring greater formalization of thelatter process.

2. Application

The models proposed earlier in this section aredesigned to be applied to multiple populations andconfigurations. Specific population requirements may be apolicy decision, but should be tempered by a consensus ofexpert concurrence in grouping for evaluation by coincidenceof variables, which of necessity requires expert decision.

RECOMMENDATION

An HRP should be established for appropriatepopulations, with application of evaluation and monitoringcomponents to meet specific mission criteria as designatedby expert professionals.

3. Evaluation

Establishment of an HRP for EREs will requiredevelopment or adaptation of a variety ofevaluative/diagnostic instruments due to the especiallyrigorous nature of such environments. At the same time, theorganizational diversity of candidate populations mandatesan extraordinary requirement for flexibility in programstructure. Process selection in a multi-populationenvironment must be a policy determination, but theflexibility afforded by the various model structures allowsready expert professional addressal of the specificinstrumentation necessary to flesh them out. Either of theprimary validation processes may be used to achieve goalsmeeting these requirements, so the following recommendationis geared toward selecting the process considered mostlikely to produce the most broadly integrated HRP in theleast time for the lowest cost and security risk.

RECOMMENDATION

Use a content and construct validation process asdescribed earlier in this section. Develop/adoptinstrumentation as deemed essential by expert opinion tomeet mission requirements.

X-24

4. Models--Front End Selection

It would be possible with any of the model structuresproposed and either validation process described toconstruct an effective HRP. Hence, the decision as toappropriate model may be made on other factors pertaining tothe particular strengths of specific structures for themissions and populations to be supported and of coursepolicy realities. As with the preceding recommendation thenthe following one is that which is calculated to give themost "bang for the buck". Of necessity, it is not geared tospecific population and mission requirements, as these arenot identified.

RECOMMENDATION

The Modular model appears most effective for meetingthe broadest range of screening concerns, capability,andcost effectiveness. It therefore appears to be the bestoption overall, though it must be noted that: 1) the Mixedmodel is very similar, and differs primarily in providingsomewhat more flexibility offset by more complexadministration, and 2) the Sliding Scale is the mostcomprehensive.

RECOMMENDATION

The non-operational model is recommended for inclusionas a first module in any of the larger structures proposedinstead of as a stand alone in order to permit readytransition in status and inhibit redundancy in screening.

5. Monitoring Models

Practical considerations must be the key determinantwhere monitoring models are concerned, because althoughthose presented all have validity under certain conditionsonly one will best suit a particular scenario application.

RECOMMENDATION

The Mixed Variable model appears to be the best choicefor a single program incorporating disparate populations,scenarios, and missions because it offers the greatestflexibility and coverage in such conditions.

6. Intervention

An effective intervention mechanism and operation canbe created using any of the options presented, and ifmonitoring (or in most circumstances, front end selectiontoo) is to be used, some form of intervention is mandatory

X-25

in order to properly evaluate observations made. Except inthe behavioral area, an investigation alone is not likely tosurface events or precipitating factors. Even if it does,an investigation takes far more time and has no strongassurance of producing usable results in the reliabilityarena. In many cases, a combination of investigation andintervention testing and/or evaluation offer the bestcapacity to obtain information necessary to resolve issuesand provide the most accurate personnel securityadjudication.

RECOMMENDATION

Obtain necessary evaluative and intervention capacityfrom existing government sources. Of these, a source withat least limited research interest and capability would bebest suited for the needs of an ERE HRP.

7. Oversight

Although a general case was made earlier in thissection for security oversight, if multiple populations fromvarious organizations and entities are to be included in anERE HRP structure, then a mechanism must be emplaced thatcan do so across agency lines. The primary orientationwould remain one of security, but a broad policyimplementation capacity would have to be added.

RECOMMENDATION

Utilize an appropriate security oriented oversightgroup for a multiple agency HRP. If the effort were toremain within a single agency, oversight should be by thesenior security official.

X-26

This page intentionally left blank.

XI-I

Section X1. Other Issuesand Recomendations

A. INTRODUCTION

Some of the real benefits of any basic research are thethrowoffs generated, those relevant but not directlypertinent or required discoveries that are made as acollateral development of the primary themes and issues.This particular research has been no exception. A number ofitems have surfaced as alternatives or beneficialconsiderations that may well serve to make an ERE HRP moreeffective.

The items cited in the following subsections are of adisparate nature due to the collateral nature of theirorigin. No attempt was made to integrate them into theprimary models or with each other. Most of them havepreviously been examined from other perspectives: thisstudy examines them from the perspective of humanreliability. They remain options for the most part, withthe exception of the last or2 of all, Community Coordinationfor Fit, some form of which is essential if a multi-agencyHRP is contemplated.

B. STRESS TRAINING

Even in populations that train as often as feasible inorder to increase and maintain proficiency with operationalprocedures, systems, and mission requirements, there isseldom training provided to increase stress tolerance. Muchof the stress tolerance training provided inmilitary-oriented programs such as the Navy's SERE(Survival, Evasion, Rescue, and Escape) or SEAL (Sea, Air,and Land) programs is directed at narrow population groupswith very specific missions. While obviously beneficial,these techniques are not suitable for all ERE populations,nor do all ERE scenarios resemble SERE or SEAL operationalenvironments.

Stress training in an operationally orientedenvironment should deal with isques other than physiologicalstress tolerance or psychological passive or active copingtechniques to deal with the results of stress. It should bea proactive process aimed at increasing resistance tosources of stress. As stated by Driskell et al (1990):

"Researchers in the stress field have identifiedthree critical components of stress training: a)

indoctri:iation, b) skills training, and c)confidence drill. The purpose of the

XI-2

indoctrination component of stress training is toprovide the individual with information about thestress environment. Knowledge provided in thismanner is relatively straightforward, and may bedelivered via classroom sessions, handbooks, andthe like. The purpose of skills training is toprovide the individual with exposure to the stressenvironment, have the individual experience skilldegradation in terms of decreased performanceaccuracy or speed, and then develop work-aroundprocedures to allow adaptation to the environmentso that performance returns to an acceptablelevel. The purpose of the confidence drill is toprovide the individual with exposure to the stressenvironment, to allow the individual to operateeffectively in that stress environment, and todevelop positive expectations regarding his or herability to perform under stress conditions."(p.33)

Sound familiar? In this model, however, systemperformance is manipulated to create, confront, challenge,and be conquered by an effective response to an operationaldegradation rather than an equipment-oriented troubleshootand reroute designed to teach only a specific fix for aspecific problem: the latter may instead be a vehicle forimprovement of the operator's ability to respond. Theintention is to create a high confidence expectancy so that:

"...in dangerous situations, individuals who havea high confidence expectancy would focus on thetask at hand, while those low in such expectancywould direct their attention to the danger stimuli

and attempt to avoid them." (Keinan, 1986, p.186)

An effective HRP can be constructed for EREs withoututilizing stress training, but to use it productively wouldenable higher confidence in all aspects of reliabilityfocused upon in this report. Selection for stability thatmight be enhanced by training provides the best hope forcontinued security and performance in an operationalenvironment. Rote or hands-on hardware training will notaccomplish the tasks effectively in an analytical ordecision-oriented environment (though it does in a militaryclose combat environment) because knowledge-related problemsare not degraded by stress, whereas computationally-orientedspatial demands are (Wickens et al, 1989).

C. CROSSTRAINING

Many operational pnpulations already do at least somecrosstraining, in order to create greater depth in availableresources, as stressed by Dr. Ginzburg in interview. It

XI-3

also has value as a stress-reduction mechanism. Involunteer-based ERE populations, if any, and even manyinvoluntary ones, crosstraining will help deal with a factof life in operational conditions: failure of at least somepersonnel to report for duty as required.

Thorough crosstraining would prevent or minimize singlepoint of failure issues, increase personal and groupconfidence in their endurance, and increase stress toleranceby adding variety to tasks and perceived usefulness. Thetradeoff is longer and/or more frequent training.

D. MOTIVATION AND RESPONSIBILITY

It reflects the highest credit on their remarkablesense of "duty, honor, and country" as the Army puts it thatpersonnel may be found to volunteer for or accept assignmentto populations responding to ERE situations. It also speaksmuch for their motivation and level of responsibility. Yetfor all this they remain human beings, and they will faceagain major decisions they thought they had already settledin their own minds if they must deploy for operations in orsupporting an ERE scenario instead of just practicing duringan exercise, or remain on station at a location likely to bedestroyed momentarily because the greater situation demands.Without a reliability program that has as a partial objectthe maintenance and enhancement of motivation and commitmentin the most extreme of stress conditions, many more will belikely to fail the challeage than would be the caseotherwise. This will create greater security risk in anumber of areas including response actions and the securingof replacements, and will obviously degrade operationalcapability.

Motivation and responsibility are primarily behavioralcharacteristics. Behavioral aspects of an HRP to supportselection and operational effectiveness should focus uponmeans of selection and screening of candidates that not onlymeet ERE requirements, but demonstrate clear likelihood tocontinue to meet patterns of conduct most able to assuresafety and operational capability for the mission,themselves, and their fellows. In addition to screening forthat and monitoring continued compliance, other actions areneeded to provide the highest confidence in candidateselection and retention. Three of these areas are exploredin the following.

1. Contracts

Contractual arrangements specifying what the Governmentexpects from operational and direct support populations area mechanism that could be used to emphasize the seriousnessof the relationship, give a better idea of what it entails,and enhance the concept of commitment for the individual.Such a contract instrument would be offered at the time of

XI-4

initial affiliation with a population operating in orsupporting operations in an ERE scenario. If securityconstraints prevent revelation of actual scenario criteriaor position requirements, it would also be possible andadvisable to warn the candidate after he/she signs aNondisclosure Agreement that they should not proceed withthe following indoctrination if they are unprepared toaccept a major personal and National operational commitment.

No contractual procedure or instrument will sway aperson determined not to honor it, and the Government may beunwilling or unable to seek to prosecute such an individual,depending on scenario conditions. In an ERE, however, itwould be counterproductive to retain the unwilling even in apre-operational, preparatory phase. It is felt though thatthe purpose of the contractual instrument should serve as ameans to focus the candidate's attention on the commitment,not as a bludgeon to enforce compliance: persons who failto honor it would be undesirable for employment under anycircumstances.

It is acknowledged that the proposed contract would notnecessarily meet the purposes of such tools before the law(if there was no certainty of prosecution for violation);it would be more broadly a "gentlemen's agreement". Assuch, having only a questionable practical validity in law,legal opinion might well hold that the instrument has novalue. Given the potential severity of some scenarios,however, the Government might well be unable to enforce thecontract even if it desired to do so. Additionally, from asecvrity perspective the agreement does have validity,purpose, and value in that it gives some evidence of apersonal commitment, a stated intention, on the part of theindividual, whose personal honor and sense of duty willultimately be the determining factors in his/her deploymentin response to an ERE scenario. Similar documents have beenand are being used by Federal agencies for certainpopulations, with the same legal and security considerationsin mind. Naturally, depending on individual and degree ofexposure to the ERE, the fight-or-flight reaction may beactivated anyway, and best intentions overridden.

2. Grandfathering

While the policy of grandfathering people already in aprogram in the event of a major policy change orimplementation does produce more acceptance from currentpopulations, it also produces inequitable application ofpolicy and allows persistence for an extended time ofproblems the change was designed to correct. In addition,in a human reliability program of the type envisioned wherea monitoring as well as screening program is recommended, itbecomes nigh unworkable.

Should the recommendations in this study be adopted toestablish an ERE HRP, it will be necessary to assemble asmall establishment to create the specialized

XI-5

instrumentation and mechanisms for the necessaryapplications. Some of the structural formats proposedrequire full establishment before use; others can beimplemented piecemeal. Regardless, there will be a periodof one or more years before effective products are likely tobe available. Announcement of the impending programinstitution over such a timeframe would permit the departureof those who wished to leave with no questions asked, whilenew accessions could be apprised at or even beforeaffiliation.

RECOMMENDATION

1. Deploy an HRP as tested and ready, either piecemealor wholesale depending upon structural compatibility.

2. Deploy a monitoring program at its widestoperational extent when tested and trained.

3. Do front end screening at individual five yearbring-up points in conjunction with a single scopebackground reinvestigation or full field reinvestigation.

3. Ergonomics

Ergonomics is an area that is often ignored to asubstantial degree in scenario planning for response to manytypes of contingency or operational plans including EREs.During short term scenarios, there may be little need toconsider the effects of discomfort produced on thepopulation responding, but where functionality issignificantly impaired, or during any longer-term (more thana few days maximum) scenarios, lack of attention toergonomics will have a profound effect on stress tolerance,causing emergence of interpersonal conflict, loweredproduction, and incorrect decisions and recommendations inan operational ERE. At the same time, care must be taken inthe selection of ergonomic devices and tools; as a verylimited example, muted coloring and posters of abstract ornear-abstract subjects may be helpful to supporting stresstolerance, where bright, clashing colors or posters ofscenic wonders or places of personal importance that mightbe lost as a result of the scenario's occurrence would havethe potential to create depression, anger, and an increasedsense of loss. Space and costs of ergonomic modificationare genuine factors in the decision, but they should bebalanced against a carefully considered review of thepotential impact of not making the modifications.

XI-6

RECOMMENDATION

Exanine (or reexamine) the issue of ergonomics inapplicable operational environments, making changesparticularly where possible without affecting or having onlyminimal affect on configurations, with an eye to stressreduction. Use of a specialized contractor for thisrelatively short term requirement may be advisable.

There are undoubtedly other mechanisms and toolsavailable to enhance motivation in personnel regardless oftheir status that have not been mentioned in this briefwork. There are also many more obvious issues to addresseven in mature operational systems that might be expected torespond to EREs. The area of ergonomics may well in theevent of an operational scenario deployment represent amissing nail in the shoe of the leader's horse.

RECOMMENDATION

Explore means of motivational enhancement inoperational EREs for personnel responding to them.

E. PHASED RESPONSE

Creation of a particular extreme risk environmentrequires an appropriately strong response by populationsdesignated to deal with the scenario. In some cases,however, it may not be obviously necessary to deploy allavailable assets at once, either to retain a reserve forextended operations, or to deal with a less than planned forcataclysm, or for precautionary purposes upon anticipationof an ERE. In the event that it is not clearly necessary tofully deploy assets to face an ERE scenario, it is suggestedthat a phased response would be the best option until an HRPdesigned to deal with EREs is established. A phasedresponse would do much to alleviate potential problems intwo areas. Such a deployment pattern would, if begun veryearly or with minimal warning upon initiation of the ERE,present in many cases fewer conspicuous threats to thepersonal concerns of responding personnel, which shouldresult in an easier decision to report for deployment andthus mean that more personnel will do so. It also allowssubstitution from later-deploying elements if desired tofill holes in earlier-departing ones. It would allowattempts to be made to substitute for losses due to changedminds if desired and appropriate personnel were available.

If the scenario requires application of operationalsecurity (OPSEC) process, a phased deployment would alsooffer a number of advantages: there would be less activityat any given time to draw adversary attention in a combatsituation, there would be a lower operational profile lesslikely to cause undue alarm to the public, and it would

XI-7

allow concentration of limited security resources to assistin the OPSEC process for each phase, rather than diffusingassets to cover multiple phases at once.

RECONMENDATI ON

Consider use of the phased deployment option if viablein a particular ERE context, especially before fielding anHRP for responding populations.

F. NONOBTRUSIVE EVALUATION TECHNIQUES:VOICE STRESS ANALYSIS

In addition to more conventional test and evaluationinstruments that might be devised for an ERE HRP, lesscommon techniques such as biometrics may be approachinglevels of statistical reliability sufficient to warranttheir consideration under certain circumstances as programcomponents. If techniques introduce sufficient interferenceinto individual response as to alter response time orquality or produce other artificialities in behavior orperformance, they lose their validity. For this reason,nonobtrusive means are considered to offer better prospectsfor useful testing and application in an HRP both as asecurity and as an operational (performance) diagnostic.Nonobtrusive means would be especially valuable inmonitoring and even in some screening applications as a wayof assessing type or degree of stress. The particulartechnique chosen for these remarks is voice stress analysis.

A promising but insufficiently proven technology formany years, voice stress analysis is beginning to accumulateenough of a body of research to begin to support someconsistent results. Further, it has now passed one majorhurdle of acceptability according to Dr. Strome in aninterview, in that the National Transportation Safety Boardwas able to get voice stress analysis accepted as evidencein court for the first time in the case involving the ExxonValdez. Research has cited seven voice characteristics(Jones, 1990) as offering potential for evaluation; in thereference cited, three of these were examined, and resultsconsistent with prior research were identified thatfundamental frequency (pitch) was the best stress indicatorfor voice.

The purpose of this subsection is not to endorse anyparticular technique or procedure, but rather to note thatthere are emerging areas in research that may offer addedvalue to an ERE HRP. Another technique that may beespecially helpful in performance evaluation for examplemight be dichotic listening, which is currently available.Some other biometrics may also be useful.

XI-8

RECOMMENDATION

Have a research effort aimed at developing tools for anERE HRP also look into valid emerging techniques aimed atnonobtrusive evaluation.

G. PERSONAL RELATIONSHIP ISSUES

Given the definition of EREs used for this study,persons with whom an individual may have close personal tiesas members of the same responding team, element, ororganization are likely to become casualties. In some EREscenarios, the magnitude of the threat may be sufficient topresent risk of loss of family members even if notcollocated with the individual. In either case, an ERE HRPwould be forced to deal with a formidable potential factorthat might result in failure. This is because of a commontendency of persons to be reluctant to see those emotionallyclose to themn assume risks they themselves might accept. Ifpersons emotionally close to an individual can be sparedsome of the degree of risk the individual calculates theywill be exposed to, he/she will generally seek to preventtheir assuming the full degree of risk. This is true evenwhen the act of sparing others is likely to place the personacting in greater personal danger themselves. At the sametime, the absence of such relationships might lead toacceptance by the acting individual of an unnecessary degreeof risk, as might be readily observed on any highway--singledrivers are far more likely to be aggressive or takeunnecessary risks than those with a family or carpool in thesame vehicle.

The Gordian knot presented by these issues is made evei,more complex by the fact that emotional ties and responsesto them are at least somewhat fluid and to at least somedegree perceptually based. Because of them, the individualmight not be aware himself exactly what his/her reactionmight be to a threat to another person--who couldfurthermore even be a non-threatening stranger.

Earlier in this study, the issue of personalrelationships was cited as being rather outside the scope ofthis immediate effort. It has been included in this Sectionbecause although it has great validity in some scenarios, ithas little or none in others. Nor is there any apparentAlexander to reso've all the many interrelated factorspresented, including this report--solutions here appear tobe only partial ones.

Having stated this, it should also be noted that acombination of approaches to the personal relationshipsissue may serve to mitigate at least some of the impact ofthe problem for those ERE scenarios that would be affected.One source interviewed for this study, Dr. Patterson, feltthat the issue of questionability of support for family orgroup members who might also be exposed to an ERE during aparticipant's involvement was a training, not a selection,

XI-9

problem. Certainly an absence of training to cope with suchissues might cause security and operational threats thatcould become unmanageable in an ERE that threatens personsemotionally close to the operational and direct supportpersonnel. Such training might (as very brief examples)appeal to immediate team/group ties over non-collocatedpersons, whose status was uncertain and not readilyascertainable, emphasize inadequacy of the individual'sefforts to mitigate effects by themselves, or address otherfactors leading to increasing personal defenses againstperceived threat of loss. Selection processes can also helplower the risk to some degree by screening for personnelpredisposed to a number of factors perhaps including thefollowing:

1. Exhibiting low central nervous system arousal.

2. Willingness to leave family or others to whomemotionally close to go do dangerous work.

3. Having a strong affinity for a "higher mission".

Reliance solely on selection will not produce adequateassurance that personnel will report at all, will not reportwith family in tow where provisions for same had not beenmade, or will not disrupt operations because of anindividually-perceived threat to others emotionally tied to;these variables lie beyond parameters assessable byscreening alone. There must also be a thorough,sophisticated, sustained training effort directedspecifically at the issue, and monitoring for effect.Spread throughout a population responding to the ERE, thiswould have the added benefit of encouraging all personnelthat their group performance can mitigate the ERE's effects,will mitigate the effects, and must do so. For those EREscenarios where possible to do so, substantial overmanningand crosstraining offer additional assurance that adequatepersonnel will be available and functional to support orsustain operations in the ERE.

Complete equity in arrangements and applicationregarding personal relationship issues is essential inresponding to an ERE scenario that threatens more than anindividual or group sharing a common likely fate. If someare allowed to make arrangements for protection of others,all must; if some are not, none can. To do otherwise woulddestroy morale and team or organizational cohesion,preventing effective response. It would also have strongpotential to inject personal s'rife and violence.

H. ALTERNATE SCENARIOS

The authorities interviewed by Mr. Nelson agreed that ascenario the proposed HRP would be used in was necessary as

XI-lO

a baseline in order to develop a workable program. Drs.Crawford and Wiskoff in particular felt that the operatingpremises for this study were inadequate for effectiveplanning, so Dr. Crawford proposed the following elements asthe most basic scenario criteria to define programparameters against:

1. Long or short buildup?2. Quick or slow response?3. Multiagency or single?4. Availability of resources?5. Screening that could be done before/after an

emergency?

Baseline criteria were set forth at the beginning ofthis study, but the framework of the structures and issuespresented in this study seek to address requirements againstas many of the parameters cited as possible in order toachieve maximum flexibility for any overall program formatultimately chosen. Should it be decided that the structuresand issues presented herein are insufficiently elasticnonetheless, the above criteria may be used for theconstruction of additional scenarios for special-purposevariants, or even other unrelated scenario and responseconstitution.

I. COMMUNITY COORDINATION FOR FIT

Each Federal agency and department has its own uniquereporting and decision making processes, and each hasspecific concerns about the formulation of new policypotentially affecting their personnel. This documentattempts to identify overall issues about which decisionswould be required should an ERE HRP be established with theintention of supporting multiple agencies or departments,but no attempt has been made to deal with department oragency-specific issues. Instead, this study deals with themost basic and essential of larger issues: whether or notan HRP would be needed for effective response to EREs;whether or not an ERE HRP could be designed to be usable ineither single or multi-agency situations; whether or not toestablish such an HRP; whether or not to participate indevelopment of an ERE HRP as postulated; and which modeland format the structure of an ERE HRP should follow. As nopurpose would be served by proposing or developing specifictools to fill the structure with until a decision is made toestablish an HRP, only processes are identified. It ishoped that the degree of flexibility built into the approachtaken in this work will both encourage and enable thoseinterested to undertake the necessary courses of action eachmust in order to participate in what could be made the mostadvanced and innovative program of its type. The vehiclesproposed are intended to be able to achieve compatibility

XI-11

with any existing regulations and policies not forbiddingparticipation in any type of HRP.

Although the approach taken in this study is for asingle program that might be useful to multiple agencies,there is obviously no requirement that it be administered asa single program for multiple agencies, either. While thereare certain financial advantages if several agencies share aprogram, since most will not have substantial numbers ofpersonnel requiring the intensity of screening provided inan HRP of this sort, an agency wishing to utilize thematerial developed for an in-house program rather thanparticipating in a multi-agency one can clearly do so.

XI-12

This page intentionally left blank.

XII-I

Section XII. Summary and Recomendations

A human reliability program seeks to satisfy two basicareas of pertinent inquiry: prediction and detection.Existing systems rely upon minimal predictive means and evenless detection. Additionally, the focus of existingprograms is chiefly behavioral suitability at selection andadministrative followup afterward.

Extreme risk environments of the types theorized forthis study require an HRP that will provide not only themost effective feasible security selection and monitoring,but also a mechanism designed to assure that the personnelmeeting security needs will also meet critically demandingoperational ones: that they will report for duty and beable to function in EREs. Existing HRPs don't offer enoughassurance of meeting such rigorous requirements.

The lack of existing HRPs to satisfy the requirementsof EREs demanded both exploration of existing research inrelated areas and development of entirely new designs andprocesses. In addition to a number of highly flexibleformats for program architecture, a number of pertinent orassociational throw-offs were generated that address areasbeyond the specific scope of this study but which offer hopefor an even better program if pursued. A less evident butno less real potential available in most of the models andthrow-offs is that though specific structures and issueswere addressed, they can also generally be combined inmultituidinous other ways and still produce workable,effective results. The architectural models can be used forsingle-agency up to Government-wide application.

The bottom line is that to assure that EREs can beeffectively dealt with requires an HRP designed to do so,that the structures are available to build an EREprogram-specific architecture, and that the individual toolsfor use in the architecture are available or can bedeveloped.

1. Recommendations Summary

Executive Summary: A human reliability program shouldbe instituted to ensure that adequate numbers of personnelcapable of meeting both the necessary security andoperational concerns are available for response to extremerisk environments.

1. Compile an unclassified version of this studywhether or not a classified version is done. (p. 11-4)

The Director, OEO, verbally authorized the developmentof an unclassified version of this report on July 16, 1991.

XII-2

2. Establish a limited HRP screen even for indirectsupport personnel for EREs (which could also be used as aninitial screen/identifier for operational selection and/ormonitoring test instrurent). (p. X-4)

3. Implement development, testing, and fielding of anHRP under Security aegis. (p. X-7)

4. Continue with the current select in(Personnel)/screen out (Personnel Security) process withinformal select in for teams/staffs not requiring greaterformalization of the latter piocess. (p. X-23)

5. An HRP should be established for appropriatepopulations, with application of evaluation and ronitoringcomponents to meet specific mission criteria as designatedby expert professionals. (p. X-23)

6. Use a content and construct validation process asdescribed earlier in this section. Develop/adoptinstrumentation as deemed essential by expert opinion tomeet mission requirements. (p. X-23)

7. The Modular model appears most effective formeeting the broadest range of screening concerns,capability, and cost effectiveness. It therefore appears tobe the best option overall, though it must be noted that:1) the Mixed model is very similar, and differs primarilyin providing somewhat more flexibility offset by morecomplex administration, and 2) the Sliding Scale is the mostcomprehensive. (p. X-24)

8. The non-operational model is recommended forinclusion as a first module in any of the larger structuresproposed instead of as a stand alone in order to permitready transition in status and inhibit redundancy inscreening. (p. X-24)

9. The Mixed Variable (monitoring) model appears to bethe best choice for a single program incorporating disparatepopulations, scenarios, and missions because it offers thegreatest flexibility and coverage in such conditions.(p. X-24)

10. Obtain necessary eval'iative and iiterventioncapacity from existing Government sources. Of these, asource with at least limited research interest andcapability would be best suited for the needs of an ERE HRP.(p. x-25)

11. Utilize an appropriate security oriented oversightgroup for a multiple agency HRP. If the effort were toremain within a single agency, oversight should be by thesenior security official. (p. X-25)

XII-3

12. 1) Deploy an HRP as tested and ready, eitherpiecemeal or wholesale depending upon structuralcompatibility.

2) Deploy a monitoring program at its widestoperational extent when tested and trained.

3) Do front end screening at individual five yearbring-up points in conjunction with a single scopebackground reinvestigation or full field reinvestigation.(p. XI-5)

13. Examine (or reexamine) the issue of ergonomics inapplicable operational environments, making changesparticularly where possible without affecting or having onlyminimal effect on configurations, with an eye to stressreduction. Use of a specialized contractor for thisrelatively short term requirement may be advisable. (p.XI-5)

14. Explore means of motivational enhancement inoperational EREs for personnel responding to them.(p. XI-6)

15. Consider use of the phased deployment option ifviable in a particular ERE context, especially beforefielding an HRP for responding populations. (p. XI-7)

16. Have a research effort aimed at developing toolsfor an ERE HRP also look into valid emerging tecbniquesaimed at nonobtrusive evaluation. (p. XI-8)

XII-4

This page intentionally left blank.

A-i

Appendix A

Version 1

1. Are any staff involved with research, test, orevaluation aspects?

May I be shown results? Particularly interested in basisfor using procedures (objectives for program, bothphilosophical and operational), selection/deselectioncriteria, whether or not there are specific-purpose variantsthat have been examined, overall effectiveness, impact oflife-threatening stressors.

. Are staff aware of research on the component elements ofc.he overall program, standing alone or together? Results?

3. Are staff aware of research into impact on an individualin program from family/close personal relationships?

4. Has research been done on predictability offunctionality in circumstances of threat to life of theindividual vice family?

5. In the staff's opinion, what specific factors must bepresent in an individual to assure ability to properlyfunction in a life-threatening environment? Can they beinculcated? What additional factors must be present toassure ability to properly function if the threat is toimmediate family or others similarly close to theindividual? What degree of predictability or confidence canbe determined, and what degree gained by application ofoutside affect?

6. If research has been done into functionality of anindividual in a threatening environment to self and/orfamily what degree of confidence is there in predictabilityfor a group under such stress, and over increasing periodsof time?

7. What impact can life-threatening stressors be expectedto have, immediately and over time, on various age groupsand both sexes in otherwise satisfactory health? Can thesebe effectively mitigated while not seriously impairingfunctionality?

8. Most regulations reinforce importance of dailyobservation. Has research been done regarding effectivenessof this aspect? Are there any viable alternatives? Can aprogram be successful without close observation?

A-2

9. Are staff aware of catastrophic failures in screening?Research re same? Results? Solutions?

10. What procedures would have to be undertaken in order toobtain substantially reliable data on questions unanswered?

11. Are staff in contact with others doing R, T & E? Anyrecommendations?

12. Are existing capabilities at your or other agenciesyou're aware of capable of handling additional personnPlscreenings, should they be sought?

13. Assuming research is available to undertake a viableprogram at X level, what type and extent of resources wouldbe necessary? Are psychiatrists or psychologists (or otherspecialists) preferred or necessary? What aboutsupport/screening staff ratios? Facilities?

A-3

Version 2

1. Are any staff involved with research, test, orevaluation aspects?

May I be shown results? Particularly interested in whetherthere are specific-purpose variants that have been examined,overall effectiveness, impact of life-threatening stressors?

2. Can I obtain a copy of the standard psychiatricquestionnaire? What other standard formats, e.g. MMPI, areused in routine or nonstandard testing?

3. Is research, testing, and evaluation ongoing in thisarea? What areas are/have been explored, and whatconclusions have been reached? Can results be quantified?

4. Has research been done on predictability offunctionality in circumstances of threat to life of theindividual vice family?

5. In the staff's opinion, what specific factors must bepresent in an individual to assure ability to properlyfunction in a life-threatening environment? Can they beinculcated? What additional factors must be present toassure ability to properly function if the threat is toimmediate family or others similarly close to theindividual? What degree of predictability or confidence canbe determined, and what degree gained by application ofoutside affect?

6. If research has been done into functionality of anindividual in a threatening environment to self and/orfamily what degree of confidence is there in predictabilityfor a group under such stress, and over increasing periodsof time?

7. Most regulations reinforce importance of dailyobservation. Has research been done regarding effectivenessof this aspect? Are there any viable alternatives? Can aprogram be successful without close observation?

8. Are staff aware of catastrophic failures in screening?Research re same? Results? Solutions?

9. What procedures would have to be undertaken in order toobtain substantially reliable data on questions unanswered?

10. Are staff in contact with others doing R, T & E? Anyrecommendations?

11. Are existing capabilities at your or other agencies

A-4

that you're aware of capable of handling additionalpersonnel screenings, should they be sought?

12. Assuming research is available to undertake a viableprogram at X level, what type and extent of resources wouldbe necessary? Are psychiatrists or psychologists (or otherspecialists) preferred or necessary? What aboutsupport/screening staff ratios? Facilities?

13. If your agency were unable or unwilling to process thenumbers of personnel involved, would you be able to trainother medical personnel in your techniques? How long wouldtraining take? Any estimates as to how long it woule taketo field such a program, and how much it might cost?

A-5

Version 3

1. Are any staff involved with research, test, orevaluation aspects?

May I be shown results? Particularly interested in whetherthere are specific-purpose variants that have been examined,overall effectiveness, impact of life-threatening stressors?

2. Can I obtain a copy of the standard psychiatricquestionnaire? What other standard formats, e.g. MMPI, areused in routine or nonstandard testing?

3. Is research, testing, and evaluation ongoing in thisarea? What areas are/have been explored, and whatconclusions have been reached? Can results be quantified?

4. In the staff's opinion, what specific factors must bepresent in an individual to assure reliability? What degreeof predicability or conficence can be determined, and whatdegree gained by application of outside affect?

5. If research has been done, what degree of confidence isthere in predictability over increasing periods of time?

6. Most regulations reinforce importance of dailyobservation. Has research been done regarding effectivenessof this aspect? Are there any viable alternatives? Can aprogram be successful without close observation?

7. Are staff aware of catastrophic failures in screening?Research re same? Results? Solutions?

9. What procedures would have to be undertaken in order toobtain substantially reliable data on questions unanswered?

10. Are staff in contact with others doing R, T & E? Anyrecommendations?

11. Assuming research is available to undertake a viableprogram at X level, what type and extent of resources wouldbe necessary? Are psychiatrists or psychologists (or otherspecialists) preferred or necessary? What aboutsupport/screening staff ratios? Facilities?

A-6

This page intentionally left blank.

A-7

Appendix B

Terminology

Back-end loading Focus of the review process ison-the-job performance afteraffiliation.

Confidence expectancy The individual's subjectivelyassigned probability ofincurring serious physicalinjury or death (Keinan, 1986).

Deviancy credits Inexcusable behavior that isexcused/deliberately overlookeddue to prior establishmentof a good record. (Dr.Glogower)

Dichotic listening Testing mechanism providing twoinformation streams, one in eachear, slightly out of sync, withalternate tasking by ear.Multiple single tasks can thenbe layered or used withvalidated testing measures.Embedded tasks may be included.(Dr. Strome)

Front-end loading Focus of the review process ison initial selection.

Locus of control An individual's enduring,cross-situational beliefs aboutcontrol over outcomes ofimportance (Barge et al, 1984).

Monitoring process Continuing review of employeesfor compliance with standards.

Selection in Approving only those personnelwith demonstrated ability orpotential to meet requirementsof a specific assignment; allothers are disapproved.

Selecting out Disapproving only thosepersonnel contraindicated bystandard; all others areapproved.

A-8

Selection process Screening of a candidate to meetstandards.

Self-reporting The reporting of informationconcerning oneself.

Self-selection The individual himself choosinga course of action for him/herto pursue.

A-9

Appendix C

Experts Interviewed

CAPT Chandler, USN XO, USNH San DiegoMAJ Ronald Chapman, USAF Psychologist, NSADr. Kent Crawford, PhD. Personnel Security Research CenterMr. Pat Dowd Crew Technology Division, School of Aerospace

MedicineDr. Douglas Eddy, PhD. NTI, Inc.COL Joe Fagan, USA Psychiatrist, Office of the Surgeon

General, USADr. Edna Fiedler Psychologist, Clinical Research Service,

Lackland AFBMAJ Edwin Gerwell, USAF Behavioral Analysis Service,

Lackland AFBDr. Harold Ginzburg Psychologist, Public Health ServiceLCDR Fred Glogower, USN Psychologist, Psychology

Department, USNH San DiegoLTCOL Neil Hibler, USAF Psychologist (Interviewed by S/A

Donald Brenno, MI, USA)MAJ Kolski, USAF Psychologist, NSADr. Martin Kurke Psychologist, Drug Enforcement AgencyMr. Richard McMurray, Medical Officer, CIAMAJ John Patterson, USAF Psychologist, Neuropsychology

Division, School of Aerospace MedicineCOL John Plevis, USA Psychiatrist, Office of the Surgeon

General, USAMr. John Reardon Public Health ServiceDr. Richard Rees Psychologist, CIAMr. William Robbins Medical Officer, CIAMAJ Walter Sipes, USAF Psychologist, Neuropsychology

Division, School of Aerospace MedicineDr. David Strome, PhD. Systems Research Laboratories, Inc.Dr. Michael Wigglesworth Psychologist, NSADr. Martin Wiskoff, PhD. BDM, Inc.

A-1O

This page intentionally left blank.

R-1

References

1. Advisory Group for Aerospace Research and Development(AGARD), (1989). Conference Proceedings No. 458, Humanbehavior in high stress situations involving aerospaceoperations.

2. Air Force Regulation (AFR) 35-99 (Cl), 1988. Nuclearweapons personnel reliability program.

3. Anderson, Harald T. (1989). "Prediction ofpersonality". In Advisory Group for Aerospace Research andDevelopment (AGARD) Conference Proceedings No. 458, Humanbehavior in high stress situations in aerospace operations.

4. Antonovsky, A. and Bernstein, J. "Pathogenesis andsalutogenesis in war and other crises: who studies thesuccessful coper?" Stress and Coping in Time of War. Ed.N. Milgram. Brunner and Mazel, 1986, pp. 52-65.

5. Barge, B. N.; Hough, L. M.; and Dunnette, M. D.(1984). Behavioral reliability: a review of academicliterature and organizational programs.

6. Barnes, I. and Lurie, 0. E. (1989). "The descent fromolympus: the effect of accidents on aircrew survivors". InAGARD Conference Proceedings No. 458, Human behavior in highstress situations involving aerospace operations.

7. Belenky, G. L.; Tyner, C. F.; and Sodetz, F. J.(1983). Israeli battle shock casualties: 1973 and 1982,Walter Reed Army Institute of Research (WRAIR) Report NumberNP-83-003.

8. Benner, A. W. "Psychological screening of policeapplicants". Psychological Services for Law EnforcementNational Symposium on Police Psychological Services, FBIAcademy, Quantico, VA, Washington, D.C., 1986.

9. Benson, H.; Kotch, J. B.; Crassweller, K. D.; andGreenwood, M. M. (1977). Historical and clinicalconsiderations of the relaxation response.

10. Berkeley, D. and Humphreys, P. (1982). "Structuringdecision problems and the "bias heuristic'". ActaPsychologica, 50: 201-252.

11. Bosshardt, M. J.; Dubois, D. A.; Paullin, C.; andCarter, G. (1988). The investigative interview: a reviewof practice and related research. Personnel DecisionsResearch Institute Report No. 160, Minneapolis.

R-2

12. Bosshardt, M. J.; Dubois, D. A.; and Crawford, K. S.Continuing assessment of cleared personnel in the militaryservices: report 4--system issues and programeffectiveness. Defense Personnel Security Research andEducation Center (PERSEREC) PERS-TR-91-004, (1991).

13. Cannon, W. B. (1914). "The emergency function of theadrenal medulla in pain and the major emotions". AmericanJournal of Physiology, 33: 356-372.

14. Carter, B. and Cammermeyer, M. (1989). "Humanresponses to simulated chemical warfare training in U.S.army reserve personnel". Military Medicine 154, 6:281-288.

15. Chappelow, J. W., (1989). "Causes of aircrew error inthe royal air force". In AGARD Conference Proceedings No.458, Human behavior in high stress situations involvingaerospace operations.

16. Chidester, T. R. and Foushee, H. C. (1989). "Leaderpersonality and crew effectiveness: factors influencingperformance in full mission air transport simulation". InAGARD Conference Proceedings No. 458, Human behavior in highstress situations involving aerospace operations.

17. Damos, D. (1989). "Training and selecting individualsfor high levels of information processing load". In AGARDConference No. 458, Human behavior in high stress situationsinvolving aerospace operations.

18. Department of Defense (DOD) 5210.42 (1985). Nuclearweapon personnel reliability program.

19. Director of Central Intelligence Directive 1/14 of 22January 1992, Personnel security standards and proceduresgoverning eligibility for access to sensitive compartmentedinformation (SCI).

20. Driskell, J. E.; Mullen, B; and Batchelor, C. L.(1990). Development of a meta-analytical technique to asessstress effects. Air Force Human Resources Laboratory.

21. Emurian, H. H. and Brady, J. V. (1981). Behavioraland biological interactions with confined micro societies ina programmed environment. Johns Hopkins University Schoolof Medicine Technical Report TR number 3 for the Office ofNaval Research (Code 452), Arlington, VA.

22. Evans, G. W.; Stokols, D.; and Carrere, S. (1987).Human adaptation to isolated and confined environments.NASA Grant number NAG 2-387, Final Technical Report,California University.

R-3

23. Federal Emergency Management Agency CPG 2-21 of May1988, "Habitability and human problems in shelters".

24. Gerwell, E. L. and Fiedler, E. R. Using the osi tomeasure stress in people in high stress jobs, June 1990paper presentation at annual meeting of the AmericanPsychological Society, Dallas, TX.

25. Gunderson, E. K. E. (1966). "Small group structureand performance in extreme environments". Navy MedicalNeuropsychiatric Research Unit, Report Number 66-3, SanDiego.

26. Helmreich, R. L. (1987). "The critical role ofpersonality and organizational factors as determinants ofreactions to restricted and stressful environments".Feature presentation at Third International Conference onRest, New York, NASA-CR-180621, Texas University.

27. Henderson, P. R.; Bakal, D. A.; and Dunn, B. E.(1990). "Cardiovascular response patterns and speech: astudy of air traffic controllers". Psychosomatic Medicine52: 27-41.

28. Herd, J. A. (1991). "Cardiovascular response tostress". Physiological Reviews, Vol. 71, Number 1, pp.305-330.

29. Holmes, T. H. and Rahe, R. H. (1967). "The socialreadjustment rating scale". Journal of PsychosomaticResearch II, pp. 213-218.

30. Inwald, R. E. "Issues and guidelines for mental healthprofessionals conducting pre-employment psychologicalscreening programs in law enforcement agencies". NationalSymposium on Police Psychological Services, FBI Academy,Quantico, VA, Washington, D.C., 1986.

31. Jayne, B. (1988). "The validity of different employeescreening techniques". Unpublished report. Chicago; JohnReid and Associates.

32. Jones, D. R. (1982). "Emotional reactions to militaryaircraft accidents". Aviation, Space, and EnvironmentalMedicine, June.

33. Jones, W. A. Jr., (1990). "An evaluation of voicestress analysis techniques in a simulated awacsenvironment."

34. Keinan, G. "Confidence expectancy as a predictor ofmilitary performance under stress". Stress and Coping inTime of War. Ed. N. Milgram. Brunner and Mazel, 1986, pp.183-196.

R-4

35. Klein, G. A. and Thordsen, M. (1989). "Decisionmaking under stress". Seventh Users' Workshop on Stress:Communit:'/National/International Cooperative Efforts, SanAntonio.

36. Labuc, S. (1984). "Training to cope withpsychological stress". The Technical Corporation Program,Technical Panel on Training Technology.

37. Lovallo, W. R.; Pincomb, G. A.; Brackett, D. J.; andWilson, M. F. (1990). "Heart rate as a predictor ofneuroendocrine responses to aversive and appetittivechallenges". Psychosomatic Medicine 52: 17-26.

38. Mandelbaum, D. G. (1954). "Psychiatry in militarysociety". Human Organization 13, pp. 5-15.

39. Marshall, S. L. A. (1947). "Island Victory". TheInfantry Journal, Washington, D.C.

40. Mickley, G. A. "Psychological factors in nuclearwarfare:. In R. Zajtchuk, D. P. Jenkins, R. F. Bellamy, andV. M. Ingram "eds." Textbook of Military Medicine, Part I:Warfare, Weaponry, and the Casualty, Vol. 2 MedicalConsequences of Nuclear Warfare, TMM Publications, Office ofthe Surgeon General. Falls Church, VA, 1989, pp. 153-169.

41. Nelson, P. D. "Compatibility among work associates inisolated groups". U.S. Navy Medical NeuropsychiatricResearch Unit, San Diego, CA, Report number 64-13, 1964.

42. Nelson, P. D. "Structural change in small isolatedgroups". U.S. Navy Medical Neuropsychiatric Research Unit,San Diego, CA, Report number 64-13, 1964.

43. Novaco, T. W.; Cook, T. M.; and Sarasan, I. G."Military recruit training". In D. Meichenbaum and M.Jaremko "eds." Stress Reduction and Prevention, PlenumPress, New York, 1983, pp. 377-418.

44. Pitz, G. F. and Sachs, N. J. (1984). "Judgement anddecision theory and application". Annual Review ofPsychology 35, pp. 139-163.

45. Pleban, R. J.; Allentoff, H. L.; and Thompson, T. J.(1989). Preliminary assessment of selected predictors ofspecial forces qualification course success. U.S. ArmyResearch Institute for the Behavioral and Social SciencesResearch Report.

46. Radiological Factors Affecting Decision Making in aNuclear Attack, NCRP Report number 42. National Council on

Radiation Protection and Measurements, Washington, D.C.,1974.

R-5

47. Rahe, R. H. CPT, (1988). "Acute versus chronicpsychological reactions to combat". Military Medicine, 153,7:365, pp. 365-372.

48. Selye, H. (1936). "A syndrome produced by diversenoxious agents". Nature London 138:32.

49. Siem, F. M. (1989). "Personality characteristics ofusaf pilot candidates". In AGARD Conference Proceedings No.458, Human behavior in high stress situations involvingaerospace operations.

50. Simonov, P. V. and Frolov, M. V. (1984)."Psychophysiological control of the functional state of thehuman operator". Zh. Vyssh. Nerv. Deyat. Pavlov, 34(2),195-206.

51. Spiller, N. B. (1989). "Human behavior in high stresssituations in aerospace operations keynote--the operator'sperspective". In AGARD Conference Proceedings No. 458,Human Behavior in high stress situations in aerospaceoperations.

52. Steinglass, P. and Gerrity, E. (1990). "Naturaldisasters and post-traumatic stress disorder: short-termversus long-term recovery in two disaster-affectedcommunities". Journal of Applied Social Psychology 20, 21,pp. 1746-1756.

53. Stephens, C. R. LTC. (1987). "Stress: an unnecessaryfact of life". Military Medicine 152, 1:19, pp. 19-24.

54. Trejo, L. J.; Lewis, G. W.; and Blankenship, M. H.(1987). "Brain activity during tactical decision-making:probe-evoked potentials and workload". Navy PersonnelResearch and Development Center Technical Note 88-12.

55. United States Army GTA 21-3-4 of June 1986, Battlefatigue: normal, common signs; what to do.

56. United States Army GTA 21-3-5 of June 1986, Battlefatigue: more serious signs; leader actions.

57. United States Army GTA 21-3-6 of June 1986, Battlefatigue: company leader actions and prevention.

58. United States Army Regulation 50-5 of August 7, 1989,Nuclear Surety.

59. Ursin, H. (1989). "Activation--positive and negativeeffects of the alarm system in the brain". In AGARDConference No. 458, Human behavior in high stress situationsinvolving aerospace operations.

R-6

60. Waller, N. K. and Burkhardt, J. F. (1965). "The combateffectiveness of various human operator controlled systems".Proceedings of the 17th U.S. Military Operations ResearchSymposium.

61. Warren, P. H. (1988). "Psychosocial accommodation togroup confinement in the advanced base habitat". Scienceand Advanced Technology Directorate, U.S. Army Natick R, D,and E Center, Report number Natick/TR-88/063.

62. Wickens, C. D.; Barnett, B.; Stokes, A.; Davis, T.Jr.; and Hyman, F. (1989). "Expertise, stress, and pilotjudgement". AGARD Conference Proceedings No. 458, Humanbehavior in high stress situations in aerospace operations.

63. Wolff, H. G.; Wolff, S. G. Jr.; Hore, C. C. "eds."Life Stress and Bodily Disease. Research Publications,Association for Research in Nervous and Mental Diseases,Vol. 29, Baltimore, Williams, and Williams, 1950, pp.118-123.

64. Yerkes, R. M. and Dodson, J. D. (1908). "The relationof strength of stimulus to rapidity of habit formation".Journal of Comparative Neurology and Psychology, 18, pp.459-482.

65. Zakay, D.; Epstein, Y.; and Shapiro, Y. "The effectsof a coacting setting and competitive motivationalperformance under heat stress". Stress and Coping in Timeof War. Ed. N. Milgram. Brunner and Mazel, 1986, pp.205-215.