ESD-TR-72-140

[ - MINUTEMAN PERSONNEL SELECTION STUDY

William H. Hendrix, Ceiptohm, U.S.A.F.

March 1972

TECHNICAL REOUIREMENTS AND STANDARDS OFFICEHQ ELECTRONIC SYSTEMS DIVISION (AFSC)L. G. Hanscorn Field, Bedford, Massachusetts 01730

ftcpraduýod byNATIONAL TECHNICALINFORMATION SERVICE

Springfield. Va. 221I1

Approved for public release; 1distribution unlimited.

nD D

APR~

iiff1~ifL~wNow

[ 4 ,

LEGAL NOTICE

When U. S. Government drawings, spe.ifications or other data are used for anypurpose other than a definitely related government procurement operation, thegovernment thereby incurs no responsibility nor any obligation whatsoever; andthe fact that the government may have formulated, fu-nished, or in any way sup-plied the said drawings, specifications, or other data is not to be regarded byimplication or otherwise as in any manner licensing the holder or any other personor conveying any rights or permission to manufacture, use, or sell any patentedinvention that may in any way be related thereto,

OTHER NOTICES

Do not return this copy. Retain or destroy.

.ICEUPJ for

0AWM0

II

4; , / --

ESD-;TR-72-140

MINUTEMAN PERSONNEL SELECTION STUDY

William H. Hendrix, Captain, U.S.A.F.

March 1972

TECHNICAL REQUIREMENTS AND STANDARDS OFFICEHQ ELECTRONIC SYSTEMS DIVISION (AFSC)L. G. Hanscom Field, Bedford, Massschusetts 51730

I Approved for public release;distribution unlimited. "

. I

Security Classifiscation

DOCUMENT CONTROL DATA. D-& D(Socurity lasslflca lion of title. bedr. of oblatart sad indexing annotc.l1on rnuat beoenfteed %4hen the overall report Is classifled)

OI 01GIN A TO N4 ACT IVIT Y (Corporal* author) i~.REPOA f SECURITY CLASSIFiICATIONTechnical Integration Division uCLAMFIED 7L. G. Hanscom Field 2b. GROUP

Bedford, Moss. 01730 WM/A3, REPORT TITLE

MINUTEMAN PERSONNEL SELECTION STUDY

4. DESCRIPTIVE "IOTES (Typo of report Cnd Inclusive dates)None -.

S AUTH4OR(St (Pita S narme. middle Iintital, Iasi namr)

Wfiliam H. Hendrix, Captain, U.S.A.F.

S. REPORT OATI 7M. TOTAL No. or PAGES 7h. NO OF REFS

March_1972 22____15_MS. CONTRACT On1 GRANT NO- Ms. ORIGINATOR-S REPORT NUM11•RISI

IN-HOUSEb. PROJECT NO S-R7-4

C. 9)b. OT?4 R REPORT NO-11 (Any otnet num bete that may be asign,.ted

10, DISTRIBUTION STATCMENT

Approved for pubiic release; distribution unlimited.

4I. SUPPL•E"MENTARY NOTES•t2. SMON$ORINGIA|LITARY ACTIVITY

Technical Requirements and Standards OfficeHq Electronic Systems Division (AFSC)

, T. L. G. Hanscom Field, Bedford, Mass. 01730

• %Thit. report is-a-pilot-study-which dea!s with the selection of personnel for assignment asMinuteman Missile Combat Crew Members. The data evaluated consists of biographical data,

Officer Efficiency Reports, and Air Training Command Standard Scores obtained on individuals

L during individual missile training. Two predictors of performance were isolated using a multipleregression technique, which incorporated the use of dummy variables. The importance of thisstudy, however, is not limited ooly to those variables found to be significart. In fact, anexample cited in the study indicates that certain variables which are not predictive of successcould have an equally important managerial implication. It is recommended that future researchbe undertaken to evaluate ORT scores, AFOQT scores, and other pre-service variables such asGrade Point Average as possible predictors. Isolation of additional predictor variables whencombined with those isolated by this study could result in a flexible selection program forMinuteman Missile Combat Crew Members.

D4 t 4 73 15 __ __ ___ _ __

~ ~z.. - -~---JA

_�_ _ _ _ _____I___

security clasuification1 4. &LINK A LINK- LINK C

KEY WGOGO -

MOLE WT R'*LZ WT RtOLE WIr

PERSONNEL SELECTION 2

gg

I

161

16 SecuntiY CI&'Ittifcatinn

- -- .- .

FOREWO,

This Technical Report is based on a pilot study dealing withthe feasibility of develcping an improved personnel selection pro-cedure for Minuteman Missile Combat Crew Members.

The author wishes to express his appreciation for the datasupplied for this study by Headquarters Strategic Air Command(DPPBM and DOTKT) and the Air Training Command (TSDM1, Chanute AFB,Illinois). The author is also deeply indebted to Lt Col Thomas Graham,AFHRL (AFSC), USAF Academy, for his assistance in computer programmingand his constructive criticism of the study; and Mrhs. Mary Batson,Frank J. Seiler Research Laboratory (AFSC), USAF Academy, for herassistance in preparing the data for computer processing. This projectwas funded by the Air Force Human Resources Laboratory, and the AirForce Systems Command through the USAF Academy, Director- of FacultyResearch.

This Technical Report has been reviewed and is approved.

CAR14INE PINTO, ChiefTech Rqmts & Stds Office

CJ

[ ii

V/

g-

ABSTrACT

This report is a pilot study which deals with the selection of -Zpersonnel for assignment as Minuteman Missile Combat Crew Members.The data evaluated consists of biographical data, Officer EfficiencyReports, and Air Training Command Standard Scores obtained on indi-viduals during individual missile-training. Two predictors of per-formance were isolated using a multiple regression technique, whichincorporated the use of dummy variables. The importance of thisstudy, however, is not limited only to those variables found to besirnificant. In fact, an example cited in the study indicates thatcertain variables which are not predictive of success could have anequally important mainagerial implication. It is recommended that

future research be undertaken tc evaluate ORT scores, AFOQT scores,and other pre-service ,'ariables such as Grade Point Average as pos-sible predictors. Isc.iation of additional predictor variables whencom;bined wizh those i zlated by this study could result in a flexibleselection program for Minuteman Missile Combat Crew Members.

'4

'1

!-z

;!N

iii A

A

TABLE OF CONTETS

SECTION PAGE

I. Introduction., . . . . . .

1. General .. . .,. . . 1

K I 2. Objectives ... . .... .... I

II. Scope and Procedure........... 2

1. General .............. 22. Criterion 2o........... 23. Data Collection ........ e . 24. Analysis-, . . . .. .

III. Recommendations .

References • . . • • • . . . .. . 6

Appendix A: Criterion and Predictor Variablesin Final Regression . . . .. . 8

Appendix B: Initial Regression Analysis . . . 1

Appendix C: Regression Analysis with Trans-generated Variables . . . . . .. 13

J

I "III'?

SECTION I

INTRODUCTION

1. GENERAL. In January, 1969, UFF Headquarters Personnel (AFMPC)expressed interest to Headquarters SAC in improving the quality ofmissile force personnel. In an attempt to improve the quality of+%.e missile force certain questions needed answers. The followingare some of the critical questions addressed in the correspondence:

Is grade an important factor?

Is bar-:ground (education and experience) important?

;b 'ength if service related to performance?

Is source of commission a factor of importance?

Are there c~her important factors which have not been addressed?

At the time of this correspondence there was no method of selectionestablished which would select those officers which had the highestprobability of being successful Missile Combat Crew Members (WiCM's).This study was an outgrcfth .,f Headquarters SAC's request for assistancein this area from the USAF Academy.

2. OB.ECTIV&s. This study was intended as a pilot study to establishthe f.-asibility of a Minuteman Missile Combat Crew selection program.The objectives of the study were: A

a. To estabLsh a criterion of job success.

b. To identify variablez which are characteristic of the success-ful Minuteman Missile Combat. Crew Member, and thair degree of importance.

c. To recommend a course of action for establishing a procedure""A-• identifying the pot'ýntially' successful Minuteman Missile CombatCrew Member from zhe total officer resource pool.

4-x

i,S .. . .. . ... .

I Nt+• + • ,•+• +.,. .% ,. _. :•• i ,+.+:-++,+,,.,, •+ ++'..,...+_. _,+-+ .. +++• .+ .-+:--+- .. •-+-.,+ .. ,+ ,+-;++++ .'"-•: • . ... +"'

SECTION II

12COPE AND PROCEDURE

1. GENERAL. The scope included:

a. Investigating officer personnel assigned as Minuteman MissileCombat Crew (04CC) Members as rf 1 July 1969.

b. A comprehensive review of the literature associated with theproblem area in order to provide a sound base for further research.

c. A review of various indices of job success for possible useas a criterion oi performance.

The procedure involved:

a. Establishing a criterion of job success (as a 1,1CC memrnerj.

Anlyib. Collecting data.

.•- c. Analyzing data.

2. CRITERION. Establish;ng a criterion of job success was accomplishedby drawing upon the expciience of former Minuteman Missile CombatCrew Members, a review of the literature on criterion development, andby analysis of those indices available. The indices reviewed consistedof OER means, tenure, Missile Combat Crew Member Evaluation Indices(standboard scores), and rank-order ratings. All of the criteria con-sidered were elimin-ted because of non-availability on a large sampleor because of deficiency (grossly skewed, restricted range of values,etc.), except for the Missile Combat Crew Mamber (MACCM) Evaluationindices. The MCCM evaluation indices were considered acceptable as acriterion of success since they were quantitative (range 0.0-5,0),were conducted on a periodic bauis throughout the year, and were obtain-dby a standardized procedure derived by personnel trained in performingMCCM evaluations.

3. DATA COLLECTION. Data were collected from rpcords at HeadquartersUSAF, Headquarters SAC, and Chanute AFB, Illinois. The data were eom-bined and stored on magnetic tape for analysis (see Appendix A Tor data

items). The data collected were on an initial random sample of 399"individuals (n=399) out of a population of approximately 20]1. The datagathered consisted of biographical and performance data.

7VI

Ii. ANALYSIS. Analysis involved the use of a multiple regression tech-nique (BMJDO2R Stepwise Regression). This technique incorporated the useof dw.•nny variables as a means of establishing the relationship of quali-tative variables to the criterion (MICCM evaluation indices), in additionto the typical regression procedure used for quantitative predictorvariables. In the initial regression analysis 38 variables were correlatedwith the criterion (see Appendix B). The sample size, as previouslymentioned, was 399. The only variable which yielded a correlation ofany practical significance was "OER Mean" (r=.3375).

A later regression analysis added a new variable for consideration(Air Training Comrimand Standard Score). The data for this new variablewas obtained from Chanute AFB, Illinois, where individual MinutemanMissile operator tra.ning was being conducted at the time of this study.R.:.cords on individuals who received training were readily accessible onlyon those individuals who had completd the school in the past two yearperiod. This factor resultel in a .,eduction of our initial sample of 399to 218. Subsequent regression anelysis revealed two significan.t variables,OER mean aaxd Air Training Command Standard Score (ATCSS), which correlated

with our criterion at r=.36 and r=.'l respectively. In addition, theintercorrelation between the two predictor variables was reasonably low(.12). Combining the two predictor variables resulted in a multiplecorrelattion of .49.

The re----ssion analysis program provided a transgeneration feature.Transgeneration refers to any transformations of an input variable whichresults in the generation of a new variable. This feature was exercisedby taking certain "romising variables to the following powers: X2 I X3

and X (see A pen ix C). Results indicated that "OER Mean" taken to thefourth power (9ERý) increasQd its correlation to r-.47, while the ATCSt-ndard Score linear relationshin was the best correlation (r-.31).When comLined the two yielded a multiple correlation of .52, with an inter-c=rrelation of .19.

In reviewing appendices A, B, and C, other interestinz- correlationscin be noted, such as the negative correlation between in,2rease in ageand performance. Depending on the level of significance establishedby an individual, a few of these may be found to be statistically signif-ican-m however, they have little practical significaix.c. Also cautionshoulu be taken when generalizing about the variables w;±i -h have beenplaced in discrete categories (dummy variables). Thi. ca egorizationresulted in very low n's in scme of the discrete catetLori s (nslO).

In addit-or to isolati:.g.. two predictor variYt I-i:. thq followringJquestions T )sed by Headqu.r-.ars USAF were answer-d: P1 certain source

Inmiivi-udlm. Minuteman, Mlsý-ie operator ,i..."z .w- _1 la;ently trans-ferred to ,• "cenberg AFB, 6 f.' .

of conmnission groups (ROTC, OTS, USAFA, etc.) perform better as MissileCrew Members than do other groups? Is there a difference in performanceas a Minuteman Missile Combat Crew Member between rated and non-ratedinputs? Do individuals with certain academic backgrounds perform betteras Minuteman Missile Combat Crew Members than do other individuals?Arc there any other variables which are significant?

In order to answer these questions, the factors of academic back-ground, marital status, rated non-rated status, source of commission,and geo-rraphic area wei e broken out into subgroups and -analyzed (ser-i.ppendices A, B, and C).

After examining the size of the n ,iithin each subgroup and thecorrelation, it was found that none of the subgroups differed signifi-cantly from the others.

'This does not mean; however, that these nonsignificant variables arenou important. In fact, certain of these variables may have an equallyimportant managerial implication. For example, take the problem ofassigning a rated vs. a non-rated officer to missile crew duty duringa period of pilot/navigator shortage. Since there was no significantdifference in perfoimance found between these officers, management wouldbe justified in establishing a predominantly non-rated missile crewforce when the rated force is critically needed in their rated specialty.This does not imply, however, that the rated officer force should nothave missile experience at all. In fact, our upper management of thefuture should have both the knowledge of missile operations as well asaircraft operations. The thesis here is only that during perioas wherethere is a drastic shortage of rated officers in rated assigrnments,;management could draw o: those rated officers earmarked for missile-crewduty without degrading the missile force's performance.

4s

I-

SECT.ION III

RECOMMIDATIONS

A request of Headquarters SAC (DPPRM) for additional data foranalysis was placed during a December, 1969, briefing on the progressof the research pr:gram. Due to a manpower shoru.age., withini Head-quarters SAC (DPPR14), the data has not been collected as of' this dateand the project is at an impasse. Reconmend tha~t a future study deal-ing with this problem obtain the data previoUSl;r requested, which wasAir~ Force Officer Qualification Test (AFOQT) scovýs and OperationalReadiness Test (ORT) scores. It is also recommTended that other pre-31service factors be investigated. such as colleee Grade Point Average r(GPA)2 -Which could prove extremely valuable wheii selecting newlycommissioned officers for missile duty.

t.present,, with only 'two significant predictors(R-a)aselection procedure is-not feasible. The reason that a selectionprocedure is not feisible is twofold. First, one of the predictors,is not available on newly commissioned officers (0ZR Mean). Secon~d,the remaining predictor (ATC-SE) is available onfly afteanidvulhas completed individual training. In future research, should addi-t-ional significant pr'ý-dictors be found annrng the AFOQT scores, ORTscores, ardler pre-ser-vice factors, it is highly likely that t~he result-ine :m-ultiple correlation would !)e increased to a level whpre a flexibleselection plan would 10e possible. This platn would include provisionsfor initial selection based on pre-service data and possibly OER data,as well as establishing a crite-rion for student elliniiat~on during ATCmissile training and SAC Operatione-1 Readiness Training (ORT). Inorder' to establish the valid-,y of the, data items selected, a cross-

"vaidation wou.Ld b~e indicated prior to implententing the selection planx.This selection plan, which provides initial selection plus a courseelimination feature, would ins-ure that only those officers who would behigh-quality Minuteman Missile Combat Crý!w Members would be _ssigrned to-missile duty. The recults: a high-quality Missile Combat Crew forceselected by a cost effective selecticn ýrograin.

~ .. A,

REEMiIES

Departwi.nt of the Air Force, Headquarters, Strategic Air Ccumnd, Offutt

AFB., NB. ICBM combat crew standardization and evaluat-xo, SAC

Manual, 55-66, Vol II, April 1, 1969.

Blum., M. L., & Naylor, J. C. Industrial psycholowj: Its theoretical

and social foundations (ReX ed.) New York: Harper & Row, 1968.

Rottenberg, R. A., & Ward, J. H., Jr, Applied multiple linear re-

gression. WISAF Technical Report PRL-Tra-63-6, March, 1963.

Cohen, -J. MUtiple regression as a general data - analytic system.

E!Xnqlogical Bulletin, 1968,-70(6). 426-443.

Fi.r, E.. E ., & Potter, N. R. C•haracteristics of basic airmen villing

to volunteer for a six-year touz in ýaissile squadrons. WA

Technical Report WADC-TH-59-35, Februaryv 1959.

Ghiselli, E. E. Dimensienal problems of criteria. Journal of Applied

Psycholog, 1956, 1-4.

G•ilon, R. M. Regression analysis: Prediction from classified variables.

Psyehological Bulletin, 19.5, 51(5), 505-510.

Gujon, R. M. Personnel selection. Annual Review of Psychology, 1967,

18, 191-216.

Himelstein, P., & Blaskovics, T. L. Prediction of an Intermediate

criterion or eombot effectiveness with a bWographical inventory.

Journial or Applied Psychology, 1960, 41 (3), 266-168.

Lingoes, J. C. The multivariate analysis of qualitative data. Multi-

variate Behavioral fesea.ch, 1968, )LQ, 61-94t

6

~ :- -

SiegrA1., L. Irzdu..trial Psych'ýIxogy (Rev ed.) Ho~mewood, Illinois:

D.~r 1xvir~n, lr'-. N969.

iTanofshy,. R., Z;heppR.?., OrNeili, P. J. Patttern analysis. of3

bicrgiaphic-al pi-edictor-i of~ succtlsr ai. an~ in.-urdnuce Salestran.

jJt:,urzral of' Applioýd P-1;yenology, 1969. L53(2), 1.3--139.

Thtr.-vi~ie., R. L. Th e utiriur test c-ompsite3 to predict a set of

I ~ ~~criteria. -a-I&!i? Teczinica2. Report A hCT5-l3,July, 1957.

4Tupecs, E. C.: & Claristal, R. E. P!F-tholor__vdi tez-,s s.ni th.~e selection

ar' -classIf ica.tion of Air Forze Officers. 1YGTF Technical Report

AFYIýw1;'ý-2 ____ ____ __1957_ ___

Valentine, L. D., Sc £Ceager, J. A. Officer selecticn and classification

tc T n:, T~r Aev.#'- 1Ptrent and use. 1USA~z' Techiiicl Repart. ASD-Tlv*

4 4

APPEEDIX A

CRiTERION AND PRDICTOR VArUALESI FINAL FMm~lssoN

VARIABIES VARIMlIE NUM

Criterion (Mean value of all MCC9 evaluation ratings) 1

Minuteman Education Program* 2

Regular Commission* 3

Age 4

•Age < 3o* "5

Years service 6-

Military obligation* 7 31

Declined Regular Commission* 8

Passed over for promotion* 9

U.S. Citizen-spouse 10

Education level* 11

, Engineering 120o Military Science* 13

- Social Science* 14

SPh.)sical Sience* 15

< Other B.S./B.A. l.jor* 16

Days in SEA 17

COKMBA Tours (NH) 18

Professilonal Military EFucat!' * (PME) 19 -f

PHE -k=8 20

?1

APPENIXfl A (cant)'

VARIABLES VARIABLE NIR4BER

RACE* 21

-r rmarried* 22 "F> Singl:e* - 23

P2ilot* -

I 43 Non-rated* 26 .

4 ated-other than pilot* 27

Officer Efficiency Report-Mean 28

So Military Academy Gmduate* 29

3 30311

'V ltfther-commission soarce* 32

U otheast* 33

Northeast* 34•..

SMidwest* 35

Southwest*

Far West* 37Number of Dependents 38

Grade 39

- ATC Standard Seore W

Age Squared 000) 41

OER MEAN-4th power (OXR 4 142

9

APPENDIX A (cont)

VARIABLES VARIABLE NUMBER

Grade Squared (Grade) 243

Grade-3rd power (Grade• 44•

• 4,

Variables placed _ :-screte categories.

3.01

APPENDIX B

INITIAL REGRESSION ANAMLSISa

Multiple0 Multiple0

b 2STEPý VARIABLELMIMBER r r r

1 OER Mean 28 .38 .38 .14

f 2 Minuteman Ed. Program 2 .09 .39 .15

3 Grade 39 -. o4 .40 .16

4 Rated-other than pilot 27 .05 .41 .17

5 Engineering i2 .08 .42 .18

6 Passed over 9 -. 08 .42 .18

7 N.E. 34 .02 .43 .18

8 Social Science 14 -. o7 .43 .19

9 Ed. Level 11 -. 10 .144 19

10 Mf!itary Obligation 7 .05 .44 .20

11 Years Service 6 -. 01 .45 .20 I1.2 Age 4 -.07 .45 .20

13 PME 19 -. O1 .45 .21

14 Married 22 -. 05 .46 .21

15 Physical Science 15 .0007 .i46 .21

16 Race 21 .08 .46 .21

17 Other-comIission source 32 .08 .46 .21

18 Days in ZEA 17 -. 07 .46 .22

19 U.S. Citizen-spouse 10 .13 .47 .22

20 Single 23 -. 05 4147 .22

21 Combat Tours(NR) 18 .06 T 7 .22

N11

- -,-

APPENDIX B (cont)

MultipleC Multiple

STEPb VARIABLW/1 ER r r r

22 POW-Metrhod 20 .01 .47 .22

23 Regular Commission 3 .03 -47 .22

24 Midwest 35 -. 06 .47 .22

25 WR Dependents 38 -. 01 .47 .22

26 Academy Graduate 29 -. ')03 .47 .22

27 ROTC 30 .01 .47 .22

28 Pilot 25 -.01 .47 .22

29 Declined RLg. Comm. 8 .02 .47 .22

30 Military Science 13 .07 .47 .22

P Only those variables which had a sufficient F-level to be included in the

regression are listed. F-level for inclusion .010, for deletion .005.

bVariables added in a stepwise manner, at each step the variable added is

"the one which makes the greatest red.uction in the error sum of squares.

0 carried to four piaees and rounded to two.

Tis

A'PEDIX C

REGRESSION ANALYSIS WUri TH ANSGENFRNTED

Mu~ltipls' MtlltipleCb 2

STEP VARIABLE/NUMBER r r r

IOER 4 42- .47 .47 .22

2 ATCSS 40 .31 -5', .27

3 Grade 39 -.14 .54 .29

4 Social Science 14 -. 16 .55 .30

5 Engineering 12 .09 .55 .30

6 Age <30 5 .13 .5.31

7 Age 2 41 .13 .55 .31

' Age 4 -.13 .56 .31

9 QER Mean. 28 .36 -56 .31

Onythose vari'nbles which had a suff"Icient, F-level to be included in

the regression are listed. F-level for inclusion .010, for deletioui .005.

'Variables added in a stepwise manner, at eac-h step the variable added Is

the one which makes the greatest reduction in the error sum of squar'es.

eCarried to f'*nar places and rounded to two.

'13