AD-AO14 790

SCREENING FOR ADAPTABILITY TO MILITARY SERVICE

Nancy Guinn, et al

Air Force Human Resources LaboratoryBrooks Air Force Base, Texas

May 1975

DISTRIBUTED BY:

National Technical Inoratn ServiceU. S. DEPARTMENT OF COMMERCE

REPRODUCED FROMBEST AVAILABLE COPY

268201

AFHRL-TR7.F;.*30

AIR FORCESCREENING FOR ADAPTABILITY TO MILITARY

- HII By

0I Nancy GulnnAllan L. Johnson, Sgt, USAFZ7 M ay E. Kantor

U PERSONNEL RESEARCH DIVISIONA Lacland Air Forc Bus, Texa 78236

"N

0 May 1975R Final Rsert for Pfriad .uly 1974 - Apil 1975

ES0 Approved (or public release- diotrlbutlon unlnmitcd.

UNATIONAL TECHNICALR INFORMATION SERVICE

C ,-p0n . hd , VA 22151

ES LABORATORY

AIR FORCE SYSTEMS COMMANDBROOKS AIR FORCF c1SZ,TEV,*,% 79235

NOTICE

When US Govermewnt drawinlnp, specifications, or othtr data atc vuedfor any purpose other than a definitely related Goverunmentprocurement )~peration, the ,Government thereby incurs nofespoeswlblllty nor iny obligation i/Imtmoever, and the fact that theGovernment may have formsulated, furnished, or in smy way suppliedthe add drawlnp, specifications, or other data .s not to~ be regarded byI ~impliation or otherwise as in any ffmnner licensing the holder or anyother person or corporation, or conveying any rights or permiission tomanut'acture, wse, or sell .Any patented invcrntun that may in any waybe related thereto.

716 finkal report was submitted by Personael Research Division, AirForce Human Reources Laboratory, LackLind Air Force Rase, TeO=78236, under project 7719, with Hq Air Force Humnas ResourcesLaboratory (AI7SC), Brooks Air Force Base, Te,ýax 78235. Dr. NancyGui~nn, Demographic and Attitudinal Researchi Srndtc, was the taskscientist.

Th1s report has be ~n reviewed and cleared for open publication and/orpublic release by the appropriate Office of InformatIon (01) inan'cordance with AI'R 190.17 and DODD 5230.9, 1There Is no objectionto unlimited distribution of this report to the public at large, or byDDC to the Nstiuoaal Technical Information Service (N7iS).

'This tchnical report han been reviewed and is approved.

LELAND D. BROKAW. Technical Di iector

Personnel Research Division

Approved for publication.

HAROLD E, FISCHER, Colonel, USAF

Cormmander.

VKA ~ udi *UIIE .

SECUMi- :LASS-FIcA'.oN OF 1,5S WAGE 'W%- r. o....

PAGE RLAD INSTR4UCT1014SREPORT COCM ENTATION PAERFR COMIPLETTNC. D-ORM

I P10ONT .OE*1 ', GOVT ACCIESSIOhs 0.0 1 REC'sEfoS CATA13,( k0WMISE

AFHRL-TR-75-30

4 TITLE Wdu611 0 Type OF REPORT II '(ROD0 1CO.9ER

SCREENING FOR ADAPTABILIT Y TO %U4UTARY SIIRVICE FinalJuly 1974 - April 1975

6 owErommlqG Ono REPORT I*,,M6CR

7 4u~~e CONTRACT 00 GRANT NUMBER.

Nancy GumniAllan L JohnsonJeffrey E. Kantor

9 Plar#oR,sor o~rA%,ZAT.or. NAME AND aDDRIESS to P0005*U: E.E!FN' PROJECT TASK

Personndl Researcht Dis *rA 6 SORE "NIT

Ntjmalin5

AKi Force Human Resources La4boratory 62070Lacklan Air Force Base. Texas 78236 771902-41

It CO"TROLLIMG OFFICE "AMC AND0 ADDRESS 12 REPORT' DATE

Hq, Air Force Human Resources Laboratory (AFSC) May 1975&ooks Air Force Base. Texas 7XZ35 12 oWumeoR of PAIAS

IA MCN;Tofti"ý AGEoC NAME A AOORESI(,1 d,11.. I.,-, to ,.Ijoff~-., IS SECuRITY .ASS *.1 Ih.. loo

Unclassified

SCMEOUI.E

IS DISRIUU7tION STAT'EMENT olt IN,. Rep-.'

Approved for public relese.distribution wshirrmed.

t? O4STRI~uTIONv STATEMENT 'at the* 5abaar 1a,,d o1I 81a'1, ;9, It difloo-Y, f,-e, P.p5,I)

14 SUPPLEMENTARY "OTES

Is KEr WORODS 'C-1tlnuon * ...*sdo I4 oto.- -sv d b.p~,tc hfck -b-) er

adaptabilitypersonality inventoriesscreening methodologymaladaptive accessionsthgh-risk personnel

20 ABSTRACT (C-tifnct .1'nr sodot of nooeserv &,,d Idoo-sIv b, 1t,1-1 nunbsor.

A sample of 1 5,252 basic airmen were admninistereo the history opinion inventory (HOI) dumi., basic militarytraining. The service careers of these subjects were monitored for two years in order to assess the ability of the HOIto predict the criterion of in/out of service. An a ptiori adaptation index developed from H01 items~ correctlyidentified as high risk 23 percent of those subjects discharged from service durnn, the two year period. whileincorrectly labeling as high risk only 6 percent of those subjects still in service afte. two years. The possibility ofincreasing the accuracy of prediction by utilizing bi ographic/ de mographic data and the operational usefulness of theH01awe discussed.

DD F0ANIP 1473 EDITION OF I NOV 615 IsOBSOLETE* UnclassifiedI.SECURITY CLASSIFICATION Or TmISI PAGE (*%*n .Dat Ertf.Itd)

if'

PREFACE

TU reearch wu oonducted under Projct 7719, Air Force Personnel SystemDevelopnent on Selection, Aaii•nment, Evaluation, Quality Control, Retention,Promotion, and Utilization; Task 771902, Exp-rat)on of Methods for Increaiag theEffectiwnew of Persormcl Prolpnu.

Appreciation is expressed to Mr, Jim FHemmn and AIC Vance Srrith of theComputational Sciences Division, Air Force Human Resouces Laboratory, for theirprofesional and te-hica istance in computer piogranuing and accomplishment ofdaired aralynes.

TABLI" OF CONTENTS

I. Intf oduclio . . ... . . . . . . . .. . . . . .. . .

Objectives of the Current AnAlyses . . . . . . . . . . . . . . . . . . . . 6

II. M ethod . . . . . . . . . . . . . . . . . . . . . . . . . . . .

II!. Results and Dlscwuson ... . . . . . . . . . . . ... . .

IV. Conclusions and Recommendations . . . . . . . . . . . . . . . . 1

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Appendix A. Distributin of n Sale Scores . . . . . . . . . . . . . . . . . . . . 23

LIST OF TABLES

Table paI Criterion Groups .......... ............. .......................... 7

2 Means and Standard Deviations of Scale/Index Scores Orilinal Sample ..... ........ )

3 Results of t-Tests Between In-Service and Loss Category Means Original Sample 10.....I

4 Zero Order Corrclatlons Originai Sample .......... .................. .. 10

5 Frequency and Percent of Original Sample in Each Criterion Group by Scale/Index Cutoff Scores as Normal or High Risk ......... 12

6 Percent of Each Criterion Group Identified by Decision Rules as High RiskOriginal Sample ................. .......................... .. 13

7 Percent Screened by ('urrent Enlistment Standards. ...... ............... ... 13

H Means and Standard Deviations of Scale/Index Scores CurTentEnlistment Standards Sample .............. ...................... ... 14

1) Results of t-Teits Between In.Service and Loss Category Means - CurrentEnlistment Standards Sample .......... ..................... 14

10 Zero Order (orrelatlons - Current Enlistment Standards Sample 14..........4

I I Frequency and P-rcent of Current Enlistment Standards Personnel in Each CriterionGroup Identified by Scale/Index Cutoff Scores as Normal or High Risk ............ 15

12 Percent of Each Criterion Group Identified by Decision Rules as High Risk --Current Enlistment Standards Sample . . . . . . .. . . . . . . . . . . . . 16

13 Multiple Correlations ......... ....... ......................... ... 16

14 Summary of Regression Analyses - Original Sample . . . . . . . . . . . . . . 17

IS Summary of Regression Analyses Current Enlistment Standards Sample ..... ........ 18

Al PEI Scale - Emotional Instability . . . . . . . . . . . . . . . . . . . . . 23

A2 PDA Scale Mrug Use Admission .... .................... 23

Preceding paig blank3

Lat of Tables (hnranned)

A3 Frequency and C'umulative Percentage Distributions of PEI Scoresby Criterion Gioup Oriinal Sample .............. . 24

A4 Frequency and Cumulative Percentag Distributions of PDA Scoresby Criterion Group -- Orillina Sample ....... ..................

A5 Frequency and Cumulative Percentage Distributions of ADI Scores

by Criterion Group foOrignal Sample. .. .. .E.n... .. ..... ......... 26A6 Frequency and Cumuiative Percentage Distributions of PEI Scores

by Criterion Group (or Individuals Qulifying Under Current Enlistment Standards .. 27A7 Frequency and Cumulative Pevoenta.e Dlstributions of PDA Scores by Criterion

Group for Individual Qualifying Under Current Enlistment Standard. ......... 28

A8 Frequency and Cumulative Percentage Distributions of AD! Scores by CriwtrionGroup for Individuak Qualifying Under Current Enlistment Standards 2q

4

SCREENING FOR ADAPTABILITY TO MI IUTARY SERVIC h

Idntification of personnel who are most likely to succeed on the job is the primary goaM of everyselection and classfication program. Hiring personnrel who wlfl later terminate employment due tonon4daptablity of unsuccessful pefforwusece represents a ooat which night be minimized if mnore preciseproaduses could be developed to identify those individuals not likely to succeed at a later date.All brandhes of the armed services have been concerned with problems of adaptation. For over 15 years,research has been directed toward the devellopmnent of a screening technique which could be used toWidnrdfy recruits who ane considered high-risk for problems in adaptation to the military environment(lyer, 1959; Mostr, Dubuheos, & de Jung. 1961; Flag & Goffntan, 1966). Some have investigated theefficiency of the psychiatric interview and general clinical aussertment procedures in identifying potentialnialadaptve accessons (Jensen, 1961; -,lag, 1964. Plag A Arthur. 1965;- Plag. Arthur, & Phelmn. 1970:Shoemaker, Drucker, & Kriner, 1974). Other investigators have focussed on the importance ofptS4nlhstmetnt/blographical variables and their relationship to Ilate performance and adjwtment in thenrilitary (Flyer, 19S9; Fischer. Ward, & Iloldrege, 1960; Gordon A Bottenberg, 3962: flag, 3962;Gundenon, 1963-. Arthur, 1971). Ib several studies, various inventories developed for screening wereevaluated for their effectiveness in predicting adjustment problemns (Danielson A Clark, 19)54-. Jensen, 196 1,flag, 11962; Larsn a Krlattasomn, 1969; Bucky A Edwards, 1974; LaChar, Sparks, & Lawsn. 1974).

Although the practical usefulness of these personality and biographicaljattltudlnal inventories has notbeen conclusively denmonstrated, findinpi from these studies do indicate consistent relationships betweenvariables such as level of education, age, and general intellectual level with overall military effectiveness(Flyer, 1959, 1963, 1964; flag A Hartlacre, 3964; Drucker & Schwartz, 1973;DBoyd & Jones, 1973). Otherfators such as problems in schooling.A family stability, and arrest history were also found to predicteffective perfornmane (Flag, 1962; Plag & Guffman, 1966; Flag, Arthur, & Goffmtatk, 1970; Arthur, 19", ).

In 1972, Air Force nudca] personnel initiated a research projct to develop a screening techniquewhich could be used to identify recruits who are considered high-risk for problerm in adaptation to themillitary environment (LaChar, Larson, & Sparks, 1974). For use in this project, LaChar et ad. developed a100-Item WISlfrport history opinion inventory (HOI) designed to tap dimensions of schoct adjuatment,family stability, social orientation, emotional stability, bodily complaints, motivation and expectations forachievement, and response toward authority.

Using the inventory and criterion data obtained on approximately 15,000 male airmen during basictraining, two predictive scales were developed from the HOI for future use in screening, The prediction ofemotional instability (PEI) "cae was designed to rmeasure characteristics associated with emoationalmaladjustmaent; the prediction of drug use admission (FDA) scale was designed to measure thosecharacteristics associated wath the acknowledgment of previous drug usage. These two scales were thencombined Into ai adaptation index (ADI), and an optimal cutoff score was determined which wouldclassify recruits into one of two categories, normal oz high-risk. Rased on this ADI cutoff score, 32 percentof the umple population was labeled as high-risk for military adaptation. According to records maintainedby LaChar et al. half of the high-risk group did, in fact, experience problems in adjustment during basictraining, although their problerns did not necessarily result in discharge from service. This high-risk groupwas composed Of Seven percent of the normal criterion group, 41 percent of the severe adjustment groupand 47 percent of the drug discharge group.

Based on these results, it was concluded that predictiont of initial adaptability to military service isPossible, and that such screening could result in substantial savings to the Mir Force in identifying personnelwho require special treatment or who should be separated from service (LaChar et al.. 3974). However, thecriterion classifications used in the initial analyses were partially based on subjective clinical And Instructorevaluations or' the individual's behavior, prior to consideration of such an instrument for use in theoperational screening of Alr Force accessions, it wus considered advisable that further investigation of thehistory opinion inventory be accomplished usinr the objective criterion of In-service versus actualseparatIon/dizchargle from service,

Mobctives of the C - Analyse

The obecties of the curmnt study were (I) to follow up the accesmons adinitered the historyopmoni mwntory in basc nulitar) training order to detemine the accuracy of the HOI scores inpredictaig the cntenon of tneout of service durng the first two years after enlintment. and (2) to determinewhether aditional apatudial and biographical data might inae the effecttveness of the screeningpiocedure.

IL VrA

Subwtrst The sample population consisted of 15 '25 basc armen who were adrministered the HOIdunng basc muiitary tramnig at Lakland AFB. Texas. horn June through August 1972.

Pmtmwhurr The data files estabished by LaChar ei al. were nutched with the airman tape filesmIntamed 1-y the Computational Soences Division. Air Force Human Reomces Laboratory, to obtainaptitudinal, biographical and discharpe data. A total of 160 case in the onlnial population dw not matchthe official data file which reduced the sample population to 15.092. It a not believed that the [on ofthen cues represents any bia in the remainiag sample which would materially affect the results obtained.

Dischag status was determined by a standard designation number (i.e.. low code) which identifiedall peonnel who had been separated of disharged from serrice during the first two years after enkstment.

Lost codes indicating a similar reason for separation or discharg from service were grouped togetheras shown in Table i. BRed on the specifc lou code indicated in an individual's official record, eachindividual in the sample population was assigned to one of the following mutually exciunv criterionPoupS:

I. hwerke - this group consated of 10.329 individuals who were still on active duty or hadextended their onginal comnuntment as of September. 1974.

2. Loss. arrewd sepmiton - this group of 658 included those individuals whose loss ( )aes did notreflect any problem in adaptation, such a sepssation and transfer to AF Reserve.

3. Loss. desi'bilirv indetrmire - this group of 364 airmen included those categories of losses forpersonuai/hardship masons, death, releas to enter an educational institution, and release for the convenienceof the Govemment.

4. Loss. ph)icd reasons - this group of 457 individuals included all separtiom/retirements due tophysical disability, obesity, and failure to meet medical fitness standards at time of enlistment.

5. Lou. uvuftabi/ti. - a total of 371 airmen comprised this classification of undesirable loss. Majorrea-ons for discharpg induded character and behavioripersonality disorders, drug abuse, and sexualdeviation.

6. Loss. moarin pmaductr~lý - this group of 853 airmen induded discharges due to minimal ormarpnal productivity and unsuitability due to apathy, defective attitudes and inability to expend effortconstructively.

7. Loss, disquaified for retention - this group of 1.828 individuals was discharged based on theirfailure to meet mninmumn requirements for retention in the Air Force.

8. Loss. wurftnes; - the 156 individuals in this group were discharged for reasons of unfitness ormisconduct; i.e., frequent involvement in incidents of a discreditable nature with civil or militaryauthorities, conviction by civil court/court martial, AWOL/desertion.

9. Loss, miscedaneous undesirable -- the remaining 77 individuals assigned to this group includedthose released for reasons which were considered under a miscellaneous category of aindesarable. e.g, beinga conscientious objector or for the good of the service.

The nine loss categories were then combined to form thrue additional criterion classifications:out-of-service, ioss-not undesirable, aid lou-undesirable. The out-of-service group was comprised of allindividuals separated or discharged from service regardless of cause. Individuals in the loss-not undesirable

6

its

P4gI-

r.44

f-l fuf

Lz -

f if

*1 ~ .CL

06 -

;0. w

II-. OS .9ac ar,

ItQi 3 -CW ý 3ow

00

I: ime2c

V7 wwf

371

3 .-3

c. E.- - 5

group were thosec asupgnd to one of thc three following loss categories: nornral sepairations, loss-desirabilityineternsinate, or loes-physicul reason. The final group, low-undesirable, included the loss categories of

unsuitability, ma~rginal productivity, disqualified for retention, unfitness, and miscellantous-undesif able .I :Scoring of the HOI response data and the cutoff s~ores used in the curreant study are those previously

established b~y LaChay at al., 1974.The actual items, scoring used in deriving the HOt scale scores, and estimates of scale reliability are

plasn.ted in Tables AlI and A2 in Appendix A. The weighted linear comnbination of the two scale used toderive the ad~aptiation index was .6568 of the PEI scale value and .7541 of the PDA scale value. The decisionrules (cutoff scores) basd on the optimum value which differentiated between recruits who would andwould not have problems lin basic training were as follows: PEI -. 7.5 wale value-, PDA - 11.5 scale value;ADI - 12.5 scale value.

Distributional analyses of HOI scale scores ifte accomplished to determine the number of individualsscoring at each score interval on the three scales, Based on cutoff scores for the 1101 scales, the percentageof indviduals in each criterion group identified as high-risk for problems in adaptation was t~abulated.

Comparison between the means of the irx-aervice group and the different lows groups wereaccompliscd. and the differences between means were tested for statistical significance by means of t-ests.Error rate fot thewe comparisons was controlled per hypothesis ; i.e., a total Type I err or rate of .05 wasconsidered acceptable

Correlations were also computed to indicate the relationships between HOI scores and the variousin/out criterion cAtegorIes. Finally, regression analyses were accomplished to determine the usefulness ofbiographical and aptitudinal data in predicting adaptability to ndlitary servicer, and whether these datasigirificantly increase accuracy in prediction over and above the use of the HOI ;cores alone.

MD. RESULTS AND ISICUSSION

The means and standard deviations of the HO! scores by criterion classification are presented in 'Table2. Results of the t-testa between the means of the in~ervioe group and the loss groups, summarized in Table3. indicate the differences between the rmwas are quite similar for the three scales, lIn a majority ofcomparisons, the in~aervice group mean differs significantly from the lou/out-of..ervice means on all sca1las.Ilowever, mean comparisons between in-service and normal separations and between in-service andlu.a-deslrahlllty indeterminate were not significant across all scales, For the PEI scale, comparisns of meandifferences between the in-service group arid the unfitness and rniscellaneous-undesif able groulps wereb alsonon4ignllfcant.

Table 2. Mleans usid Standard Deviations of Scale/Index Scores - Original Sample

put PDA Adaatati-ansaeal Scale Index

Valld -____ _____ _______criterion Groupean so Mea" so Mean SD

InI Service 10,329 3.03 2.42 5.20 3.46 5.91 3.80Out of Service 4,764 4.16 3.51 7.46 5.25 8.35 5.93

Loss, Not Undesirable 1,479 3.35 2.98 5.56 4.14 6.39 4.78Normal separation 658 3.15 2.60 5.33 3.56 6.09 4.03Desirability Indeterminate 364 3.14 2.83 5.40 3.81 6.13 4.41Physical Rermons 457 3.80 3.52 6,03 5.04 7.04 5.83

Loss, U~ndesirable 3,285 4.53 3.67 8.32 5.47 9.24 6.18Unsuitability 371 4.05 2,95 7.37 4.09 8.22 4.55Marginal Productivity 853 3.30 2.66 7.05 4.07 7.48 4.40Disqualified -Retention Stds 1,828 5.37 4.08 9.25 6.21 10.50 7.02Unfitnes 156 3.19 2.47 7,15 3.81 7.49 4.11Miscellaneous - Undesirable 77 3.10 2.32 7.,1' 4.54 7.44 4.55

9

I

Tjbhlc 3. Results of t-Tests lBetween ln-Service and Loss CategoryMeans - Oviginal Sample

MeahComparisons PEI POA AOI

In/out of service 23.00** 31 .35#4 30.43"*in/loss, not undesirable 4.62** 3.o2"* 4,39**In/loss, normal separations 1 .26" ,8 8 a 1,12aIn/loss, desirability lndetm .85a 1.052 i.07aIn/loss, physical reasons 6.53** 4.870 6.04 *In/loss, undesirable 26.97w* 38.480* 36.940*In/lass, unsuitability 7.92** 11.76"* I 1.38*11In/loss, marginal productivity 3,16* 14.76"* 11.46**In/loss, disqualified for retention 33.71 ** 39.87** 40.70**In/loss, unfitness .84a 6.980* 5.14*0In/loss, misc undesirable ,28a 4 .9 5 "* 3.52"*

2Not significant,"•Significant at .05 level,

"•*Significant at .01 level.

The corelations of the HOI scales for the various criterion group classifications are presented in Table4. Thosa undesirable categories containing a sufficient number of individuals to assure some stability ofresults were used separately to indicate the effectiveness of the scales in differentiating between thosein-%ervice aid those discharged for a specific reason. All correlations are statistically significant at or beyondthe .01 level, The absolute value of the correlations reported miy be somewhat inflated since a portion ofthe sample had been previously used for scale construction, However, the degree of inflation can beconsidered minimal based on the large sample size and the fact that the present criterion groups were notused in the actual scale development. Although significant, the observed relationship between HOI scoresand the criterion groups comprising the marginal producers or unsuitable personnel appears negligible froma practical standpoint. A definite but low to moderate relationship is evident for the remaining criteriongroups. It should also be noted that the correlations obtained on a sample population previously screentby operational selection tests are somewhat lower than if they had been computed on an unrestrictedpopulation.

Table 4. Zero Order Correlations" - Original Sample

H4o0 Seaom

Criterion Groups PaI POA ADI

In/out .2022 .2735 .2659In/total loss, undesirable .2241 .3122 .3013In/loss, marginal productivity .0296 .1378 .1072In/loss, disqualified for

retention .2910 .3387 .3448In/loss, unsuitability .0763 .1131 .1093

aAll correlations significant at .01 level.

10

The statistical significance of a measure often fails to reflect the practical usefulness of any screeningdevice. An assessment of -101 utility can be made by a comparison of the number of personnel correctlyidentified (i.e., hits) versus the number of individuals incorrectly dassifled (i.e., false positives and misses).Hits include all personnel Identified as normal who are still in service and those identified as high-risk whohave been discharged from service. False positives include those individuals still in service who wereidentified by the HO! as high-risk and miseL include those losses classified as normal. Table 5 shows thefrequency and percent of each criterion group identified as normal or high4isk using the decision rulesestablished by LaChar in 1974. Cumulative percentage dlistributiom indicating the number of individualsat each score interval for the three HO! scales are also presented in Tables A3 through AS in Appendix A.Overall, I I percent of the total sample used in these analyses was identified as high-risk by the PDA andADI scales; nine pcrcent by the PEI scale. Six percent of the in-aervice group was identified as high-risk byeach of the three decision rules for the HOl scales. The scales vary somewhat in the percentage of the losscategories identified as high-risk. Using the PEI scale, 18 percent of all lones and 21 percent of theundesirable losses were identified as high-risk; with the PDA scale, 22 percent of all losses, 28 percent of theundesirable lones;, with the ADI Index, 23 percent of all losses, 28 percent of the undesirable category. Acloser review of the high-risk subgroup identified by the PDA or ADI scales shows that over 60 percentwere actually discharged from service and over 55 percent for reasons of undesirability (Table 6). It appearsthat the PDA is almost as effective as the ADI Index in identifying personnel who are separated ordischarged from service. If similar results are found In future validation of the HOI, consideration should begiven to simplifying the scoring process by using a single scale score for screening instead of the weightedAD! index.

Since the sample population entered service, enlistment standards have become more stringent.Today's accenid must meet three criteria: (1) each individual must obtain a total score of 170 or higheron the fouw combined aptitude indextv of the Armed Services Vocational Aptitude Battery; (2) theirGeneral Aptitude Index score must be 45 or higher; and (3) if they receive a mental classification ofCategory IIl or IV on the Armed Forces Qualifying Test, they must be a high school graduate, To give someinsight into the effectiveness of the personality scales of the HO! on a population similar to current enlistedaccessions, identical analyses on only those recruits meeting the new enlistment standards were performed.The actual number in each category qualifying on the multiple standards is shown in Table 7.

Descriptive statistic: for the HOI scales on this restricted population ae presented in Table 8 with theresults of the comparisons between the means in Table 9. Results of the analysis of differences between themeans between in-service perbonnel and out-of-service categories were simtilar to those in the originalsample. Statistical comparisons of PEI mean differences between in-service personnel and each of the Ionscategories reflected significant differences in all comparisons except those involving normal separations,losses with desirability indeterminate, marginal productivity, unfitness and miscellaneous-undesirable losses.For PDA and ADI mean comparisons, differences between in-service personnel and two of the lons groups,normal separations and loues-deslrablity indeterminate were not significant. In addition, the in-service ADImean did not differ significantly from the miscellaneow-undetdrable loses. All other comparisons of meanson the three scales were significant at or beyond the .05 level.

Some decrease in the abs.olute magnitude of' the correlations between HOI scores and criterioncategories is also evident in the restricted population (Table 10). Although the observable relationships areattenuated by the restriction of range imposed by the new enlistment criteria, the low to moderatecorrelations are still statistically &!,nificant.

The proportion of individuals identified as high-risk by the decision rules of the HO! scales differslightly from the originil sample as shown in Table 1I. The more detailed frequency and cumulativepercentage distributions Are contained in Tables A6 through AB in Appendix A. Due to the more stringeatselection standards, only 26 percent of the undesirable loss category was identified as high-risk compared to28 percent in the original sample. However, it should he noted that 35 percent of the undesirable loncategory would have been rejected prior to enlistment had %he new zriteria been prerequisite for entry intothe Air Force in 1972. Of the number identified as high-risk by the PD. or AD! scale, over 50 percent wereactualh\ discharged for reasons of undesirability (Table 12).

Since certain biographical and aptitudinal data are available from an individual's official records at thetime of entry into the Air Force, the use of these data would eliminate the administration of the HO! ifsuch data were as effective as the HOI scales in identifying personnel who are discharged from service.

Il

0 r , - 0 r- 0 0 1 ý C

I -

t €",4 0 ,- ' *'

zV -r- -IV-

-00

.Z .- o, ,v .ri e•

*1 r-o'c -o O.*.4M I,

r•J-

SA•ba. t

en e C' w~ %0 -I-...W, I-

8A

c A

12

V

Table 6. Pecent of Each Criteulon Group Identified by DIeclton Rulesa tUg% iRk - OrWh Saunp

Pat Sele IPOA gas AOl Imslax

H1g96 RNOk WOggs RIsk W141 Risk

Crnteen o eunp > 7.s > 11. s>2.5

In Service 41 35 36Out of Service 59 65 64

L.•s, Not Undesrable 10 8 9Nornmal Separation 3 2 2Deirability Indeterminate 2 2 2Physical Roasons 5 4 5

LoU, Undesirable 49 57 55Uniuitsbility 3 4 4Margnal Productivity 5 9 7D6quallfled.Retention Stds 40 42 42Unfitness I I IMlscelianeous.Undesirable 0 1 1

Total 100 100 100

Table 7, Percent Sceened by Current Enlistment Stdand•s

1nMItment Standard Compositets

crHeulen Groupap Number De~eIttlen N % N

I In Service 8.125 79 2,189 21Out of Service 3,233 68 1,523 32Lous, Not Undesirable 1,105 75 372 25

2 Normal Separation 525 80 132 203 Desirability Indeterminate 266 73 97 274 Physical Reasons 314 69 143 31

Lou, Undesirable 2,128 65 1,151 355 Unsuitability 231 62 140 386 Marginal Productivity 541 63 311 377 Disquallfled.Ratention Stds 1,219 67 604 338 Unfitness 90 58 66 429 Misc- Undesirable 47 61 30 39

Total 11,358 75 3,712 25

aro be qualified, individual must have a total of 170 for his combined aptitude index scores (M,A,G,E), a GeneralAptitude Index sawre of 45 or better, and Cat III and IV personnel must be high school Fdustes,.

bInformation required to determine enlistment stundard composite was not available for 23 caes.

13

Table 8. Meas and Standard Deviations of Scale/Index Scores - CurrentEnlktment Stadad. Sample

Pal PDA AdaptationVoa soero Index

Vellll

Criterion Group N Mean sD Muan So Mean So

In Service 8,125 2.93 2,36 • 5.00 3.33 5.69 3.67Out of Service 3,233 3.98 3,42 7,06 5.01 7,94 5.69

Losa, Not Undesirable 1,105 3.16 2,85 5,33 3.93 6.09 4.49Normal Separation 525 3.06 2.60 5 1(, 3.53 5.92 3.99Desirability Indeterminate 266 2.83 2,62 5.lu 3.63 5.68 4.12Physical Reasons 314 3.60 3.34 5.77 4.71 6.71 5.44

Lou, Undesirable 2,128 4.41 3.61 7.96 5.27 8.12 4.51Unsuitability 231 4.04 2.92 7.22 4.02 8.52 4.49Marginal Productivity 541 3.14 2,61 6.76 3.95 7,17 4.27Dlaquallfled-Retention Stds 1,219 5,18 3,99 8,78 5.93 10.03 6.76Unfitness 90 3.23 2,66 6.72 3.86 7,18 4.28Mtmcelaneous.Undesirable 47 2,79 2,19 6.51 4.05 6,72 4.08

Table 9. Results of t.Teste Between In.Servlce and Lou CateloryMeans - Cutent Enlitment Standards Sample

Main aiCeompartion Pal PDA ADO

In/out uf service 18.b9"* 25.50"* 24,890*ln/ou, not undesirable 2,964 9.040* 3.28*

In/loss, normal separations 1.19& 1.276 1,40AIn/loss, desirability indetm ,63& .47' .05'In/loss, physical reasons 4,88** 3.92"* 4.71 **In/loss, undesirable 22,74" 31.83"* 30.92"*In/lons, unsuitability 7,030* 9.89** 9,85**

In/lon, marginal productivity 2,041 11.73"* 8.95*0In/loss, disqualified for retention 27,924* 32.6100 33.56*0In/lon, unfitnes 1.22' 4.8600 3.81"*In/loss, misc undesirable .41' 3.090 1.92'

*Non-tlgniflcant.*Signiflcant at the .05 level,

*Signiflcant at the .01 level.

Table 10. Zero Order Correlationa -. CurrentEnlistment Standards Sample

Criterion Groups PIl PDA AOl

In/out .1933 .2611 .2547In/total los, undesirable .2188 .2997 .2915In/lon, marginal productivity .0219* .1250 .0952In/lon, disqualified for

retention .2771 .3192 .3276In/lons, unsuitability .0767 .1076 .1060

2AH correlations except where noted significant at .01 level.

*Signiflcant at the .05 level.

14

cl -ý N - t- t^-

00

- (4

*1 15

Table 12. Percent of Each Criterion Group Identified by Decision Rulesas High Rhk - Curent EnhImesnt Stainderds Sample

Pal Ismic PDA meal ADI Ikdes

1419fgls HIP Risk with Risk

Criterlo eup >7. •.5 11.a > 1.

In Service 43 37 38Out of Service 57 63 62

Loss, Not Undesirable 10 9 9Normal Separation 4 3 3Desirability Indeterminate 2 2 2Physic.a Reasons 4 4 4

Lous, Undesirable 47 54 53Unsuitability 3 4 3Marginal Productivity 4 7 6Disquallfled-Retention Stds 38 41 42Unfitness I I IMlscallaneous.Undeirable 1 I I

Total 100 100 100

Therefore, a series of tegression analyses were accomplished on the orlginhal and current enlistmentstandards samples to determine the usefulnem of biogamphical and aptitudinal data alone or in combinationwith the HOI scales in predicting Ion from active duty. Baned on the similar percentage of out-of-servlceand undesirable lou personnel identified by the PDA scale in comparison to the PEI scale and the weightedAD! index, another series of analyses were accomplished to see if any significant loss in predictive accuracywould occur by using the PDA scale alone. Two criterion goupings were used for the regresson analyses-in4ervice and out.of4ervioe; in-service and total loss-widemlrable. Multiple correlstiom for the variousgroups of predictors are given in Table 13. Summaries of these regression analyses are presented in Tables14 and 15. The first regression comparison indicates that the aptitudinal and bioraphical data do addsignificantly to prediction over above the AD! index. On the other hand, however, the aptitudinal andbiographical data cannot be used in lieu of the AD[ scale; I.e., the ADi index does make a uniquecontribution in both criterion groupings, The comparison to determine whether the PDA scale contributes

Table 13, Multiple Correlations'

Aptitulindl m4 Aptitudinal. Apltiudlnal, Aptlttdloal,CrItstiOn drumplos 0 remPhlis daal" Ias and ALI ae, PgA, PEI We, PDA

Original Sampleln4ervice/out of service .:149 .3127 .3158 ,3150ln4erviceflos.undesirable .2303 .3459 .3506 .3500

Current Enlistment Standardsln.service/out.of.service .1475 .2778 .2816 .2805In 4ervice/lot-sundesirable .1600 .3130 .3176 .3165

t ADI correlation silnifkant at or beyond .01 level.bAttudn data includes four aptltude index sý.ores, APQT scorei biopaphical data Includes age at enlistment and

yomrs of education.

16

W) -I'. r#

00

0 0 -0 -0

v00 90 0r m t Nr

17

I. ,

- - -

I 1'8

as mudih as the .unmbuiation ot all IR1! scalc, uidi aics the PD)A scale L.na be efiectively uw.d alone. If thepredictive efficiency of the PDA scale over the other HOI measures is found in future validation studies, thecontinued use of the PEI scale to form the weighted ADI index appears redundant and unnecessarily morecomplicated than a single score cutoff. The final comparison of squared multiple correlations wu made todetermine if the aptitudinal and biographical data still made a sigtificant and unique contribution if thePDA i,:ale were used alone Resits of this comparison indicate the aptitudinal and biographical data makea signiticant contrihti',n to the PDA scale also.

In general, the value of implementing any screening procedure based on biographical, aptitudinal, orpersonality data must be carefully evaluated by considering the savings which would be accrued by earlyidentification of maladaptive personnel versus the loss to the Air Force of potentially sutcessful personnelwho might be denied enlistment or separated prematurely from service, When the quantit) and quality ofthe prospective rucruit manpower pool are high, a coarse screening rnethodolugy can be cost-effective insaving the expenses of training, counseling, tteatment, and administration associated with personnel whohave adjustment problems even though it also identifies a sizeable proportion of potentially productivepersonnel. On the other hand, if the volume of prospective recruits is low, the number of potentiallysuccessful persornel identified as maladaptive becomes a critical Isue.

Should a screening mensure such as the HOI be considered for operational use, several additionalprocedures should be incorporated to safeguard against identifying and possibly rejecting a large number ofpotentially productive and successful military personnel, For example, counseling could he scheduled for allpersonnel exhibiting symptoms of initial maladjustment. Many problems might he transitory if professionalguidance were made available during basic training. Secondly, additional in-depth assessment proceduresshould also be administered to high-risk personnel in an effort to identify those with majorpsychkitric/Jmotional prohlerns who should be separated fro, n-rvct' as s ,oin as possible.

IV. CONCLUSIONS AND ReCOMMENDATIONS

The correltions obtained for the 1OI scales appear to indicate that the self-report data contained inthe HO1 has some practical usefulness as a rough, preliminary screenip, device, However, the smallmagnitude of the observed relationships necessitates careful review by professional personnel of allpersonnel identified as high-risk. In no instance should identification as high-risk by the HOI be used as thesole basis for any adverse personnel action. On the positive side, the HOI does identify a slzeantle proportionof recruits who were actually discharged from service as undesirable during the first two years of activeduty. Even under current enlistncnt standards, over 25 percent of the widesirable losses would have beenlabeled as high-risk. However, 'he overall savings which mnight be accrued from early Identification of thehigh-risk group might be obscured by the costs of iMtplcmenting a secondary assessment and counselingphase which is considered necessary with the use of a rough screening device such as the HOI,

Prior to the use of the HOI in an operational setting, the following recommended courses of actionare considered mandatory.

a. Revalidate the 1101 on accessions under current enlistment standards to deternube itseffectiveneu and stability on a new population. In the original sample, a large number of personnel whowere discharged for admission to prior drug usage were used for scale construction. Although the PDA scaleappears to be quite effective in identifying all types of undesirable losses, the appropriateness of the originalscales or cutoff scores developed on that population may he questionable if a decrease in the number ofdrug discharges has occurred during the past two years.

b. While results obtained on the original sample suggest that such a screening procedure might beused effectively, the population consisted of rnale accessions only. Prior to using the HOI as a screeningdevice on a female population, additional research must be accomplished to establish the applicability ofthe scales and cutting scores on a WAF population,

c, It is further recommended that use of this screening device should be limited to preliminaryscreening only and that additional psychometric and/or psychiatric assessment be mandatory before anypersonnel action is recommended. Every effort should be made to retain as many of the potentiallysuccessful personnel in the high-risk category as possible.

d. To uimplify the administrative scoring of the HOt, it is reommended that further research on thisinstrument investigte the possibility of developing one scale rather than the com.lex weigl~ng proem usedfor deriving the ADI index score.

e. Sued on the pr*linauy repemion analyses, additional aptitudinal and biogaphical data which isavallable on all recruits should be considered in combination with HOI data to improve the accuracy inidentification of malad•ptive personnel.

UEFERENCUE

Arthur, RJ, Success Is predictable. M~litwy Medick.I, 1971, 13, 539.545,

Boyd, K.N,, & Jones, HA, An andysb of fIaton relted to desertion among FY 1968 md FY 1969 Armyacrasion. AFHRL.TR.73.63, AD.-772 731. Alexandria, VAý Manpower Development Divslon, AirForce Human Resources Laboratory, January 1973,

Ducky, S.F.. & Edwrds, D. The Recruit Temperament Survey (RTS) as it discrininates between psychoses,neuroses, and penrotality disorden. Jouma of(Clincal l.vrhod•gv., 1974, 30,195-199.

Dnilsom, J.R., & Clark, J.H, A personality inventory for induction screening. Journal of lincalPsycholog,, 1954,10, 137.143.

IDucker, E.H, & Schwartz, S, The prcdictlon of A WOL, military skills, and leadership polcntiai.HumRRO.TR.73,1, Alexandria, VA; Hunman Resources Research Orpsnization, January 1973,

Fisher, W.E., Wald, J.H., Jr., Holdreg, F.E,, & Lawrence, HG. Prediction of umuitabillo, dischages,WADD.TN.60-260, AD-248 077, Lackland AFB, TX: Personnel Laboratory, Wright Air DevelopmentCenter, October 1960,

Flyer, ES. Factors relating to dischage for unsuitability among 1956 airmen accessions to the Air Morce.WtDC.TN.59.201, AD.230 758. Lackland AFB, TX: Personnel Laboratory, Wright Air DevelopmentCenter, December 1959.

Flyer, E.S. Prediction of unsuitabiity among fint. term airmen ftom aptitude indexes, Ngh school referencedata. and baic training epaluations, PRL-TDR-63-17, AD.420 530. Lackland AFB, TX: 6570thPersonnel Research Laboratory, Aerospace Medical Divlslnn, June 1963,

Flyer, ES. Prediction by career field of fint-term airmen performance from selection and basic trainingvariables. PRL-TDR.64.5, AD.600 781. Lackland AFB, TX: 6570th Personnel Research Laboratory,Aerospace Medical Division, March 1964.

Gordmn, MA., & Bottenberg, R.A. Prediction of unfavorable dischade by separate educational levels.PRL.TDR.62.5, AD-284 802. Lackland AFB, TX: 6570th Personnel Research Laboratory, AerospaceMedical Division, April 1962.

Gundeuson, EK. Biographical indicators of adaptation to Naval service. Report Number 63.19. San Diego,Califorida: U.S, Navy Medical Neuropsychiatric Research Unit, November 1%3,

Jensen, M.S. Adjuative aid nonedjustive reactions to basic training in the Air Force. Joumal of SocialPsychology, 1961,56,3341.

Kle:.. W.A., Dubauson, AU., & de Juug, J.E. Prediction ofunavceptable performance in the Army. TechResearch Note 113, Washington, D.C.: Human Factors Research Branch. TAG Research and Devel.opment Command, June 1961.

20

L•lhr, 0)., Sparks. .K., & l..sen, R.M. PKyehuciiw, ic pirediction iif hhld i.,,, c1lt0iic•.'a 01 adapta•uioI lotUSAF baic trainccs.Joumal oj Conamunit' PIsvcholgky, 1974, 2(3), 268-277.

LAwa, El., & KIistlamsn, D.M. Predlction of dacipllnzrv offense erlyv in Army service. Tech ResearchNote 210. Arington, VA: US, Army Behavtoral Science Research Laboratory, April 1969.

Kis, J.A. Pre nlistnent variables related to the performance and adjustment of Navy recruits.Journal ofCOiical Psychoklgv, 1962, 19, 168.171.

fM, J A. The p trwtcal mime of a psyhatilc screening Interview in pnrdicting mfiitaty ineffnitienets, ReportNumber 64-7, San Diego, CA: U.S. Navy Medical Neuropgychiatric Research Unit, April 1964.

Fa, JA,, & Arthur, RJ, Psychiatric re-examination of unsuitable Naval recruits: a two.year follow-up.American Joumn of Psyehtiary, 1965, 122, 534.541.

fa, J.A., & Goffmn , J.M. The prediction of four-year ndilitaty effectiveness from characteristics of Navaliecruits. Military Atdirne', 1966, 131, 729.735.

flag, J.A., & Hda•re, L.E. The validity ofa e, vdutwtlon, and ;CT sonv' as prrdid'ors of two..veaattrition anumg Naval enilstees. Repoit Number 64.15,. San Diego, CA: U.S. Navy MedicalNeuropsychiatric Research Unit, June 1964.

Plag. J.A., & Hardsve, LW, Age. vvars of'/scliolPig, and Intelllgenr•' as predlcton of nfilitate, efjicth, enessfor Naval enlistees. Report Number 65.19. San Diego, CA: IUS. Navy Medical NeuropsychiatricResearch Unit, July 1965.

Ph& J A., Arthur, RJ., & Goffman, J.M. Dimension of psychlatric lllnem among fint-term enliswees in theUnited States Navy. Milltarv M.,didne, 1970, 135, 665.67";,

Plag, JA., Arthur, RJ., & Phelan, J.D. An evaluation of psychiatric wlection at Naval training centers.Paper presented at the 119th Annual convention (if the American Medical Association, Chi ago, June1970.

Shoemaker, W.B,, Dnicker, E.H., & Kriner, RE. Irdklcttin ofi delinquency among Army enlisted men: Aiultivaoiate apdivsis. Hum RRO.TR.74.3, Alexandria, VA: Human Resources Research Oganlatton,

February 1974.

21

APPENDIX A. DISTRIBUTION OF SCALE SCORES

Table 4 1. PM Scae - Emotioni Instabity

ptlam ne e.41

I, I have needed help for erotional problems. T2. At one time I needed nedication to stay calm. T3. For a long time I have had difficulty deeping. T4, 1 often have headaches. T5. have cried several times this put year T6. 1 usually take things hard, T7. 1 enjoyed physical education. F8. 1 have had more than my share of illness. T9. 1 needed special help with my school studies. T

10. 1 am joining the Air Force to get a better education. F1I. i would rather work by myself than with others. T12, 1 was a dow learner in school. r13, I would rather read than be with people. T14. 1 entered the service (AF) because there was nothing else to do. TI. 1 do not mind orden and being told what to do, F16. As a child I was loner. T17. My father is (was) a nervous man. T18. I never cared much for school. T

w te.- KKL20 scak reliability; .716

Table A 2. PDA Soe - Dsg Use Adunslon

I. I often played hookey from school. T2, I quit school because I lost interest, T3. I feel better when I drink. T4. For a long time I have had diff-ulty sleeping. T5. I think I will make the Air Force a career. F6. 1 am joining the Air Force to get a better education. F7. I never cared much for school. T8. 1 have been in trouble with the police. T9. 1 wu suspended from school more than two times. T

10. 1 have often gone against my parents' wishes, T1I, 1 do not mind orders and being told what to do. F12, 1 often have headaches, T13. 1 entered the service (AF) because there was nothing else to do. T14, 1 had my share of trouble with teachers, TIS, 1 was expelled or suspended from school. T16. At one time I needed medication to stay calm. T17. 1 have never done any heavy drinking. F18. I enjoyed physical education, F19. I quit school because I was falling, T20. I have been expelled from school more than once. T21. I have been arrested more than twice. T22. ligh school wu boring, T23. I often cure and swear, T24. 1 plan to attend college. F25. I sometimes wanted to run away from home. T26. 1 have needed help for emotional pioblernu. T

Nets. - KR-20 cake teliabiity: .801

Preceding page blank 23

Mvo mf~ Mwl" 0 cc 00 00 0 01-

a ' ~ ' aN -h m 0' '

z mm MUS 'D C', 'D 0 0

W) r4 ON k.M0 V a -- -e

0rii ~~~~ r4,N~Oa f4f

~ ~ ~c en1 ~C ' M-ri

rn 14

ae3 e'r

-U f-

z0 fM 4 V

.4e NN

0 rCý ON N f4k

.00 a 00NIn C4

A U

24

, "I t- W)t

E

ow~

on '-r o c'

3 Er

M'D00

at 0 r t1'T e

00 00 t O I 4C

10 -0 I-' Ef, 1N0' W II m0E 0 00 10 c 0 't t- EN CD M-0 n W

8' 0 No 0 0 ~ N I W) w0 C- - -)t

E0 00 00 0,00r

EN a, 000 t-E EIn0 It0CC4

ZO t- 'D N' 00 t- 00 t-. I W1 0 O Z) or0 0W)-t- -oo 10 10 -* EN In W)

t- I-

wn: ,,7 u (, D O 1 ^ - m 0 -25,

*V N

tt% -- mo -CD

I $*m

P,' M. .4r-f - W0

41-M E0 -00---

~' C C -00 00- 0

-n -

hn t 44 - 4. M O N 'fumC r~~ -

a;000 oc 0- f"4 1 0"

26

z t-*= oo'

It

@0 a) 10 f'4 WI C t-:- W,4 f4 -

r- w I-t 0N m 0 -ýj4 W~'-

v0 1f-fI -PI

E.

'Z C10~s44

IFI

mlr

51 InC go 4 00 C 0~~,I un- a.,

00

@ i 10 Ot- 00 01 -I

3 ~N ~ I ~ f-

a,~ ~ ~ ~ DW fo C.I~ W04 M0W 'W

M1 f4 --

'n ~ 0~ W) 0o() W

II~ ~~~~~~~~I Wý.-A~C0N0'0..N..O O O

* 0000400 'q 1-0WI - - ') m' N0

M t- U '0 I- w" 4WI f' , N .

N N N NP .0 Nt-.It~I-.00 .N~~i~N~r'..0 q~l~28

8t 0 t- -ý %a N& -A M In N f

t - g Q A S-

In PW1I N . 22 OIO V" v AMN#

kn a 0M w tý 4~ 1.- 00 0 -

~~ 51"0~O0 4 'N '40 *O00 cc I

W)t~M0 gc~0 Chr~4 -- P4- nM f-

U ~~~~1- fIn f~-.r -f- -

Ru~ ~ ki M In -0 M 4 -0-00o~0 ý

#- ~rI -00*W , , I 1

e) In,* N rOdi MI n- 4Q0IUen a 0 )Mr - -- - -

U .0-',t 4 In

N ~ ~ $ fn V) 0 -

29

a