8/2/2019 1996-costa

1/18

APPLIED PSYCHOLOGY: AN INTERN ATIONA L REVIEW, 1996,45 (3), 225-241

Work and Personality: Use of the NEO -PI-R inIndustriaVOrganisational Psychology

Paul T. Costa, Jr., Ph.D.National Institute on Ag ing, U SA

Linventaire NEO Rtvist de la Personnalitt (NEO-PI-R) est une mesureactuelle d e 30 traits qui renvoient aux cinq facteurs de base de la perso nna littnormale. I1 a fourni des preuves de sa fidtli tt e t de sa validitt aussi bien d ansdes recherches thtoriques quappliqukes et plusieurs ttudes montrent quilcontribue i rtd ire la performance professionnelle. Le domaine et les facettesdu N EO -PI-R sont an alysts avec quelques rtsu ltats issus de son application enpsychologie des organisations.The Revised NEO Personality Inventory (NEO-PI-R) is a contemporarymeasure of 30 traits that define the five basic factors of normal personality. Inboth research and applied samples it has shown evidence of reliability andvalidity, and several studies suggest that it has utility in the p rediction of jobperformance. Th e domain and facet featu res of the N EO -PI-R are discussedalong with some issues in its use in industrial/organisationalpsychology.

INTRODUCTIONPersonality traits a re pervasive styles of thinking, feeling, an d b ehaving, a ndas such they ar e likely to affect vocational interests an d choices, work styles(Hoeks t ra , 1993),jo b satisfaction, and th e effectiveness of job perform ance.Although personality traits were once viewed as a key to selection and

Requests for reprints or for the NEO Job Profiler should be sent to Paul T. Costa, Jr.,Laboratory of Personality and Cognition, Geron tology Research Cen ter, National Institute onAging, NIH, 4940 Eastern Avenue, Baltimore, MD 21224, USA. Email:PAULC@MVX.GRC.NIA.NIH.GOV.Portions of this paper were presented at the Symposium, Work andpersonality (N. Nicholson& M. Frese, C hairs), 23rd meeting of the Interna tional Congress of Applied Psychology, 17-22July 1994, Mad rid, Spain.

Tha nks are due to David A. Dye, Leatta M. Houg h, Ralph L. Piedmont, Ann Ma ne Ryan,Paul R. Sackett, Mark Schmit, and Robert J . Schneider for helpful comments on an earlierversion.~~~~

0 1996 International Association of App lied Psychology

8/2/2019 1996-costa

2/18

226 COSTAplacement, they have been out of fashion in academic research andpublication in industriaYorganisationa1 I/O) psychology for the past 20years. Now, however, there is a renewed enthusiasm about the value ofpersonality assessment (Schmit, Guion, & Raymark, 1994) to which thisSpecial Issue attests.One of the reasons for the reconsideration of personality is a majoradvance in personality psychology: the discovery of th e five-factor m odel(FFM ) of personality or th e Big Five. The five factors-Neuroticism (N),Extraversion (E), Openness to Experience (0) , greeableness (A), andCon scientiousness (C)-are thou ght to account for most of the commonvariance in virtually all personality traits, from Adlers (1938/1964) socialinterest to Zuckermans (1979) sensation seeking. Within th e occupationalarena, Saville, Holdsworth, Nyfield, Cramp, and Mabey (1984) haveidentified 30 substan tive personality traits specifically for th e world of work.Many of these have direct parallels with the scales of the Revised NEOPersonality Inventory (NEO-PI-R; Costa & McCrae, 1992a) (e.g. OPQChange-orientated with NE O- PI-R Openness to Actions; OPQ W orry withNEO-PI-R Anxiety). Further, Matthews and Stanton (1994, p.741)concluded th at the results of the factor analysis of the 31 OP Q scales werebroadly supportive of the Big Five structural model of personality super-factors. The FFM thus provides a comprehensive framework for theorganisation of occupationally relevant personality traits, an d meta-analysesusing this framework have d emo nstrated the im portance of t he five factorsfor work performance, particularly in situations in which employees haveconsiderable autonom y (Barrick & Mo unt, 1991,1993). Similarly, Tenop yr(1993) has argued that th e FFM can be useful in interpreting research o n jo bsatisfaction and work adjustment.Increasingly, I/O psychologists (Adler & Weiss, 1988; Ho ugh & Paullin,1994; Jackso n & Rothstein, 1991, Robertson, 1993, Robertson & Kinder,1993; Schneider & Hough, in press; Schmit et al., 1994) have come torecognise that personality assessments can be valuable only if they aremeaningfully match ed to occupational criteria. As Schneider and Hough (inpress, p.22) sagely stat e, instead of hurling all pred ictors a gainst criteria inthe hop e that som e will stick-which G ui on and Go ttie r (1965) criticised asa broadsid e approach-a theoretically and empirically driven con struct-oriented approach should guide research in this area. An excellentillustration of this other approach is provided by Houghs Project Apersonality research. Jo b performance criterion taxonomies were d evelopedalong with a nine-construct personality predictor taxonomy (H ou gh e t al.,1990) in the contex t of a sophisticated a ppre ciation fo r discriminant validityas the cornerstone of a construct validity approach to advancing ourunderstanding of the links between personality traits an d jo b perform anceconstructs.

8/2/2019 1996-costa

3/18

NEO-PI-R AT WORK 227This new attention to differentiated criteria and to more precisespecification of the linkages between personality an d jo b constructs has ledto interest in more detailed personality profiles than the five broadfactors-the Big 5 -o ff e r. Th e issue of the proper level with which tomeasure personality is currently being heatedly discussed in I/O circles.Some advoca tes urge measurem ent a nd assessment at th e very highest levelof th e hierarchy, emphasising extremely broad and su perordinate constructslike Integrity (Ones, Viswesvaran, & Schmidt, 1993; Schmidt, Ones, &Hunter , 1992). Others advocate a lower but still broad level (e.g. thatafforded by the Big 5, Barrick & Mount , 1993) or Houghs (1992) nine-factortaxonomy. We (Costa & McCrae, 1995) join othe rs (Mershon & Gorsuch,1988) in advocating a bottom-up approach. It is important to note thatmeasuring 30 or so specific facets or traits at the lowest level of the traithierarchy allows one to combine the m into five broad er domains, o r even tocombine the five dom ains into bro ad er constructs such as Integrity (O ne s etal., 1993) or Hogans Sales Potential (Hog an & Hogan, 1992) o r Goughs(1984) Managerial Po tential (Costa & McCrae, 1995).Th e point is not to beforced by ones instrument to stay at only o ne level of the hierarchy, but to b eable to go up and down the hierarchy d epe ndin g on t he assessment purposes,goals, and objectives.

For many applied purposes, the detailed information provided fromfirst-order facets o r more specific traits is crucial. As noted in Co sta, McC rae,and Kay (1995, p.124), b ot h intellectual curiosity and aesthe tic sensitivityare asp ects of the dom ain of Op enn ess to Experience, but the fo rm er is abetter predictor of investigative vocational interests, the latter of artisticinterests. This trade-off between bro ad , sup ero rdi na te personalityconstructs and narrower, homogeneous and more specific traits is oftenreferred to as the band w idthh delity dilemma (Cronbach & Gleser, 1965).As Ones and Viswesvaran (in press) write, essentially the debate ... iswhether broadly defined personality traits are better in predicting jobperformance as well as in explaining behaviors, than narrowly definedpersonality traits.Th e NE O-P I-R provides validated measures of both the five factors and30 specific traits (see Tab le 1) that define them, a nd thus should be a usefultool in I/O psychology. Th e NE O-P I-R has been a dop ted by a n um ber ofprofessionals in this field, and several studies have reported supportingevidence. This article reviews basic psychometric properties of theinstrument, summarises some recent studies using the NEO-PI-R inorganisational contexts, and discusses some practical issues in its use inapplied settings.

8/2/2019 1996-costa

4/18

TABLE 1Internal Consistency a nd Factor Structure of NEO-PI-R Scales.CoefficientAlpha FactorNEO-PI-R Scale F o r m s Form R N E O A C

DomainsNeuroticism (N)Extraversion (E)Openness (0)Agreeableness (A)Conscientiousness(C)N1: AnxietyN2: Angry HostilityN3: DepressionN4: Self-ConsciousnessN5: ImpulsivenessN6: Vulnerability

Extraversion facetsEl : WarmthE2: GregariousnessE3: AssertivenessE4: ActivityE5: Excitement-SeekingE6: Positive Emotions

Neuroticism facets

Openness facets01: Fantasy02: Aesthetics03: Feelings04 : Actions05: Ideas06 : ValuesAgreeableness facetsAl: TrustA2: StraightforwardnessA3: AltruismA4: ComplianceA5: ModestyA6: Tender-Mindedness

C1: CompetenceC2:OrderC3: DutifulnessC4: Achievement StrivingCS: Self-DisciplineC6: Deliberation

Conscientiousnessfacets

0.920.890.870.860.90

0.780.750.810.680.700.770.730.720.770.630.650.730.760.760.660.580.800.670.790.710.750.590.670.560.670.660.620.670.750.71

0.930.900.890.950.920.820.860.810.730.690.810.810.790.760.770.740.820.720.810.690.600.870.690.900.840.800.780.830.690.730.710.700.700.820.73

0.81 0.02 -0.01 -0.01 -0.100.63 -0.03 0.01 -0.48 -0.080.80 -0.10 0.02 -0.03 -0.260.73 -0.18 -0.09 0.04 -0.160.49 0.35 0.02 -0.21 -0.320.70 -0.15 -0.09 0.04 -0.38

-0.12 0.66 0.18 0.38 0.13-0.18 0.66 0.04 0.07 -0.03-0.32 0.44 0.23 -0.32 0.320.04 0.54 0.16 -0.27 0.420.00 0.58 0.11 -0.38 -0.06

-0.04 0.74 0.19 0.10 0.100.18 0.18 0.58 -0.14 -0.310.14 0.04 0.73 0.17 0.140.37 0.41 0.50 -0.01 0.12-0.19 0.22 0.57 0.04 -0.04-0.15 -0.01 0.75 -0.09 0.16-0.13 0.08 0.49 -0.07 -0.15

-0.35 0.22 0.15 0.56 0.03-0.03 -0.15 -0.11 0.68 0.24-0.06 0.52 -0.05 0.55 0.27-0.16 -0.08 0.00 0.77 0.010.19 -0.12 -0.18 0.59 -0.080.04 0.27 0.13 0.62 0.00

-0.41 0.17 0.13 0.03 0.640.04 0.06 -0.19 0.01 0.70-0.20 -0.04 0.01 0.29 0.68

-0.09 0.23 0.15 -0.13 0.74-0.33 0.17 -0.08 0.06 0.75-0.23 -0.28 -0.04 0.22 0.57From Costa& McCrae (1992a).N =1539 for Form S alphas, 277 for Form R alphas, and lo00for factor loadings.

228

8/2/2019 1996-costa

5/18

NEO-PI-R AT WORK 229RELIABILITY AN D VALIDITY OF THE NEO-PI-R

The FFM is currently the most widely accepted model of personalitystructure (Go ldberg, 1993), with a history going back at least to Fiskes 1949publication. A n operationalisation of the FFM , the NEO -PI-R is the resultof ov er 15 years of research on volunteer samples. Details on thedevelopment of the instrument ar e given in the manua l (C osta & McCrae,1992a). This section summ arises psychometric prope rties in bo th v oluntee rand 110 samples.ReliabilityThe first two columns of Table 1 report internal consistency for theself-report (Form S) and observer rating (Form R ) version of t h e NE O-PI-R. Th e five factors are approximated by dom ain scales that a re the sum ofsix subscales (called facets). Re liabilities for th e 48-item dom ain scales ar eexcellent; reliabilities for the brief (%item) facet scales are acceptable.Similar values have recently been re po rte d in I/O sam ples (Costa e t al., 1995;Piedmont & Weinstein, 1993). Traits measured by the NEO-PI-R showimpressive long-term stability (Costa & McCrae, 1992b), implying highretest reliability.Content ValidityIn the NEO-PI-R, content validity is addressed by identifying six distinctfacets to sample each domain, and by selecting nonredundant items tomeasure each facet. Because facets were selected by surveying thepersonality literature r ath er than the I/O literature, m any I/O psychologistsmight not immediately grasp the content validity of the facet scales. It iscommonly the case that the trait names of personality questionnairesderived from a literature different fro m the I/O literature d o not easily ordirectly m ap o n to the kind of comp etencie s or attributes in which selectionand assessment tasks are often described. Nevertheless, as Schmit et al.(1994) suggest, many of their 1 9 Jo b equirement dimensions can be relatedto individual differences in personality, as measured by the FFM. Forexample, a particular position may require a dimension th at they identifiedas Trustworthiness with Information or Coope rative H elping Tendencies.The former is related to C and th e latter is related t o A, but if they a re no tidentified and measured they will be ov erlook ed for selection. T he 30 facetsof the N EO -PI -R should cover many traits of interest to I/O psychologists.Convergent and D iscriminant ValidityConvergent validity is seen in the fact that NEO-PI-R facet scales arecorrelated with alternative me asures of similar constru cts (Costa & McCrae,

8/2/2019 1996-costa

6/18

230 COSTA1992b). For example, N1:Anxiety is related to Anxiety as measured bySpielbergers State-Trait Personality Invento ry (r =0.55) and Tension asmeasured by the Profile of Mood States (r=0.54).A1:Trust is relatedpositively to th e Trusting scale of the Interpersonal Style Inventory (r=0.68)and negatively to th e Suspicion scale of the Buss-Durkee HostilityInventory (r= -0.46).Discriminant validity is seen by contrasting the correlates of differentfacets, particularly within th e sa me dom ain. Consider the E facet correlatesof scales in Jacksons Personality Research Form (PRF): E1:Warmth andE2:Gregariousness a re strongly related to P R F Affiliation; E3:Assertivenessis related chiefly to PRF Dominance; E5:Excitement-Seeking is related toP R F Ha rm A voidance; and E6:Positive Em otio ns is related t o PRF Play.Consensual ValidationO ne of the attractive features of the NEO-PI-R is the availability of bothself-report and observer rating forms. There is strong evidence ofconvergence on domain scores between self-reports an d spouse ratings, andbetween self-reports and mean peer ratings, with correlations typicallyranging from 0.5 to 0.6 (Costa & M cCrae, 1992a). Significant an d substan tialcross-observer agreement is also seen on t he 30 facets of the NEO-PI-R.Spouses tend to agree w ith self-reports more strongly than do single peers ,perhaps because spouses disclose mo re of themselves to each oth er than tofriends and neighbours.Factorial ValidityThe NE O- PI-R is intended to represent the FFM, so a key test of its validityis how well its internal structu re co rrespon ds t o th e predictions of the model.As shown in the last five columns of Table 1, a factor analysis of theNEO-PI-R facets in the normative sample of 500 men and 500 womenconfirmed hypotheses. Each facet scale had its highest loading on theintended factor, and where secondary loadings appeared, they wereappropriate and m eaningful. Th e sa m e five-factor structure has been foundin self-reports and peer ratings, and in men and women, whites andnonwhites, and young adults and older adults (Costa & McCrae, 1992a).However, all the research cited was conducted o n volunteer samples whohave relatively little reason to distort their responses. The NEO-PI-R willfrequently be used in selection and placement contexts, in whichrespondents might wish to present themselves in an especially favourablelight. Would such a bias affect the factor structure? Some evidence insuppo rt of tha t view was provided by Schm it an d Ryan (1993, p.969), whoexamined the sh ort form of the NE O-PI-R , the N EO Five Factor Inventory(N EO -FF I), in a sample of 293 jo b applicants seeking emp loyment

8/2/2019 1996-costa

7/18

NEO-PI-R AT WORK 231assistance at a midwestern branch of the United States EmploymentService. The original 12 items for each of the five domain scores wererandomly divided into three 4-item subsets. T he 15 item subsets (3 subsetsper factor x 5 factors) were factor analysed. They reported that the re was asixth or general factor they interpreted as reflecting evaluation concerns.However, they did not report a 5-factor solution. Subsequently, MarkSchmit kindly provided the results of a five-factor solution shown in Ta ble 2.As T able 2 shows, the A greeableness factor was not very clearly defined, butthe othe r four factors were easily recognised. This analysis suggests that anyeffects of evaluation bias on the structure of the NEO-FFI are relativelymodest.Two oth er studies have studied the factor structure of th e full NEO-PI-Rin applicant samples. Montag and Levin (1994) examined a Hebrewtranslation of the NEO-PI-R and extracted five varimax factors in a largefemale applicant sample (n= 539) and a smaller male applicant sample(n= 396). Th e congruence coefficients of the varim ax-rotated factors for thefemales, with the Am erican-normative structure for N, E, 0,A, and C were0.97, 0.92, 0.96, 0.95, 0.92, respectively. For males, the congruencecoefficients with the American-normative structure were 0.94, 0.89, 0.87,0.90, and 0.92 for N, E , 0, A, and C, respectively. Montag and Levin

TABLE 2NEO Five-factor Inventon/ (NEO-FFI) Item Parcel Factors in an ApplicantSampleNEO-FFI Varim ax-Roiaied Max imum Likelihoo d FactorItem Parcel N E 0 A CN (1)N (ii)N (iii)E ( 9E (ii)E (iii)0 (i)0 (ii)0 (iii)A (1)A ( i i )A (iii)c (i)C (ii)C (iii)

0.720.800.76

0.43

0.490.670.560.590.530.62

0.85(0.34)0.450.54 0.440.46 0.520.75

N =293. All loadings greate r than 0.40 in absolute magnitude are reported.Source: Personal communication, M . Schmit, 6 July 1994.

8/2/2019 1996-costa

8/18

232 COSTAconcluded (1994, p.8) that their study exte nds th e factorial validity of theNEO-PI-R to non-volunteer subjects in selection settings, where thesituation evokes a strong m otivation to obtain positive results, which mightslant the response pattern. Similar results were obtained by Costa et al.(1995) in a po lice selection sample.

ILLUSTRATIVE I/O APPLl CAT10NSPersonality and Vocational InterestsIn 1984 Co sta, McC rae, and Holland published an analysis of personalityand vocational interests using an early version of the NEO Personali tyInventory which measured only Neuroticism, Extraversion, an d O penness.The measure of vocational interests was Hollands Self-directed Search(Holland, 1985), which is based o n a hexag onal m ode l for defining thepsychological resemblances among types and environments and theirinteractions.N was largely unrelated to interests as measured by the Self-directedSearch, but E was strongly related to interest in Social and Enterprisingvocations, and 0 was positively related to A rtistic a nd Investigative interestsand negatively related to C onven tional interests. W hen th e sam e da ta werereanalysed using measures of Agreeableness and Conscientiousnessobtained later on the sam e adult sample, A w as found t o be related to Social(but not Enterprising) interests. C, like N, was no t related t o occupationalinterests.T he associations between 0 and vocational interests was replicated byHolland, Johnston , Hughey, an d Asam a (1991), who also showed tha t 0 wasassociated with a m easure of creativity, the Precon scious Activity Scale, inboth men ( r=0.50,P < 0.01) an d w omen ( r=0.61, P< 0.001). This suggeststhat individuals who score high in 0 will be m or e successful in occupationsthat req uire original thinking.The partial overlap between measures of personality and measures ofvocational interest suggest that these two kinds of instruments may servecomplementary functions in career assessment and counselling. A fullerdiscussion of the use of the N EO -PI -R in care er assessment is provided byCosta et al. (1995).Validity in Vocational SettingsW hen em ployees or jo b applicants complete a personality questionnaire,there is always som e possibility that they will distort their responses to makethemselves look good. If all respondents shifted equally in a favourabledirection, there would be n o net effect o n th e validity of the scores (althoughnew, applicant norms might be needed). But if some respondents report

8/2/2019 1996-costa

9/18

NEO-PI-R AT WORK 233honestly whereas other s grossly falsify their answers, th e validity of the testwould suffer.There is reason to believe that concerns about the fakeability ofpersonality tests have been exaggerated. Simulation studies, in whichsubjects ar e asked to fake good o r fake bad give some information o nhow responses might be distorted, but no information whatsoever onwhethe r or how frequently they actually ar e distorted in real life situations.T he few studies that have exam ined this question find little evidence that jo bapplicants seriously misrepresent themselves on personality inventories(Hou gh e t al., 19 90 Michaelis & Eysenc k, 1971). Th us, there is reason tobelieve that personality scores will be valid even in selection settings. Twostudies have examined t he validity of N EO -PI -R self-reports in I/ O settings.Piedmont and W einstein (1993) administered the N EO -PI-R t o a sampleof 52 women and 159 m en employed in a wide range of occupations includingsales, custom er service, man agem ent positions (lower, middle, an d u pper),and finance. They also asked supervisors to com plete th e Adjective C heckList (ACL; Gough & Heilbrun, 1983) to describe the employeespersonality. Correlations between A and C facet scales and supervisorratings on the ACL showed evidence of cross-method validity. Overallself-reported A was positively associated with rated Self-Control,Abasement, and Deference, and negatively correlated with Aggression,Auton om y, and Creative Personality; employ ees who described themselvesas high in A w ere thus seen by their supervisors as being m ild-mannered an dcompliant, though perhaps lacking independence and originality. Overallself-reported C was positively associated with supervisor-rated needs forAchievement, Endurance, and O rde r, and negatively related to rated needfor Succorance.Although these data from Piedmont and Weinstein do not directlyaddress the susceptibility or lack of susceptibility to faking, theynevertheless show that th e consensual validation of employees self-reportsand supervisors ratings cannot be d ue to a share d metho d variance, nor canwhatever degree of faking in the NEO-PI-R self-reports influence ormanipulate the supervisors ratings. If people who describe themselves asbeing cooperative and agreeab le were merely faking, then the re ou gh t not tobe agreement with supervisors ratings. If the self-reported NEO-PI-Rscores in the Piedmont and W einstein study were fak ed , then su pervisorsshould not have rated high Agreeableness scorers as mild-mannered andcompliant. Inde ed , if they w ere faking, they should be rated a s having theopposite characteristics. However, we must acknowledge that these datacome fro m incumbents who presumably have less motivation o r pressure tofake their responses.A t th e level of individual adjectives, supe rvisor s described em ploy ees asbeing depen dable, logical, eficient, mature, clear-thinking, alert, mannerly,

8/2/2019 1996-costa

10/18

234 COSTAresourceful, andprecise-traits that perhaps explain why Barrick and Mount(1991,1993) found C to be a predictor of superior performance across a widerange of jobs.

Costa et al. (1995) reported a study of police selection. Entry-level policeofficer candidates completed the NEO-PI-R and were also interviewed bytrained psychologists (blind to NEO-PI-R scores) who subsequently ratedthem as Highly Recommended or Recommended (n=188) orRecommended with Reservations or Not Recommended (n=31). Acomparison of NEO-PI-R scores for these two groups showed significantdifferences on 23 of the 30 facet scales. In particular, candidates who wererecommended scored higher on all six NEO-PI-R Conscientiousness facets,and lower on all six Neuroticism facets; they also reported themselves to behigher on E1:Warmth and A3:Altruism.

Personality and Job PerformanceCosta (1992) reported correlations between supervisor ratings of jobperformance and NEO-PI-R scores in a national sample of over 1500 menand women (see Gandy, Dye, & MacLane, 1994). The subjects were ingeneral successful, long-term employees. A single supervisor providedratings, so interrater reliability could not be estimated. Both the restrictedrange of actual job performance and the probably limited reliability of thesingle supervisor ratings tended to restrict the magnitude of the significantcorrelations, which ranged from 0.05 to 0.12 (P

8/2/2019 1996-costa

11/18

NEO-PI-R AT WORK 235Th e strongest patte rn of correlations was with Conscientiousness, whichwas related to the amount, quality, and accuracy of work, and to overalljudgements of competence. It was also related to supervisors appraisal ofabilities in the areas of oral expression and written comprehension,reasoning and ability to recall job-related information, conformance withrules, ability to overcome obstacles, and ability to adapt to new workdemands. Further, the association of C with superior performance wassignificant in both men and women, and was significant in the full sampleeven after con trolling for age, sex, and years of education. Five of th e six Cfacets-Com petence, Order , Dutifulness, Achievement Striving, and Self-discipline-were related to superior perform ance ratings.In a later analysis of their 1993 da ta, Piedm ont and W einstein (1994)examined NEO-PI-R scores as predictors of job performance ratings.Respondents supervisors were asked to rate the employees on a 5-pointscale ranging from unsatisfactory to excellent on 12 relevant perform anceitems that constituted three performance scales, labelled InterpersonalRelations (communicates ideas clearly, relates well to supervisors, teamplayer and service minded), Task Orientation (self-starter, hard-working,detail skills, and gets things done ), and Ada ptive Capaciry (learns and adaptsreadily, copes effectively with setbacks, functions well in unstructuredsituations, and plans, coordinates, and follows up o n the work of others).Alphas were good.Piedmont and Weinstein (1994) found a number of significantcorrelations. In this group of sales and service employees, E scores weresignificantly correlated with success ratings on the Interpersonal Relations( r=O.20) , Task Orientation (r=0.16), and Adaptive Capacity (r=0.19)scales. N emerged as a significant predictor of ratings concerningInterpersonal Relations ( r = -0.16) and Adaptive Capacity (r=-0.17). Thestrongest predictor ofj ob performance was C and its facets (n 0.15 to0.28).The authors noted that competence, achievement striving, and self-discipline appea r to be qualities that underlie successful perform ance in allthe areas rated.McDaniel (1992) conducted a study of personality and organisationalchange. H e adm inistered the N EO -PI t o 62 individuals identified as changeleaders in a large electronics firm. Th e success of the change e ffort was ratedby the change leader and by between one and three other knowledgeableraters. As hypothesised, the most effective change leade rs were those whowere highest in Openness. Individuals who were rated as being flexible and

innovative, and preferred novelty and variety, were the most able tocoord inate structural changes in a large organisation.Finally, Salgado and Rumbo (1994), using a Castilian version of theNEO -FF I, found that Conscientiousness predicted ratings of job asp iration,attitude, and performance among financial service managers in

8/2/2019 1996-costa

12/18

236 COSTASpain. Such data suggest that the same traits may be related to occupationalsuccess in many different (Western or Westernised) cultures.

INDUSTRIA40RGANISATIONAL APPLICATIONSOF THE NEO-PI-RStrategies for Selection and PlacementGiven that the NEO-PI-R may provide valid assessments of personalityeven in job selection conditions, how should the instrument be used? Themost frequent application is likely to be in placement and selection, that is,infinding the optimal match between the person and position. Both empiricaland rational strategies might be used.

A purely empirical strategy, useful especially when many employees areto be placed on an ongoing basis, would be to administer the NEO-PI-R to agroup of candidates and subsequently evaluate the performance of thosewho were hired. Ratings of job performance could then be predicted fromNEO-PI-R scores using multiple regression or a similar technique; theprediction equation (ideally cross-validated on a second group) could thenbe applied to the NEO-PI-R scores of new candidates to estimate their likelyfuture performance. (Other indicators, such as cognitive ability tests andbiographical data, might also be included in the prediction equation.) TheNEO-PI-R is ideally suited for this purely empirical approach because thecomprehensiveness of the FFM ensures that any personality trait relevant tojob performance is likely to be measured by one or another of the 30 facetscales.

Rational strategies require that the I/O psychologist determine theoptimal personality profile for a particular position, and seek candidateswho have such a profile. This strategy is perhaps most appropriate when toofew candidates are to be screened to allow empirical approaches. Based onmany studies reviewed earlier, it would be reasonable to look for high scoreson Conscientiousness in candidates for almost any job. Beyond that,however, the particular personality traits relevant to a particular position arelikely to vary with the nature of the job, and a conceptual job analysis isneeded. Extraversion might be desirable in a flight attendant; introversionmight be more desirable in a night security guard.

Beyond the simple identification of traits, more sophisticated profileinterpretations are possible and indeed often necessary to fully exploit thepersonality information derived from most assessments. The profile, orpatterning, of the various traits condition and contextualise the meaning of aparticular trait score when considered in isolation. For example, a T-score of58 on the Achievement Striving facet of the NEO-PI-R may lead to greaterexpectations of industriousness and drive to succeed when accompanied byequally high activity level (E4) and high self-discipline (C5). Conversely, the

8/2/2019 1996-costa

13/18

NEO-PI-R AT WORK 237same A chievement Striving score seen in context of high anxiety ( N l) andvulnerability to stress (N6), low warm th ( E l ) and high exciteme nt-seeking(E5), low trust (Al) and modesty (A5) would lead to quite differentpredictions about success and/or th e expression of tha t achieveme nt strivingdisposition. To take a no the r example, a person m ay have generally low oraverage scores on N euroticism facets except for high N2 o r Ang ry Hostilityscore, indicating that such a person is impatient and quick to becomefrustrated and irritated especially when things are not to their liking.However, if they have low assertiveness (E4) or high Tendermindedness(A 6) scores they ar e unlikely to express their frustration directly. Obviouslythere are any number of such impo rtant interactions among facets or traitsthat could be taken in to consideration in m aking specific predictions ab ou tjob behaviour and performance.The selection of desirable traits might be based on a job analysis or areview of th e published literature. O ne alterna tive is to consult local experts:supervisors, successful jo b incumbents, perhaps customers. T h e N E O J o bProfiler (available from the au tho r) was designed t o elicit this informationfrom lay judges. For example, judges are asked if it is very undesirable,undesirable, desirable, o r very desirable for a candidate for th e position to bedescribed as energetic, lively, high activity level; may find sed entary workunappealing (Costa e t al., 1995). If th at trait is desirable, th en ind ividualsscoring high on NEO-PI-R E4:Activity should be favoured in selectionprocedures. T he N E O Jo b Profiler contains descriptions of each of the 30facets, with an emphasis on their work-related implications; in a study ofpolice selection, high inter-judge agreement was found and job profilesmatched traits found in candidates whom interviewers recomm ended (Costaet al., 1995). Schmit et al. (1994) have offered an altern ative instrum ent forpersonality-related j ob analysis that is also keyed to N EO -P I-R facet scales.It is, of course, possible to combine ration al and empirical app roach es, byconducting empirical validation studies on the su bset of NEO-PI-R facetscales hypothesised to b e relevant to jo b perform ance prediction.Using the Revised NEO-PI for FFM MeasurementThe distinctive fact that accounts for the success of the FFM is that itaccounts for dimensions of personality found in almost every majorinstrument. And one consequence of this is that research using anyparticular instrument can be generalised to other instruments. This isexceptionally good news for I/O psychologists who have a diverse litera tureon personality tests that can be integrated by the FFM: it allowspsychologists to optimise th e use of whatever instrument they choose. Forexample, th e J-P scale of the Myers-Briggs Type In dicator (M BT I; Myers &McCaulley, 1985) is known to be related to measures of C, which are

8/2/2019 1996-costa

14/18

238 COSTApredictors of vocational success in many occupations. Consequently, thisMBTI dimension might be a useful predictor of vocational success or jobperformance. Consider another example. Jacksons (1984) PRF, one of thebest measures of Murrays system of needs, has two scales, Achievement andOrder, which measure the C factor. In addition, other PRF scales mark eachof the remaining Big 5 dimensions (Costa & McCrae, 1988).

The traits of the FFM are ubiquitous, so many different personalityquestionnaires might be used effectively in I/O applications. Although theNEO-PI-R is my preferred instrument, other FFM-based personalitymeasures that might be considered include the Hogan Personality Inventory(Hogan & Hogan, 1992) and Goldbergs adjective markers (Goldberg,1992). There are, however, some strengths of the NEO-PI-R thatrecommend its use, beyond the fact that it is a well-validatedoperationalisation of a comprehensive model. It is relatively brief,acceptable to job applicants, available in both self-report and observerrating versions, and increasingly available in translation.

The full NEO-PI-R has 240 items and requires about 30 minutes tocomplete; it can of course be administered to groups. Where time is animportant factor, the 60-item NEO-FFI might be used to gain globalinformation on the five factors. Alternatively, professionals might wish tolicense use of a subset of relevant facet scales from the publisher. TheNEO-PI-R provides a large yield of information for a relatively smallinvestment of time.I/O psychologists are properly concerned about how job applicants willreact to psychological tests; questionnaires that seem to invade therespondents privacy are likely to interfere with a productive rapportbetween applicant and evaluator. A recent study by Rosse, Miller, andStecher (1994) showed that most applicants for seasonal jobs in a propertymanagement firm had generally favourable reactions to completing ashortened form of the NEO-PI-R as part of the selection process, a resultthat the authors attributed to the fact that the instrument was designed tobe as nonoffensive as possible (1994, p.990).

Although data suggest that in most cases self-reports of personalityobtained in I/O settings are valid, there may certainly be some instances inwhich that conclusion may be questioned. The use of the observer ratingform of the NEO-PI-R provides an alternative in such cases. When verysensitive positions are being filled, background checks with informants areoften utilised; Form R of the NEO-PI-R might be a useful part of thisprocess.Finally, in the context of an increasingly global world economy, it isreassuring to know that research on personality and vocational behaviour isnot entirely culture-bound. The structure of personality itself seems to beuniversal, as studies using translations of the NEO-PI-R have suggested (e.g.

8/2/2019 1996-costa

15/18

NEO-PI-R AT WORK 239McCrae, Costa, & Yik, 1996;Montag & Levin, 1994),and th e sam e qualitiesthat predict successful job perform ance in America (B am ck & Mount , 1991)predict job performance in Spain (Salgado & Rumbo, 1994). W ith minormodifications (vacation becomes holiday) the NEO-PI-R has beenadapted for use in the United Kingdom by The Test Agency. A Germanedition of the NEO-FFI by Borkenau and Ostendorf (1993) has beenpublished, and translations of the NEO -PI-R in to a dozen oth er languagesar e in progress. These translations pro vide new tools for unde rstanding therole of personality in t he world of work, and the w ork of the w orld.

Manuscript received March 1995Revised manuscript received No vemb er 1995REFERENCES

Adler, A. (1964). Social interest: A challenge to mankind. New York: Capricorn Books.(Original work published 1938).Adler, S., & Weiss, H.M. (1988). Recent developments in the study of personality andorganizational behavior. In C.L. Coo per & I.T. Robertson (Eds.), International review ofindustrial and organisationalpsychology. Chichester, U K: John Wiley.Barrick, M.R., & Mount, M.K. (1991). The Big Five personality dimensions and jobperformance: A meta-analysis. Personnel Psychology, 44, 1-26.Barrick, M.R., & Mount, M.K. (1993). Autonomy as a moderator of the relationshipsbetween the Big Five personality dimensions and job performance. Journal of Applied

Borkenau. P.,& Ostendorf, F. (1993). NEO-Fiinf-Faktoren nvenrar (NEO-F FI) nach Cosraund McCrae: Handanweirung [N EO Five-factor Inventory (NEO -FFI) of Costa andMcCrae: Manual]. Gottingen: Hognefe.Costa, P.T., Jr. (1992, June) . The Big 5: Personality and wor k. Paper presented at the SixthEuropean Conference on Personality, G roning en, Th e Netherlands.Costa, P.T., Jr., & McCrae, R.R. (1988). From ca talog to classification: Murrays needs andthe five-factor model. Journal of Personality and Social Psychology , 55,25&265.Costa, P.T., Jr., & McCrae, R.R. (1992a). Revised NEO Personaliry Inventory (NEO-PI-R)and N E O Five-Factor Inventory (N EO -F FI ) professional manual. Odessa, FL:Psychological Assessment Resources.Costa, P.T., Jr., & McCrae, R.R. (1992b). Trait psychology comes of age. In T.B.Sonderegger (Ed.), Nebraska Symposium on Morivarion: Psychology and aging (pp.169-204). Lincoln, NE: University of Nebraska Press.Costa, P.T., Jr., & McCrae, R.R. (1995). Domains and facets: Hierarchical personalityassessment using the Revised NEO Personality Inventory. Journal of PersonalityAssessment, 64.21-50.Costa, P.T., Jr., M cCrae, R.R., & Holland, J.L. (1984). Personality and vocational interests inan adult sample. Journal of Applied Psychology, 69,390-400.Costa, P.T., Jr., McCrae, R .R., & Kay, G.G. (1995). Persons, places, and personality: Ca ree rassessment using the Revised N EO Personality Inventory. Journal of Career Assessment, 3,123-139.Cronbach, L.J., Gleser, C.C. (1965). Psychological tests andperson nel decisions (2nd edn.).Urb ana, IL: University of Illinois Press.Fiske, D.W. (1949). Consistency of factorial stru cture s of personality ratings fro m differe ntsources. Journal of Abnormal and Social Psychology, 44,329-344.

Psychology, 78, 111-1 18.

8/2/2019 1996-costa

16/18

240 COSTACandy, J.A., Dye, D.A., & MacLane, C.N. (1994). Federal government selection: Theindividual achievement record. InG.S. Stokes, M.D. Mumford, &W .Ow ens (Eds.), Biodatahandbook: Theory, research, and use of biographical information in selection and

performance prediction (pp.275-309). Palo Alto, CA: Consulting Psychologists Press.Goldberg, L.R. (1992). The development of markers for the Big-Five factor structure.Psychological Assessment,4,2&42.Goldberg, L.R. (1993). The structure of phenotypic personality traits. AmericanPsychologist, 48,2634.Cough, H.G. (1984). A Managerial Potential scale for the California PsychologicalInventory. Journal of Applied Psychology, 69,233-240.Cough, H.G., & Heilbrun, A.B., Jr. (1983). Adjective Check List manual. Palo Alto, CA:Consulting Psychologists Press.Guion, R.M.,& Gottier, R.F. (1965). Validity of personality measure s in personnel se lection.Personnel Psychology, 18,135-164.

Hoekstra, H. (1993). Work and personality. Paper presented at the 6th meeting of theInternational Society for the Study of Individual Differences, Baltimore, Maryland.Hogan, R., & Hogan, J. (1992). Hogan Personality Inventory manual. Tulsa, OK: HoganAssessment Systems.Holland, J.L. (1985). Self-directed search-I985 edition. Odessa, F L PsychologicalAssessment Resources.Holland, J.L., Johnston, J.A ., H ughey, K.F., & Asama, N.F. (1991). Som e explorations of atheory of careers: VII. A replication and some possible extensions. Journal of CareerDevelopment, 18,91-100.Hough, L.M. (1992). T he "Big Five" personality va riab les- con stru ct confusion: Descriptionversus prediction. Human Performance, 5,139-155.Hough, L.M., Eaton, N.K., Dunnette, M.D., Kamp, J.D., & McCloy, R.A. (1990).Criterion-related validities of personality constructs an d the effect of response distortion onthose validities. Journal of Applied Psychology, 75,581-595.Hough, L.M.. & Paullin, C. (1994). Construct-oriented scale construction: The rationalapproach. In G.S. Stokes, M.D. Mumford, & W.A. Owens (Eds.), The biodata han dbook:Theory, research, and application. Palo Alto, CA: Consulting Psychologists Press.Jackson, D.N. (1984). Personality Research Form manual (3rd edn.). Port Huron, MI:Research Psychologists Press.Jackson, D.N., & Rothstein, M . (1991). Personality put to the test. Nature, 35, 100.Mat thew, G., & Stanton, N. (1994). Item and scale factor analyses of the Occupational

Personality Questionnaire. Personality and Individual Differences, 16,733-743.McCrae, R.R.. Costa, P.T., Jr.. & Yik, M.S.M. (1996). Universal aspects of Chinesepersonality structure . In M.H. Bond (Ed.), Handbook of Chinesepsychology (pp. 189-207).Hong Kong: Oxford University Press.McDaniel, R.N. (1992). The relationship between personality and perceived success oforganizational change. Unpublished doctoral dissertation, The Fielding Institute.Mershon, B., & Gorsuch, R.L. (1988). Number of factors in the personality sphere: Doesincrease in factors increase predictability of real-life criteria? Journal of Personality andSocial Psychology, 55,675680.Michaelis, W., & Eysenck, H.K. (1971). The determination of personality inventory factorpatterns and intercorrelations by changes in real-life motivation. Journal of GeneticMontag, I., & Levin, J. (1994). The five-factor personality model in applied settings.Myers, I.B., & McCaulley, M.H. (1985). Manual: A guide to the developm ent and u se of the

Psychology, I18.223-234.European Journal of Personality, 8 , 1-1 1 .Myers-Briggs Type Indicator. Palo Alto , CA : Consulting Psychologists Press.

8/2/2019 1996-costa

17/18

NEO-PI-R AT WORK 241Norman, W.T. (1963). Toward an adequate taxonomy of personality attributes: Replicatedfactor structure in peer nomination personality ratings. Journal of Abnormal and SocialPsychology, 66,574-583.Ones, D.S., & Viswesvaran, C. (in press). Bandwidth-fidelity dilemma in personalitymeasurement for personnel selection. Journal of Organizational Behavior.Ones, D.S., Viswesvaran, C., & Schmidt, F.L. (1993). Comprehensive meta-analysis ofintegrity test validities: Findings and implications for personnel selection and theo ries of jobperformance [Monograph]. Journal ofApplied Psychology, 78,679-703.Piedm ont, R.L., & Weinstein, H.P. (1993). A psychometric evaluation of the new NEO-P I-Rfacet scales for Agreeableness and Conscientiousness. Journal of Personality Assessment,Piedmont, R.L., & Weinstein, H.P. (1994). Predicting supervisor ratings of job performanceusing the N EO Personality Inventory. Journal of Psychology, 128,255-265.Robertson, I.T. (1993). Personality assessment and personnel selection. European Review of

Applied Psychology, 43,187-194.Robertson, I.T., & Kinder, A. (1993). Personality and job competences: The criterion-related validity of som e personality variables. Journal ofOccupational and OrganizationalRosse, J.G., Miller, J.L., & Stecher, M.D. (1994). A field study of job applicants' reactions topersonality and cognitive ability testing. Journal of Applied Psychology, 79,987-992.Salgado, J.F., Rumbo, A. (1994). Personality and job perform ance in financial servicesmanagers. Unpublished manuscript, University of Santiago de Compostela.Saville, P., Holdsworth, R., Nyfield, G., Cramp, L., & Mabey. W. (1984). OccupationalPersonality Questionnaires manual. Esher, UK: Saville & Holdsworth Ltd.Schmit, M.J., Guion, R.M., & Raymark, P.H. (1994). Development of an instrument to

identify personality-related position requirements. Unpublished manuscript, University ofFlorida.Schmidt, F.L., Ones, D.S., & Hunter, J.E. 1992). Personnel selection. Annual Review ofPsychology, 43 ,627670 .Schmit, M.J.. & Ryan, A.M. (1993). Th e big five in personnel selection: Factor structure inapplicant and nonapplicant populations. Journal of Applied Psychology, 78,966-974.Schneider, R.J., & Hough, L.M. (in press). Personality and industriaUorganizationa1psychology. In C.L. Coo per & I.T. Robertson (E ds.), International review of ndustrial andorganizational psychology. Chichester, UK: Wiley.Tenopyr, M. (1993). Construct validation needs in vocational behavior theories. SpecialIssue: The theory of work adjustment. Journal of Vocational Behavior , 43, 84-89.Zuckerman. M. (1979). Sensation seeking: Beyond!h e optimal level ofarousal. Hillsdale, NJ:Lawrence Erlbaum Associates Inc.

60,302-318.

Psychology, 66,225-244.

8/2/2019 1996-costa

18/18