a quantitative review of the relations between the “big 3 ... · tellegen’s test construction...

Journal of Research in Personality 40 (2006) 250–284

www.elsevier.com/locate/jrp

A quantitative review of the relations between the “Big 3” higher order personality dimensions

and antisocial behavior

Ellison M. Cale

South Carolina Department of Juvenile Justice, 1711 Shivers Road, Columbia, SC 29210, USA

Available online 17 March 2005

Abstract

In this meta-analytic review, Hans J. Eysenck’s theory of criminality (Eysenck, 1964, 1977)serves as a theoretical framework for examining the relations between higher order personalitydimensions and antisocial behavior (ASB). The three higher order dimensions examined arereferred to as extraversion/sociability, neuroticism/emotionality, and impulsivity/disinhibition(see Sher & Trull, 1994), and they are likened to Eysenck’s dimensions of extraversion (E), neu-roticism (N), and psychoticism (P), respectively (Eysenck & Eysenck, 1975) and Tellegen’sdimensions of positive emotionality (PEM), negative emotionality (NEM), and constraint(reversed) (CON), respectively (Tellegen, 1982). Ninety-seven samples, from 52 published andunpublished studies, were reviewed. The results indicated that among the “Big 3” personalitydimensions, impulsivity/disinhibition is most strongly related to ASB and extraversion/socia-bility is least strongly related to ASB. Additional variables, including age and methodologicaldiVerences, were found to moderate the associations between the personality dimensions andASB.© 2005 Elsevier Inc. All rights reserved.

Keywords: Eysenck; Personality theory; Personality dimensions; “Big 3”; Antisocial behavior, Meta-analysis

E-mail address: [email protected].

0092-6566/$ - see front matter © 2005 Elsevier Inc. All rights reserved.doi:10.1016/j.jrp.2005.01.001

mailto: [email protected]

mailto: [email protected]

E.M. Cale / Journal of Research in Personality 40 (2006) 250–284 251

1. Introduction

Understanding individuals who engage in antisocial behavior (ASB) has long beena challenge to researchers in the social sciences (Wilson & Herrnstein, 1985). ASBcomprises a number of socially deviant behaviors, typically investigated in terms ofcriminality, delinquency, and other relevant clinical syndromes [e.g., antisocial per-sonality disorder, conduct disorder; American Psychiatric Association (APA), 1994,2000]. Because personality features are relatively stable and predict overt behaviors,personality’s role in predicting ASB has been extensively researched (Caspi et al.,1994; Krueger et al., 1994).

Many researchers have used personality models to generate hypotheses about thecauses of ASB, and Hans J. Eysenck’s theory has been one of the most inXuential inthis body of literature (Furnham, 1984; Raine, 1997b). Although Eysenck’s “Big 3”dimensions have been the most widely researched, other personality constructs maponto his dimensions to a substantial extent and are pertinent to the study of ASB. Abetter understanding of the degree to which these dimensions are associated withASB and of variables (e.g., biological and social variables) that moderate these asso-ciations may open doors for better predicting and treating ASB.

1.1. Eysenck’s personality theory of crime

Positing that individuals inherit predispositions to behave in certain ways underspeciWc environmental conditions, Eysenck’s theory has been referred to as a bioso-cial model of behavior (Lane, 1987; Raine, 1997b). Though acknowledging that indi-vidual diVerences are shaped partly by environmental factors, Eysenck argued thatgenetic factors largely account for biological diVerences that inXuence personality(Eysenck, 1977, 1996b). He also asserted that personality features are manifested andmeasurable via behaviors in laboratory and social settings (Eysenck, 1964, 1996b;Eysenck & Eysenck, 1970). Eysenck used factor analytic techniques to developinstruments for assessing personality dimensions [e.g., the Eysenck Personality Ques-tionnaire (EPQ) Eysenck and Eysenck, 1975] and to advance his theory (Eysenck,1972; Eysenck & Eysenck, 1976). Early in his career, he asserted that the two dimen-sions of extraversion and neuroticism underlie normal personality functioning.

Eysenck’s conceptualization of extraversion (E) was derived from research onbrain functioning and Pavlov’s learning theory (Eysenck, 1964). He reasoned thathigh E individuals have relatively low levels of cortical arousal and seek excitementto increase arousal levels (Eysenck, 1977, 1996a). He further argued that high E indi-viduals are underaroused, require high levels of stimulation to learn, and therefore,are less conditionable than low E individuals (Eysenck, 1964; Eysenck & Eysenck,1970). He postulated that the conditioning of a conscience is integral to refrainingfrom ASB and that criminals do not develop consciences capable of doing so(Eysenck, 1996a).

Predictions concerning neuroticism (N) extended from Hull’s learning theory andEysenck’s hypotheses about emotional reactivity (Eysenck, 1964, 1972). Eysenck pre-sumed that N was related to autonomic nervous system functioning, involving

252 E.M. Cale / Journal of Research in Personality 40 (2006) 250–284

primarily the hypothalamus and limbic system, and that high N individuals haveheightened emotional drives. Throughout his career, he proposed that criminals haverelatively high levels of E and N, signifying that they have central nervous systemsthat condition poorly and autonomic nervous systems that overreact and augmentfailures in refraining from ASB (Eysenck, 1996a). His hypotheses further suggest aninteraction between E and N that accounts for especially high levels of ASB. Eysenckhimself stated that a “reason for N acting as a predictor is probably related to itsdrive properties, which multiply the action tendencies present,” (Eysenck, 1996a, pp.149–150). Despite some ambiguity regarding N’s relation with E, Eysenck also con-tended that high levels of N directly predict criminal oVending (Eysenck, 1996b).

“Subtraits” were theoretically predicted to each lie hierarchically below one higherorder personality dimension and be distinct from the remaining dimension(s). Factoranalyses of EPQ item responses have conWrmed this hierarchy. The subtraits that fallunder the E dimension are sociable, lively, active, assertive, sensation seeking, care-free, dominant, surgent, and venturesome, and the subtraits that fall under the Ndimension are anxious, depressed, guilt feeling, low self-esteem, tense, irrational, shy,moody, and emotional (Eysenck, 1996a).

Years after developing E and N, Eysenck and colleagues developed conjecturesregarding psychoticism (P). Considering evidence that family members of psychoticindividuals tend to exhibit high levels of ASB, Eysenck and Eysenck (1976) con-structed self-report personality items to distinguish psychotics and criminals fromnormals. Eysenck and colleagues regarded P as a “polygenetic” dimension, indicatingthat it contains numerous genetics inXuences, each of which has a distinct predisposi-tion and bears additive eVects on personality (Eysenck, 1996b; Zuckerman, 1997).Based on this, Eysenck and Eysenck (1976) posited a threshold eVect for P, wherebypast a certain threshold of environmental stresses, one exhibits psychosis and justbelow the threshold, one exhibits ASB. The P dimension has not, however, proven tobe an adequate measure of psychosis because inmates and individuals with psychiat-ric symptoms other than psychosis score higher on this dimension than psychoticindividuals (Farrington, Biron, & LeBlanc, 1982; Zuckerman, 1997).

Eysenck (1977) also asserted that high P individuals inherit nervous systems pre-disposing a “tough-minded” nature, and the P subtraits include aggressive, cold, ego-centric, impersonal, impulsive, antisocial, unempathic, creative, and tough-minded(Eysenck, 1996a). In later years, he stated that P and E are both characterized by lowcortical arousal (Eysenck & Gudjonsson, 1989). Eysenck acknowledged that Eincludes both sociability and impulsivity features and that impulsivity is linked morewith conditioning and a proclivity towards engaging in ASB than is sociability.Results of eye-blink conditioning studies (e.g., Eysenck & Levey, 1972) and item-analyses of the E scale (e.g., Eysenck & Eysenck, 1971b) supported this distinction.Although Eysenck and colleagues excluded impulsivity items from E and incorpo-rated them in the P dimension, Eysenck maintained that high E individuals are pre-disposed to engage in ASB due to sensation seeking tendencies (Eysenck, 1996a).Eysenck ended his career defending the same predictions concerning E and N as hehad in 1964, but he added that high P individuals were especially prone to be antiso-cial (Eysenck, 1996b).


1.2. Other personality models used to predict ASB

Using an exploratory approach involving rationally, factor-analytically derivedscales, Tellegen developed the Multidimensional Personality Questionnaire (MPQ;Tellegen, 1982). The MPQ assesses three higher order dimensions, which include 10lower order dimensions (and an 11th lower order dimension, termed absorption,which does not fall under one of the three higher order dimensions). Analyses inTellegen’s test construction of the MPQ revealed that it assesses three dimensions,which are conceptually similar to Eysenck’s (Fowles, 1987; Tellegen & Waller, 1994).The MPQ dimensions are positive emotionality (PEM), negative emotionality(NEM), and constraint (CON).

PEM and NEM indicate dispositions toward experiencing positive and negativeaVective experiences, respectively, whereas CON refers to self-restraint and responseinhibition (Tellegen & Waller, 1994; Waller, Lilienfeld, Tellegen, & Lykken, 1991).PEM’s lower order factors are well-being, social potency, achievement, and socialcloseness; NEM’s lower order factors are stress reaction, alienation, and aggression;and CON’s lower order factors are control, harm-avoidance, and traditionalism.PEM, NEM, and CON are somewhat broader than but overlap substantially withEysenck’s dimensions of E, N, and P (reversed), respectively. Factor-analytic devel-opment of the MPQ dimensions revealed that PEM’s, NEM’s, and CON’s factorloadings were .78, .69, and .50 for E, N, and P (reversed), respectively (Tellegen &Waller, 1994). The MPQ has been used in a number of studies examining personalitycorrelates of ASB.

Factor analyses of MPQ data have also yielded a four-factor solution. PEMdivides into Agentic PEM, which includes primarily well-being, social potency, andachievement, and Communal PEM, which includes primarily well-being and socialcloseness. Church (1994) described these factors as reXecting social/work eVectivenessand interpersonal connectedness, respectively. Research examining the relationshipsbetween the MPQ scales and other personality measures supports the distinctionbetween Agentic PEM and Communal PEM (e.g., Patrick, Curtin, & Tellegen, 2002).

In their review of personality’s relations with externalizing conditions (e.g., alco-holism, antisocial personality disorder), Sher and Trull (1994) described three broad-band personality trait dimensions that have been extensively investigated in studiesof ASB. The authors referred to these dimensions as extraversion/sociability, neurot-icism/emotionality, and impulsivity/disinhibition. What Sher and Trull termed extra-version/sociability denotes what Eysenck referred to as E and Tellegen referred to asPEM. Similarly, N and NEM characterize neuroticism/emotionality, and P and CON(reversed) characterize impulsivity/disinhibition.

It is important to acknowledge that many authors regard the Five Factor Model(FFM), which was based on a lexical hypothesis (i.e., that all important personalitytraits have been encoded in language), to be the primary model of personality traits.The FFM includes Neuroticism/Emotional Stability, Extraversion, Openness toExperience, Agreeableness, and Conscientiousness. Compared with Eysenck’s model,there are signiWcantly fewer studies examining the FFM’s use in predicting antisocialbehavior (e.g., Heaven, 1996; see also Miller & Lynam, 2001). Eysenck’s model is


broadly comparable to the FFM in that E appears related to the FFM Extraversionfactor and that N appears related to the FFM Neuroticism factor. Eysenck arguedthat P is negatively related to Agreeableness, and further, that the Agreeableness andOpenness to Experience dimensions are better classiWed as lower order traits, whichhe incorporated in P (Eysenck, 1990, 1991). There is some empirical evidence sup-porting these claims, and the FFM Conscientiousness dimension also appears to benegatively related to P (Costa & McCrae, 1995; Zuckerman, Kuhlman, Joireman,Teta, & Kraft, 1993). Regarding Tellegen’s model and the FFM, there is some evi-dence that Agentic PEM is related to both the FFM’s Extraversion and Conscien-tiousness factors, whereas Communal PEM is more closely related to Extraversion.Moreover, NEM appears positively related to Neuroticism and negatively related toAgreeableness, and CON appears positively related to Conscientiousness (e.g.,Church, 1994).

1.3. Prior reviews of the relations between personality dimensions and ASB

Reviewing early studies of Eysenck’s theory of crime, Passingham (1972) con-cluded that support for Eysenck’s predictions, especially his predictions concerningE, was weak. Cochrane (1974) reviewed 20 studies and reported that only one studysupported Eysenck’s predictions concerning E and that about half of the studies sup-ported his predictions concerning N. Although both Passingham and Cochranenoted that P is integral to Eysenck’s theory of crime, there were virtually no data onthe P dimension when these authors conducted their reviews.

Eysenck and Gudjonsson (1989) reviewed numerous studies of the EPQ and ASB,and they reported that P appeared strongly related to ASB, regardless of age. In addi-tion, they noted that E appeared related to ASB primarily at younger ages and that Nappeared related to ASB primarily at older ages. The authors proposed that E mightbe more relevant to ASB at younger developmental stages because the conditioningof a conscience, which is integral to Eysenck’s conceptualization of E, occurs moreduring childhood than in adulthood. Likewise, N’s role in amplifying previouslyconditioned tendencies would become relatively more important in later years. Nev-ertheless, methodological diVerences across studies that could have accounted for theage-related Wndings were not examined in Eysenck and Gudjonsson’s review.

Other reviews have provided equivocal support for Eysenck’s three-factor theoryof ASB. In their review of 14 studies, Farrington et al. (1982) reported mixed evidencefor an association between E and ASB. Similarly, some reviews (e.g., Feldman, 1993;Furnham & Thompson, 1991) have indicated that support for a relation between Eand ASB in adults and for a relation between N and ASB in juveniles was mixed.

Miller and Lynam (2001) used meta-analytic techniques to examine ASB’s associ-ations with Eysenck’s model, Tellegen’s model, and the FFM (as well as Cloninger’smodel of temperament and character). The authors reported Wndings for 37 studiesusing the EPQ and two studies using the MPQ, and they tested variables (e.g., age,biological sex) for moderating eVects. Weighted mean eVect sizes (rs) for ASB’s asso-ciations with N, P, NEM, and CON (reversed) ranged from .23 to .39, whereas theweighted mean eVect sizes for ASB’s associations with E and PEM were .10 and .01,


respectively. Miller and Lynam found that dimensions reXecting primarily lowAgreeableness and low Conscientiousness of the FFM were most consistently relatedto ASB.

1.4. Methodological issues to consider when reviewing studies of the “Big 3” and ASB

Among studies of personality and ASB, one critical methodological issue involvesthe operationalization of ASB. A known groups approach was predominantly usedin early studies of Eysenck’s theory (see Eysenck & Gudjonsson, 1989). Typicallywith this approach, researchers determine whether mean levels of personality scoresfor a known antisocial group (e.g., incarcerated oVenders) are signiWcantly diVerentfrom mean levels of personality scores for a nonantisocial group (e.g., nonincarcer-ated participants). One limitation of this approach is that known groups of antisocialindividuals may be conservative representations of individuals who engage in ASB(Henry, Caspi, MoYtt, & Silva, 1996). Some comparison participants probably com-mit antisocial acts but avoid being arrested or convicted. A few authors have alsoproposed that inmates do not represent the antisocial population at large becausetheir personalities might be altered by incarceration. For example, there is provi-sional evidence that E levels decrease with incarceration (Allsopp & Feldman, 1974;Forrest, 1977; Haapasalo, 1990).

Another limitation of many known groups design studies is that antisocial andcomparison groups are often not matched on demographic variables (e.g., race, socio-economic status), and unexamined third variables may account for personality diVer-ences found between groups. On the other hand, researchers should carefully selectvariables for matching antisocial and nonantisocial groups because matching proce-dures can be disadvantageous, such that with matching, researchers may control forvariance that is important to explaining personality and ASB relations (Meehl, 1971).

Some have argued that using self-report measures of ASB circumvent some limita-tions of known groups approaches (Furnham & Thompson, 1991; Powell, 1977).However, most self-reported ASB design studies have used samples of nonincarcer-ated individuals, which limits the generalizability of their Wndings. Another limitationof self-reported ASB designs involves self-report measures’ susceptibilities to impres-sion management, malingering, and other response styles, which can inXuence associ-ations between personality dimensions and ASB. For example, participants may failto report levels of ASB because ASB is socially undesirable (Eysenck & Gudjonsson,1989). Conversely, there is some evidence that non-ASB individuals tend to report alarge number of relatively trivial ASB symptoms (Henry et al., 1996).

Another methodological issue to consider involves the measures used to assesspersonality. There is concern that some EPQ or MPQ items directly assess ASB ten-dencies, which may inXate correlations between some personality dimensions andASB. Caspi et al. (1994) and Krueger et al. (1994) identiWed items from the aggressionsubscale of the MPQ’s NEM dimension as describing violent tendencies. However,after removing these items, the researchers did not Wnd that the items substantiallyaVected the correlations between the MPQ and ASB measures. It has been arguedthat predictor-criterion overlap is a concern when using Eysenck’s P dimension to


examine personality—ASB relations, but it does not seem to be problematic withregard to the MPQ CON dimension (Caspi et al., 1994; Krueger et al., 1994).

2. Present review

Eysenck’s theory has played a major role in research investigating personality cor-relates of ASB. Rooted in studies of biological systems, this theory predicts that E, N,and P are positively related to ASB. Overall, support for Eysenck’s claims has beenmixed, and prior reviews of the relations between E, N, and P and ASB have not beencomprehensive in addressing important methodological issues. Tellegen’s MPQrelates closely to Eysenck’s three-factor model, but nearly all the prior reviews ofEysenck’s theory have not extended Wndings to this more recent personality concep-tualization. Three higher order personality dimensions, broadly referred to as extra-version/sociability, neuroticism/emotionality, and impulsivity/disinhibition (Sher &Trull, 1994), and assessed by the EPQ and MPQ, have been examined in numerousstudies of ASB.

Although there is a large body of data on the relations between these “Big 3” per-sonality dimensions and ASB, the relative strengths of these relations are largelyunknown, and potential moderating eVects of methodological and demographic vari-ables have received little attention. A more comprehensive investigation of the rela-tions between these three higher order personality dimensions and ASB appearswarranted. Given the overlap among Eysenck’s and Tellegen’s models, there is rea-son to integrate them in order to better understand personality’s role in predictingASB. Miller and Lynam (2001) provided a theoretical and meta-analytic review ofthese issues. However, in their quantitative review, they examined EPQ and MPQstudies separately. In addition, they did not obtain unpublished data, and there areadditional published studies that were not examined.

The primary questions of the present meta-analytic review are whether the positedrelations between the “Big 3” personality dimensions and ASB are substantiated byempirical Wndings and whether the magnitudes of these relations diVer. Strong con-sideration is placed on the methods by which researchers have operationalized ASB.Also, it is examined whether associations diVer depending on demographic variables,such as age and biological sex.

3. Meta-analytic review

Meta-analytic procedures provide a systematic aggregation of studies’ results andexamination of moderating variables (Rosenthal, 1991). Some potential problems ofmeta-analytic reviews (e.g., inclusion of primarily published Wndings, variable studyquality among selected studies, and generalizing across unlike studies), if not care-fully considered, also threaten the validity of qualitative reviews. These potentialproblems were considered throughout the sample selection and analyses for thisreview.


3.1. Search strategy and sample selection

The primary method for locating studies involved searching computerized dat-abases (PsycINFO, MEDLINE, ERIC, Sociological Abstracts, and DissertationAbstracts International) with 27 keywords related to personality dimensions andASB (e.g., EPQ, MPQ, extraversion, positive emotionality, antisocial behavior, andcrime). Second, the ancestry method was used, whereby references listed in studiesand review articles were surveyed for additional studies. Third, given that H.J.Eysenck founded the International Society for the Study of Individual DiVerences(ISSID) and that studies of his theory are often presented at ISSID conferences, theconference programs of the ISSID biennial meetings dating back to 1995 wereretrieved. Authors of relevant presentations were asked if they would provide datafor the purposes of this review. Fourth, 20 researchers of personality and ASB werecontacted with inquiries about unpublished or in-press studies that might be perti-nent to this review.

The construct of ASB includes a broad cluster of socially undesirable, often illegal,behaviors that are harmful to the welfare of others. ASB has typically been deWned interms of legal concepts (e.g., criminality and delinquency) and categorical psychiatricdisorders [e.g., antisocial personality disorder (ASPD) and conduct disorder (CD)].For this review, ASB included several operationalizations. Studies that comparedincarcerated oVenders and juvenile delinquents with nonantisocial participants wereincluded as were studies in which individuals self-reported levels of ASB. In addition,studies using assessments of ASPD and CD were included because their criteria areprimarily behavioral in nature and because individuals with ASPD and CD typicallyengage in a variety of criminal acts (Abram, 1989; American Psychiatric Association(APA), 1994, 2000; Farrington, 1994, 1995). Also included were teacher, school, andparent reports of ASB and self-reported bullying behavior because these measuresserve as useful ASB operationalizations for young populations (Farrington, 1994).Sample characteristics are displayed in Tables 1 and 2.

In addition, the following self-report measures were considered valid operational-izations of ASB. Studies using the Personality Diagnostic Questionnaire—RevisedASPD and CD scales (Hyler & Rieder, 1987), the Minnesota Multidimensional Per-sonality Inventory-2 (MMPI-2) Antisocial Practices content scale (Butcher, Graham,Williams, & Ben-Porath, 1990), and West and Farrington’s (1977) adolescent behav-ior problems scale were included. Also included were the MMPI-2 PsychopathicDeviate clinical scale (McKinley & Hathaway, 1944) and the California Psychologi-cal Inventory Socialization scale (scored in reverse) (Gough, 1960) because they cor-relate moderately to strongly with other well validated ASB measures (Hare, 1985).The delinquent behaviors scale of the Youth Self-Report Form (Achenbach, 1991)was also considered a valid ASB measure, although no studies that met the otherinclusion criteria examined this self-report measure.

Studies examining psychopathy as a marker of ASB and studies of speciWcoVender types were excluded. Psychopathy (see Cleckley, 1941/1988) includes a vari-ety of personality features (e.g., lack of empathy, manipulativeness, and impulsivity)in addition to behavioral manifestations, and in this respect, it is broader than ASB

258E

.M. C

ale / Journal of Research in P

ersonality 40 (2006) 250–284

Age Sex EVect size

Ext/Soc Neu/Emo Imp/Dis

11–14 Male ¡.05 .12 .06

19.14 Male .16 ¡.03 .4224.64 Female ¡.09 .24 .4525.20(Adult) Male (¡.21)

12–13 Male ¡.07 .13 .1712–13 Male ¡.04 .20 .2218 Male .05 .48 .4418 Female .01 .34 .4417 Male .13 .41 .3713.26 Male .28 .14 .5514.09 Female .31 .27 .5116–19 Male (¡.06) (.41) (.36)

20–29 Male (.03) (.44) (.36)

30–39 Male (.04) (.48) (.35)

40–49 Male .13 .43 .37

50–59 Male .28 .44 .41

Table 1Studies selected for review

Reference Sample description Personality measure N

Af Klinteberg et al. (1992) Criminals with 2+ crimesa EPQ-R short version 199Criminals with 1 crimea

Nonclinical participantsAleixo and Norris (2000) Convicted oVenders EPQ-R short version 101Barack and Widom (1978) Participants awaiting trial EPQ 111

Nonclinical participantsBartol and Holanchock (1979) Inmates EPQ 585

Nonclinical participantsCaspi et al. (1994) Nonclinical participants MPQ Q-Sort version 216

Nonclinical participants MPQ Q-Sort version 181Caspi et al. (1994), Krueger et al. (1994) Nonclinical participants MPQ short version 440

Nonclinical participants MPQ short version 417Dåderman (1999) CD juvenile oVenders EPQ 47Eysenck (1981) Schoolchildren Junior EPQ 101

Schoolchildren Junior EPQ 306Eysenck and Eysenck (1977) Inmates EPQ 927

Nonclinical participantsInmates EPQ 1816Nonclinical participantsInmates EPQ 868Nonclinical participantsInmates EPQ 545Nonclinical participantsInmates EPQ 282Nonclinical participants

E.M

. Cale / Journal of R

esearch in Personality 40 (2006) 250–284

259Eysenck et al. (1976) Inmates EPQ 126 25.4 Male .01 .29 .27

25.627.2 Female ¡.05 .22 .5027.514.6 Male .08 .07 .0314.613.89.8 Male .09 .32 .209.716.91 Mixed .13 ¡.19 .5115 Mixed .24 .38 .66

19.9 Mixed .04 ¡.01 .5211.14 Mixed .08 .32 .6410.217.7 (Mixed) .01 .17 .2722.96 Male .09 .42 .4430.8329.9730.729.00 Female .02 .43 .4527.7526.4929.7124 Female .53 ¡.15 ¡.06

28.24 Male (.22) (¡.18) (¡.17)26.9727.04 Female .19 ¡.24 .0525.89

(continued on next page)

Nonclinical participantsInmates EPQ 126Nonclinical participants

Fonseca and Yule (1995) Aggressive delinquentsa Junior EPQ 64Nonaggressive delinquentsa

Nonclinical participantsCD participants Junior EPQ 60Nonclinical participants

Furnham (1984) College students EPQ 210Furnham and Barratt (1988) Delinquents Junior EPQ 54

Clinical adolescentsb

Nonclinical participantsb

Furnham and Thompson (1991) Undergraduates EPQ 100Gabrys et al. (1988) CD participants Junior EPQ 684

Clinical participantsGe and Conger (1999) Survey participants MPQ 406Gomà-i-Freixanet (1995) Inmates EPQ 622

Risky sports participantsRisky jobs participantsNonrisky participants

Gomà-i-Freixanet (2001) Inmates EPQ 227Risky sports participantsRisky jobs participantsNonrisky participants

Gossop (1978) Criminal drug addicts EPQ 37Clinical participants

Gossop and Eysenck (1983) Inmates EPQ 1089Clinical participantsInmates EPQ 137Clinical participants

260E

.M. C


ersonality 40 (2006) 250–284

Age Sex EVect size


33.78 Male ¡.25 .17 .65

16 Male .09 .4516 Female .06 .4816 Male .20 .4216 Female .23 .3313–15 Male .4013–15 Female .3316 Male .15 .22 .4016 Female .28 .14 .6616 Male .17 .01 .6416 Female .39 .25 .5313 Male ¡.10 .28 .57(Child) Male (.31) (.10) (.58)(Child) Female (.41) (.09) (.59)31.4 Mixed .15 .28 .5033 Male .11 .28 .27(Adol.) Mixed .25 ¡.23 .58

(Adol.) Mixed .07 ¡.31 .54

18.41 Male ¡.17 .43 .2318.9 Female .15 .43 .2318.49 Male .22 .06 .4918.73 Female ¡.21 .29 .3418.58 Male .02 .21 .1918.2 Female ¡.15 .33 .36

Table 1 (continued)

Reference Sample description Personality measure N

Haapasalo (1990) Inmates EPQ 429Nonclinical participants

Heaven (1993) Nonclinical participants EPQ-R short version 126Nonclinical participants EPQ-R short version 141High school students EPQ-R short version 175High school students EPQ-R short version 211

Heaven (1994) High school students EPQ-R short version 136High school students EPQ-R short version 146

Heaven et al. (2000) High school students EPQ-R short version 172High school students EPQ-R short version 45High school students EPQ-R short version 169High school students EPQ-R short version 40

Heaven and Virgen (2001) High school students EPQ-R short version 110Jamison (1980) Secondary school students Junior EPQ 781

Secondary school students Junior EPQ 501Jang et al. (1999)c Nonclinical participants EPQ-R 200Krueger et al. (2001) Twin study participants MPQ short version 673Lane (1987) Severe conduct studentsa EPQ 120

Conduct problem studentsa

Comparison studentsDelinquents EPQ 120Nonclinical participants

Lilienfeld (2000)c Undergraduates MPQ short version 43Undergraduates MPQ short version 57Undergraduates MPQ short version 26Undergraduates MPQ short version 49Undergraduates MPQ short version 79Undergraduates MPQ short version 48

E.M



261Ma et al. (1996) Secondary school students Junior EPQ—Chinese 425 14.25 Mixed .22 .27 .51

12–13 Female .13 .05 .38

7–9 Male .05 .16 .3142 Male .06 .13 .1242.917.95 (Mixed) .05 .27 .0617.92

10.7 Mixed .09 .18 .19

31 Male .04 .47 .16

29 Female ¡.05 ¡.02 .148.39 Male .17 .34 .328.47 Female ¡.04 .38 .319.29 Male ¡.17 .09 .519.19 Female .26 .17 .5810.21 Male .19 ¡.04 .4210.32 Female ¡.02 .07 .4711.26 Female .04 .33 .1313.4 Male .52 .34 .5213.33 Female .38 .13 .5515.33 Male .41 .32 .5115.48 Female .12 .39 .5713–15 Male ¡.01 .02 .34

29.16 Male .10 .25 .21

(continued on next page)

Maliphant et al. (1990) Disruptive students Junior EPQ 44Moderate studentsb

Well-behaved studentsb

Maliphant et al. (1990) Primary school students Junior EPQ 50McGue et al. (1997) Alcoholics w/ASPD sxms MPQ short version 304

Alcoholics w/o ASPD sxmsMcGurk and Bolton (1981) Inmates EPQ 200

Nonclinical participants

Mynard and Joseph (1997) Bullies onlya Junior EPQ 179Bullies and victimsa

Victims onlyb


Nagoshi et al. (1992) Drug txt study participants EPQ-R 95

Newman (2000)c Inmates MPQ 263Powell (1977) Junior school students Junior EPQ 46

Junior school students Junior EPQ 57Junior school students Junior EPQ 70Junior school students Junior EPQ 85Junior school students Junior EPQ 79Junior school students Junior EPQ 77Senior school students Junior EPQ 72Senior school students Junior EPQ 62Senior school students Junior EPQ 72Senior school students Junior EPQ 60Senior school students Junior EPQ 64

Putnins (1982) IdentiWed delinquentsa EPQ 179Delinquents: post-testa

Nonclinical participantsRahman (1992) Inmates EPQ—Bengali 636


262E

.M. C


ersonality 40 (2006) 250–284

e Sex EVect size


45 Female ¡.04 .37 .42

3 Male (.67) (.21) (.70)75

45 Male .01 .13 .3749 Female .11 .18 .2616 Male ¡.09 .00 .35

Mixed .34 ¡.16 .30Mixed .33 ¡.31 .31Mixed .40 .01 .33

ult) (Mixed) .10 .25 .34ult) Mixed ¡.23 ¡.17 .13ult) Mixed .34 ¡.16 .35

11 Male ¡.26 .16 .39

15 Male .32 .20 .63ol.) (Mixed) ¡.05 .01 .4411 Male ¡.26 .16 .39

7 Male .07 .15 .519 Female .21 .29 .3211 Male ¡.18 .22 .27

Table 1 (continued)

Reference Sample description Personality measure N A

Rahman and Husain (1984) Inmates EPQ—Bengali 601 19Nonclinical participants

Rangaswami and Arunagini (1982) Inmatesa EPQ—Tamil 145 37Mentally ill criminalsa 39Nonclinical participants 36

Romero et al. (2001) H.S./College Students EPQ—Spanish 435 15H.S./College Students 529 15Delinquents 95 17

Rushton and Chrisjohn (1981) Undergraduates EPQ 124 18Undergraduates EPQ 31 31Undergraduates EPQ 42 20Undergraduates EPQ 31 (AUndergraduates EPQ 25 (AUndergraduates EPQ 41 (ANonclinical participants

Saklofske (1977) Badly behaved students Junior EPQ 74 10Well behaved students

Saklofske and Eysenck (1980) Junior school students Junior EPQ 147 13Seisdedos (1982) Antisocial delinquents Junior EPQ 181 (ASaklofske (1977) Badly behaved students Junior EPQ 74 10Silva et al. (1986) Community sample Junior EPQ—Spanish 174 13

Community sample Junior EPQ—Spanish 183 13Delinquents Junior EPQ—Spanish 145 13Nonclinical participants

g

–

.

.

.

.

.

.

ddd

–

–d–...

E.M



263

EPQ, Eysenck Personality Questionnaire; EPQ-R,lt), sample of adults, but age estimate is unknown; is unknown; mixed, both males and females in thetheses were for samples identiWed as outliers.

Male (¡.11) (.30) (.01)

Male .28 .11 .37

Male ¡.26 .61 .02

Male .02 .51

Note. EVect sizes listed are Pearson rs. CD, conduct disordered; ASPD, antisocial personality disorder; Eysenck Personality Questionnaire—Revised; MPQ, Multidimensional Personality Questionnaire; (Adu(Child), sample of children, but age estimate is unknown; (Adol.), sample of adolescents, but age estimatesample; (Mixed), both males and females in the sample, but proportions are unknown. EVect sizes in paren

a Group statistics were combined to form one ASB group.b Group statistics were combined to form one non-ASB group.c Authors provided unpublished data.

Sipos et al. (1998) Delinquents Junior EPQ—Hungarian 756 16.03At-risk studentsb 16.24Hungarian normsb 15

Slee and Rigby (1993) Bullies Junior EPQ 87 10.9Victimsb


Stewart and Hemsley (1984)c Criminals EPQ 32 31.16Nonclinical participants 28.5

Taylor (1999) Early starter delinquentsa Informant MPQ 129 11.32Late starter delinquentsa



(Hare, 1996; Lilienfeld, 1994). Regarding oVender types, while there appear to be sub-tle diVerences among the personalities of criminal oVenders (Eysenck, 1996b), there isalso considerable homogeneity found among antisocial individuals’ personality fea-tures (Eysenck & Gudjonsson, 1989; Wilson & Herrnstein, 1985). Although researchon personality diVerences among oVender types is important to understanding ASB,the homogeneity among antisocial individuals was most important for this review.

Data were typically reported as mean levels of personality dimensions and com-pared across ASB and non-ASB groups or as correlations between personality scoresand continuous measures of ASB. When studies reported results separately for diVer-ent samples, categorized by age or sex, the samples were examined separately. Fornearly all of the known groups design studies, mean levels of personality dimensionscores of ASB and non-ASB groups were compared. It is important to note that forsome studies, school reported ASB was considered “known” because the ASB datacame from an external source. In some cases (e.g., Lane, 1987; Maliphant, Hume, &Furnham, 1990; Saklofske, 1977), participants were separated into ASB and non-ASB groups, based on teacher or school reports. In one case (Maliphant et al., 1990),school reported ASB was a continuous variable, and correlations between EPQscores and ASB were provided. For self-reported ASB designs studies, correlationcoeYcients were usually reported. However, in some studies (e.g., McGue, Slutske,Taylor, & Iacono, 1997; Mynard & Joseph, 1997), researchers grouped participantsbased on self-reported ASB scores and compared mean levels of EPQ or MPQ scoresacross groups.

The assumption of independence is that each sample included in a meta-analyticreview is composed of individuals who are not participants in the other includedsamples (Hedges & Olkin, 1985; Rosenthal, 1994). When selecting studies, samplecharacteristics were carefully reviewed to prevent analyzing overlapping samples as ifthey were independent. Including an eVect size for each personality dimension andASB relation (a maximum of three eVect sizes per sample) did not violate theassumption of independence because each personality dimension was considered anindependent variable. On the other hand, many studies contained data for multipleoperationalizations of ASB, and using each of these eVect sizes would have violatedthe assumption of independence. Only eVect sizes for the ASB operationalization thatincorporated the greatest number of participants were included. If the same partici-pants were assessed across multiple operationalizations, then data for the ASB oper-ationalization that was the study’s primary focus were used.1

For studies in which researchers examined only one or two personality dimensions,the Wrst author or author for whom there was an available address was contacted andasked to provide data for the other personality dimension(s). For Eysenck’s dimen-

1 The ASB operationalizations selected by following these rules were quite comparable with other stud-ies (e.g., self-reported delinquency). However, for four samples (Caspi et al., 1994; Krueger et al., 1994), re-searchers focused on informant reports of delinquency, parent reports of delinquency, police contacts, andcourt convictions, in addition to self-reported delinquency. Only self-reported delinquency data were in-cluded in the analyses because examining the other, less frequently used, ASB operationalizations wouldhave limited comparisons across samples.


sions, only studies using the EPQ (Eysenck & Eysenck, 1975), the EPQ—Revised(Eysenck, Eysenck, & Barrett, 1985), or translations of this measure were included.This criterion was used because the P dimension was not incorporated in earlier mea-sures and, as noted previously, E was revised considerably after its initial conceptual-izations. Some versions of the EPQ contain a Lie (L) scale (i.e., an ostensible measureof “faking good”), but it is unclear whether the L scale adequately detects positiveimpression management (see Eysenck & Eysenck, 1994). Due to this scale’s question-able validity, studies that controlled for L scores (or excluded data based on high Lscores) were excluded. For Tellegen’s dimensions, studies using the MPQ and variantsof this measure (e.g., MPQ Q-sort version; Caspi et al., 1994) were included.

Studies’ results were reported in suYcient detail to permit the calculation of eVectsizes between personality dimension scores and measures of ASB. If studies did notinclude adequate information for this purpose, the Wrst author or author for whomthere was an available address was contacted with a request for additional data.Three studies were excluded because additional data could not be obtained. A totalof 97 samples, from 52 published and unpublished studies, had suYcient data forbeing included in this meta-analytic review.

Table 2Sample characteristics

Note. Total numbers of participants are in parentheses. Samples excluded from the Wnal analyses areexcluded from this table. ASB, antisocial behavior; EPQ, Eysenck Personality Questionnaire; MPQ, Mul-tidimensional Personality Questionnaire. “Antisocial individuals” in the ASB operationalization categoryincluded adult inmates, juvenile delinquents, children or adolescents with conduct disorder, and individu-als awaiting trial. “Other self-reported ASB” included the various self-report measures listed in the text.

All samples Known groups Self-reported ASB

ASB operationalizationAntisocial individuals vs. comparison participants 22 (5077) 22 (5077) —School reported ASB 4 (288) 4 (288) —Self-reported delinquency 47 (7239) — 47 (7239)Other self-reported ASB 13 (2533) — 13 (2533)Self-reported bullying 2 (266) — 2 (266)

Age of participantsChild participants 15 (1813) 5 (997) 10 (816)Adolescent participants 39 (6490) 10 (1306) 29 (5184)Adult participants 34 (7100) 11 (3062) 23 (4038)

Sex of participantsMales only 44 (8085) 14 (2950) 30 (5135)Females only 26 (4145) 6 (1056) 20 (3089)Males and females 18 (3173) 6 (1359) 12 (1814)

Personality measurementEPQ dimensions 73 (12072) 25 (5236) 48 (6836)MPQ dimensions 15 (3331) 1 (129) 14 (3202)

Publication statusPublished 80 (14709) 25 (5236) 55 (9473)Not published 8 (694) 1 (129) 7 (565)


3.2. Moderator variables

As emphasized throughout this review, the following methodological issues wereof central importance: operationalization of ASB, sample demographics, and person-ality measurement (cf. Caspi et al., 1994; Krueger et al., 1994). Known groupsapproaches were compared with self-reported ASB approaches. The demographicvariables examined were age and biological sex. Eysenck suggested that ASB’s asso-ciation with P was consistent across age, that ASB’s association with E was strongerin children than in adults, and that ASB’s association with N was stronger in adultsthan in children (Eysenck, 1996a; Eysenck & Gudjonsson, 1989). With regard to sexdiVerences, some have hypothesized that females need more genetic and/or environ-mental factors relevant to personality to place them above their relatively highthresholds and become antisocial than do males (see Cloninger, 1978). Therefore, it isalso possible that the magnitudes of the associations between certain personality fea-tures and ASB diVer between males and females. Potential diVerences in the corre-lates of Eysenck’s and Tellegen’s conceptualizations were important to identify aswell. Although these measures overlap considerably, they are not considered inter-changeable measures of personality, and it is possible that certain facets within oneor more dimensions of one measure predict antisocial manifestations better thanthose of the other measure. Moreover, the problem of item overlap between EPQ’s Pdimension (and possibly MPQ’s NEM dimension) and ASB assessments can be fur-ther examined by comparing associations with those of alternative personalitydimensions.

3.3. Calculation of eVect sizes

The Pearson product–moment correlation (r) was used as the overall eVect sizeindex. For each sample, eVect sizes of the relations between each personality dimen-sion and ASB were calculated. When available, correlations were used as direct esti-mates of eVect sizes. In cases where only means and standard deviations ofpersonality dimension scores were provided for an ASB group and a non-ASB group,a t-value was calculated, and t-values were converted to rs (see Rosenthal, 1991). Forstudies in which there were more than two ASB or non-ASB groups, weighted meansand variances (Schwarzer, 1989) were calculated to represent Wndings for ASB andnon-ASB groups. Magnitudes of correlations were interpreted such that an r of atleast .10 is small, an r of at least .24 is moderate, and an r of at least .37 is large (seeCohen, 1988; Lipsey, 1990).

4. Results

The results reXect essentially three meta-analyses, whereby each procedure wasconducted separately for each personality dimension. There were 97 samples, whichincluded 94 samples for the extraversion/sociability dimension, 90 samples for theneuroticism/emotionality dimension, and 96 samples for the impulsivity/disinhibition


dimension. For extraversion/sociability, 73% of the samples had positive eVect sizes;for neuroticism/emotionality, 84% of the samples had positive eVect sizes; and forimpulsivity/disinhibition, 98% of the samples had positive eVect sizes.

The eVect size distributions (see Rosenthal, 1991) are displayed in Fig. 1. The meanrs of the distributions suggest that the association between personality and ASB islarge in magnitude for impulsivity/disinhibition and small in magnitude for bothextraversion/sociability and neuroticism/emotionality. After considering the widerange in samples sizes (25–1816) and the fact that the results of larger samples tend tobe more stable than the results of smaller samples (Hedges & Olkin, 1985), only meta-analytic procedures incorporating the weighting of sample sizes were performed.

4.1. Primary analyses

The primary analyses were conducted based on a random eVects model approachto meta-analysis (see Hedges & Vevea, 1998). In random eVects models, each study ispresumed to be from a population that may have an eVect size diVerent from anyother study in the analysis, whereas in Wxed eVects models, the population eVect sizeis assumed to be the same for all studies analyzed. In the present review, randomeVects analyses were more defensible for examining the relative strengths of the asso-ciations between the “Big 3” dimensions and ASB because they provided for strongergeneralizations, beyond the studies examined here.

Following Hedges and Olkin (1985), Hedges and Vevea (1998), and Rosenthal(1991), eVect sizes were combined to form a weighted average of the correlationsbetween personality dimensions and ASB. Because a distribution of rs becomes moreskewed as the population value of r gets further from zero and because there was alarge number of participants in many samples, each eVect size r was transformed intoFisher’s z (see Hedges & Olkin, 1985; Rosenthal, 1994). Averaged weighted z-valueswere transformed back into rs.

Fig. 1. Stem and leaf plots of eVect sizes.


Meta-analytic procedures were also used to test for eVect size heterogeneity. A sig-niWcant heterogeneity estimate (Q-value; Hedges & Olkin, 1985) indicates that theeVect sizes are more heterogeneous than would be expected due to sampling error(Schwarzer, 1989). Given the wide range of sample sizes, it was imperative to examinefor outliers, and the inXuence of outliers was determined by calculating Q-values(Hedges & Olkin, 1985). Heterogeneity statistics were computed, such that each sam-ple was omitted in a separate analysis, to determine which samples contributed themost unexplained heterogeneity. Outliers that increased the heterogeneity statistic byat least 50% were identiWed for each dimension. None of the identiWed outliers con-tributed substantially to the heterogeneity for all three dimensions. However, becausethe outlier samples were considered to be substantially diVerent from the others forat least one dimension, these samples were omitted from all the analyses. This processresulted in nine samples being excluded from the following analyses and reduced theoriginal heterogeneity statistic for each dimension by 48–54%.

As shown in Table 3, the estimated eVect size for the association between extraver-sion/sociability and ASB was negligible. For neuroticism/emotionality, the estimatedeVect size was small in magnitude, and for impulsivity/disinhibition, the estimatedeVect size was large in magnitude. Examination of the conWdence intervals for theweighted mean eVect sizes indicated that these estimates were signiWcantly diVerentfrom each other. The overall weighted eVect sizes were examined separately for theknown groups design and self-reported ASB design studies. For the known groupsdesign studies, the magnitudes of the estimated eVect sizes for the extraversion/socia-bility, neuroticism/emotionality, and impulsivity/disinhibition associations were neg-ligible, small, and moderate, respectively. For the self-reported ASB design studies,the magnitudes of the estimated eVect sizes for extraversion/sociability, neuroticism/emotionality, and impulsivity/disinhibition were small, small, and large, respectively.For both design types, the association between impulsivity/disinhibition and ASB

Table 3Relations between the “Big 3” personality dimensions and ASB

Note. k, number of samples in analysis; N, total number of participants in analysis; r, weighted mean r(eVect size); CI, conWdence interval for weighted mean r; Q, heterogeneity statistic.

Personality dimension k N R 95% CI Q

All samplesExtraversion/sociability 85 14,992 .09 (.06–.12) 279.96Neuroticism/emotionality 82 14,468 .19 (.15–.23) 454.86Impulsivity/disinhibition 88 15,403 .39 (.35–.42) 533.65

Known groups design samplesExtraversion/sociability 25 5,236 .04 (¡.03–.10) 119.75Neuroticism/emotionality 26 5,365 .17 (.09–.25) 200.38Impulsivity/disinhibition 26 5,365 .36 (.27–.44) 293.37

Self-reported ASB design samplesExtraversion/sociability 60 9,756 .12 (.08–.15) 160.20Neuroticism/emotionality 56 9,103 .19 (.15–.24) 254.48Impulsivity/disinhibition 62 10,038 .40 (.36–.43) 240.28


was signiWcantly stronger than the associations between both neuroticism/emotional-ity and extraversion/sociability and ASB.2

4.2. Moderator analyses

Following Hedges (1994), Hedges and Olkin (1985), and Rosenthal (1991), formu-las for examining moderator eVects were based on a Wxed eVects model approach.Accordingly, the results for these analyses are limited in that they support generaliza-tions to studies identical to the ones examined here. Categorical and continuousmoderator analyses were conducted, and all eVect sizes were weighted by samplesizes. For categorical moderators, eVect sizes for each level of a moderator variablewere tested for between-groups and within-groups heterogeneities (see Hedges &Olkin, 1985). A signiWcant between-groups Q (Qb) indicates that the groups’ averagedeVect sizes are signiWcantly diVerent from each other. A signiWcant within-groups Q(Qw) for one level of the moderator indicates that the heterogeneity among thatlevel’s eVect sizes is greater than would be expected due to sampling error, and thatanother variable (or variables) could be systematically inXuencing the magnitudes ofthe eVect sizes for that group of samples. Therefore, when Qw values were signiWcant,the Wndings were interpreted with caution. Weighted least squares regression proce-dures were used to test for moderating eVects of continuous variables (see Hedges,1994). Correlations between the moderator and eVect sizes (R Model) were inter-preted based on Cohen’s (1988) recommendations. To test the signiWcance levels ofthe continuous moderators, Z-tests (i.e., two-sided tests of the null hypothesis thatthe regression coeYcients equal zero) were conducted.

4.2.1. Operationalization of ASBThe association between extraversion/sociability and ASB was signiWcantly stron-

ger for self-reported ASB design samples than known groups design samples(p < .001). However, the associations between both neuroticism/emotionality andimpulsivity/disinhibition and ASB were not signiWcantly diVerent across ASB opera-tionalization (p valuesD .456 and .186, respectively). Nearly all of the subsequentmoderator analyses were conducted separately for known groups and self-reportedASB designs for the following reasons. For all three dimensions, eVect sizes withinthe two ASB operationalization groups were heterogeneous, suggesting that othervariables within these groups were systematically inXuencing the magnitudes of theeVect sizes. In addition, because there was a relatively large number of samples

2 A limitation in using correlation coeYcients for known groups design studies is that these studies donot provide dichotomizations of ASB that are the same across studies, which contributes to excess hetero-geneity in the analyses. Three exploratory analyses were conducted, using known groups design studies’natural metric Cohen’s d, which can be converted to a correlation coeYcient, were conducted [omittingMaliphant et al.’s (1990) study, which reported correlation coeYcients]. The Wndings from these analyseswere comparable with the primary analyses’ weighted mean rs [extraversion/sociability (d D .08,QD 93.52), neuroticism/emotionality (d D .37; QD 209.02), and impulsivity/disinhibition (d D .86,QD 316.21)].


available for each ASB operationalization, having a small number of samples to ana-lyze was not a concern.

4.2.2. Age (Table 4)For 82 samples, participants’ ages were reported, and either mean or median age

in years was coded as a continuous moderator. SigniWcant moderating eVects diVeredacross ASB operationalization. For known groups design samples, the associationbetween extraversion/sociability and ASB was signiWcantly stronger as age increased,but for self-reported ASB design samples, the association between extraversion/sociability and ASB was signiWcantly weaker as age increased. For both ASB opera-tionalizations, the association between neuroticism/emotionality and ASB was sig-niWcantly stronger as age increased. Age did not signiWcantly moderate the

Table 4Age moderator analyses

Note. k, number of samples in analysis; R, model r; Z, signiWcance test of continuous moderator; Qw,within groups heterogeneity statistic; Qb, between groups heterogeneity statistic. Qw data in bold indicatehomogenous eVect sizes.

Personality dimension k R Z P Qw

Known groups design samplesExtraversion/sociability 22 .24 2.86 .004 95.96Neuroticism/emotionality 23 .46 5.84 <.001 95.74Impulsivity/disinhibition 23 .07 ¡1.18 .239 280.51

Self-reported ASB design samplesExtraversion/sociability 57 .15 ¡2.19 .028 145.53Neuroticism/emotionality 53 .18 2.55 .011 237.52Impulsivity/disinhibition 59 .34 ¡6.36 <.001 211.05

Age group k Mean r Qw Qb p

Known groups design samplesExtraversion/sociability Children 4 .05 7.45 0.39 .822

Adolescents 10 .02 17.93Adults 11 .04 83.80

Neuroticism/emotionality Children 5 .27 11.59 55.08 <.001Adolescents 10 .06 47.75Adults 11 .31 85.61

Impulsivity/disinhibition Children 5 .66 29.70 66.77 <.001Adolescents 10 .32 70.35Adults 11 .40 125.41

Self-reported ASB design samplesExtraversion/sociability Children 10 .10 13.45 8.43 .015

Adolescents 27 .14 90.98Adults 23 .07 40.68

Neuroticism/emotionality Children 10 .17 11.19 26.33 <.001Adolescents 23 .18 58.32Adults 23 .29 158.43

Impulsivity/disinhibition Children 10 .38 22.39 13.30 .001Adolescents 29 .44 126.24Adults 23 .36 77.74


association between impulsivity/disinhibition for known groups design samples, butfor self-reported ASB design samples, the association between impulsivity/disinhibi-tion and ASB was weaker as age increased.

Age was also examined as a categorical moderator for the following reasons. Evenfor studies that did not provide information regarding the ages of participants, theauthors reported whether the sample included child, adolescent, or adult participants.Therefore, coding age categorically allowed for all the samples to be included. Inaddition, childhood, adolescence, and adulthood represent qualitatively diVerentdevelopmental stages in a variety of ways. Furthermore, a number of ASB research-ers have examined how ASB manifestations diVer across these age groups (e.g., Far-rington, 1995; MoYtt, 1993).

Comparing the categorical moderator analyses with the continuous age analysesrevealed that the inXuence of age on the associations between personality and ASB iscomplex and diYcult to determine with the present sample of studies. It is alsoimportant to note that signiWcant within-group heterogeneity was found among mostof the age groups’ eVect sizes. For the known groups design samples, the associationbetween extraversion/sociability and ASB did not diVer signiWcantly across agegroups. However, the associations between ASB and both neuroticism/emotionalityand impulsivity/disinhibition diVered across age groups. Post hoc contrast analysesindicated that these associations were signiWcantly weaker for adolescents than forchildren and adults. For the self-reported ASB design samples, the associationsbetween both extraversion/sociability and impulsivity/disinhibition and ASB diVeredacross age groups, such that these associations were signiWcantly stronger for adoles-cents than adults. The association between neuroticism/emotionality and ASBdiVered signiWcantly across age groups in self-reported ASB design samples, and theWndings were similar to the results when age was examined as a continuous modera-tor. More speciWcally, the association was signiWcantly stronger for adults than chil-dren and adolescents.3

4.2.3. Biological sexBiological sex was reported for 82 samples, 12 of which were samples including

both males and females. Although samples were typically coded as zero (all male) orone (all female), coding sex as a continuous moderator variable (i.e., the proportionof females in the sample) enabled more samples to be included in the analyses. Bio-logical sex did not exhibit a signiWcant moderating eVect for any of the associationsbetween the personality dimensions and ASB, although one Wnding came relativelyclose to reaching statistical signiWcance. For the self-reported ASB design samples,the association between extraversion/sociability and ASB increased as a function ofthe proportion of females in the samples (R modelD .12, ZD 1.46, pD .144). Theother analyses’ p values ranged from .319 to .844.

3 Follow-up regression equations, testing the quadratic eVects of age on the relations between the per-sonality dimension and ASB, were computed. For each dimension, the quadratic eVect of age was nonsig-niWcant and was weaker than the linear eVects of age, as reported in Table 4.


4.2.4. Age and biological sex interactionsInteraction eVects were examined for those studies that reported participants’ ages

and biological sex. For both known groups samples (NsD21–22) and self-reportedASB design samples (NsD 56–58), weighted least squares regression analyses wereconducted for each personality dimension. Age and sex (proportions of females) wereentered Wrst, and the product of age and sex (with the partialled product term repre-senting the interaction) was entered second (see Jaccard, Turrisi, & Wan, 1990).

Interaction eVects came relatively close to reaching statistical signiWcance only forextraversion/sociability. For both the known groups design studies and the self-reported ASB design studies, the interaction between age and biological sex yielded anegative eVect on the association between this dimension and ASB (p valuesD .056and .136, respectively). This suggested that the strengths of the relationship betweenextraversion/sociability and ASB diVer among groups categorized by age and sex(i.e., older males, older females, younger males, and younger females). For example,the association between extraversion/sociability and ASB may be particularly weakamong older females, compared with the other groups. Nevertheless, these Wndingswere not statistically signiWcant, and they should be considered provisional.

4.2.5. Measurement of personality (Table 5)Personality measure (EPQ versus MPQ) was examined as a categorical moderator.

The eVect sizes within nearly all of the EPQ and MPQ groups were heterogeneous.For all the samples, the association between extraversion/sociability and ASB wassigniWcantly stronger for EPQ studies than MPQ studies. The association betweenneuroticism/emotionality and ASB was not signiWcantly diVerent between EPQ andMPQ studies. The association between impulsivity/disinhibition and ASB was

Table 5Personality measure moderator analyses

Note. No analyses were conducted for known groups design samples. EPQ, Eysenck Personality Question-naire; MPQ, Multidimensional Personality Questionnaire; k, number of samples; r, weighted mean r(eVect size); Qw, within groups heterogeneity statistic; Qb, between groups heterogeneity statistic. Qw datain bold indicate homogenous eVect sizes.

Personality dimension Measure k r Qw Qb p

All samplesExtraversion/sociability EPQ 71 .10 250.32 12.53 <.001

MPQ 14 .03 17.07Neuroticism/emotionality EPQ 67 .23 373.63 1.80 .180

MPQ 15 .25 79.43Impulsivity/disinhibition EPQ 73 .44 424.83 50.51 <.001

MPQ 15 .30 58.31

Self-reported ASB design samplesExtraversion/sociability EPQ 46 .15 105.22 31.24 <.001

MPQ 14 .03 17.07Neuroticism/emotionality EPQ 42 .21 175.97 5.83 .016

MPQ 14 .26 72.48Impulsivity/disinhibition EPQ 48 .46 131.61 58.47 <.001

MPQ 14 .29 49.59


stronger for EPQ studies than MPQ studies. Because only one MPQ sample used aknown groups design approach (Taylor, 1999), personality measurement was notexamined as a categorical moderator within known groups design samples. Amongself-reported ASB samples, the association between extraversion/sociability and ASBwas stronger for EPQ studies than MPQ studies, the association between neuroti-cism/emotionality and ASB was stronger for MPQ studies than EPQ studies, and theassociation between impulsivity/disinhibition and ASB was stronger for EPQ studiesthan MPQ studies.

4.2.6. Publication statusAn inherent problem of literature reviews is that studies are typically retrieved

from published sources (Rosenthal, 1991; Schwarzer, 1989). For this review, publica-tion status was examined as a categorical moderator. Among all the samples, thereappeared to be a moderating eVect of publication status for all three dimensions,such that published samples had signiWcantly larger eVect sizes than unpublishedsamples. Because all of the unpublished samples had MPQ assessments of personalityand because personality measurement was found to be a signiWcant moderator, per-sonality measurement was speculated to be confounding the results of the publica-tion status analyses. Nevertheless, there was a signiWcant moderating eVect ofpublication status for neuroticism/emotionality, which would not be expected if per-sonality measure was confounding the Wndings because MPQ studies (most of whichwere not published) had higher associations between neuroticism/emotionality andASB than did EPQ studies. The eVects of publication status were also examined foronly the samples that had MPQ assessments of personality. These Wndings revealed asigniWcant moderating eVect of publication status only for neuroticism/emotionality,such that published samples had signiWcantly larger eVect sizes than unpublishedsamples.4

5. Discussion

This review bolsters the assertion that among the “Big 3” personality dimensions,impulsivity/disinhibition is most strongly related to ASB. The overall Wndings sug-gest that the association between impulsivity/disinhibition (P and CON—reversed) ismoderate to large in magnitude. Neuroticism/emotionality (N and NEM) also

4 To further examine the potential inXuence of publication bias, weighted least squares regression analy-ses were conducted, whereby personality measure was entered Wrst and publication status was entered sec-ond. For neuroticism/emotionality, publication status appeared to contribute a moderating eVect over andabove personality measure, but there appeared to be no such moderating eVect for extraversion/sociabilityand impulsivity/disinhibition. R2 change statistics for the publication status variable were examined to de-termine the magnitude of these Wndings. For neuroticism/emotionality, the R2 change statistic revealedthat the eVects of publication status, over and above personality measure, accounted for 2.7% of the vari-ance in eVect sizes, suggesting that the eVect of publication bias was relatively small. For extraversion/so-ciability and impulsivity/disinhibition, the eVects of publication status, over and above personalitymeasure, accounted for 1.0 and 0.1% of the variance in eVect sizes, respectively.


appears related to ASB, but overall, this association seems to be small in magnitude.Perhaps the most theoretically interesting Wnding, because it contradicts a centraltenet of Eysenck’s personality theory of ASB, is that extraversion/sociability (E andPEM) appears, at best, weakly related to ASB.

5.1. Extraversion/sociability Wndings

Like prior reviews (e.g., Cochrane, 1974; Farrington et al., 1982; Feldman, 1993;Furnham & Thompson, 1991; Miller & Lynam, 2001; Passingham, 1972), the presentWndings indicate a weak to negligible association between E and ASB. However, theydo not necessarily negate Eysenck’s hypotheses anchored in arousal theory. Manyresearchers have found that suboptimal arousal is related to ASB (Ellis, 1987; Raine,1997a). Whether suboptimal arousal is linked to E, as Eysenck asserted, should befurther examined. Eysenck (1990) reviewed studies of electroencephalogram, skinconductance, and other laboratory measures’ associations with E and argued thatsupport for an association between low arousal and E is strong, but a closer look atthese Wndings (see Zuckerman, 1997) suggests that the evidence is mixed. Eysenckacknowledged that the impulsivity components of E’s original conceptualizationwere most related to poor conditionability (Eysenck & Levey, 1972; Eysenck &Eysenck, 1971a, 1971b), and there is evidence that impulsivity is more related to lowarousal than is E (see Zuckerman, 1991, 1997).

The association between extraversion/sociability and ASB was stronger for self-reported ASB design samples than known groups design samples, and this was con-sistent with Miller and Lynam’s (2001) Wndings. Empirical evidence suggesting thatincarceration decreases levels of E, which might limit the variance of E in these popu-lations, is one explanation for this diVerence because incarcerated individuals areoften participants in known groups deigns studies. It is important to consider thatmany extraversion/sociability items may be irrelevant in forensic settings becausethey apply to social activities and interactions that are prohibited, thus restricting therange of these scores for incarcerated samples. Longitudinal studies, which examinepersonality features before and after incarceration, can help researchers ascertainwhether levels of extraversion/sociability change during incarceration.

It is also possible that some individuals who engage in ASB avoid being arrestedand convicted due to E tendencies (e.g., assertive and engaging). Moreover, highextraversion/sociability antisocial individuals may be more likely than low extraver-sion/sociability antisocial individuals to engage in antisocial activities that do notlead to arrests. For example, individuals may lie, manipulate, and take advantage ofothers, often times in very covert ways. Individuals who engage in such behaviorsmight have a lower risk of incarceration, compared with individuals who exhibitmore overt violent and/or destructive behaviors. Eysenck, Rust, and Eysenck (1977)examined EPQ scores among diVerent criminal types and found that “con men” (i.e.,inmates who committed nonviolent oVenses, such as fraud) had higher E scores thanproperty oVenders. In addition, many self-report delinquency questionnaires used inself-reported ASB design studies (see Furnham & Thompson, 1991), include itemsassessing behaviors that are relatively less likely to lead to arrests and convictions


(e.g., using illegal substances before the age of 16, lying, isolated incidents of creditcard fraud). It is important to bear in mind that such behaviors do not necessarilydistinguish nonincarcerated from incarcerated individuals, as many incarceratedindividuals also have a history of these behaviors. However, compared with otherantisocial behaviors, the relatively subtle, covert antisocial behaviors may be moreassociated with E.

Among known groups design samples, extraversion/sociability was signiWcantlymore related to ASB as age increased. However, when age was examined categori-cally, it was not a signiWcant moderator. Because the within groups heterogeneity washighest for the adult samples, it is possible that diVerences among the adult samplesaccount for this discrepancy. Perhaps because overall rates of ASB decline with age(Farrington, 1994, 1995), older prisoners may require more factors, including person-ality features and other predictors of ASB, to contribute to their propensities towardsASB than do younger prisoners. Among self-reported ASB design samples, extraver-sion/sociability was less related to ASB as age increased, and the association betweenextraversion/sociability and ASB was highest among the adolescent samples. Theselatter Wndings provided some support for Eysenck’s prediction that E is morestrongly related to ASB in young populations than in adults (Eysenck & Gudjons-son, 1989).

Developmental trajectories of ASB have been posited to diVer between males andfemales (Keenan & Shaw, 1997; Silverthorn & Frick, 1999), and these diVerences sug-gest possible age by biological sex interaction eVects. The interaction analyses in thepresent review suggest that the association between extraversion/sociability and ASBmay be especially strong for younger males compared with older males and/orfemales. However, this interpretation is speculative, and further investigation of theseinteraction eVects is warranted.

The association between extraversion/sociability and ASB was stronger for Emeasures than PEM measures, and it is possible that these two scales assess slightlydiVerent features. E includes not only sociability and assertiveness (as does PEM),but also sensation seeking (Eysenck, 1996a), and sensation seeking is particularlyrelated to a proclivity towards ASB (Zuckerman, 1997). Data indicating that PEMdivides into Agentic PEM and Communal PEM also oVer some hypotheses concern-ing this diVerence. While aspects of Agentic PEM (e.g., social potency and achieve-ment) may not account for the diVerences observed across personality measures, it isplausible that social closeness, a predominant characteristic of Communal PEM, isnegatively related to ASB. Thus, the social closeness component of PEM mayweaken the association found between PEM and ASB.

5.1.1. Neuroticism/emotionality WndingsThe present Wndings are largely consistent with prior reviews (e.g., Feldman, 1993;

Furnham & Thompson, 1991; Miller & Lynam, 2001), which reported mixed supportfor an association between N and ASB. Consistent with Eysenck’s assumption that Nplays more of a role in adulthood ASB than in childhood ASB, age signiWcantlymoderated the association between neuroticism/emotionality and ASB, such that thisassociation was stronger as age increased for nearly all the analyses. Among known


groups designs samples, however, the association between neuroticism/emotionalityand ASB was negligible for adolescents, but it was moderate in magnitude for chil-dren. Researchers may consider examining whether this dimension is a particularlystrong predictor of ASB (e.g., school reported ASB) in young populations. Eysenck(1996a) and Eysenck and Gudjonsson (1989) reviewed studies in which N was moreassociated with ASB in adults than in juveniles. However, in these reviews, most ofthe studies of adults used known groups design approaches and most studies of juve-niles used self-reported ASB design approaches. The diVerences in neuroticism/emo-tionality and ASB relations across these groups appeared to be confounded by theoperationalization of ASB in these reviews. Moreover, these reviews did not examinechildren and adolescent samples separately.

The association between neuroticism/emotionality and ASB was stronger forNEM than for N. Alienation and aggression, which are two NEM subscales, are par-ticularly relevant to ASB because they represent tendencies toward hostility (Telle-gen & Waller, 1994). As noted previously, other researchers (e.g., Caspi et al., 1994;Krueger et al., 1994) did not Wnd that aggression items of the NEM scale substan-tially aVected the correlations between MPQ and ASB measures. Nonetheless, thesigniWcant moderator eVect of personality measure on the association between neu-roticism/emotionality and ASB, found among the self-reported ASB design samplesin this review, suggests that further evaluation of this issue is warranted.

5.2. Impulsivity/disinhibition Wndings

The association between impulsivity/disinhibition and ASB was moderate in mag-nitude for the known groups design samples and was large in magnitude for the self-reported ASB design samples, although the diVerence between these associations wasnonsigniWcant. One explanation for the slight diVerence across ASB operationaliza-tions is that, compared with low impulsivity/disinhibition individuals, high impulsiv-ity/disinhibition individuals are more likely to self-report levels of ASB because theyare less inhibited when self-reporting socially undesirable features. Consequently, thisresponse bias could inXate the correlations between impulsivity/disinhibition andASB (see also, Furnham, 1984; Furnham & Thompson, 1991).

Additional results indicated that among self-reported ASB design samples, impul-sivity/disinhibition was less related to ASB as age increased. When age was examinedcategorically, this association was strongest for the adolescent samples. This suggeststhat in nonincarcerated populations, impulsivity/disinhibition exhibits a strong inXu-ence on ASB manifestations in adolescence. Moreover, because impulsivity tends todecline with age (Farrington, 1994, 1995; Hare, 1996), impulsivity/disinhibition mightplay less of a role in ASB patterns of nonincarcerated individuals as they age.

The results also suggest that the association between impulsivity/disinhibition andASB is stronger for the P dimension than the CON (reversed) dimension. The factthat some P scale items directly assess ASB tendencies may account for this diVer-ence. Based on the consistency of the Wndings in this review and Eysenck’s descrip-tion of the P scale (e.g., Eysenck, 1996a), the robust association found betweenimpulsivity/disinhibition and ASB does not appear to merely reXect tautological


assessments of these constructs. It is also important to note that the MPQ’s CONscale does not have overtly antisocial items. To address the issue of predictor-crite-rion contamination, researchers should select personality measures that do not con-tain explicit ASB items or should statistically control for such items.

5.3. Limitations

Several limitations indicate that the moderator analyses results should be consid-ered provisional. One limitation is that because the known groups design studiesdiVered widely in their dichotomization of ASB, it was diYcult to compare relationsamong the known groups design studies and across ASB operationalizations.Another limitation is that most of the self-reported ASB design studies included non-incarcerated individuals, whereas a few samples included only identiWed criminals.These diVerences likely contributed to the heterogeneity found in these groups ofstudies. It has been suggested that researchers consider using a broad range of antiso-cial individuals to examine the extent to which personality and ASB relations gener-alize across all populations (see Rutter, Giller, & Hagel, 1998). In addition toincluding both known groups and self-reported ASB operationalizations within stud-ies, researchers may consider operationalizing ASB in relatively speciWc units (e.g.,number of arrests or convictions).

In this review, the Wndings for EPQ studies were largely congruent with Wndingsfor MPQ studies, which supported the presumption that extraversion/sociability,neuroticism/emotionality, and impulsivity/disinhibition represent common higherorder personality dimensions. The results from the personality measure moderatoranalyses also corroborated the results of Miller and Lynam’s (2001) review.Although there are advantages to examining the relations of broader personalitydimensions, one drawback is that lower level factors may exert unique, even oppos-ing, eVects in predicting ASB. Therefore, important lower level factors may be hiddenwhen examining only the relationships for the broader dimensions.

Another limitation of this review is that only one of the 14 MPQ studies used aknown groups design. It is important to consider whether the inXuence of personalitymeasure accounted for diVerences observed across ASB operationalizations, andconversely, whether the ASB operationalization accounted for diVerences observedbetween EPQ and MPQ studies. Examination of the moderator results does not sug-gest that a confound between ASB operationalization and personality measurementsigniWcantly inXuenced the Wndings. The weighted mean eVect sizes for extraversion/sociability and impulsivity/disinhibition were larger for the self-reported ASB designstudies than the known groups design studies. If personality measure was accountingfor these diVerences, then the weighted mean eVect sizes would be expected to besmaller for the self-reported ASB design studies than the known groups design stud-ies because the weighted mean eVect sizes for the MPQ studies were smaller thanthose for the EPQ studies. For neuroticism/emotionality, the weighted mean eVectsizes were larger for the self-reported ASB design studies than for the known groupsdesign studies and the weighted mean eVect sizes were larger for the MPQ studiesthan for the EPQ studies. However, these diVerences were nonsigniWcant, as indicated


by the moderator analyses for ASB operationalization and personality measurementfor all the samples.

Compared with prior reviews, the moderator analyses were relatively speciWcbecause age was examined as both a continuous and a categorical moderator variableand because children and adolescent samples were examined separately. However,for the child and adolescent samples, the age ranges were typically narrow (i.e., withina couple of years), whereas the age ranges were often quite large in the adult samples.To disentangle potential age diVerences in personality and ASB relations, researchersshould attempt to include a narrow age range in a given sample. Another limitation isthat there were few child samples compared with the number of adolescent and adultsamples. In addition, studies of children also used diVerent methods of operationaliz-ing ASB compared with other studies (e.g., school-reported ASB versus incarcera-tion), which complicated the interpretations of Wndings across age groups.

In this review, biological sex was not a signiWcant moderator of the associationsbetween the personality dimensions and ASB. However, there were relatively fewstudies of females, particularly among the known groups design studies, which lim-ited these analyses as well as the age by biological sex interaction analyses. Addition-ally, many researchers did not report the number of males and females in theirsamples.

Unfortunately, race could not be examined in this review because over half of thestudies did not report this information. DiVerences in personality and ASB relationsacross races and ethnicities are largely unknown, particularly in adult populations(Farrington, 1987), but a few studies of younger populations reveal mixed Wndings inthis regard. Caspi et al. (1994) compared the correlations between MPQ scores andASB measures of Black adolescents with those of White adolescents, and did not Wndsubstantial diVerences. Cooper, Wood, Orcutt, and Albino (2003) investigated per-sonality variables as predictors of risky or problem behaviors (including delin-quency) in a large sample of Black and White adolescents, and found that thrillseeking was signiWcantly related to a higher order problem behavior factor for Whiteadolescents but not Black adolescents. These Wndings suggest that researchers shouldconsider the eVects of race when investigating personality and ASB relations.

5.4. Future model testing and implications

Eysenck (1964, 1977) proposed linear relations between his higher order dimen-sions and ASB, and he viewed the dimensions as having independent, additive eVectson behavior (Eysenck & Eysenck, 1971b). As noted previously, he also described N asan ampliWer of E, suggesting an interactive eVect. Although there have been someeVorts at interpreting additive eVects of personality on ASB (e.g., Allsopp & Feld-man, 1976), interactive eVects among personality dimensions in predicting ASB arerelatively unknown. Burgess (1972) reported preliminary evidence for the hypothesisthat that the product of E and N predicts ASB better than the independent eVects of E orN. Caspi et al. (1994) recommended that personality variables be simultaneously exam-ined as predictors of ASB, and they further implied that NEM and CON may exhibit aninteractive eVect in this causal relationship. Related to this, Cooper et al., (2003)


examined interactions among the personality variables in their study, and they foundthat impulsivity and avoidance coping (with negative emotions) exhibited a margin-ally signiWcant interaction eVect to predict a higher order problem behavior factor. Itis recommended that researchers of ASB examine the interactive eVects of higherorder personality dimensions, in addition to the independent eVects of thesedimensions.

Causal model testing is further recommended to advance the understanding ofpersonality and ASB relations, and social variables (e.g., SES, childhood experiences)should be incorporated in causal model testing. Eysenck (1996a) explained that socialfactors can inXuence behavior diVerently, based on individual diVerences in personal-ity. For example, people respond to a given social environment diVerently, due todiVerences in personality. Moreover, because personality factors can directly inXu-ence one’s social conditions (e.g., personality features can inXuence one’s choice ofexperiences and social environment), they can act through such variables to inXuencebehavior.

Eysenck stated that although ASB individuals have inherited central and auto-nomic nervous systems that are ASB-prone, some eVorts may eVectively reducecriminality, and he discussed the importance of considering personality factorswhen outlining treatment interventions for antisocial individuals (Eysenck, 1964,1977, 1996b). Because high levels of impulsivity/disinhibition appear to increaseone’s risk for ASB, interventions may be designed to identify individuals withthese features and teach them ways to eVectively control their impulses. In addi-tion, high neuroticism/emotionality individuals may be taught ways in which tocope with negative emotions. Presently, numerous treatment interventions con-centrate on developing eVective problem solving and coping skills. The primaryWndings of this review support the use of such interventions in treating antisocialindividuals.

The impact of Eysenck’s theory of personality on the study of ASB cannot beoveremphasized. His biosocial perspective has generated a number of hypothesesconcerning the interplay among personality, methodological, and demographic vari-ables in predicting crime. The Wndings and unanswered questions of this review canspur future model building and hypotheses about personality and ASB relations.Furthermore, modiWcations of Eysenck’s comprehensive framework may bear theo-retical and practical importance in the study of ASB.

Acknowledgments

I thank Brian Agnetta, Arin Connell, Frans de Waal, Marshall Duke, Molly Emb-ree, Eugene Emory, Denise Glanville, Erin McClure, and Lynne Nygaard for theirhelpful feedback on earlier drafts of the manuscript. I also thank Anna Dåderman,David Hemsley, Patrick Heaven, Kerry Jang, Joe Newman, and Don Saklofske forproviding data and/or information for studies they conducted. In particular, I thankmy doctoral advisor Scott Lilienfeld for providing data for this review and for hisguidance in its formulation and writing.


References

References preceded by an asterisk indicate studies that were included in the meta-analyses.

Abram, K. M. (1989). The eVect of co-occurring disorders on criminal careers: Interaction of antisocialpersonality, alcoholism, and drug disorders. International Journal of Law and Psychology, 12, 133–149.

Achenbach, T. M. (1991). Youth self-report for ages 11–18. Burlington, VT: University of Vermont.*Af Klinteberg, B., Humble, K., & Schalling, D. (1992). Personality and psychopathy of males with a his-

tory of early criminal behaviour. European Journal of Personality, 6, 245–266.*Aleixo, P. A., & Norris, C. E. (2000). Personality and moral reasoning in young oVenders. Personality and

Individual DiVerences, 28, 609–623.Allsopp, J. F., & Feldman, M. P. (1974). Extraversion, neuroticism, psychoticism and antisocial behavior in

schoolgirls. Social Behavior and Personality, 2, 184–190.Allsopp, J., & Feldman, M. (1976). Personality and antisocial behaviour in schoolboys: Item analysis of

questionnaire measures. British Journal of Criminology, 16, 337–351.American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders. (4th ed.).

Washington, DC: Author.American Psychiatric Association (2000). Diagnostic and statistical manual of mental disorders. (4th ed.

Text Revision). Washington, DC: Author.*Barack, L. I., & Widom, C. S. (1978). Eysenck’s theory of criminality applied to women awaiting trial.

British Journal of Psychiatry, 133, 452–456.*Bartol, C. R., & Holanchock, H. A. (1979). Eysenck’s theory of criminality: A test on an American pris-

oner population. Criminal Justice and Behavior, 6, 245–249.Burgess, P. K. (1972). Eysenck’s theory of criminality: A new approach. British Journal of Criminality, 12,

74–82.Butcher, J. N., Graham, J. R., Williams, C. L., & Ben-Porath, Y. S. (1990). Development and use of the

MMPI-2 content scales. Minneapolis: University of Minnesota Press.*Caspi, A., MoYtt, T. E., Silva, P. A., Stouthamer-Loeber, M., Krueger, R. F., & Schmutte, P. S. (1994). Are

some people crime prone? Replications of the personality–crime relationship across countries, genders,races, and methods. Criminology, 32, 163–195.

Church, A. T. (1994). Relating the Tellegen and Five-Factor models of personality structure. Journal ofPersonality and Social Psychology, 67, 898–909.

Cleckley, H. (1941/1988). The mask of sanity. St. Louis, MO: Mosby.Cloninger, C. R. (1978). The link between hysteria and sociopathy: An integrative model of pathogenesis

based on clinical, genetic, and neurophysiological observations. In H. S. Akiskal & W. L. Webb (Eds.),Psychiatric diagnosis: Exploration of biological predictors (pp. 189–218). New York: Spectrum.

Cochrane, R. (1974). Crime and personality: Theory and evidence. Bulletin of the British PsychologicalSociety, 27, 19–22.

Cohen, J. (1988). Statistical power analysis for the behavioral Sciences (2nd ed.). Hillsdale, NJ: Erlbaum.Cooper, M. L., Wood, P. K., Orcutt, H. K., & Albino, A. (2003). Personality and the predisposition to

engage in risky or problem behaviors during adolescence. Journal of Personality and Social Psychology,84, 390–410.

Costa, P. T., Jr., & McCrae, R. R. (1995). Primary traits of Eysenck’s P-E-N system: Three- and Wve-factorsolutions. Journal of Personality and Social Psychology, 69, 308–317.

*Dåderman, A. M. (1999). DiVerences between severely conducted-disordered juvenile males and normaljuvenile males: The study of personality traits. Personality and Individual DiVerences, 26, 827–845.

Ellis, L. (1987). Relationships of criminality and psychopathy with eight other apparent behavioral mani-festations of suboptimal arousal. Personality and Individual DiVerences, 8, 905–925.

Eysenck, H. J. (1964). Crime and personality. Boston: Houghton MiZin.Eysenck, H. J. (1972). Human typology, higher nervous activity, and factor analysis. In V. D. Nebylitsin &

J. A. Gray (Eds.), Biological bases of individual behavior (pp. 165–181). New York: Academic Press.Eysenck, H. J. (1977). Crime and personality (3rd ed.). London: Routledge & Kegan Paul.Eysenck, H. J. (1990). Biological dimensions of personality. In L. A. Pervin (Ed.), Handbook of personality:

Theory and research (pp. 244–276). New York: Guilford.


Eysenck, H. J. (1991). Dimensions of personality: 16, 5, or 3? – criteria for taxonomic paradigm. Personal-ity and Individual DiVerences, 12, 773–790.

Eysenck, H. J. (1996a). Personality and crime: Where do we stand. Psychology, Crime & Law, 2, 143–152.Eysenck, H. J. (1996b). Personality theory and the problem of criminality. In J. Muncie & J. McLaughlin

(Eds.), Criminological perspectives: A reader (pp. 81–98). London: Sage.Eysenck, H. J., & Eysenck, S. B. G. (1975). Manual of the Eysenck personality questionnaire (junior and

adult). London: Hodder & Stoughton.Eysenck, H. J., & Eysenck, S. B. G. (1976). Psychoticism as a dimension of personality. London: Hodder &

Stoughton.Eysenck, H. J., & Eysenck, S. B. G. (1994). Manual of the Eysenck personality questionnaire. San Diego:

Educational and Industrial Testing Services.Eysenck, H. J., & Gudjonsson, G. H. (1989). The causes and cures of criminality. New York: Plenum.Eysenck, H. J., & Levey, A. (1972). Conditioning, introversion-extraversion and the strength of the nervous

system. In V. D. Nebylitsin & J. A. Gray (Eds.), Biological bases of individual behavior (pp. 206–220).New York: Academic Press.

*Eysenck, S. B. G. (1981). Impulsiveness and antisocial behaviour in children. Current PsychologicalResearch, 1, 31–37.

Eysenck, S. B. G., & Eysenck, H. J. (1970). Crime and personality: An empirical study of the three-factortheory. British Journal of Criminology, 10, 225–239.

Eysenck, S. B. G., & Eysenck, H. J. (1971a). A comparative study of criminals and matched controls onthree dimensions of personality. British Journal of Social and Clinical Psychology, 10, 362–366.

Eysenck, S. B. G., & Eysenck, H. J. (1971b). Crime and personality: Item analysis of questionnaireresponses. British Journal of Criminology, 11, 49–62.

*Eysenck, S. B. G., & Eysenck, H. J. (1977). Personality diVerences between prisoners and controls. Psycho-logical Reports, 40, 1023–1028.

Eysenck, S. B. G., Eysenck, H. J., & Barrett, P. (1985). A revised version of the psychoticism scale. Personal-ity and Individual DiVerences, 6, 21–29.

Eysenck, S. B. G., Rust, J., & Eysenck, H. J. (1977). Personality and the classiWcation of adult oVenders.British Journal of Criminology, 17, 169–179.

*Eysenck, S. B. G., White, O., & Eysenck, H. J. (1976). Personality and mental illness. PsychologicalReports, 39, 1011–1022.

Farrington, D. P. (1987). Epidemiology. In H. C. Quay (Ed.), Handbook of juvenile delinquency (pp. 33–61).New York: Wiley.

Farrington, D. P. (1994). Human development and criminal careers. In M. Maguire, R. Morgan, & R.Reiner (Eds.), The Oxford handbook of criminology (pp. 511–584). New York: Oxford University Press.

Farrington, D. P. (1995). The development of oVending and antisocial behaviour from childhood: KeyWndings from the Cambridge study in delinquent development. Journal of Child Psychology and Psychi-atry, 360, 929–964.

Farrington, D. P., Biron, L., & LeBlanc, M. (1982). Personality and delinquency in London and Montreal.In J. Gunn & D. P. Farrington (Eds.), Abnormal oVenders, delinquency, and the criminal justice system(pp. 153–201). New York: Wiley.

Feldman, P. (1993). The psychology of crime: A social science textbook. New York: Cambridge UniversityPress.

*Fonseca, A. C., & Yule, W. (1995). Personality and antisocial behavior in children and adolescents: Anenquiry into Eysenck’s and Gray’s theories. Journal of Abnormal Child Psychology, 23, 767–781.

Forrest, R. (1977). Personality and delinquency: A multivariate examination of Eysenck’s theory withScottish delinquent and non-delinquent boys. Social Behavior and Personality, 5, 157–167.

Fowles, D. C. (1987). Application of a behavioral theory of motivation to the concepts of anxiety andimpulsivity. Journal of Research in Personality, 21, 417–435.

*Furnham, A. (1984). Personality, social skills, anomie and delinquency: A self-report study of a group ofnormal non-delinquent adolescents. Journal of Child Psychology and Psychiatry and Allied Disciplines,25, 409–420.

*Furnham, A., & Barratt, L. (1988). Moral judgment and personality diVerences in “problem children”and delinquents. Personality and Individual DiVerences, 9, 187–188.


*Furnham, A., & Thompson, J. (1991). Personality and self-reported delinquency. Personality and Individ-ual DiVerences, 12, 585–593.

Gabrys, J. B., Peters, K., Robertson, G., Utendale, K. A., Schumph, D., Laye, R. C., et al. (1988). Personal-ity attributes of children with conduct disorders: The discriminant power of the Junior Eysenck Per-sonality Questionnaire. Psychological Reports, 62, 63–69.

*Ge, X., & Conger, R. D. (1999). Adjustment problems and emerging personality characteristics from earlyto late adolescence. American Journal of Community Psychology, 27, 429–459.

*Gomà-i-Freixanet, M. (1995). Prosocial and antisocial aspects of personality. Personality and IndividualDiVerences, 19, 125–134.

*Gomà-i-Freixanet, M. (2001). Prosocial and antisocial aspects of personality in women: A replicationstudy. Personality and Individual DiVerences, 30, 1401–1411.

*Gossop, M. (1978). Drug dependence, crime and personality among female addicts. Drug and AlcoholDependence, 3, 359–364.

*Gossop, M. R., & Eysenck, S. B. G. (1983). A comparison of the personality of drug addicts in treatmentwith that of a prison population. Personality and Individual DiVerences, 41, 207–209.

Gough, H. G. (1960). Theory and measurement of socialization. Journal of Consulting and Clinical Psychol-ogy, 24, 23–30.

*Haapasalo, J. (1990). Sensation seeking and Eysenck’s personality dimensions in an oVender sample. Per-sonality and Individual DiVerences, 11, 81–84.

Hare, R. D. (1985). Comparison procedures for the assessment of psychopathy. Journal of Consulting andClinical Psychology, 53, 7–16.

Hare, R. D. (1996). Psychopathy: A clinical construct whose time has come. Criminal Justice and Behavior,23, 25–54.

*Heaven, P. C. L. (1993). Personality predictors of self-reported delinquency. Personality and IndividualDiVerences, 14, 67–76.

*Heaven, P. C. L. (1994). Family of origin, personality, and self-reported delinquency. Journal of Adoles-cence, 17, 445–459.

Heaven, P. C. L. (1996). Personality and self-reported delinquency: Analysis of the “Big Five” personalitydimensions. Personality and Individual DiVerences, 20, 47–54.

*Heaven, P. C. L., Caputi, P., Trivellion-Scott, D., & Swinton, T. (2000). Personality and group inXuenceson self-reported delinquent behaviour. Personality and Individual DiVerences, 28, 1143–1158.

*Heaven, P. C. L., & Virgen, M. (2001). Personality, perceptions of family and peer inXuences, and males’self-reported delinquency. Personality and Individual DiVerences, 30, 321–331.

Hedges, L. V. (1994). Fixed eVects models. In H. Cooper & L. V. Hedges (Eds.), The handbook of researchsynthesis (pp. 285–299). New York: Sage.

Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. San Diego: Academic.Hedges, L. V., & Vevea, J. L. (1998). Fixed- and random-eVects models in meta-analysis. Psychological

Methods, 3, 486–504.Henry, B., Caspi, A., MoYtt, T. E., & Silva, P. A. (1996). Temperamental and familial predictors of violent

and nonviolent criminal convictions: Age 3 to age 18. Developmental Psychology, 32, 614–623.Hyler, S. E., & Rieder, R. O. (1987). Personality diagnostic questionnaire—revised. New York: New York

State Psychiatric Institute.Jaccard, J., Turrisi, R., & Wan, C. K. (1990). Interaction eVects in multiple regression. Newbury, CA: Sage.*Jamison, R. N. (1980). Psychoticism, deviancy and perception of risk in normal children. Personality and

Individual DiVerences, 1, 87–91.*Jang, K. L., Livesley, W. T., & Vernon, P. A. (1999). The relationship between Eysenck’s P-E-N model of

personality traits delineating personality disorder. Personality and Individual DiVerences, 26, 121–128.Keenan, K., & Shaw, D. (1997). Developmental and social inXuences on young girls’ early problem behav-

ior. Psychological Bulletin, 121, 95–113.*Krueger, R. F., Hicks, B. M., & McGue, M. (2001). Altruism and antisocial behavior: Independent ten-

dencies, unique personality correlates, distinct etiologies. Psychological Science, 12, 397–402.*Krueger, R. F., Schmutte, P. S., Caspi, A., MoYtt, T. E., Campbell, K., & Silva, P. A. (1994). Personality

traits are linked to crime among men and women: Evidence from a birth cohort. Journal of AbnormalPsychology, 103, 328–338.


*Lane, D. A. (1987). Personality and antisocial behaviour: A long-term study. Personality and IndividualDiVerences, 8, 799–806.

Lilienfeld, S. O. (1994). Conceptual problems in the assessment of psychopathy. Clinical PsychologyReview, 14, 17–38.

*Lilienfeld, S. O. (2000). MPQ and ASB data in undergraduate samples. Unpublished raw data.Lipsey, M. W. (1990). Design sensitivity: Statistical power for experimental research. New York: Sage.*Ma, H. K., Shek, D. T. L., Cheung, P. C., & Lee, R. Y. P. (1996). The relation of prosocial and antisocial

behavior to personality and peer relations of Hong Kong Chinese adolescents. Journal of Genetic Psy-chology, 157, 255–266.

*Maliphant, R., Hume, F., & Furnham, A. (1990). Autonomic nervous system (ANS) activity, personalitycharacteristics and disruptive behaviour in girls. Journal of Child Psychology and Psychiatry and AlliedDisciplines, 31, 619–628.

*Maliphant, R., Watson, S., & Daniels, D. (1990). Disruptive behaviour in school, personality characteris-tics and heart rate (HR) levels in 7- to 9-year-old boys. Educational Psychology, 10, 199–205.

*McGue, M., Slutske, W., Taylor, J., & Iacono, W. G. (1997). Personality and substance use disorders: I.EVects of gender and alcoholism subtype. Alcoholism: Clinical and Experimental Research, 21, 513–520.

*McGurk, B. J., & Bolton, N. (1981). A comparison of the Eysenck Personality Questionnaire and the Psy-chological Screening Inventory in a delinquent sample and a comparison group. Journal of ClinicalPsychology, 37, 874–879.

McKinley, J., & Hathaway, S. R. (1944). The MMPI: Hysteria, hypomania, and psychopathic deviate.Journal of Applied Psychology, 28, 153–174.

Meehl, P. E. (1971). High school yearbooks: A reply to Schwarz. Journal of Abnormal Psychology, 77,143–148.

Miller, J. D., & Lynam, D. R. (2001). Structural models of personality and their relation to antisocialbehavior: A meta-analytic review. Criminology, 39, 765–798.

MoYtt, T. E. (1993). Adolescence-limited and life-course-persistent and antisocial behavior: A develop-mental taxonomy. Psychological Review, 100, 674–701.

*Mynard, H., & Joseph, S. (1997). Bully/victim problems and their association with Eysenck’s personalitydimensions in 8 to 13 year-olds. British Journal of Educational Psychology, 67, 51–54.

*Nagoshi, C. T., Walter, D., Muntaner, C., & Haertzen, C. A. (1992). Validation of the Tridimensional Per-sonality Questionnaire in a sample of male drug users. Personality and Individual DiVerences, 13, 401–409.

*Newman, J. P. (2000). MPQ and ASB data in a female inmate sample. Unpublished raw data.Passingham, R. E. (1972). Crime and personality: A review of Eysenck’s theory. In V. D. Nebylitsin & J. A.

Gray (Eds.), Biological bases of individual behavior (pp. 342–371). New York: Academic.Patrick, C. J., Curtin, J. J., & Tellegen, A. (2002). Development and validation of a brief form of the Multi-

dimensional Personality Questionnaire. Psychological Assessment, 14, 150–163.*Powell, G. E. (1977). Psychoticism and social deviancy in children. Advances in Behaviour Research and

Therapy, 1, 27–56.*Putnins, A. L. (1982). The Eysenck Personality Questionnaires and delinquency prediction. Personality

and Individual DiVerences, 3, 339–340.*Rahman, A. (1992). Psychological factors in criminality. Personality and Individual DiVerences, 13,

483–485.*Rahman, A., & Husain, A. (1984). Personality and female criminals in Bangladesh. Personality and Indi-

vidual DiVerences, 5, 473–474.Raine, A. (1997a). Antisocial behavior and psychophysiology: A biosocial perspective and a prefrontal

dysfunction hypothesis. In D. M. StoV, J. Breiling, & J. D. Maser (Eds.), Handbook of antisocial behav-ior (pp. 289–304). New York: Wiley.

Raine, A. (1997b). Classical conditioning, arousal, and crime: A biosocial perspective. In H. Nyborg (Ed.), ThescientiWc study of human nature: Tribute to Hans J. Eysenck at eighty (pp. 122–141). Oxford: Pergamon.

*Rangaswami, K., & Arunagini, S. (1982). Personality dimensions of criminals and mentally-ill criminals.Indian Journal of Psychiatry, 24, 88–91.

*Romero, E., Luengo, M. A., & Sobral, J. (2001). Personality and antisocial behaviour: Study of tempera-mental dimensions. Personality and Individual DiVerences, 31, 329–348.


Rosenthal, R. (1991). Meta-analytic procedures for social research. Newbury Park: Sage.Rosenthal, R. (1994). Parametric measures of eVect size. In H. Cooper & L. V. Hedges (Eds.), The handbook

of research synthesis (pp. 231–244). New York: Sage.*Rushton, J. P., & Chrisjohn, R. D. (1981). Extraversion, neuroticism, psychoticism and self-reported

delinquency: Evidence from eight separate samples. Personality and Individual DiVerences, 2, 11–20.Rutter, M., Giller, H., & Hagel, A. (1998). Antisocial behavior by young people. Cambridge: Cambridge

University Press.*Saklofske, D. H. (1977). Personality and behavior problems of schoolboys. Psychological Reports, 41,

445–446.*Saklofske, D. H., & Eysenck, S. B. G. (1980). Personality and antisocial behavior in delinquent and non-

delinquent boys. Psychological Reports, 47, 1255–1261.Schwarzer, R. (1989). Meta-analysis programs, version 5.3. Freie Universitat Berlin: Author.*Seisdedos, N. (1982). La coducta antisocial de los adolescentes. In D. H. Saklofske & S. B. G. Eysenck

(Eds.), Individual diVerences in children and adolescents 1998 (pp. 209–226). New Brunswick: Transac-tion.

Sher, K. J., & Trull, T. J. (1994). Personality and disinhibitory psychopathology: Alcoholism and antisocialpersonality disorder. Journal of Abnormal Psychology, 103, 92–102.

*Silva, F., Martorell, C., & Clemente, A. (1986). Socialization and personality. Study through question-naires in a preadult Spanish population. Personality and Individual DiVerences, 7, 355–372.

Silverthorn, P., & Frick, P. J. (1999). Developmental pathways to antisocial behavior: The delayed-onsetpathway in girls. Development and Psychopathology, 11, 101–126.

*Sipos, K., Bujdos, L., Bangó, L., Zakar, M., & Sipos, M. (1998). Some psychological characteristics ofinstitutionalized delinquent teenage boys in Hungary. In D. H. Saklofske & S. B. G. Eysenck (Eds.),Individual diVerences in children and adolescents 1998 (pp. 86–98). New Brunswick: Transaction.

*Slee, P. T., & Rigby, K. (1993). The relationship of Eysenck’s personality factors and self-esteem to bully-victim behaviour in Australian schoolboys. Personality and Individual DiVerences, 14, 371–373.

*Stewart, C. H. M., & Hemsley, D. R. (1984). Personality factors in the taking of criminal risks. Personalityand Individual DiVerences, 5, 119–122.

*Taylor, J. E. (1999). Early and late starting delinquency: Correlates, outcomes, and inXuences. DoctoralDissertation, University of Minnesota, Minneapolis.

Tellegen, A. (1982). Brief manual for the Multidimensional Personality Questionnaire. Minneapolis: Univer-sity of Minnesota Press.

Tellegen, A., & Waller, N. G. (1994). Exploring personality through test construction: Development of theMultidimensional Personality Questionnaire. In S. R. Briggs & J. C. Cheek (Eds.), Personality mea-sures: Development and evaluation (Vol. 1, pp. 133–161). Greenwich: JAI.

Waller, N. G., Lilienfeld, S. O., Tellegen, A., & Lykken, D. T. (1991). The Tridimensional Personality Ques-tionnaire: Structural validity and comparison with the Multidimensional Personality Questionnaire.Multivariate Behavioral Research, 26, 1–23.

West, D. J., & Farrington, D. P. (1977). The delinquent way of life. London: Heinemann.Wilson, J. Q., & Herrnstein, R. J. (1985). Crime and human nature. New York: Simon & Schuster.Zuckerman, M. (1991). Psychobiology of personality. New York: Cambridge University Press.Zuckerman, M. (1997). The psychobiological basis of personality. In H. Nyborg (Ed.), The scientiWc study

of human nature: Tribute to Hans J. Eysenck at eighty (pp. 3–16). Oxford: Pergamon.Zuckerman, M., Kuhlman, D. M., Joireman, J., Teta, P., & Kraft, M. (1993). A comparison of three struc-

tural models of personality: The Big Three, the Big Five, and the Alternative Five. Journal of Personal-ity and Social Psychology, 65, 757–768.

a quantitative review of the relations between the “big 3 ... · tellegen’s test construction...

Documents