A Genetic and Environmental Analysisof the California PsychologicalInventory Using Adult Twins RearedApart and TogetherTHOMAS J. BOUCHARD, Jr.1* MATT MCGUE1, YOON-MI HUR2,and JOSEPH M. HORN31Department of Psychology and Institute of Human Genetics,University of Minnesota, Minneapolis, MN 55455, USA2Dongjak-Ku, Shindaebangl-Dong, Hyundai APT 102±1707,Seoul 156-010, Sejopng University, Seoul, Korea3Department of Psychology, University of Texas, Austin,TX 78712, USAAbstractThe California Psychological Inventory (CPI) was administered to a sample of71 pairs of monozygotic and 53 pairs of dizygotic twins reared apart (MZA, DZA,adult twins) and 99 pairs of monozygotic and 99 pairs of dizygotic twins rearedtogether (MZT, DZT, adult male twins). The twin reared apart data was age and sexcorrected. The twin reared together data represented one sex and a narrow age rangeand was not standardized. The CPI was scored using the 1996 scori ng keys, for its 20folk scales, three vector scales, and 11 special purpose scales. The correlations for thefour groups were subjected to model-®tting and the following four parametersestimated: (i) additive genetic variance (Va), (ii) dominance variance (Vd), (iii)shared environmental variance (Vc), and (iv) idiosyncratic environmental variance(Ve). This design has considerable power to detect Vc. The average estimate of Vcforall variables was essentially zero. The average estimate of genetic in¯uence (Va Vd)for all variables was 0.46. Consistent with these results either the MZT or the MZAcorrelations alone would have provided quite good estimates of the heritability of thetraits. Measures of contact between the twins reared apart were unrelated to twinsimilarity. These ®ndings are highly consistent with the larger behavior geneticliterature on genetic and environmental in¯uences on personality. A mean spousalCCC 0890±2070/98/050307±14$17.50 Received 18 June 1998#1998 John Wiley & Sons, Ltd. Accepted 30 July 1998European Journal of PersonalityEur. J. Pers. 12, 307±320 (1998)*Correspondence to: Thomas J. Bouchard Jr., Department of Psychology & Institute of Human Genetics,University of Minnesota, Minneapolis, Minnesota, 55455, USA.Contract grant sponsors: Pioneer Fund, The Seaver Institute; The University of Minnesota GraduateSchool; The Koch Charitable Foundation; The Spencer Foundation.Contract grant sponsor: The National Science Foundation.Contract grant number: BNS-7926654.Contract grant sponsor: Harcourt Brace Jovanovich Publishing Co.Contract grant sponsor: NSF.Contract grant number: GU-1598.correlation of 0.20 for the 34 scales based 111 pairs suggests that there is only modestassortative mating for the CPI scales.#1998 John Wiley & Sons, Ltd.INTRODUCTIONThe California Psychological Inventory (CPI; Gough, 1996), one of the most widelyused psychological assessment instruments (Aiken, 1997), has been analyzed in anumber of previous behavioral genetic studies of personality (Bouchard and McGue,1990; Gottesman, 1996; Horn, Plomin and Rosenman, 1976; Loehlin, 1986; Loehlinand Nichols, 1976; Nichols, 1966). When averaged across studies, heritabilityestimates for the CPI scale s are remarkably consistent and very similar to thoseobtained with other personality measures. For example, Loehlin and Nichols (1976)in a sample of 490 reared together monozygotic (MZT) and 317 reared togetherdizygotic (DZT) high school twin pairs found average twin correlations of 0.49 and0.29 for the MZT and DZT samples, respectively. Horn et al. (1976) in a sample of99 pairs each of MZT and DZT middle aged male twins reported average twincorrelations of 0.44 and 0.19 for the MZT and DZT samples, respectively. Althoughthe adu lt twins in the study of Horn et al. (1976) had lived apart for an average of25 years they were only a little less similar than the adolescent twins in the study ofLoehlin and Nichols (1976). Bouchard and McGue (1990) reported age- and sex-corrected intraclass correlations of 0.45 and 0.18 for an adult sample of monozygoticand dizygotic twins reared apart (MZA and DZA), ®gures virtually identical to thoseof Horn et al. (1976).Twin studies with the CPI are similar to the results of a meta-analysis of thetwin personality data presented by Nichols (1978), as well as to model ®tting ofreared-apart and reared -together twin data for the Multidimensional PersonalityQuestionnaire (MPQ) (Tellegen, Lykken, Bouchard, Wilcox, Segal and Rich, 1988)and ®ndings with the Eysenck Personality Questionnaire (Eaves, Eysenck andMartin, 1989). Bouchard (1997) recently reviewed the behavior genetic literature onpersonality.In the current study we report a biometric analysis of the CPI in a sample whichcombines the reared-together male twins from the study of Horn et al. (1976) with thereared-apart twins from the Minnesota Study of Twins Reared Apart (MISTRA).The MISTRA sample has nearly doubled in size since the previous report byBouchard and McGue (1990). In addition to the new standard scales we report®ndings for the new vector scales and the special purpose scales described in the mostrecent CPI manual (Gough, 1996).Loehlin and Gough (1990) reported estimates of heritability, shared environ-mental in¯uence, and nonshared environmental in¯uence on the vector scalesbased on the adolescent twin sample from the National Merit Twin Study. Their®ndings are of interest because, while Vector 1 (Internality) and Vector 3 (Self-Realization) yielded results that are typical for personality scales, heritabilities of0.66 and 0.58 with virtually no shared family environmental in¯uence, Vector 2(Norm-Favoring) yielded a heritability of only 0.18 and a shared environmentalcomponent of 0.32. This shared environmental component is by far one of thelargest reported for a personality variable from a twin study with a large sample308 T. J. Bouchard Jr et al.#1998 John Wiley & Sons, Ltd. Eur. J. Pers. 12, 307±320 (1998)size. Loehlin and Gough (1990) suggest, however, that their results may notgeneralize to adult twins who have lived apart for many years. Our data will allow atest of that speculation.METHODSubjectsThe samples of twins come from two independent studies. The reared-apart mono-zygotic (MZ A) and dizygotic (DZA) twins are from MISTRA. They participated inthat study between 1979 and 1997. Bouchard and McGue (1990) previously reportedCPI results for a subset of this sample. The twin sets in the present reared-apartsample include 37 same-sex dizygotic (SS-DZA) pairs, 16