Sources of Structure: Genetic,Environmental, and Artifactual Influenceson the Covariation of Personality TraitsRobert R. McCraeNational Institute on AgingKerry L. Jang and W. John LivesleyUniversity of British ColumbiaRainer Riemann and Alois AngleitnerUniversität BielefeldABSTRACT The phenotypic structure of personality traits has been welldescribed, but it has not yet been explained causally. Behavior genetic covari-ance analyses can identify the underlying causes of phenotypic structure;previous behavior genetic research has suggested that the effects from bothgenetic and nonshared environmental influences mirror the phenotype.However, nonshared environmental effects are usually estimated as a residualRobert R. McCrae, Personality, Stress and Coping Section, Gerontology ResearchCenter, National Institute on Aging, National Institutes of Health; Kerry L. Jang andW. John Livesley, Department of Psychiatry, University of British Columbia, Vancouver,British Columbia, Canada; Rainer Riemann and Alois Angleitner, Department of Psy-chology, Universität Bielefeld, Bielefeld, Germany.We thank Erik Mutén for providing self/spouse data from Sample B, and John C.Loehlin and Jerry S. Wiggins for helpful comments on the manuscript.Correspondence concerning this article should be addressed to Robert R. McCrae,Box 03, Personality, Stressand Coping Section, NIA Gerontology Research Center,5600Nathan Shock Drive, Baltimore, MD 21224-6825. Electronic mail may be sent [email protected] of Personality 69:4, August 2001.Copyright © 2001 by Blackwell Publishers, 350 Main Street, Malden, MA 02148,USA, and 108 Cowley Road, Oxford, OX4 1JF, UK.term that may also include systematic bias, such as that introduced by implicitpersonality theory. To reduce that bias, we supplemented data from Canadianand German twin studies with cross-observer correlations on the Revised NEOPersonality Inventory. The hypothesized five-factor structure was found in boththe phenotypic and genetic/familial covariances. When the residual covariancewas decomposed into true nonshared environmental influences and method bias,only the latter showed the five-factor structure. True nonshared environmentalinfluences are not structured as genetic influences are, although there was somesuggestion that they do affect two personality dimensions, Conscientiousnessand Love. These data reaffirm the value of behavior genetic analyses for researchon the underlying causes of personality traits.For decades, personality psychologists worked to define the phenotypicstructure of personality—the pattern of covariation of observed person-ality traits (Goldberg, 1993). The present near-consensus on the Five-Factor Model (FFM; McCrae & John, 1992) has allowed researchers toproceed to a new question: What accounts for the observed structure?Personality traits are known to be substantially inherited (Loehlin, 1992),but can genetics alone explain the covariation of traits, or is the pheno-typic FFM the net result of combined genetic and environmentalinfluences?The answer to that question has always been relevant to developmen-talists and clinicians who seek to understand the origins of personality,but it has taken on particular importance in the era of molecular genetics(Plomin & Caspi, 1998). Identifying the particular genes associated withany single trait has been likened to finding “many tiny needles in thehaystacks” (Plomin, 1990, p. 187), and this daunting search must beguided by an understanding of the origins of personality structure.Cloninger, Adolfsson,and Svrakic (1996) haveargued that “traits derivedby factor analysis at the phenotypic level, such as extraversion . . . arelikely to be composites of aetiologically heterogeneous facets” that maynot be optimal guides for “unraveling the genetics and neurobiologyunderlying human personality” (p. 3). Indeed, it would make little senseto look for the genes responsible for extraversion if extraversion weremerely a cluster of genetically unrelated traits inculcated by certainpatterns of child rearing.But are the facets of extraversion (and other phenotypic factors) in factgenetically unrelated? It might seem that molecular genetics itself wouldoffer the most direct answers to such questions, but, to date, progress inthat field has been halting. Early reports of specific genes linked to512 McCrae et al.personality traits (Benjamin et al., 1996; Lesch et al., 1996) have not beenconsistently replicated (Flory et al., 1999; Vandenbergh et al., 1997). Atpresent, it appears more realistic to rely on the well-established methodsof behavior genetics.Genetic Covariance AnalysesConventional behavior genetics analyses attempt to determine the pro-portion of variance in a trait attributable to genetic and environmentalinfluences. More recently, attention has been focused on the genetic andenvironmental sources of covariance among different traits (e.g., Carey& DiLalla, 1994; Jang & Livesley, 1999). For example, Eley (1997)investigated the possibility that the well-known correlation betweenanxiety and depression might be due to shared genes that influence boththese dispositions. Such analyses can be extended to examine the inter-correlations among many traits, and factor analysis (e.g., Livesley, Jang,& Vernon, 1998) can summarize the structure of genetic and environ-mental influences on personality just as it summarizes the structure oftraits themselves. Genetic covariance analyses might be used to askwhether and to what degree the FFM is the result of genetic and environ-mental influences.Personality psychologists have frequently assumed that different traitsand trait factors had different sources (Strelau, 1987). In particular, theclassic distinction between temperament and character assumed thatsome traits (such as Neuroticism and Extraversion) were primarily theresult of nature, whereas others (such as Agreeableness and Conscien-tiousness) reflected socialization and personal experience (cf. Cloninger,Przybeck, Svrakic, & Wetzel, 1994). This appealing dichotomy has beencalled into question by recent evidence that traits from all five majorpersonality factors are moderately to strongly heritable (Jang, Livesley,& Vernon, 1996), but the heritability of individual traits does not speakdirectly to the heritability of their patterns of covariance. Traits such asdutifulness, order, and achievement striving might each individually bepartly heritable, but their covariation in