Journal of Applied P s y c h o l o g y199J, Vol. 78, No. 4, 649-661Copyright 1993 by the American Psychological Association, Inc.0021-9010/93/$3.00Heritability of Interests: A Twin StudyD. T. Lykken, T. J. Bouchard, Jr., M. McGue, and A. TellegenThe authors administered inventories of vocational and recreational interests and talents to 924pairs of twins w h o h a d been reared together and to 92 pairs s ep ar a te d in infancy and reared apart.Factor analysis of all 291 items yielded 39 identifiable f actors and 11 superfactors. The data indi-cated that about 50% of interests variance (about two thirds of the stable variance) w a s associatedwith genetic variation. The authors show t h at heritability c a n be conservatively e s t i m a t e d from thewithin-pair correlations of adult monozygotic twins reared together. Evidence for nonadditivegenetic effects on interests may explain why he ri ta bi li ty estimates based on family studies are somuch lower. The authors propose a model in which precursor traits of aptitude and personality, i npart genetically determined, guide the development of interests through the mechanisms o f gene-environment correlation and interaction."Although there is evidence for the contribution of geneticfactors to interests, these factors account for less than 5% ofinterests variance" (Gati, 1991, p. 312). This summary ap-praisal, in a recent and authoritative article, seems to representthe prevailing v i e w among professionals involved in the studyof interest measurement and its applications in vocational se-lection and counseling. This view, however, is m is ta k en . S e v e r a learlier studies not cited by Gati (e.g., Loehlin & Nichols, 1976;Roberts* J o h a n s s o n , 1974; V a n d e n b e r g & S t a f f or d , 1967), s t u d -ies of twins reared together who were e it h er college students or1 Ith-grade Merit Scholarship aspirants when tested, suggestedheritabilities on the order of 36% f o r the major interest factorsin these populations (cf. N i c h o l s , 1978).More recently, Moloney, Bouchard, and Segal (1991) exam-ined interest-test scores of 45 pairs of monozygotic (MZ) twinsseparated in infancy and reared apart (MZA twins) and of 22pairs of same-sex dizygotic ( D Z ) t w i n s s e p a r a t e d in infancy andreared apart (DZA twins). These separated twins, who rangedin age from 12 to 70 years when tested, had completed theStrong Vocational Interest Blank (SVIB) and also the JacksonVocational Interest Survey ( J V I S ) . For the 34 scales of the JVISand the 23 basic interest scales of the SVIB, t h e mean heritabil-ity estimates were .37 and .44, respectively. F o r the 10 principalcomponents underlying the total set of 57 scales, the age- andsex-corrected MZA intraclass correlations, which directly esti-mate broad heritability, r a n g e d from .25 (SE = . 1 4 ) t o .73 (SE =.07; median r = .50, SE = . 1 1 ) . The DZA correlations on theseD. T . L y k k e n , T. J . B o u c h a r d , J r ., M . McGue, a n d A. T e l l e g e n , Depart-ment of Psychology, University o f Minnesota.This research was supported by National Institute o f Mental HealthGrant MH37860-07, by a University of Minnesota Graduate SchoolGrant to the Minnesota Twin Family Registry, and by the PioneerFund. We acknowledge with thanks the careful scrutiny and helpfuladvice of several anonymous reviewers.Correspondence concern ing t h i s a r t i c l e s ho u ld be addressed to D. T.Lykken, Department of Psychology, University o f Minnesota, ElliottHall, Minneapolis, M i n n e s o t a 5 54 55 -034 4.superfactors of occupational interest ranged from —.07 (SE =.21) to .30 (SE = .20; median r = .03, SE = .21).These results suggest that variation in occupational interestsmay, in fact, be quite strongly associated with genetic variation.Moreover, for such factors a s Artistic, for which t h e MZA corre-lation was .52 but the DZA value was only —.07, these datasuggest that certain interests may be "emergenic" (Li, 1987;Lykken, 1982). An emergenic trait is one for which t h e relevantpolygenes contribute configurally (i.e., epistatically), ratherthan in a n additive f as h io n, with the result t h a t , although geneti-cally determined, emergenic traits may not run in families(Lykken, McGue, Tellegen, & Bouchard, 1992).Because of the rarity of twins reared apart, these sampleswere small. Yet, because of the power of the MZA design forestimating heritabilities, the statistical reliability of these f i n d -ings is comparable to those based on nearly 20 times as ma nypairs of MZ and DZ twins r ea red together (Lykken, Geisser, &Tellegen, 1981). Nevertheless, like all findings with importantimplications, these results require constructive replication(Lykken, 1968) with additional samples and different instru-ments.Waller, Lykken, and Tellegen (in press) a d m i n i s t e r e d new in-ventories of occupational and leisure-time interests to a largesample of adult participants in the Minnesota Twin FamilyRegistry (Lykken, Bouchard, McGue, & Tellegen, 1990). Themean age of these subjects when tested was 39.5 years (SD =11.3 years) and 37.6 years (SD = 10.4 years) for the men andwomen, respectively. There were 1,768 men and 2,311 womenin the tested sample, of whom about two thirds were twins; theremaining one third were spouses or first-degree relatives o f thetwins. Because the inventories used by Waller et al. were alsoused in the present research, they are more fully described laterin this article. Factor analyses based on the more than 4,000subjects revealed 17 occupational interest factors a n d 18 leisure-time interest factors t h a t were c o n g r u e n t for both sexes. Herita-bilities for the occupational interests ranged from .00 to .56(median r = .42). These values were very similar to those re-ported for MZA twins b y Moloney et al. (1991) for the scales of649650LYKKEN, BOUCHARD, McGUE, AND TELLEGENthe SVIB and the JVIS. For the leisure-time interests, Waller etal. found heritabilities r a n g i n g f r o m .20 to .63 (median r = . 4 7 ) .Estimating Heritability From MZ-Twin CorrelationsTwin studies co mm o nl y use the Falconer (1960) formula (orits maximum likelihood equivalent), F al co ne r h2 = 2(RMzr -RDZT), where h2 is an estimate of narrow-sense heritability (theproportion of observed variance associated with additive ge-netic variance), w h e r e a s /?Mzx and /?DZT are the within-pair in-traclass correlations for MZT and DZT twins, r esp ec tiv el y, a n