Tellegen Et Al 1988 Personality Similarity in Twins Reared Apart and Together

7/30/2019 Tellegen Et Al 1988 Personality Similarity in Twins Reared Apart and Together

http://slidepdf.com/reader/full/tellegen-et-al-1988-personality-similarity-in-twins-reared-apart-and-together 1/9

Jour nal of Personality and Social Psychology Copyright 1988 by the Am erica n Psychological Association, Inc.1988, Vol. 54, No. 6, 1031-1039 0022-3514/88/$00.75

Personality Similarity in Twins Reared Apart and Together

Auke TellegenUnive rs i ty o f Minneso ta

Thoma s J. Bouchard, Jr., Kimerly J. Wilcox, andNan cy L. Segal

Unive rs i ty o f M inneso ta

Dav id T. LykkenDepar tm ents o f Psychology and Psychia t ry

Unive rs i ty o f Minneso ta

Stephen Ric hDepa r tment o f Pa tho logy and L abora tory Medic ine and

Ins t i tu te o f Hum an Gene t ic sUnive rs i ty o f Minneso ta

We administere d the Multidimensional Personality Ques tionnaire (MPQ) to 217 monozygotic and114 dizygotic reared-together adu lt twin pairs and 44 m onozygotic a nd 27 dizygotic rea red-a partadul t twin pairs. A four-parameter b iome tric model (incorporating genetic, additive versus nonaddi-tive, shared family-environment, and u nshared environment components) and five reduce d modelswere fitted through maximum-likelihood techniques to da ta obtained with the 11 prim ary MP Qscales and its 3 higher order scales. Solely environmental models did not fit any o f the scales. Al-

though the o ther redu ced models, including the simple additive model, did fit many of the scales,only the full mod el provided a satisfactory fit for all scales. Heritabilities estima ted by the full modelranged from .39 to .58. Consistent with previous reports, but contra ry to widely held beliefs, theoverall contrib ution of a co mm on family-environment com ponent was small and negligible for allbut 2 of the 14 personality measures. Evidence of significant nonadditive genetic effects, possiblyemergenic (epistatic) in nature, was obtained for 3 o f the measures.

Unti l recently, a lmost a l l knowledge regarding environmen-ta l and genetic causal inf luences on stable personali ty tra i ts hascome from studies of twins rea red together. The f indings havebeen both rem arka ble and puzzling, On the genetic side , regard-less of the tra i t s tudied, the intrac lass corre la t ion for f ra ternal ,

or dizygotic (DZ), twins has a pproa che d .25; that for identica l ,o r monozygot ic (MZ) , twins has approached .50 (Goldsmi th ,1983; Nichols, 1978) . Applica tion of the simp lest geneticmodel, the Falconer (1960) formula for her i tabil i ty, [h2 =2(RMz -- RDZ)], to tho se re sults yields a herita bility of abou t .50.This leaves 50% of the var ianc e to systematic environm entalinf luences, mea surem ent error, and tempo ral instabil ity.

Pa r t icu la rly puzz l ing , and cont ra ry to wha t many psycholo-gists would predic t , is the f inding that a lmost none of the envi-

This research has been supported by grants from the University ofMinnesota Graduate School, the Koch Charitable Foundation, theSpencer Foundation, the Na tional Science Foundation (BNS-7926654),the Pioneer Fund, and the H arc ourt Brace Jovanovich Publishing Com-pany.

We thank the following people for the tim e and effort they have givento testing the twins: Margaret Keyes, Jeff McHenry, Elizabeth Rengel,Susan Resnick, Joy Fisher, Jan Englander, and An n Riggs. We are in-debted to o ur colleagues and collaborator s on the M innesota Study o fTwins Reared Ap art project, Elke Eckert and Le onard Heston for theirhelp and advice. We owe special thanks to our colleagues, Greg Ca reyand Matt McGue, for their valuable biometric advice and assistance.We also thank J ack D arley for his resolute a nd dedica ted support andthe M inneap olis War Memorial Blood Bank, Herbert Polesky, Director,for the b lood testing.

Correspondence concerning this artic le should be addressed to AukeTellegen, Dep artm ent o f Psychology,Elliott Hall, University of Minne -sota, 75 East River Road, Minneapolis, Minnesota 55455.

ronmenta l va r iance is due to sha r ing a comm on fami ly env i ron-ment. Loehlin and Nichols (1976) reached this conclusion onthe basis of personali ty data c ollec ted on a large National MeritScholarship twin sample . Eaves and Young (198 l) ca rr ied o utan informative addit ional analysis of Extraversion-In troversion

and Neuro t ic i sm scores f rom the same twin sample , us ing abiome tr ic mode l s imilar to the one to be re por ted in this ar t ic le .They f i t ted three models to the data . The f irst model a l lowedfor unshare d e nvironm ent effects and addit ive gene effects. Thesecond a l lowed for two environmental effec ts: shared family-environment effects and unshared environment effects , but nogenetic effects . The third specif ied the join t effec t of a l l three .The results were stra ightforward: The second model, whichonly specif ied environ men tal param eters, did not f i t the da tafor e i ther sex or for the two sexes considered jointly. The f irstmode l , which a ssumes no com mon fami ly-envi ronment e ff ec t,f i t a l l three cases and was as capa ble o f f i tt ing both sexes simul-taneously (with identica l param eter e st imates) as i t was of f i tt ing

the results of each sex separa te ly. Addin g the co mm on family-envi ronment e ffec t d id no t improve the f i t o f the th i rd mode lover the f irst . These results essentia lly conf irm ed L oehlin andNichols 's (1976) ear l ier analysis.

Loehl in and Nichols (1976) ca r r ied ou t many addi t iona lanalyses of their data . In pa r t icular, they a t te mp ted to identifythe effects of systematic within-fam ily var ia t ions. Because iden-t ica l twins share a l l of their genes, any differences between the mmust be e nvironm ental in or igin. By re la t ing within-pair zygos-i ty-group differences in exper iences to personali ty differences,they could answer questions such as, "Do identica l twins whowere dressed a l ike turn out to be more similar in personali tythan identica l twins who were not?" T he results were essentia lly

negative . Greate r s imilar i ty in the twins ' exper iences cou ld ac-

1031



10 32 TELLEGEN, LYKKEN, BOUCHARD, WILCOX, SEGAL, RICH

count for only a smal l fract ion of the M Z twins ' s imilarity inpersonality. Loehlin and Nichols (1976) concluded the fol-lowing:

Thus, a consistent--thou gh perplexing --pattern is emergin g fromthe data (and it is not purely idiosyncratic to o ur study). Environ-

ment carries substantial weight in the determination of personal-ity-- it appears to account for at least half the variance~IbU t thatenvironment is one for which twin pairs are correlated close tozero. Attempts to treat twins alike do no t lead to greater similaritybetween them. In short, in the personality domain we seem to seeenvironmental effects that operate almost randomly with respectto the sorts of variables that psychologists and other people) havetraditionally deemed imp ortant in personality developm ent. (p. 92)

However, ad option data (Scarr, Weber, Weinberg, & W itting,1981 a; 1981 b) and n ontw in family data (A hern, Joh nson , W il-son, McClearn, & Vandenberg, 1982; Loehlin, Horn, & Wilier-man , 1981 ) complicate the picture. Scarr et al . (1981 a) foun dnot only modest correlations between nonbiological relatives(.04 between adoptive parent and child and .07 between adop-tive siblings), confir ming the slight influence of a com mo n fam-ily environment on personality, but also low correlations forfirst-degree biological relatives (. 15 between paren t and childand .20 between siblings). These latter figures are n ot un likethose c omp iled b y Loe hlin et al . (1981) for first-degree relatives.In addi t ion, Ahern et al . (1982) reported a parent-offsp ring cor-relation o f abou t . 12 and a s ib ling correlation o f abou t . 10, onthe basis of 54 scales. A hern et al. suggested that the heritabili tyof personal i ty es t imated from nontwin family data approaches.20. Scarf et al . (1981 a) suggested a figure betw een. 14 and .22.

Price, Vandenberg, Iyer, and Williams (1982) fit ted a ran domeffects l inear model to six personality measures obtained fromthe extended famil ies of MZ and DZ twins in the Swedish twin

registry. Their analysis incorporated 13 consanguineous rela-tionships (e.g., MZ twin pairs, DZ twin pairs, siblings, andmother-chi ld pai rs) and 9 nonconsanguineous relat ionships(e.g., spouses, siblings-in-law through MZ twins). The analysisyielded es t imates of addi t ive v ariance of nearly zero for fourt rai ts and very modest es t imates for the remaining two t rai ts(.04 and .07, respectively). Hewitt (1984) h as argued, however,that the Price et al. (1982) analysis does not adeq uately sup porttheir conclusions because the data also fit a simple additivemod el that y ields heritabili ty estimates of abou t .50 (cf. alsoHeath , Martin , & Eaves, 1984). With the exception o f Hewitt 'sreanalysis , conclusions based on twin data and nontwin familydata app ear to be at odds . Even Hewit t 's cr it ique is not so much

a refutation of the Price et al. conclusion s as a dem onstr ationthat their da ta fi t contrastin g mo dels and are no t decisive. Ho ware we to proce ed given these seeming discrepancies?

Firs t , parent-o ffspring data have the disadvantage of beingcollected when subjects are at different ages. I f personality traitsare expressed differently at different ages (i.e., some personalityt rai ts m ay be age dependent for genetic or environmental rea-sons) , then the parent-offsp ring data s imply do no t ref lect thesame phen ome na as the twin data. Th e s ibling data are no t sub-ject to th is problem to the same degree, and they do show aslightly higher correlation in some studies. Only twin d ata havethe o bvious advantage o f involving pairs o f individuals of iden-tical age, bu t the twin design has its own pro blems. Twin meth -

odology depends on the assumption of equal environmental in-

f luences on both MZ and DZ twins and unbiased sampling ofMZ and DZ twins from the twin populat ion. Al though there isconsiderable evidence to sup port the assump tion of equal t rai t -relevant environments (Bouchard, 1984), many invest igatorsstill find it unpalatab le. This is one instance in wh ich the stu dyof twins reared apa rt can contr ibute cri tical new informat ion,a point to which we return short ly.

Second, most previous analyses of personal i ty data have sug-gested that a simple additive model, reflecting the contributionof an unspecif ied num ber of independent ly segregat ing geneticfactors, explains the d ata adequately. If genetic effects are en-t i rely addi t ive and there are no shared family-environmenteffects, then th e DZ correlation s should be ab out hal f the sizeof the M Z co rrelations. It has been observed , however, tha t forsome t rai ts the DZ correlat ions , l ike nontwin famil ial corre-lations, are clearly less than one h alf the size of MZ correla tions(e.g., Lykken, 1982). This suggests nonadditive genetic effects,part icularly dom inance, or i f DZ correlat ions are very low, epis-tasis (i .e., interaction between nonallelic genes).

Lykken and his colleagues (Lykken, 1982; Lykken & B ouch-ard, 1983/84; Lykk en, Tellegen, & Iaco no, 1982) have suggestedthat a co mple x form of epistasis, calledemergenesis, may ac-coun t for a s ignif icant port ion o f the discrepancy between twindata and Parent X Offspring and s ibl ing data. Part icularly inthe case o f epistasis, genetic effects cou ld be b oth substantialand nonfamil ial . In those ins tances , rather than exaggerat ing"the degree to which personal i ty di fferences are explained bygenetic differences" (Scarr et al . , 1981a, pp. 896-897), twindata m ay tel l a t ruer s tory, and the data based on nontwins mayunderes t imate the actual imp act of genet ic factors .

Another s t r ik ing phenomenon in the personal i ty domain isthe ap parent lack of d i fferences in heri tabi li ty am ong personal-

ity variables. Loehlin and Nichols 's (1976) findings are a casein point . Their twin personal ity data involved a wel l -known in-s t rument , the Cal i fornia Psychological Inventory (CPI; Meg-argee, 1972), and revealed tha t the scales of this multim easu reinstrument have highly similar heritabili t ies. The differentialheritabili ty of personality variables (or the lack thereo f) hassubsequent ly become a subject of some controversy (Carey,Goldsmith, Tellegen, & Gottesman, 1978; Carey & Rice, 1983;Loehl in , 1982; Zonderman, 1982). What on the surface mayappear to be lack of d i fferent ial heri tabi l i ty amo ng psychologi-cally diverse scales may in reality reflect substantial content(and i tem) overlap and high correlat ions amo ng scales in inven-tories such as the CP I (Care y et al. , 1978).

Even factor analytical ly const ructed mul t iscale inventoriesusually reflect substantial saturation with only a small numberof higher order superfactors, particularly Positive Em otio nality(extravers ion) an d N egat ive Emo tional i ty (neurot icism). Wat-son and Cla rk (1984) have discussed the diverse nam es given towhat are essentially Negative Emotionality measures, reflectingthe wide scope of that d imension. If , in addi t ion, the superfac-tors themselves do not differ mu ch in heritabili ty, then only theaddition of scales that are no t excessively super factor sa turatedcould lead to ev idence of differential heritability. L oeh lin

Environment,as defined by these authors, presumably also includes

measurement error and nonsystematic changes over time.



PERSONALITY SIMILARITY IN TWINS 1033

(1982) has adduced evidence that this is in fact the case. Givenall of the problem s involv ing subject samples and choice ofmeasures, it is desirable to cross-check the various findings dis-cussed earlier using a design and i nstr umen ts that avoid thesedifficulties.

Our study makes use of a design that has never, to our knowl-edge, been implemented: the simultaneous use of MZ and DZtwins reared together (MZTs and DZTs, respectively) andreare d apart (M ZAs and DZAs, respectively). Altho ugh severalpersonality findings have been reported on M ZAs (cf. Bouch-ard, 1984, Table 10), we know of no one who has ever reportedpersonality data gathered on all four groups using the same per-sonality inventory. This design circum vents many of the prob-lems discussed earlier. First, the use of twins avoids the p roble mof testing at different ages. Second, the use of twins reared apart,although possibly subject to other flaws (Bouchard, 1984), re-lieves us of the assu mptio n of equal environ ment al similarityfor MZ and DZ twins. Because they were placed for adoptionat such a youn g age, pl aceme nt practices are unlikel y to havebeen different for MZAs a nd DZAs (cf. Bouchard, 1984). Fi-nally, with respect to the role of a com mon en viro nmen t, it isdifficult to believe that twins reared apart (even if placed insomewhat simila r homes) would have the s imilarity in trait-rel-evant environments that are experienced by twins reared to-gether. U nder a purely c omm on- env iro nme nt hypothesis, MZtwins and DZ twins would be expected to show the same degreeof within- twin similarity, but twins reared together should ex-hibit greater similarity than twins reared apart. The biometricmodel allowed us to test specific hypotheses of this type.

To test the hypothesis of differential heritability of personalityvariables, we found it necessary to use an i nst rum ent composedof relatively indepe ndent scales. We used the Mu ltidi mensi onal

Personality Question naire (MPQ), formerly called the Differ-ential Personality Questionnaire (DPQ), in this study. TheMPQ (discussed in more detail in the Method section) was de-veloped using a factor a naly tic strategy (Tellegen, 1982). Itsscales, compar ed with those in other multiscale inventories, arerelatively indep ende nt but can still be me aningfull y integratedinto higher order factor measures. The latter corresp ond closelyto the superfactor scales that were developed by Eysenck an dEysenck (1975) and an alyzed by Eaves an d Young (1981) andman y others. The MP Q th us perm itted us to assess differentialherit abili ty on lower as well as higher ord er fac tor levels.

Me thod

Samples

The MZ and DZ twins reared together participated in the MinnesotaTwin Study between 1970 and 1984. Details of their recruitment arereported in Lykken (1982) and Lykken et al. (1982). The zygosity diag-nosis of all twins included in this study was based on analyses of eightblood group systems, four serum proteins, six red blood cell enzymes,fingerprint ridgecount, ponderal index, and cephalic index. Probabilityofmisdiagnosis is less than i001 (Lykken, 1978).

The MZA and DZA twins participated in the Minnesota Study ofTwins Reared Apart between 1979 and 1986. Details of their recruit-ment are reported in Bouchard (1984) and references cited therein. Thetwins' age at separation ranged from birth to 4.5 years, with a median

of 0.2 years. Separation time (number of years from separation to first

Table 1Sam ple Sizes, Percentage Females, and Means and StandardDeviations or Age o r MZA , DZA , MZ T, and D Z T Twins

Descriptive statistics MZA DZA MZT DZT

Total sample size (sets) 44 a'b 27 b 217 114Percentage females 66.7 75.5 65.0 64.9Mean age (years) 40.7 41.1 23.5 19.8SD of age 12.0 11.3 11.1 4.3

Note.MZA = Monozygotic wins reared apart, DZA = dizygotic twinsreared apart, MZT = monozygotic twins reared together, and DZT =dizygotic wins reared together.a Two MZA male triplet sets entered as 1 set each. b One reared-aparttriplet set (two MZA females and one male) entered as 1 MZA pair and2 DZA pairs.

contact) ranged from 0 to 64 years, with a median of 33.8 years, z Zygos-ity was determined for all reared-apart twins in the same way as forreared-together twins. All twin pairs used in this study were of the samesex, except for 4 opposite-sex DZA twin pairs. Descriptive statistics onthe samples are presented in Table 1.

Personality Inventory

The MPQ is a factor analytically developed self-report instrument. Itsscales represent 11 primary personality dimensions and 3 higher ordertraits; alpha coefficients ange from .76 to .89, with a median of.85; 30-day test-retest correlations range from .82 to .92, with a median of .89(Tellegen, 1982, 1985). Particular care was taken dur ing the develop-ment of the MPQ to achieve relatively independent primary scales.Thus, unlike the CPI used in the Loehlin and Nichols (1976) study, wehad a number of scales that are not primarily superfactor markers. We

did, however, essentially ap these factors at the higher order level.The higher order MPQ scales are interpreted as self-report dimen-sions describing basic parameters of emotional and behavioral regula-tion. Factor 1, Positive Emotionality, is primarily associated with theMPQ Well-Being, Social Potency, Achievement, and Social Closenessscales and has clear "extraverted" features. High scorers on these scalespresent themselves as being engaged in active, pleasurable, and effica-cious transactions with their environment and as being ready to experi-ence the positive emotions congruent with these involvements. Lowscorers report few of these pleasurable transactions, a lower degree ofself-efficacy, a higher threshold for experiencing positive affect, and atendency toward depressive, nonpleasurable disengagement. PositiveEmotionality is related to Eysenck's Extraversion (E; Eysenck & Eys-enck, 1975) and clearly resembles the Norman-Goldberg Surgency di-mension (Goldberg, 1981 ).

Factor 2, Negative Emotionality, is primarily associated with theMPQ Stress-Reaction, Alienation, and Aggression scales. High scorersdescribe themselves as being unpleasurably engaged, stressed and ha-rassed, and prone to experiencing strong negative emotions such as anx-iety and anger. Low scorers convey a less "catastrophizing" picture, ahigher threshold for negative affect, greater resiliency under stress(Block, 1964), and a tendency toward phlegmatic, non-unpleasurabledisengagement. This dimension shares important features with Eys-enck's Neuroticism (N) and Norman-Goldberg's (reversed) Agreeabil-ity and (reversed) Emotional Stability factors, particularly the latter.

2 Separation time was arbitrarily set at zero for twin pairs who werereared in different homes but who had had periodic contact during

childhood.



10 34 TELLEGEN , LYKKEN, BOUCHARD, WILCOX, SEGAL, RICH

The affective interpretatio n of higher order MPQ Trait Factors 1 and2 is supported by th eir convergen t-discriminan t relations to the statedimensions of Positive and Negative Affect, respectively (Tellegen,1982), wh ich dom inate measures o f current moo d (Watson & Tellegen,1985). The same p attern of relations to Posit ive and Negative Affecthas been reported for Eysenck's Extraversion and Neuro ticism scales,

respectively (Costa & McCra e, 1980; Wart, Barter, & Brownbridg e,1983).The thir d higher order MPQ dimension , Constraint , is most strongly

associated with the MPQ Control, Harm Avoidance, and Tradit ional-ism scales. High scorers on these scales describe themselves as beingrestrained, cautious, avoiding dangerous kinds of excitement andthril ls, deferential, and con ventional. Low scorers present a picture o fimpulsiveness, fearless sensation seeking, and rejection of conve ntionalstrictures on the ir behavior. Constraint is most clearly related to Eyse-nck's (reversed) Psychoticism (P) and the Norma n-G oldb erg Conscien-tiousn ess factor.

A more detailed exposit ion of this three-dimensio nal scheme and i tsconceptual relations, particularly to Eysenck and Eysenck's (1975),Gray's ( 1973, 1981), and ps ychody namic views, ha s appeared elsewhere(Tellegen, 1985). Depue , Krauss, an d Spoo nt (1987) have pres ented anentirely consonan t interpretation of Posit ive Emotionali ty as a "behav-ioral engagement" dimension and have emphasized i ts particular rele-vance to euphoric and depressive diatheses. Watson and Clark (1984)have provided a comprehensive interpretive review of Negative Emo-tionali ty, docu mentin g the breadth of i ts manifestations.

Analytic Procedures

Age- andsex-correction ofpersonality scores.Raw scores on the MPQwere correcte d for age, age2, and Age X Sex using the com bined samplesof MZT, DZT, MZA, and DZA twins. Scores were then standardized(M = O, SD = 1). The regression procedures are described in McGu eand Bouc hard (1984).

Biometric models.Although intraclass correlations are often infor-mative and are prese nted here, for analytic purposes they can be mis-leading if MZ and DZ variances differ. Biometric geneticists, followingJinks and Fulker (1970), therefore, prefer analyzing variances over cor-relations. We also take this approac h in this art icle. The expec ted valuesof the between (B) and w ithin (W) mea n squares of each twin group areexpressed as appropriately weighted sums of the particular environ-mental and genetic var iance compone nts that are incorporated in toone's chosen model. T he weights (coefficients) are estima ted from theobserved mean squares. For presentations of biometric approaches, seeEaves (1978, 1982); Eaves, Last, Young, and Martin (1978); Jinks andFulker (1970); Martin, Eaves, Kearsey, and Davies (1978); Nance andCorey ( 1976); and Rose et al. (1980).

Our full model decomposes the phe notypic variance of each trait intothree main components: genetic, shared family environment, and un-shared en vironm ent, each estimated as a parameter. The genetic com-ponent i s fu r ther analyzed in terms of an add i tive subcomponent ( re-f lect ing the con t r ibu t ion o f an unspecif ied num ber o f independent lysegregating genetic factors) and a nonadditive component. Specifically,the model, as developed by G. Carey (personal communica tion, 1986),exploits the joi nt availabil i ty of twins reared together and apa rt to esti-mate an "add i t iv i ty" parameter. Th is four th parameter permi ts us inprinciple (albeit with low power, due to the small nu mbe r of DZAs) todistinguish traits that are prima rily additive from ones that are primar-i ly nonadditive. Marked no nadditivity is at tributa ble to epistasis, a ge-netic effect of particular interest to us, as explaine d earlier. As for envi-ronmental components , the shared fami ly-env i ronment componen trepresents the contribution of those dist inctive family characterist ics(dist inguishing families from one another) that influence individuals

reared in the same family in the same way. The un share d enviro nme nt

Ta b l e 2Biometric Model Specifying Genetic and Enviro nmenta lComponents of Variance of Between (B) and Within (IV) Mea nSquares of Twins Reared Together and Apart

E n v i ro n men t a l

Mean square Genet ic Shared fami lia l Unshared

B(MZT ) 2 2 1W(MZ T) 0 0 1B(MZA ) 2 1 1W(MZA) 0 1 1B(DZT) 1 + C 2 1W(D ZT) 1 - C 0 1B(DZ A) 1 + C 1 1W(DZ A) 1 - C 1 1

Note.MZ T = M onozygotic twins reared together, MZ A = mo nozygotictwins reared apart , DZT = dizygotic twins reared together, andDZA = dizygotic twins reared apart . C is a paramete r ranging between0 and .5, reflecting degree of addit ive versus nona dditive genetic control .

If C equals .5 , the genetic effects are entirely additive; ifC approa ches 0,the genetic effects are increasingly nonadditive, reflecting an un kno wnmixtu re of additive, dom inant, a nd epistatic effects; and i fC equals 0,the genetic effects are entirely epistatic.

com pone nt reflects, amo ng other things, mea surem ent error and statefluctuations.

The full model is displayed in Table 2. The table shows how the Band W mean squares can be expressed in terms of the four parametersmentioned earlier: genetic, addit ive, shared family environment, andunshared env i ronment . The model can be examined by determin inghow it specifies expected genetic and en viron men tal twin correlations,as follows.

Using the formula for the intraclass correlation, R = (B - W)/(B +W), we learn from Table 2 that the genetic correlation for MZ twinsequals (2 - 0)/(2 + 0) = 1, as required (see the Genetic colum n, rows 1and 2 or 3 and 4). For DZ twins, if the additivity parame ter C is set ati ts maximu m value of.5, the genetic correlation equals [( 1 + .5) - (1 -

. 5 ) ] / [ ( 1 + .5) + (1 - .5)] = .5 (see Gen etic colu mn, ro ws 5 and 6 or 7and 8), in accordance with the theoretical expectation tha t the additivegenetic correlation between DZ cotwins is half that of MZ cotwins. IfC is set at i ts minim um value of 0, then the genetic DZ correlation,comp uted in the same way, equals (1 - 1)/(1 + 1) = 0, showing thatcomplete epistasis eliminates any genetic correlation between DZ cot-wins. Turning to environmental correlations, we can see from the ap-propr iate rows in the Shared Famil ia l -Envi ronmental co lumn of Tab le2 that for twins reared together and apart , the shared family-environ-men t correlati on equals 1 and 0, respectively. The unsh ared environ -men t correlation equals 0 by definit ion, which is the value entailed bythe entries in the last colum n of the table.

The fou r parameters are estimated and statist ically evaluated with themaximum likelihood method described by Martin et al . (1978). Ex-pressed as proportions of the total variance, these estimates are the heri-tabil i ty (in the broad sense, incorporating both the additive and no nad-dit ive compone nts), addit ivity versus nonadditivity, shared fam ily-"en-v i ronmental i ty" and unshared "env i ronm ental i ty" values that , g iventhe m odel, best fi t the da ta for a given scale.

In addition to th e full model show n in Table 2, one can also evaluatesimpler (reduced) versions. This is done b y assigning a priori values toone or more o f the parameters whi le f i tt ing the remain ing ones to thedata. In this analysis, six models were investigated: (1) unsh ared envi-ron men t effects; (2) shared family -environ ment and un share d environ-

men t effects; (3) addit ive genetic and unsh ared env iron men t effects; (4)



PERSONALITY SIMILAR ITY IN TWIN S 103 5

Table 3

Intraclass Correlations or 11 Primary Scales and 3 HigherOrder Scales or the Multidimensional PersonalityQuestionnaire or Four Kinship G roups

Scale MZA DZA MZT DZT

PrimaryWell-Being .48 .18 .58 .23Social Potency .56 .27 .65 .08Achieveme nt .36 .07 .51 .13Social Closeness .29 .30 .57 .24Stress Reaction .61 .27 .52 .24Alie nati on .48 .18 .55 .38Aggression .46 .06 .43 .14Control .50 .03 .41 -. 06Ha rm Avoidance .49 .24 .55 .17Traditiona lism .53 .39 .50 .47Ab sorp tion .61 .21 .49 .41

Higher orderPositive Em otionality .34 -. 07 .63 .18

Negative Em otio nali ty .61 .29 .54 .41Constra int .57 .04 .58 .25

Note.MZA = Monozygotic twin pairs reared apart (n = 44), DZA =dizygotic twin pairs reared apar t (n = 27), MZT = m onozygotic twinpairs reared together (n = 217), and DZT = dizygotic twin pairs rearedtogether (n = 114).

additive genetic, shared family-environment, and unshared environ-ment effects; (5) additive a nd nonadditive genetic effects, and unshare denvironment effects; and (6) the c omple te four-parameter model (addi-tive and nonadditive genetic effects, as well as shared family- and un-shared environment effects). For each of the six models, the estimateswere made through maximum likelihood techniques and were evalu-

ated by means o f appro priate residual chi-square computations. Accep-tance o f a mo del required a nonsignificant chi-square. 3

R e s u l t s

Descriptive Statistics

We inspec ted the m eans and s tanda rd dev ia t ions o f the M PQraw scores o f the four twin groups and found them to be com pa-rable to those of other samples. The scale intercorre la t ions,com pute d separa te ly for wome n and men, were l ikewise similarto those found in other samples. Th e results conf irm the re la t iveindependence of the p r imary sca le s and jus ti fy our cho ice o fthis instrum ent to evaluate differentia l her i tabil i ty of personal-

i ty var iables. Table 3 shows intrac lass corre la t ions for the fourtwin groups. The mode st s ize of the DZ A sa mple is reason forcaution in com paring results for this group with the other twinsamples. The DZ A tw ins are , however, c lear ly much less similarthan the M ZA twins. Addit ionally, he corre la t ions for the MZ Aand MZ T samples a re ove ra ll h igh ly s imi lar : The m edian cor re -la t ions for the 11 pr im ary M PQ scales are .49 and .52, respec-t ively. This patte rn suggests that a pure ly env ironme ntal mode lwill not provide an adequate explanation. The B and W meansquares on which ou r analyses are base d are shown in Table 4.

Biometric Analysis

Of the reduced models, ne itber Mode l 1 nor M odel 2, the two

purely enviro nme ntal mod els, came c lose to f i t t ing the data for

any of the 14 scales: The re levant chi-squares were a l l highlysignif icant and large . The other four m odels fare d substantia l lybetter. Model 3, the simple addit ive model, y ie lded an adeq uatefit for 11 of 14 scales, with th e exc eptio n of the Well-Be ing, Con-trol , and Posit ive Emo tionali ty scales. Model 4, which adds thesha red f ami ly-envi ronment compo nent a s a pa ram e te r toModel 3, f it a ll sca les except Well-Being and Control . Mode l 5adds the nonaddit ive parameter to Model 3 and f i t a l l sca lesexcept Well-Being. Finally, Model 6, the com plete four -param e-ter model, provide d a good f i t for a ll 14 scales. The c omple te se to f pa rame te r e s t ima te s o f Mode l 6 i s p re sen ted in Table 5 .

As Table 5 shows, the genetic parameter est imates ( i .e . , theheritabilitie s) range from .39 to .58, with an average of .48.Given the good f i t of the simple addit ive genetic model (Model3) , i t is not surpr ising that Table 5 does not show many in-s tances o f apprec iab le nonaddi t iv i ty o r sha red f ami ly env i ron-menta li ty. Never theless, for three s cales wSo cia l Potency, Con-t ro l, and Pos it ive Em ot iona l i ty - - the e s t ima te s o f the C pa ram e-ter indicate sta t is t ica l ly signif icant nonadd it ivi ty and are lowenough to be consistent with an emergenic (epista t ic) interpre-ta t ion. We saw already that of these three scales, Control andPosit ive Emotionali ty did not f i t the simple addit ive geneticModel 3. The third scale , Socia l Potency, a l though accommo-dated b y a l l four genetic models, f it Mode ls 5 and 6 (which con-ta in the [non]addit ivi ty param eter) better than M odels 3 and4 . With inc reased power tha t can , fo r example , be a t ta ined byinc reas ing the s ize o f the DZ A sample , add i t iona l ev idence ofnonaddit ivi ty may be for thcoming. Table 5 shows signif icantfamilia l effec ts for just two scales, Socia l Closeness an d Posit iveEmotionali ty.

D i s c u s s i o n

These results conf irm and extend ear l ier f indings and havesevera l implica tions for curre nt thinkin g and future researchconcerning dete rmina nts o f var ia t ion in personali ty. First , ouranalyses indicate that , on average, abou t 50% of me asure d per-sonali ty diversi ty can be a t tr ibute d to genetic diversity. This re-

3 It is still customa ry practice to a rrive at conclusions abou t data byaccepting or rejecting null hypotheses on the basis of statistical signifi-cance tests. We have also followed his approac h. One reason is that nullhypothesis testing, given our sam ple sizes and the power of the analyses,appears to lead to reasonable com parisons of the extent to which theseveral models adequately represent ma jor features o f the data . In addi-tion, we did not wish to deflect attention from our p resent topic, whichis substantive and not methodological. A nticipa ting future discussions,however, we wish to be on record as viewing null hypothesis testing asinherently flawed. Given large enough samples, one can b e sure t hateven our full mod el will "significantly" fail to fit the data. All models aresimplica tions and their falsehood can be assumed without c onductingsignificance tests. Psychologists need not spend time deciding the tr uthor falsehood of their strictly false models. We can do be tter by attem pt-ing to assess the degree to which alternative models o f high c ontent fitthe data and to evaluate which model or models fit the data best. Al-though these issues have been widely discussed by struc tural-mo delingspecialists (e.g., Bentler & Bonett, 1980; Cudeck & Browne, 1983), theyappe ar not to have been considered in applic ations of biome tric geneticmodels to psychological data. For critica l discussions of significance

testing, see also Lykken (1968) and Mee hl (1978).



1036 TELLEGEN, LYKKEN, B OUC HAR D, W ILC OX, SEGAL, R IC H

Ta b le 4Between (B) and Within (W) Mean Squares of Twins Reared Apart and Togetherfor Multidimensional Personality Questionnaire Data

M Z A D Z A M Z T D Z T

Scale B W B W B W B W

Well-Being 1.62 0.56 1.85 1.28 1.52 0.4 0 1.13 0.71Social Poten cy 1.90 0.53 1.10 0.64 1.67 0.35 1.00 0.86Ach ievem ent 1.33 0.61 1.25 1.09 1.58 0.52 1.02 0.78Social Closeness 1.32 0.71 1.53 0.83 1.46 0 .40 1.37 0.84Stress Rea ctio n 1.95 0.46 1.03 0.59 1.50 0.47 1.16 0.72Ali enat ion 1.21 0.42 1.36 0.94 1.58 0.46 1.36 0.61Aggression 1.09 0.40 1.13 1.00 1.49 0.59 1.13 0.85Con trol 1.87 0.61 1.08 1.02 1.32 0.55 0.94 1.06Ha rm Avoi dance 1.25 0.42 0.79 0.49 1.63 0.47 1.25 0.89Trad itio nalis m 1.66 0.51 1.78 0.78 1.39 0.46 1.55 0.56Abs orptio n 1.90 0.46 1.25 0.82 1.50 0.51 1.34 0.57Positive Em otio nali ty 1.29 0.62 0.95 1.09 1.70 0.38 1.11 0.77Negative Em otio nali ty 1.58 0.37 1.44 0.80 1.50: 0.45 1.39 0.58Con stra int 1.43 0.39 0.82 0.76 1.63 0.44 1.28 0.77

Note. MZA = M onozygotic twins reared apart, DZ A = dizygotic twins reared apart, M ZT = monozygotic twins reared together, and DZ T =dizygotic twins reared together.

s u i t c on f i rms p re v ious f ind ings a nd re p re s e n t s a n e x te ns ion toa w ide ra nge o f d i s t inc t ive pe rs ona l i ty c ha ra c te ri s t i c s . T he re -ma in ing 50% i s t e c hn ic a l ly c l a s s i f i e d a s a l l e nv i ronme n ta l a ndinc lude s the uns ha re d e nv i ron me n t e ffe c t t ha t , i n tu rn , i nc lude sme a s u re me n t e r ro r a nd non t ra i t s c o re f luc tua t ions tha t r e f l e c tthe in f lue nc e o f t r a ns i e n t s t a te s on s e l f - re por t t r a i t me a s u re s .I f one c ons ide rs bo th the in t e rna l c ons i s t e nc y a nd s t a b i l i t y o fpersonal i ty sca les (c f . Conley, 1984), i t i s not unreasonable to

gue s s tha t n o t m ore tha n a bou t 70 to 85% o f the obs e rve d va r i -a nc e re p re s e n t s t r a i t va r i a nc e . B e c a us e the 15 to 30% non t ra i tv a r i a n c e is p re s u m a b l y c o n f i n e d t o t h e e n v i r o n m e n t a l 5 0% o fthe obs e rve d va r i a nc e , t he e nv i ron me n t a l ly ba s e d t ra i t va r i a nc e

m a y n o t a m o u n t t o m o r e t h a n 2 0 t o 3 5 % , c o m p a r e d w i t h t h erough ly 50% tha t i s ge ne t i c in o r ig in . Imp rov in g the c ons i s t e nc ya nd s t a b i l i t y o f a t r a i t s c al e s hou ld inc re a s e i t s me a s u re d ge ne t i ca n d t r a i t- r e le v a n t e n v i r o n m e n t a l c o m p o n e n t s , b u t n o t t h e i r r el -a t ive ma gn i tude . I t s e e ms re a sona b le , t he re fo re, t o c onc ludetha t pe r s ona l i ty d i ffe re nc e s a re more in f lue nc e d by ge ne t i c d i -ve rsi ty tha n the y a re by e nv i ronm e n ta l d ive rs i ty.

C ons i s t e n t w i th the s e obs e rva t ions i s t he a bs e nc e o f s ubs ta n -

t i a l s ha re d fa m i ly -e n v i ron me n t e ffe c t s in ou r da ta . Pos i t iveEmo t iona l i ty a n d Soc ia l C los e nes s , howe ve r, y i e lde d s ign i f i c a n te s t ima te s o f . 22 a nd . 19, r e s pec t ively. The s e re s u l t s wa r ra n ts ome pa r t i a l ly s pe c u la t ive c om me n t .

Ta b le 5

Estimates of Genetic and Environmental Variance Components From a Biometric Model Applied to MultidimensionalPersonality Questionnaire Data of Twins Reared Apart and Together

Variance component

Scale Genetic C parameter Shared familial Unshared

Well-Being .48 ~ (.08) .29 (.16) .13 (.09)Social Pote ncy .54 ~ (.07) .05 a (.21) .10 (.08)Ach ieve me nt .39 a (.10) .13 (.27) .11 (.11)Soc ial Closeness .40 a (.08) .19 (.22) .19 a (.09)Stress Rea ction .53 ~ (.04) .49 (.17) .00bAli enat ion .45 a (.13) .50 b .11 (.12)Aggression .44 a (.05) .27 (.19) .00bCon trol .44 a (.05) .00 a'b .00 bHa rm Avo idanc e .55 ~ (.04) .31 (.15) .00 bTrad itio nalis m .45 a (.10) .50 b .12 (.10)Abs orptio n .50 a (.10) .50 b .03 (.10)Positive Em otio nal ity .40 a (.08) .00 ~'b .22 ~ (.07)Negative Em otio nali ty .55 ~ (. 11) .50 b .02 (. 11)Con strai nt .58 a (.04) .40 (.14) .00 b

.40 a (.04)

.36 a (.04)

.51 a (.05)

.41" (.05)

.47 a (.04)

.44 a (.04)

.56 a (.05)

.56 a (.05)

.45 a (.04)

.43 a (.04)

.47" (.04)

.38 a (.04)

.43 a (.04)

.43 a (.04)

Note. Standard errors are in parentheses.

a Significantly different from null value at p < .05. b Bound ary solution; therefore, no standard erro r computed .



103 8 TELLEGEN, LYKKEN, BOUCHARD, WILCOX, SEGAL, RICH

study by Rushton, Fulker, Neale , Nias, and Eysenck (1986) ,who fou nd that measu res of a l truism and aggression ( tra i ts ex-pected a pr ior i to be heavily inf luenced by socia l iza tion pro-cesses) both had h er i tabil i t ies of about .50 and no com monfamily-environment inf luence.

I t should be kept in mind th at her i tab il i ty est imates can varydepending on the genetic model that is being f i t to the data .Clar if ica tion of the differentia l-her i tabil i ty ssue wil l , therefore ,par t ly depend on better unde rstanding of the genetic architec-ture under lying different personali ty tra i ts . This leads us to ourf inal ma jor point .

Although most previous analyses of personali ty data have ledto the a cceptanc e of a s imp le addit ive genetic model, this viewhas recently been challenged (Carey & Rice , 1983; Lykken,1982; Lykken & Bouchard, 1983/84; Lykken et al. , 1982).Carey an d Rice ( 1983 ) conclud ed that the "genetic and en viron-menta l a rch i tec ture o f pe r sona li ty may be complex and t r a i t -specif ic" (p. 54) . The Ca rey an d R ice analysis was carr ied outon three MPQ scales, Socia l Potency, Socia l Closeness, andContro l (p rev ious ly named impulsivity and scored in the re-verse direc tion) . Our own results a lso suggest that the simpleaddit ive mod el has to be re jec ted in some cases. Only the com -ple te four-pa rame te r mode l accom moda ted a l l o f the da ta , andit indicated nonadd it ivi ty for three scales: Socia l Potency, Con-trol , and Posit ive Emotionali ty. The nonad dit ivi ty patterns en-counte red in these cases are consistent with an e pista t ic inter-pre ta t ion a nd in this sense support the conc ept of emergenesisdiscussed in the introduc tion.

In this s tudy, pure ly add it ive genetic models were never the-less found to f i t the dat a well in m any cases. Larger sam ple sizes,pa r t icu la r ly more DZAs, and addi t iona l f ami ly g roups ( inc lud-ing offspr ing of twins) would better enable us to dete rmin e the

compara t ive adequacy of add i tive , nonaddi t ive , and combinedaddit ive -nona ddit ive genetic models; to isola te dom inan t andepista t ic components of the nonaddit ive genetic var iance; andto examine fur ther the effect of systematic environmental in-f luences on M Z and D Z twin s imi la r i ty th rough envi ronmenta lassessments. We hope to expand our data base in these direc-tions.

R e f e r e n c e s

Ahern, E M., Johnson, R. C., Wilson, J. R., McClearn, G. E., & Van-denberg, S. G. (1982). Family resemblances in personality.BehaviorGenetics, 12,261-280.

Bentler, P. M., & Bonett, D. G . (1980). Significance tests and goodnessof fit in the analysis of covariance structure.Psychological Bulletin,88, 588-606.

Block, J. (1964). The challenge o f response sets.New York: Apple ton-Century-Crofts.

Bouchard, T. J., Jr. (1984). Twins reared apa rt and together: What the ytell us abo ut hu man diversity. In S. Fox (Ed.),The chemical and bio-logical bases of ndividuality(pp. 147-184). New York: Plenum Press.

Carey, G., Goldsm ith, H. H ., Tellegen, A., & Gottesm an, I. I. (1978).Genetics and persona lity inventories: The limits of replication withtwin data.Behavior G enetics, 8,299-313.

Carey, G., & Rice, J. (1983). Gene tics and personality temperament:Simp licity or complexity?Behavior Genetics, 13,43-63.

Conley, J. J. (1984). The hierarchy of consistency: A review and m odel

of ongitudinal findings on ad ult individual differences in intelligence,

personality and self-opinion.Personality and Individual Differences,5, 11-25.

Costa, P. T., & McCrea, R. R. (1980). Influence of extraversion andneuroticism on subjective well-being: Happy and unhappy people.Journal o f Personality and Social Psychology, 38,668-678.

Cudeck, R., & Browne, M. W. (1983). Cross-validation of covariance

structures. Multivariate Behavioral Research, 18,147-167.Depue, R. A., Krauss, S., & Spoont, M. R. (1987). A two-dimensional

threshold mode l of seasonal bipola r affective disorder. In D. Magnus-son & A. Ohrm an (Eds.),Psychopathology: An interactional perspec-tive(pp. 95-123). New York: Academ ic Press.

Eaves, L. J. (1978). Twins as a basis for the causal analysis of hu manpersonality. In W. E. Nance, G. Allen, & P. Parisi (Eds.),Twin re-search: Psychology and methodology(pp. 151 - 174). New Y ork: AlanR. Liss.

Eaves, L. J. (1982). The utility of twins. In V. E. Anderson, W. A.Hauser, J. K. Pe nry, & C. E Sing (Eds.),Genetic basis of he epilepsies(pp. 249-276). New York: Raven Press.

Eaves, L. J., Last, K . A., Young, P. A., & Martin, N. G . (1978). Modelfitting approaches to the analysis of hum an behavior.Heredity, 41,

249-320.Eaves, L. J., & Young, P. A. (1981). How sta ble are pe rsonality traits?In L. Gedd a, P. Parisi, & W. E. Nance (Eds.),Twin research 3: PartB. Intelligence, personality and development(pp. 87-97). New York:Alan R. Liss.

Eysenck, H. J., & Eysenck, S. B. G. (1975).Manual o f he Eysenck Per-sonality Questionnaire. San Diego, CA: Educational and IndustrialTesting Service.

Falconer, D. S. (1960).Introduction t o quantitative genetics.New York:Ronald Press.

Feshbach, S., & Weiner, B. (1986).Personality (2nd ed.). Lexington,MA: Heath.

Goldberg, L. R. (1981). Language and individual differences: Thesearch fo r universals in personality lexicon s. In C. Wheeler (E d.),Re -

view ofpersonality and socialpsychology(Voi. 2, p p. 141 - 165). Bev-erly Hills, CA: Sage.Goldsmith, H. H. (1983). Genetic influences on personality from in-

fancy to adulthood .Child Development, 54,331-335.Gray, J. A . (1973). Causal theories o f personality and how to te st them.

In J. R. Royce (Ed.),Multivariate analysis and psychological theory(pp. 409-463). New York: Academ ic Press.

Gray, J. A. (1981). A critique of Eysenck's theory of personality. InH. J. Eysenck (Ed.),A model fo r personality(pp. 233-252). Berlin:Springer-Verlag.

Heath, A. C., Martin, N. G ., & Eaves, L. J. (1984). Sense and nonsensein genetic epidemiology: A critiqu e of the statistical model of Wil-liams and Iyer.Acta Geneticae Medicae et Gemellologiae, 33,557-563.

Hewitt, J. K. (1984). Normal components of personality variation.Journal of Personality and Social Psychology, 47,671-675.Jinks, J. L., & Fulker, D. W. (1970). A com parison of the biom etrical

genetical, MAVA,and classical approa ches to the analysis of huma nbehavior.PsychologicalBulletin, 73,311-349.

Kiesler, D. J. (1983). The 1982 interpersonal circle: A taxonomy forcomplementarity in human transactions.Psychological Review, 90,185-214.

Liebert, R . M., & Spiegler, M. D. (1987).Personality: Strategies andissues (5th ed.). Homewood, IL: Dorsey Press.

Loehlin, J. C . (1982). Are personality trai ts differentially heritable?Be-havior Gene tics, 12,417-428.

Loehlin, J. C., Horn, J. M., & Wil lerm an, L. ( 1981). Personality resem-blance in adoptive families.Behavior Genetics, 11,309-330.

Loehlin, J. C., & Nichols, R. C. (1976).Heredity, environment andper-



PERSONALITY SIMILARITY IN TWINS 103 9

sonality: A study of 850 sets of twins.Austin: University of TexasPress.

Lykken, D. T. (1968). Statistical significance in psychological research.Psychological Bulletin, 70,151-159.

Lykken, D. T. (1978). The di agnosis of zygosity in twins.Behavior Ge-netics, 8,437-473.

Lykken, D. T. (1982). R esearch with twins : The concep t ofemergenesis.Psychophysiology, 19,361-373.

Lykken, D. T., & Bouch ard, T. J., Jr. (1983/84). Genet ische Aspektemenschlicher Individualit~it [Genetic aspects of hum an individual-ity]. In H. V. Ditfurth (Ed.),Mannheimer Forum 1983/84(pp. 79-117). Hamburg : Hoffman an d Cam pe Verlag.

Lykken, D. T., Tellegen, A., & Iacono, W. G. (1982). EEG spectra intwins: Evidence for a neglected mecha nism o f genetic determ ination.Physiological Psychology, 10,60-65.

Martin, N. G., Eaves, L. J., Kearsey, M. J., & Davies, P. (1978). Thepower of the classical twin study.Heredity, 40,87-116.

McG ue, M., & Bouchard, T. J. (1984). Adju stment of twin data for theeffects of age and sex.Behavior Gen etics, 14,325-343.

Meehl, P. E. (1978). Theoretical risks and tabular risks: Sir Karl, Sir

Rona ld, and the slow progress o f soft psychology.Journal o f Consult-ing and Clinical Psychology, 46,806-834.Megargee, E. I. (1972). The California Psychological Inventory hand-

book.San Francisco: Jossey-Bass.Mischel, W. (1981). Introduction to personality(3rd ed.). New York:

Holt , Rinehart & Winston.Nance , W. E., & Corey, L. A. (1976). Ge netic mode ls for the analysis of

data from the families of identical twins.Genetics, 83,811-826.Nichols, R . C. (1978). Tw in studies o f ability, personality, a nd interests.

Homo, 29,158-173.Plomi n, R., & Daniels, D. (1987). W hy are children in the same family

so different from one anothe r?Behavioral and Brain Sciences, 10, 1-16.

Price, R. A., Vandenberg, S. G., Iyer, H., & Wil liams, J. S. (1982). C om-ponents o f variation in norm al personality.Journal ofPersonality andSocial Psychology, 43,328-340.

Rose, R. J., Boughman, J. A., Corey, L. A., Nance, W. E., Christian,J. C., & Kang, K. W. (1980). Data from kinships of monozygotictwins indicate maternal effects on verbal intelligence.Nature, 283,375-377.

Rowe, D. C., & Plom in, R. (1981). The imp ortance of nonshared (E l)enviro nmen tal nfluences in behavioral development.DevelopmentalPsychology, 17,517-531.

Rus hton , J. P., Fulker, D. W ., Neale, M. C., N ias, D. K . B., & Eysenck,H. J. (1986). Altruis m and aggression: To what extent are individ ualdifferences inherited?Journal of Personality and Social Psychology,50, 1192-1198.

Scarr, S., & McC artney, K. (1983). How people m ake their own en viron-ments: A theory of geno type-e nviro nmen t effects.Child Develop-

ment, 54,424-435.Scarr, S., Weber, P. L., Weinberg, R . A., & W ittig, M , A. (198 la ). Person-ality resemblance amo ng adolescents and their parents in biologicallyrelated an d adoptive families.Journal o f Personality and S ocial Psy-chology, 40,885-898.

Scarr, S., Weber, P. L., Weinberg, R. A., & Wittig, M. A. (1981b). Per-sonality resemblance among adolescents and th eir parents in biologi-cally related and adoptive families. In L. Gedda, P. Parisi, & W. E.Nance (Eds.), Twin Research 3: Part B. Intelligence, personality a nddevelopment(pp. 99- 120). New York: Ala n R. Liss.

Singer, J. L . (1984). The human personality,New York: Harco urt BraceJovanovich.

TeUegen, A. (1982). Brie f manual for the Differential Personality Ques-tionnaire. Unpublished manuscript, University of M innesota, M in-neapolis.

Tellegen, A. (1985). Structure of mo od and p ersonality and their rele-vance to assessing anxiety, with an em phasi s on self-report. In A. H .Tuma & J. D. Maser (Eds.),Anxiety and the anxiety disorders(pp.681-706). Hillsdale, NJ: Erlbau m.

Warr, P., Barter, J., & Brownbridge, G . (1983). On t he indep ende nce ofpositive and negative affect.Journal o f Personality and So cial Psy-chology, 44,644-651.

Watson, D., & C lark, L. A. (1984). N egative Affectivity: The disp ositio nto experience aversive emotional states.Psychological Bulletin, 96,465-490.

Watson, D., & Tellegen, A. (1985). Toward a consensual structure ofmood. Psychological Bulletin, 98,219-235.

Wiggins, J. S. (1982). Circump lex models o f interpersonal behavior inclinical psychology. In P. C. Kendall & J. N. Butcher (Eds.),Hand-

book of research methods in clinical psychology(pp. 183-221). NewYork: Wiley.

Zo nde rma n, A. B. (1982). Differential herita bility and consistency: Areanalysis of the NMSQTCPIdata. Behavior Genetics, 12,193-208.

Received Ma rch 21, 1986Revisio n received Augus t 31, 1987

Accep ted Oc tobe r 30 , 1987 9

Tellegen Et Al 1988 Personality Similarity in Twins Reared Apart and Together

Documents

Transcript of Tellegen Et Al 1988 Personality Similarity in Twins Reared Apart and Together