PERSONALITY PROCESSES AND INDIVIDUALDIFFERENCES

Differential Experiences of Siblings in the Same Family as Predictorsof Adolescent Sibling Personality Differences

Denise DanielsInstitute for Behavioral Genetics, University of Colorado, Boulder

In the area of personality development, environmental influences operate to make siblings in thesame family different rather than similar to each other. The goal of the present study was to deter-mine whether differential experience of siblings can be used to explain the marked personalitydifferences of siblings. The Sibling Inventory of Differential Experience (SIDE) along with personal-ity information (the EAS Temperament Inventory and questions about career expectations) wasadministered to 50 biological sibling pairs and 98 adoptive sibling pairs in adolescence and youngadulthood. The results indicated that differential sibling interaction and differential peer characteris-tics as self-reported on the SIDE explain 6%-26% of the variance in sibling personality differencescores. For example, the sibling who reports more sociability as compared to his sibling also experi-ences more sibling closeness and more peer popularity as compared to his sibling. Comparison be-tween adoptive and biological siblings indicates that the SIDE relations are mediated environmen-tally rather than genetically.

Environmental variance affecting any behavioral trait can besubdivided into a component for which family members aresimilar (shared or common family environmental variance) anda component for which family members are different (non-shared or differential family environmental variance). It is thelatter of these influences that is often the most salient in psycho-logical development. Behavioral genetic studies across many do-mains of behavior show that environmental variance is differ-ential rather than common for two siblings growing up in thesame family (Rowe & Plomin, 1981). It is particularly in thearea of personality development that nonshared environmentalvariance is so striking (Goldsmith, 1983). After correcting forerror variance (20%) there is an estimated 40% differential sib-ling environmental variance (Daniels, 1985). Sibling corre-lations range from only. 10 to .30 on standard personality ques-tionnaires. For example, in the Hawaii Family Study (Ahem,Johnson, Wilson, McClearn, & Vandenberg, 1982), the averagecorrelation for adolescent sibling pairs (N - 64-449) over 54personality traits (as measured by the Adjective Checklist, theEysenck Personality Inventory, the Comrey Personality Scales,

This article was written while the author was supported in part byNational Institute of Mental Health Grant MH-16 744.

Grateful appreciation is extended to the families who participated inthis study and to Lutheran Social Services of Colorado who helped incontacting the adoptive families.

Correspondence concerning this article should be addressed to De-nise Daniels, who is now at the Social Ecology Laboratory, TD-114,Department of Psychiatry and Behavioral Sciences, Stanford UniversityMedical Center, Stanford, California 94305.

the Sixteen Personality Factor Questionnaire, and the HawaiiAlcohol Questionnaire) was .10. This low sibling resemblanceis unsettling when one reflects that, on the average, siblingsshare 50% of the same genes, they are part of the same family,they are brought up by the same parents, and they grow up inthe same community environments. Furthermore, twin andadoption studies show that this small sibling resemblance ofabout a. 1 to .2 correlation can be accounted for by genetic fac-tors (Rowe & Plomin, 1981). Thus, none of the environmentalvariance is common for personality of siblings; all is differential.

The data that support this conclusion most clearly come fromadoption studies. Adoptive siblings grow up in the same family,yet are unrelated genetically, hence their resemblance directlyestimates shared sibling environmental variance (Plomin, De-Fries, & McClearn, 1980). Two large adoption studies with sib-ling pairs—the Minnesota Adoption Study (Scarr, Webber,Weinberg, & Wittig, 1981) and the Texas Adoption Project(Loehlin, Willerman, & Horn, 1982) have reported averageadoptive sibling correlations that are not significantly differentfrom zero—.02 and .01, respectively. This study is designed toexamine whether measured differential experiences of siblingscan be used to predict their personality differences.

Few previous studies on differential sibling environment-be-havior relations have been conducted. Family constellationvariables including birth order and gender of siblings have beena popular set of environmental variables that differ for siblings.Although extensively studied, the effects of birth order and gen-der of siblings are only weakly associated with sibling behavioraldifferences (Dunn, 1983; Scarr & Grajek, 1982). Gender of achild explains from 1% to 5% of the variance in both the areas

Journal of Peraonality and Social Psychology. 1986, Vol. 51. No. 2, 339-346Copyright 1986 by the American Psychological Association. Inc. 0022-3514/86/$00.75

339

340 DENISE DANIELS

of cognitive and personality development (Hyde, 1984; Jacklin,1981; Plomin & Foch, 1981). The research on birth order iseven more discouraging. Ernest and Angst (1983) in their re-view of over 1,000 birth-order studies concluded that the rela-tion between birth order and IQ, school achievement, occupa-tional status, and personality approaches zero when back-ground variables are controlled and when siblings from thesame family are compared.

In only one study of adolescent siblings (Daniels, Dunn, Furs-tenberg, & Plomin, 1985) has a search begun for variables, otherthan birth order and gender, within the family that might pre-dict the striking behavioral differences of siblings. In the Dan-iels et al. study 348 adolescent sibling pairs, as well as their par-ents, were interviewed regarding their perceptions of the familyenvironment. The major finding was that the sibling whoshowed more psychological adjustment (emotional stability,obedience, life satisfaction, and school adjustment) also experi-enced more maternal closeness, more sibling friendliness, moresay in family decisions, and more parental chore expectations,as compared to the other sibling. The Daniels et al. study dem-onstrated for the first time that sibling personality differencesdo systematically relate to sibling environmental differences.However, the authors point out that,

The main limitation of the'study is that it cannot separate causeand effect in the associations between differential experience withina family and sibling differences in emotional adjustment. A longi-tudinal study or one that controls for genetic influence, such as atwin or adoption study, is needed to resolve the issue of directionof effects, (p. 534)

In other words, a study that controls for genetic influence isneeded simply because nonshared sibling "environments"could relate to sibling personality differences for genetic, notenvironmental, reasons. Experience that differs for siblings,such as parents being more protective toward one sibling overthe other, could be mediated by genetic differences, rather thanby environmental differences between the siblings (Plomin,1986). Thus, differential sibling experience could then relate topersonality for genetic reasons as well.

There are two measurement weaknesses with the Daniels etal. (1985) study. First, measures of nonshared sibling experi-ence were created by using sibling difference scores. For exam-ple, each sibling was asked how many chores they were expectedto complete, and then one sibling's score was subtracted fromthe other sibling's score. A study that asks siblings to comparerelative differences in their family is in order (e.g., Who is ex-pected to complete more chores, you or your sibling?). Also, theDaniels et al. study mainly tapped parental treatment dimen-sions; other areas of sibling differential experience such as non-mutuality of sibling interaction or peer-group differences of sib-lings could be important variables that relate to sibling person-ality differences.

The present study of adolescent and young adult siblings ex-tends the Daniels et al. (1985) study. First, adoptive and non-adoptive siblings were studied to control for genetic influenceand to address the issue of cause and effect in differential siblingenvironment-behavior relations. Second, the Sibling Inventoryof Differential Experience (SIDE; Daniels & Plomin, 1985) was

used for the first time to relate to sibling personality differences,because it has the advantage of directly assessing relative differ-ences within the family and it covers a wide variety of siblingdifferential experience.

A sample of both biological (50% genetically related) siblingsand adoptive (genetically unrelated) siblings completed ques-tionnaires on their personalities (the HAS Temperament Inven-tory, Buss & Plomin, 1984, and questions about careerexpecta-tions) and, also, the SIDE. The SIDE asks each sibling to com-pare his or her experiences to those of one of his or her siblingsin the domains of sibling interaction, parental treatment, peercharacteristics, and events specific to the individual. Previouswork (Daniels & Plomin, 1985) shows that from 40% to 65% ofa sibling sample reported different environments across eachof these SIDE domains. In addition, the SIDE scales show asubstantial amount of variance.

Another aspect of the SIDE is that it shows little genetic in-fluence, implying that the experiences it taps are primarily envi-ronmental in origin. Any relation between the SIDE and behav-ior should be primarily an environmental one. In the presentstudy, adoptive and nonadoptive siblings are used to verify envi-ronmental mediation of SIDE-sibling personality difference re-lations. Finally, it should be mentioned that the SIDE showslittle relation to birth order and gender of siblings (Daniels &Plomin, 1985). Hence, SIDE-personality relations themselvesare not likely to be a function of the birth order or gender of thesiblings.

Method

Sample

The sample includes 50 biological and 98 adoptive sibling pairs. Theadoptive siblings consist of pairs in which both siblings were adopted(56) and pairs in which one sibling was adopted and the other is a biolog-ical child of the parents (42). Both types are considered to be unrelatedgenetically. The sample, from the Denver metropolitan area, is of Cau-casian ancestry. The recruitment procedures are described in greaterdetail elsewhere (Daniels & Plomin, 1985).

The mean years of education for the fathers and mothers of the sib-lings were 15.97 and 14.88, respectively. Based on Siegel Prestige Scores(Hauser & Featherman, 1977), the mean socioeconomic status score ofthe fathers was 54.02, with a standard deviation of 13.99. The meanscore is approximately 1 SD above the mean score for the white laborforce in the United States; however, the standard deviation in this sam-ple is comparable to the national sample (Siegel, 1971).

The siblings ranged in age from 12 to 28 years; the mean age was17.96 years, with a standard deviation of 4.00 years. In families withmore than two children, pairs were selected based on their adoptive sta-tus and gender, to ensure a representative amount of each pair type. Ofthe 148 sibling pairs, 61 were opposite sex and 87 same sex (35 brotherpairs and 52 sister pairs). Family size ranged from two to seven children(M =3.10, SD = 1.45). Most of the sibling pairs (85) were first-secondborns, 13 were first-third borns, 24 were second-third corns, 5 weresecond-fourth borns, 9 were third-fourth borns, and 12 were othercombinations (fourth-fifth, third-sixth, etc.).

Measures

Siblings were mailed two questionnaires—the Sibling Inventory ofDifferential Experience and the EAS Temperament Inventory.

SIBLING DIFFERENCES IN EXPERIENCE AND PERSONALITY 341

The SIDE is a 73-item instrument, tapping sibling differential experi-ence in the areas of sibling interaction, parental treatment, peer charac-teristics, and events specific to the individual. There are 11 SIDE scalesthat describe sibling differential experience: differential sibling antago-nism, differential sibling caretaking, differential sibling jealousy, differ-ential sibling closeness, differential maternal affection, differential ma-ternal control, differential paternal affection, differential paternal con-trol, differential peer college orientation, differential peer delinquency,and differential peer popularity. Each of these 11 scales are used in thisstudy; however, items on events specific to the individual (accidents,effects of teachers, etc.) are not used.

The development of the SIDE questionnaire, including factor analy-ses, scale construction, scoring procedures, and psychometric proper-ties are described in detail by Daniels and Plomin (1984,1985). A cou-ple of these characteristics of the SIDE should be briefly mentioned.First, the intercorrelations among the SIDE scales are low to moderate.Second, test-retest reliabilities range from .77 to .93. Third, the SIDEuses two scoring systems, relative and absolute scoring. The relativescoring is used in this study because it provides information concerningthe amount and direction of differential experience, A sample question,"In general, who has been more willing to help the other succeed overthe years?" can be answered my sibling has been much more this waythan I have (1), my sibling has been a bit more this way than I have (2),my sibling and I have been the same to each other in this way (3), / havebeen a bit more this way than my sibling (4), and / have been much morethis way than mysibling(5). Each SIDE question, for the present sampleof 12- to 28-year-olds, is phrased so that individuals respond by averag-ing over the years when they were growing up and living at home. Conse-quently, developmental stage or age variance is negligible in the presentsample (see Daniels & Plomin, 1985).

In families with more than two children, pairs were selected on thebasis of adoptive status and gender. Hence, each sibling was instructedto rate the other sibling of his or her pair (i.e., Sibling A was instructedto rate Sibling B, and Sibling B was instructed to rate Sibling A). Bytelling each sibling who to rate, bias in sibling differential experiencecould be controlled, in that subjects were not able to choose the mostdifferent sibling to rate.

The siblings also completed the EAS Temperament Inventory (Buss& Plomin, 1984), a 30-item questionnaire assessing the following scales:emotionality-anger, emotionality-fear, emotionality-distress, activity,sociability, and shyness. This questionnaire and its psychometric prop-erties are described elsewhere (Buss & Plomin, 1984). In addition, theadolescents were asked how many years of education they expected tocomplete and what occupation they expected to achieve. Occupationswere coded based on Siegel Prestige Scores (Hauser & Featherman,1977).

Overview of Analyses

Descriptive information. For each SIDE scale, each EAS scale, andeach item on career expectations, means and standard deviations willbe examined for the whole sample and for boys, girls, adopted children,and nonadopted children.

Sibling resemblance for personality and career expectations. A sec-ond set of analyses will determine how similar or different siblings arefor the personality measures. Correlations are conducted for all siblingpairs, biological sibling pairs, adoptive sibling pairs, brother pairs, sisterpairs, and brother-sister pairs.

Family constellation variables related to personality and career expec-tations. Birth order, age, and gender of the siblings will be correlatedwith the personality measures to determine the extent to which familyconstellation variables are a source in creating sibling personalitydifferences.

Table 1Means and Standard Deviations for the EAS TemperamentScales and the Career Expectation Items

Personality measures Mean score SD

Emotionality-angerEmotionality-fearEmotionality-distressActivitySociabilityShyness

Expected educationExpected occupation

13.112.913.017.417.115.3

16.653.0

3.83.53.73.63.93.13.0

14.7

Sibling differential experiences related to sibling personality differ-ences. Before relating the SIDE to sibling personality differences, sib-ling difference scores for the personality measures need to be created.Thus, for each of the EAS scales and for the career expectation items,one siblings' score is subtracted from the scores of the others. The signeddifference (+ or -) assesses both the amount and direction of the siblingdifference. For example, if one randomly selected member of a siblingpair (Sibling A) reports that he or she is much more sociable comparedto the other sibling's report (Sibling B), subtracting Sibling B's scorefrom Sibling A's score will yield a large positive difference score.

Another preliminary step before relating the SIDE to sibling person-ality differences is to create sibling average scores for each of the SIDEscales. Sibling agreement on the SIDE scales ranges from .23 to .73, themedian agreement correlation is .49. Thus, because the siblings agreeto a moderate extent regarding their differential experiences, aggregat-ing their scores will increase the reliability of the SIDE scores. Further-more, this procedure simplifies the analysis in that there will be just oneset of SIDE scales (the sibling averages) rather than two sets (Sibling A'sreport and Sibling B's report) to relate to personality differences.

To reduce the large number of possible correlations (88), multipleregressions will be conducted systematically within each domain—differential sibling interaction, differential parental treatment, anddifferential peer characteristics—to predict the eight sibling personalitydifference scores. Because the SIDE shows little genetic variance, SIDE-sibling personality difference relations should not show much geneticmediation. Nonetheless, multiple regressions will be conducted foradopted versus biological siblings to verify environmental mediation ofthe relations.

ResultsDescriptive Information

Means and standard deviations for the personality measuresare shown in Table 1. The EAS means and standard deviationsare similar to previously reported means and standard devia-tions for adolescents and young adults (Buss & Plomin, 1984).Regarding the career-expectation items, it can be seen that thesample, on the average, expects to obtain a college education(i.e., 16 years of schooling). The mean score for expected occu-pation is 52.96, based on Siegel Prestige Scores, which is equiva-lent to a manager or technician's score.

Means next were examined for boys versus girls and biologi-cal versus adoptive children. A few significant mean differencesfor the personality measures arose, but they only accounted forfrom 1% to 2% of the variance; hence, gender differences and

342 DENISE DANIELS

Table 2Means and Standard Deviations for the SIDE Scales

SIDE scale Mean score SD

Differential sibling antagonismDifferential sibling jealousyDifferential sibling caretakingDifferential sibling closeness

Differential maternal affectionDifferential maternal controlDifferential paternal affectionDifferential paternal control

Differential peer college orientationDifferential peer delinquencyDifferential peer popularity

2.933.042.603.10

2.963.133.053.09

3.263.103.20

0.630.790.750.74

0.560.660.590.71

0.660.860.76

Note. SIDE = Sibling Inventory of Differential Experience.

whether one is adopted or not shows little impact on the person-ality measures.

Table 2 presents means and standard deviations for the 11SIDE scales (based on the relative scoring system). Most of themean scores for this sample center around 3, the average re-sponse (me and my sibling experience the same). However, thestandard deviations for the SIDE scales (e.g., 0.66, 0.79) indi-cate that many siblings are experiencing their environmentsdifferently (from sibling more than me to me more than sibling).

Mean relative SIDE scores were examined for boys versusgirls and adoptive versus biological children. Significant meandifferences occurred on a few scales, again accounting for onlyfrom 1% to 2% of the variance. Note that comparing adoptiveand nonadoptive siblings on relative SIDE scores is not a testfor genetic influence. To examine genetic influence on SIDEscales, absolute scoring must be used. Sibling Inventory ofDifferential Experience scores based on the absolute scoringsystem, have been reported elsewhere for a sample of 222 bio-logical and 174 adoptive siblings (Daniels & Plomin, 1985).With this large sample size, significant genetic influence wasdetected for the SIDE absolute scores; however, it accounted foronly from 4% to 12% of the SIDE variance. Hence, the SIDEappears to be virtually uncontaminated by genetic influence,even though substantial variance in SIDE sibling experience ex-ists overall.

Sibling Resemblance for Personality andCareer Expectations

Table 3 presents sibling correlations for the personality mea-sures for all sibling pairs, biological sibling pairs, adoptive sib-ling pairs, brother-brother pairs, sister-sister pairs, andbrother-sister pairs. The results in Table 3 indicate that siblingsdo not resemble each other much in the temperament traits.The only significant sibling resemblance for temperament oc-curs for biological sibling pairs and for brother pairs on the traitemotionality-distress. Some of the sibling correlations for thecareer expectation items are significant; however, they too arelow. Biological sibling resemblance tends to be higher than

adoptive sibling resemblance, suggesting that some genetic vari-ance is involved in the personality traits.

The correlations presented in Table 3, although generally lowand nonsignificant, are not atypical of previous reports of sib-ling resemblance. The median personality correlation is. 12 forbiological siblings. Prior studies have reported average biologi-cal sibling correlations of. 10 (Ahern et al., 1982),. 18 (Daniels,1985), .20 (Loehlin et al., 1981), and .19 (Scarr et al., 1981).In this study, the median personality correlation for adoptivesiblings is -.04. Prior studies have reported average adoptivesibling correlations of .02 (Loehlin et al., 1981) and .01 (Scarret al., 1981). The main point to be gathered from Table 3 isthat sibling resemblance is low, implying that nonshared siblingenvironmental variance must be important for this sample andthis set of personality data.

Family Constellation Variables Related to Personalityand Career Expectations

Before relating the SIDE scales to the personality measures,family constellation variables were evaluated as possiblesources of differential sibling experience on behavior. Table 4shows associations between personality and career expectationsand gender, age, and birth order of the siblings. A few corre-lations are significant between gender and personality: Girls re-port more sociability and emotionality-fear than do boys; boysreport more shyness and more expected years of education thando girls. Age shows significant negative correlations with socia-bility and emotionality-anger: Younger siblings report greatersociability and emotionality-anger. Birth order relates signifi-cantly to sociability: Later born siblings report more sociability.These significant correlations indicate that only from 1% to 5%of the variance can be explained by family constellation vari-ables.

Sibling Differential Experiences Related to SiblingPersonality Differences

The SIDE scales within each domain (differential sibling in-teraction, differential parental treatment, and differential peercharacteristics) were used to predict each of the personalitydifference scores (the 6 sibling temperament differences and the2 sibling career-expectation differences) in multiple regressions.These 24 regressions, 8 within each domain, were carried outfor the whole sample, biological siblings and adoptive siblings.As expected, the multiple regression results (R, R2, and adjustedR2) did not differ for biological and adoptive siblings indicatingenvironmental mediation of the relations. Because the resultsfor biological and adoptive siblings were nearly identical to re-sults for the whole sample, results based on the whole sampleare presented.

Table 5 shows regression of sibling personality differences onthe SIDE differential sibling interaction scales. The table in-cludes SIDE scales that were significant (or significant incre-mental) predictors of the sibling personality differences, the R,R2, and the adjusted R2 for each regression. Four of the eightmultiple regressions are significant, each in the realm of tem-perament rather than career expectations. The adjusted R2s in-

StBLING DIFFERENCES IN EXPERIENCE AND PERSONALITY 343

Table 3Sibling Correlations for the Personality Measures

Siblings

Personality measures

Emotionality-angerEmotionality-fearEmotionality-distressActivitySociabilityShynessExpected educationExpected occupation

N

Biological

.13

.20

.32*

.04-.12-.24

.11

.31*

45-50

Adoptive

-.11-.01-.04-.02-.14-.04

.19*

.04

84-98

Brother

.01

.11

.29*-.05

.05

.23

.42*

.29

24-35

Sister

-.14.18.02.11

-.21-.20

.11

.51*

42-52

Pairs

Brother-sister

.06-.09

.07-.04-.20-.19

.11-.17

53-61

All

-.02.05.08.00

-.14-.11

.21*

.19*

129-148

*p<.05.

dicate that from 6% to 11% of the variance in the sibling person-ality differences can be predicted by the SIDE scales. The firstregression, in which both of the predictors bear a positive rela-tion, indicates that the sibling who reports more emotionalityanger as compared to his or her sibling also experiences moresibling antagonism and more sibling jealousy as compared tohis or her sibling. The other regressions can be interpreted in asimilar manner. For instance, the sixth regression indicates thatthe sibling who reports more shyness as compared to his or hersibling, experiences less sibling closeness and less sibling care-taking as compared to his or her sibling.

Table 6 presents the results of regression of sibling personalitydifferences on the SIDE differential parental treatment scales.The only significant regression involves the prediction of siblingdifferences in what they expect their occupation to be. The sib-ling who expects to achieve more in his or her occupation ascompared to his or her sibling, also experiences more paternalaffection. This SIDE prediction explains 9% of the variance inthe sibling personality difference.

Finally, Table 7 shows regression of sibling personalitydifferences on SIDE differential peer characteristic scales. Fiveof the eight regressions are significant. The adjusted J?2s indi-cate that from 6% to 26% of the variance in the sibling personal-

Table 4Associations Between Personality Measuresand Gender, Age, and Birth Order

Personality measures Gender" Age Birth order

Emotionality-angerEmotionality-fearEmotionality-distressActivitySociabilityShyness

Expected educationExpected occupation

-.08.18*.04.05.13*

-.06

-.16*-.01

-.16*-.05-.01

.09-.23*-.09

.07

.10

.05

.06

.10

.01

.12*

.10

-.02.02

* Gender coded as males (1), females (2).*p<.05.

ity differences can be explained by the SIDE scales. As an exam-ple, the second regression, in which the two predictors bear anegative relation to the sibling personality difference, indicatesthat the sibling who reports more emotionality-fear as com-pared to his or her sibling, experiences less peer delinquencyand less peer college orientation. The remaining significant re-gressions can be interpreted in a similar fashion. For example,the last regression shows that the sibling who expects to achievemore in his or her occupation, as compared to his or her sibling,also experiences a more college-oriented peer group.

Discussion

Sibling differential experience as tapped by the SIDE wasused for the first time to predict sibling personality differences.The findings, based on a sample of both biological and adoptivesibling pairs, suggests that relations between the SIDE and sib-ling personality differences are environmentally mediated. TheSIDE differential sibling interaction scales in multiple regres-sion analyses explain from 6% to 11% of the variance in siblingpersonality difference scores. For example, the sibling who re-ports more sociability as compared to his or her sibling alsoexperiences more sibling closeness, and the sibling who reportsmore shyness as compared to his or her sibling experiences lesssibling antagonism.

Regression of sibling personality differences on SIDE differ-ential parental treatment scales showed that the sibling who ex-pects to achieve more in his or her occupation also experiencesmore paternal closeness, as compared to his or her sibling. Thiswas the only significant prediction from the differential parentaltreatment scales, and it explained 9% of the variance. With re-gard to the SIDE differential peer characteristic scales, from 6%to 26% of the variance can be explained in sibling personalitydifferences. For example, the sibling who reports more emo-tionality-anger as compared to his or her sibling, experiences aless college-oriented peer group, and the sibling who reportsmore sociability as compared to his or her sibling also experi-ences a more popular peer group as compared to his or her sib-ling.

The question that would most obviously arise from the re-

344 DENISE DANIELS

Table 5Regression of Sibling Personality Differences on SIDE Differential Sibling Interaction Scales

Sibling personalitydifference

Emotionality-anger

Emotionality-fear

Emotionality-distressActivitySociabilityShyness

Expected educationExpected occupation

Significant SIDE scale predictors

Differential sibling antagonism (+)Differential sibling jealousy (+)Differential sibling caretaking (-)Differential sibling antagonism (+)Differential sibling jealousy (+)

—Differential sibling closeness (+)Differential sibling closeness (-)Differential sibling caretaking (-)Differential sibling jealousy (-)

—

R

.30

.36

.18

—.36

.37

.25

—

R1

.09

.13

.03

—.14

.14

.06

—

Adjusted R2

.06

.07

.01

—.11

.11

.03

—

Overallsignificance

p<.0l

p<.0lnsns

jx.OOl

p < .001nsns

Note. N = 144 for the HAS Temperament traits; n = 127 for expected education; n = 83 for expected occupation. SIDE = Sibling Inventory ofDifferential Experience. + = positive correlation; - = negative correlation; ns = nonsignificant.

suits just summarized is, Do differences in sibling experience rations. This is actually not so questionable, inasmuch as thelead to differences in sibling personality differences, or visa SIDE instrument shows little genetic influence and the person-versa? For example, Do differences in the popularity of the sib- ality traits in the present study showed only a moderate amountling's peer groups cause one to be more sociable than the other, of genetic variance. For the personality traits, sibling resem-or Do differences in the sibling's sociability cause one to join a blance in this study was indeed low for both biological and adop-more popular peer group than the other? The present study can- live siblings. Although not atypical of prior reports of siblingnot fully answer this question; however, it could at least elimi- resemblance, the low and negative sibling correlations implynate the reasonable hypothesis that genetic differences of the that contrast effects between siblings may be important (Buss &siblings are involved. Thus, it could be that environmentally Plomin, 1985). The fact that subjects knew the study was aimed(not genetically) influenced differences in sibling sociability at examining sibling differences could have built some contrastcould lead siblings to join different peer groups, or environmen- effects into their self-reports of personality. However, this shouldtally (not genetically) influenced differences in siblings peer equally effect biological and adoptive siblings, resulting in angroups could lead to differences in sibling sociability. To the increase of nonshared environmental variance but not a neces-extent that genetic influences are not involved, the latter state- sary decrease in genetic variance.ment becomes more plausible simply because differences in Turning to the SIDE scales, it was known prior to this studypeer groups offers an environmental mechanism leading to that genetic variance is minimal. Thus, a much larger sampledifferences in sociability. In terms of the cause-effect dilemma, would have been necessary to detect very small differences (. Iit is important in future research to focus on environmental to .2 correlations) between biological and adoptive siblings, reli-processes and theories in nonshared environment-personality ably. However, the fact that correlational differences were essen-relations. tially zero between adoptive and nonadoptive siblings for SIDE-

The next question to be considered is why genetic influences personality associations suggests that sample size is not the onlyare not important in the SIDE-sibling behavioral difference re- consideration. Alternatively, the subtle "micro" environments

Table 6Regression of Sibling Personality Differences on SIDE Differential Parental Treatment Scales

Sibling personalitydifference

Emotionality-angerEmotionality-fearEmotionality-distressActivitySociabilityShynessExpected educationExpected occupation

Significant SIDE scale predictors

Differential paternal control (-)

Differential maternal affection (+)Differential maternal control (-)Differential maternal affection (+)Differential paternal affection (+)

R

.14

.19

.24

.21

.36

R2

.02

.04

.06

.04

.13

Adjusted R*

.00

.01

.02

.01

.09

Overallsignificance

nsnsnsnsnsnsns

p<.0l

Note. N = 143 for the EAS Temperament traits; « = 126 for expected education; n = 82 for expected occupation. SIDE = Sibling Inventory ofDifferential Experience. + = positive correlation; - = negative correlation; ns = nonsignificant.

SIBLING DIFFERENCES IN EXPERIENCE AND PERSONALITY 345

Table 7Regression of Sibling Personality Differences on SIDE Differential Peer Characteristic Scales

Sibling personalitydifference

Emotionality-angerEmotionality-fear

Emotionality-distressActivitySociabilityShynessExpected educationExpected occupation

Significant SIDE scale predictors

Differential peer college orientation (— )Differential peer delinquency (-)Differential peer college orientation (-)

— .—

Differential peer popularity (+)

—Differential peer college orientation (+)Differential peer college orientation (+)

R

.34

.27

——.53

—.35.34

R1

.12

.07

——.28

—.12.12

Adjusted/?2

.09

.06

——.26

—.09.08

Overallsignificance

p<.01

p<.02nsns

p<.001ns

p<.0lp<M

Note. N = 102 for the EAS Personality traits; n = 87 for expected education;n = 55 for expected occupation. SIDE = Sibling Inventory of DifferentialExperience. + = positive correlation; - = negative correlation; ns = nonsignificant.

that siblings experience could in fact be relatively insensitiveto genetic influence. In other words, the realm of nonsharedenvironmental influences may be one in which "environmen-tal" interactions go beyond the genetic makeup of individualfamily members.

Another limitation of the present study is its reliance on serf-reports of personality and the environment. This is a weaknessinasmuch as parental reports, interview techniques, and directobservation could yield different results. Nonetheless, the ado-lescent and young-adult perceptions are reliable (test-retest forthe measures in this study are about .8) and show some validity(interrater agreement for the measures in this study centersaround .4). Moreover, the opinion of adolescents and youngadults should be recognized as interesting information standingon its own.

The relations that have emerged in this study, using the SIDE,have explained between 6% and 26% of the variance in siblingpersonality differences. Why so little? Note that a similaramount (4%-13%) of the variance was explained in the Danielset al. (1985) study. Furthermore, in studying the relation be-tween nonshared environmental variance and sibling differ-ences, one is not dealing with 100% of the variance. Some ge-netic variance is included, and all error variance is included, sothat from 6% to 26% of what is left after subtracting genetic anderror variance suggests that the cup is half full rather than halfempty.

Another issue is that differential parental treatment as as-sessed by the SIDE showed little relation to sibling personalitydifferences, whereas the domains of sibling differential interac-tion and peer group differences showed a greater relation. Dan-iels and Plomin (1985) reported that in the domain of parentaltreatment (as tapped by the SIDE), only 40% of siblings re-ported differences. However, for the areas of differential siblinginteraction and peer characteristics, 65% of the sample reporteddifferences. Hence, the domain of differential treatment by par-ents may show fewer relations to sibling personality differencessimply because there is less variance.

If restriction of range is not a problem, it is possible that thedomains of differential sibling interaction and peer characteris-tics are more important sources creating sibling personalitydifferences than are parents. On the other hand, a different set

of parental treatment variables (closeness, chore expectations,and child's say in decisions, rather than control and affection)were successful in predicting sibling personality differences inthe Daniels et al. (1985) study. At this time more work identify-ing domains and variables of differential sibling experienceneeds to be carried out before certain domains or variables, orboth, are excluded. At the least, sibling interaction and peer-group characteristics should not be overlooked in future studiesof nonshared sibling experiences, because they offered severalsignificant predictors in this study.

In conclusion, this study demonstrated for the first time thatthe SIDE is an effective predictor of sibling personality differ-ences. Personality and other behavioral differences of siblingsare known to be striking. As more studies begin to address be-havioral differences of siblings growing up in the same family,the SIDE should prove to be a useful environmental instrument.By using adoptive and nonadoptive sibling pairs, we found thatthe SIDE-personality relations were indeed environmentallymediated. This suggests that the SIDE, which assesses micro-environments of siblings in the same family, goes beyond thegenetic makeup of individual family members.

References

Ahern, F. M., Johnson, R. C, Wilson, J. R. McClearn, G. E., & Vanden-berg, S. G. (1982). Family resemblance in personality. Behavior Ge-netics, 12, 261-281.

Buss, A. H., & Plomin, R. (1984). Temperament: Early developing per-sonality traits. Hillsdale, NJ: Erlbaum.

Daniels, D. (1985). Understanding the family environment: A study ofadoptive and nonadoptive infant siblings. Doctoral dissertation, Uni-versity of Colorado, Boulder.

Daniels, D., Dunn, J., Furstenberg, F. F, & Plomin, R. (1985). Environ-mental differences within the family and adjustment differenceswithin pairs of adolescent siblings. Child Development, 56, 764-774.

Daniels, D., & Plomin, R. (1984). The Sibling Inventory of DifferentialExperience. Boulder, CO: University of Colorado, Institute for Behav-ioral Genetics.

Daniels, D., & Plomin, R. (1985). Differential experience of siblings inthe same family. Developmental Psychology, 21,747-760.

Dunn, J. (1983). Sibling relationships in early childhood. Child Devel-opmenl, 54,787-881.

346 DEMISE DANIELS

Ernst, C., & Angst, J. (1983). Birth order: Us influence on personality.New York: Springer-Verlag.

Goldsmith, H. H. (1983), Genetic influences on personality from in-fancy to adulthood. Child Development, 54, 331-355.

Hauser, R. M, & Featherman, D. C. (1977). The process of stratifica-tion: Trends and analyses. New York: Academic Press.

Hyde, J. S. (1984). How large are gender differences in aggression? Adevelopmental meta-analyses. Developmental Psychology, 20. 722-736.

Jacklin, C. N. (1981). Methodological issues in the study of sex-relateddifferences. Developmental Review, 1.266-273.

Loehlin, J. C, Willerman, L., & Horn, J. M. (1982). Personality resem-blances between unwed mothers and their adopted-away offspring.Journal of Personality and Social Psychology, 42, 1089-1099.

Plomin, R. (1986). Genetics, development, and psychology. Hillsdale,NJ: Erlbaum.

Plomin, R., & Foch, T. (1981). Sex differences and individual differ-ences. Child Development, 52, 383-385.

Plomin, R., DeFries, J. C, &McClearn, G. (1980). Behavioral genetics:A primer. San Francisco: Freeman.

Rowe, D. C, & Plomin, R. (1981). The importance of non-shared (E,)environmental influences in behavioral development. DevelopmentalPsychology, 17, 517-531.

Scarr, S., & Grajek, S. (1982). Similarities and differences among sib-lings. In M. E. Lamb & B. Sutton-Smith (Eds.), Sibling relationships:Their nature and significance across the lifespan (pp. 357-381). Hills-dale, NJ: Erlbaum.

Scarr, S., Webber, P. I., Weinberg. R. A., & Wittig, M. A. (1981). Person-ality resemblance among adolescents and their parents in biologicallyrelated and adoptive families. Journal of Personality and Social Psy-chology, 40, 885-898.

Siegel, P. M. (1971). Prestige in the American occupational structure.Doctoral dissertation, University of Chicago.

Received August 16, 1985Revision received October 31,1985 •