Children's Perceptions of the Qualities of Sibling ... · Children's Perceptions of the Qualities of Sibling Relationships ... yielded a list of 15 salient qualities. In the second

Children's Perceptions of the Qualities ofSibling Relationships

Wyndol Furman and Duane Buhrmester

University of Denver

FURMAN, WYNDOL, and BUHRMESTER, DUANE. Children's Perceptions of the Qualities of SiblingRelationships. CHILD DEVELOPMENT, 1985, 56,448-461. Although many studies of family constella-tions exist, only recently have investigators begun to examine the qualities of sibling relationships.The purpose of the present investigation was to develop a systematic fl"amework for describing andassessing such relationship qualities. In the first study, upper elementaiy school children wereinterviewed about their perceptions of the qualities of their sibling relationships. These inteiviewsyielded a list of 15 salient qualities. In the second study, a self-report questionnaire that assessedtheir perceptions of these qualities was administered to a sample of 198 fifth- and sixth-gradechildren. A principal components analysis yielded 4 underlying factors: {a) Warmth/Closeness, {b)Relative Status/Power, (c) Conflict, and {d) Rivalry. Relative Status/Power was found to be sb'onglyrelated to the relative ages of the child and sibling. The other 3 factors were also related to variousfamily constellation variables, but these relations were modest in size. Because family constellationvariables and the qualities of sibling relationships ai'e not isomoiphic with one another, it is impor-tant to study relationship qualities directly, rather than simply examining family constellation vari-ables. Some of the determinants of such qualities are discussed.

Siblings are an integral part of most cbil- ling relationships in order to understand tbedren's social worlds (Furman & Buhrmester, infiuence siblings bave.1982; Lamb & Sutton-Smith, 1982). The emo- . . i, i - i i-^i utional ties between siblings are commonly , As yet, however, relatively little researchsecond in strength only to those between pa- ^^' been conducted on the qualities of siblingrents and children (Irish, 1964). Brothers or relationships. Certainly, many investigatorssisters can be a source of frequent companion- \^Xf examined the effects of stmctural or con-ship, help, or emotional support. Older sib- stellataon variables, such as ordinal positaoo,lings can serve as caretakers, teachers, or ' ^^ f ' ' ^^^"^ 'T Jo^n^m""^ ^'% ^ Wmodels; in some instances they can even help ^J^'^ ^ ^ ° ' ? ± T t . ' Wagner, Schubert,compensate for absent or distant parents. In % J ^ t 'i ^ ®. ', "* '""'T "^ "^ .Ttheir interactions with each other, siblings "^ stmctural vanables and research on themay acquire many social and cognitive skills ^^^^^^^^ °f sibling relationsbips are not one

that are cenfral to healthy social development. ^f ^f 'l^}""' ^ ^ T™' ^ '^ '^^ '^ ^^} ^^^ Tlties 01 sibling relationships are exclusively or

Whereas siblings may bave major effects even primarily determined by family constel-on one another's development, the specific lation variables. Differences in sibling rela-nature of tbe infiuence can vary considerably. tionships may occur even when the same fam-One reason is that tbere is marked diversity in ily constellation exists. Additionally, it wouldtbe qucJities or characteristics of sibling rela- be incorrect to explain family constellation ef-tionsbips. Sibling relationships can be egali- fects solely in terms of their impact on thetarian or asymmetrical in terms of power and qualities of sibling relationships. Structural orstatus. The affective tone can also vary; rela- constellation variables may affect the qual-tionships may be close or distant, barmonious ities of parent-child or marital relationships asor confiicted, cooperative or competitive. It is well as those of sibling relationsbips. Tbus,essential tbat we examine the qualities of sib- although studies of family constellation vari-

This research was supported by a grant no. lROlHD 16142 from the National Institute of ChildHealth and Human Development (Wyndol Funnan, Principal Investigator). Portions of this researchwere presented at the International Conference on Personal Relations, Madison, Wisconsin, 1982,and at the Society for Research in Child Development, Detroit, Michigan, 1983. Appreciation isexpressed to Wendy Ritz for her assistance in the data collection. We are also indebted to the facultyand students of St. Louis Elementary School and Van Dellen Elementary School. Reprints of thispaper can be obtained from Wyndol Furman, Department of Psychology, University of Denver,Denver, CO 80208.

[Child Development, 1985, 56, 448-461. © 1985 by the Society for Research in Child Development, Inc.All rights reserved. 0009-3920/85/5602-0017$01.00]

Furman and Buhrmester 449

ables can be valuable, they should not be sub-stituted for studies of the qualities of siblingrelationships.

A few investigators have begun to studythe qualities of sibling relationships, particu-larly in early childhood (see Dunn, 1983;Lamb & Sutton-Smith, 1982). For example,some researchers have examined siblings'perceptions and attitudes toward each other(Bowerman & Dobash, 1974; Koch, 1960).Similarly, observational studies have beendone on tutoring by siblings (Cicirelli, 1972,1973), cooperative and competitive behavior(Bryant & Crockenberg, 1980; Minnett, Van-dell, & Santrock, 1983), and naturalistic pat-tems of interaction in tbe bome (Ab-ramovitcb, Pepler, & Corter, 1982; Dunn &Kendrick, 1982).

The results of these studies are encourag-ing, but the field still lacks a systematic way tocharacterize the qualities of sibling relation-ships. Most investigators have focused on aparticuleir aspect of sibling relationsbips, sucbas rivalry, witbout trying to capture the mul-tifaceted nature of the relationships. Simi-larly, many researchers have focused on sib-ling interactions in circumscribed contexts,sucb as during tutoring or when the mother ispresent. The veiriables incorporated in ques-tionnaires or observational coding schemeshave varied from study to study. We need asystematic framework for portraying the qual-ities of sibling relationsbips if we are tounderstand their infiuence on development.

The metbods of data collection have alsovaried and have included interviews, self-report questionnaires, structured tasks, andnaturalistic observations. Olson (1977) pro-posed two distinctions that are useful for con-ceptualizing how the various methods yielddifferent information about relationships. Hefirst distinguishes between insiders' and out-siders' descriptions of relationships. In histerminology, an insider is a participant ormember of the relationship being studied (inthis case, a sibling), whereas an outsider issomeone not involved in the relationship(typically a social scientist). An insider's de-scription of a relationship, as might be ob-tained in an interview, can provide a ricb pic-ture of the history and current status of arelationship. An insider is also sensitive tothe private meaning of a behavior and caninterpret behaviors witbin the broad contextof tbe relationship. On the other hand, anoutsider can provide a detached perspectiveand may be in a better position to compareand contrast the qualities of different relation-ships.

Olson also distinguished between objec-tive and subjective forms of data. The formerare measures of what actually occurs in a rela-tionship (e.g., observational coding systems),wbereas the latter are measures of percep-tions that involve some interpretation or judg-ment by a rater (e.g., most questionnaires orrating scales). Both forms of data can be ob-tained from either an insider or outsider's per-spective (see Furman, 1984, for further dis-cussion).

Olson persuasively argued that it is nec-essary to incorporate multiple perspectivesand multiple forms of data in order to obtain acomprehensive picture of relationships. Tbedifferent types of data each provide valuableand some^vbat different infonnation about thequalities of relationships. To obtain sucb acomprebensive picture of sibling relation-sbips, bowever, we first need a common set ofrelationsbip qualities tbat can be assessedfrom different perspectives or in differentforms of data.

The purpose of the present study was tobegin tbe process of developing such a com-mon framework. As a first step, we identifiedthe set of qualities that refiected insiders' sub-jective perceptions of tbeir relationships. Inparticular, cbildren were interviewed abouttbe qualities of their relationsbips witb sib-lings. Tbe descriptors the cbildren used wereexpected to fall into meaningful categoriesand reveal tbe subjectively important qual-ities of their relationsbips.

It would bave been possible to start witbanotber form of data, but there are several ad-vantages to beginning with insider subjectivedata. The qualities used by children are likelyto reflect natural ways of encoding and ag-gregating information about relationships.Additionally, the list of qualities is likely toinclude those that occur in a broad range ofsettings, many of which are not accessible tooutside observers. Certain qualities, sucb asexpressions of overt rejection, may occur in-frequently and may seldom be witnessed byoutsiders, but nevertheless may be importantfacets of tbe relationsbip.

After the list of relationship qualities hadbeen derived from individual interviews withchildren (Study 1), self-report rating scaleswere developed to assess each of these qual-ities (Study 2). The scales were administeredto a large sample of children, and factor anal-yses were used to identify the general dimen-sions underlying sibling relationships. Thisstrategy yielded a multitiered descriptiveframework in which botb general dimensionsand molecular qualities were incorporated.

450 Child Development

It was hypothesized that tbree general di-mensions would be found: (a) Relative Status/Power, {b) Warmtb/Closeness, and (c) Con-fiict. The first dimension of Relative Statusand Power refers to the degree and directionof asymmetry in the relationship. One endwould be characterized by greater power bythe child, whereas tbe otber end would becbaracterized by greater power by tbe sibling.Egalitarian relationsbips would fall in tbemiddle of tbis dimension. This dimension ofRelative Status/Power bas consistendyemerged in taxonomic studies of interper-sonal traits or types of relationships (Wiggens,1979; Wisb, Deutsch, & Kaplan, 1976). Al-though investigators have usually focused ondominance, we expected that positive qual-ities, sucb as admiration and nurturance(caretaking), would also be manifestations ofthis dimension.

In tbese taxonomic studies, relationshipshave also been found to vary along a dimen-sion of positivity-negativity (e.g., Wisb et al.'s[1976] Cooperation/Friendly vs. Competitive/Hostile; Wiggens's [1979] Warm/Agreeablevs. Cold/Quarrelsome). Although positive andnegative behaviors bave often been consid-ered to be bipolar opposites, tbey sbould ac-tually be treated as separate dimensions. Inseveral recent studies of young adults' rela-tionsbips, investigators have found separate,independent factors for relationship qualitiesrefiecting warmth or acceptance and thosereflecting conflict (Braiker & Kelley, 1979;Shaver, Furman, & Buhrmester, in press).Similarly, in observational studies of siblinginteractions, rates of positive and negative be-havior have been found to be essentially un-correlated (Bryant & Crockenberg, 1980; Min-nett et al., 1983). In light of this research, wehypothesized that Warmth/Closeness andConflict would emerge as separate factors.Warmth/Closeness was expected to be mani-fested in a range of molecular qualities, sucbas intimacy, prosocial behavior, companion-ship, admiration, nurturaoce, perceived simi-larity, and affection, whereas confiict wasexpected to be manifested in terms of quarrel-ing, antagonism, competition, and perceivedparental favoritism.

Altbougb principally concemed with therelationship qualities themselves, we werealso interested in how and to what degreefamily constellation variables may be relatedto tbese qualities. Previous investigators havefound that children ascribe greater levels ofpower to older siblings than younger ones(Bigner, 1974a; Bragg, Ostrowski, & Finley,1973; Sutton-Smith & Rosenberg, 1968). Ac-

cordingly, we expected that greater status/power would be assigned to the older child intbe dyad, particularly if the difference in agewas gieat. These relations were expected tobe relatively large in magnitude. On tbe basisof previous research (e.g., Bowerman & Dob-ash, 1974; Kocb, 1960), we also hypothesizedthat feelings of warmth or closeness would begreater in same-sex dyads than in opposite-sex ones, whereas perceived conflict was ex-pected to be greater in narrow-spaced rela-tionships than in wide-spaced ones. Tbeseeffects were expected to be relatively modestin size, bowever.

Study 1

PurposeThe first task was to develop a list of the

primary qualities of sibling relationsbips. Assome investigators of personality traits bavedone (e.g., Allport & Odbert, 1936), we beganwith the language of the layperson by askingcbildren to describe tbeir sibling relation-sbips.

MethodSubjects.—The subjects were 49 fifth-

and sixtb-grade children (20 boys and 29girls). They ranged in age from 11- to 13-year-olds. The children were enrolled in a paro-chial school in a large metropolitan area; mostwere from middle- to upper-middle-classCaucasian families.

Procedure.—Each child was individuallyadministered an open-ended interview thatlasted approximately 20 min. Previous inves-tigators have asked cbildren what sibling rela-tionships should generally be like (e.g., Big-ner, 1974b). In the present case, however,children were asked a series of five basicquestions about their relationsbip with aspecific sibling: (a) "Tell me about your rela-tionship with [name]"; {b) "What is it likehaving a brotber [sister]?" (c)"Tell me asmany good things as you can about your rela-tionship with [name]"; {d) "How about someof the not so good tbings?" and (e) "How im-portant is the relationship to you? Whatmakes it important?" In order to foster exten-sive descriptions, three standardized probeswere included after each question (e.g., "Tellme more"). If a subject had more than onesibling, be or she was interviewed about arandomly selected sibling. Sex, sex of sibling,and relative age (i.e., older vs. younger sib-ling) were distributed relatively evenly.

Coding of responses.—The coding sys-tem was not based on an a priori theory, butinstead it was derived from an examination of


the children's responses. First, 15 interviewswere transcribed. Answers were divided intounits of thought, and each unit was written ona separate card. Three research assistants in-dependently sorted the cards into categoriesbased on the perceived similarity of the state-ments. The three sets of proposed categorieswere very similar; differences were resolvedthrough discussion.

Next, a coding manual was developed,and two assistants coded a second group of 15interviews. Minor modifications in the codingrules were made prior to the final coding ofthe interviews. The list of categories is pre-sented in Table 1. References to either thepresence or absence of a relationship qualitywere scored as instances of a descriptive cate-gory; for example, both "We play a lot to-gether" and "We hardly ever play together"were coded as companionship. Responsesthat could not be classified in any of the cate-gories were coded as miscellaneous. Twonaive research assistants recoded all of theprotocols. Interrater agreement was cal-culated on each of the 16 categories using Co-hen's kappa. Coefficients ranged from .66 to1.00 (mean kappa = .93).

ResultsThe children's descriptions of their rela-

tionships were rich in content. The subjectsreferred to an average of 8.2 different relation-ship qualities in their descriptions. Table 1presents the percentage of descriptions inwhich each quality was reported. The mostcommonly mentioned positive qualities werecompanionship (93%), admiration of sibling(81%), prosocial behavior (77%), and affection(65%). The negative qualities of antagonism(91%) £ind quarreling (79%) were also fre-quently reported. Additionally, most (89%)provided some general evaluation of their re-lationship (e.g., "We have a good relation-ship").

Most of the other qualities in the listwere reported by a significant number of thechildren as well. However, neither parentalpartiality nor competition were discussed of-ten (20% and 10%, respectively). Perhapschildren are reluctant to discuss these qual-ities, although they were w illing to discussother negatively valenced aspects of their re-lationships.

Finally, it should be noted that tbese cat-egories incorporated almost all the children'scomments about their relationships. Less than

TABLE 1

RELATIONSHIP QUALITIES REPORTED INOPEN-ENDED INTERVIEWS

Qualities Percentage

Intimacy 55Prosocial behavior 77Companionship 93Similarity 46Nurturance by sibling 48Nurturance of sibling 34Admiration by sibling 8Admiration of sibling 81Affection 65Dominance by sibling 18Dominance over sibling 8Quarreling 79Antagonism 91Competition 10Parental partiality 20Ceneral relationship

evaluation 89

NOTE.^Numbers indicate percentage of children\vho referred to the quality.

5% of the comments w ere coded as miscel-laneous. Cenerally, the miscellaneous com-ments referred to interactions with a thirdperson (e.g., "be plays more with my youngerbrother than the rest of us do").

In summaiy, tbe list of relationsbip qual-ities appears to be a relatively comprebensiveone. Additionally, almost all the qusditiesseem to be salient ones as evidenced by thefact that they are commonly reported. Thosethat were mentioned infrequently warrantfurther examination as well because of theirtheoretical significance.

Study 2Overview

The next step was to develop and vali-date a structured self-report questionnaire toassess children's perceptions of the qualitiesof their sibling relationships. Scales were de-veloped to assess each of the qualities identi-fied in Study 1- General dimensions or factorsthat underlay these specific qualities were de-termined through principal components anal-yses. Finally, the pattem of relations betweenthe relationship qualities and family constel-lation variables was examined.

MethodSubjects.—Subjects were 198 fifth- and

sixth-grade children (95 boys and 103 girls).

^ Copies of the interview coding system and Sibling Relationship Questionnaire are availableupon request from the Hrst author.


Tbe sample was drawn from two parochialscbools and one public school. Forty-threepercent of the children were from two-childfamilies, 27% were from three-cbild families,and 28% from families with four or more chil-dren. Most children were from middle- to up-per-middle-class Caucasian families.

Development of questionnaire.—TheSibling Relationship Questionnaire contained17 scales. These scales measured 15 of tbe 16qualities identified in Study 1. The remainingquality, "general relationship evaluation,"served as the basis for two scales: (a) satisfac-tion with the relationship, and {b) importanceof the relationship.

Each scale consisted of three items. Afive-point Likert format (1 = Hardly at all to 5= Extremely much) was used for all scedesexcept the parental partiality scale. In thatcase, response choices ranged from "Almostalways him/her [favored]" to "Almost alwaysme [favored])," and scores were based on de-viations from tbe midpoint of "About tbesame." Thus, the scale was a measure of abso-lute partiality rather tban a measure of tbedirection of partiality for tbe subject or sib-ling.

The wording of items was based on tbedescriptive pbrases that were commonly usedduring interviews. To minimize the percep-tion that some response alternatives are moresocially desirable than otbers, we adoptedHarter's (1982) "structured altemative for-mat." In particular, every fifth item was pref-aced with a statement intended to make allresponse alternatives seem equally accept-able. For example, tbe first item reads: "Somesiblings really care about eacb other, whereasothers don't care that much. How much doyou care about eacb other?" The question-naire was initially administered to 42 cbil-dren, and several items were subsequendyrephrased.

Procedures.—In the primary sample,questionnaires were administered to groupsof cbildren at the schools. Each question wasread aloud by a trained researcb assistant dur-ing a session lasting 20—25 min. In a subse-quent session, children completed question-naires about tbeir relationsbips witb othermembers of their social network. Those re-sults are reported elsew^here (Furman &Bubrmester, in press).

If a cbild had more than one sibling, atarget sibling was selected for bim or ber todescribe. Selections were made so as toachieve and approximately equal distributionof subjects in tbe different combinations of

tbe following variables: sex of subject, sex ofsibling, relative age of the sibling (older/younger), the number of siblings in the family(two, three, or more than three), and the abso-lute age difference between the subject andsibling (less tban 4 years vs. 4 or more years).Tbe criterion of 4 years difference in age wasselected because it permitted the most equaldivision of subjects.

ResultsPsychometric analyses.—Scores were

computed for each of the 17 scales by averag-ing the three items designed to assess thequality. The internal consistency coefficients(Cronbach's alpha) for tbese composites allexceeded .70 except for the competition scale(.63) (M = .80). It should also be noted thattbe questionnaire bas been administeredtwice 10 days apart to another sample of 94cbildren, and test-retest reliabilities for tbetbree-item scales were found to be high,mean r = .71, ranging from .58 to .86. In thissupplementary sample, correlations betweenSRQ scale scores and Crandall, Crandall, andKatkovsky's (1965) Children's Social Desira-bility Questionnaire were also found to bevery low, mean r = .14 in the socially desir-able direction. Only two of the 17 correlationswere significant.

Principal components analysis.—In or-der to identify underlying dimensions, a prin-cipal components analysis was performed onscores on 15 scales of tbe SRQ. Satisfactionand importance scores were not included inthese analyses because they assessed generalevaluations of a relationship rather than spe-cific qualities. Four factors were extracted, ac-counting for 71% of the common variance.Table 2 presents tbe pattem weight matrixfrom an oblique Promax rotation. Tbe factorswere labeled "Warmtb/Closeness" (account-ing for 38% of tbe common variance after rota-tion), "Relative Power/Status" (24%), "Con-fiict" (27%), and "Rivalry" (10%). The fourwere minimally correlated with each other(r's = —.08 to —.16), except for Conflict andRivalry, which were moderately correlated (r= .35). Exact factor-score coefficients werecalculated from the rotated factor pattem ma-trix using Kaiser's (1962) procedure.

It should be noted that the eigenvalue oftbe fourth factor (.95) was below the conven-tional root one cutoff. Parental partialityscores were, however, almost exclusively ac-counted for by this fourth factor. Rather thanexclude the fourth component and, in effect,ignore the parental partiality scores, it was re-tained in the analyses and cautiously viewedas an underrepresented dimension.


TABLE 2

FACTOR PATTERN COEFFICIENTS OF SIBLING RELATIONSHIP QUESTIONNAIRE SCALES

FACTORS

Warmth/ RelativeQUALITIES Closeness Status/Power Conflict Rivalry

Intimacy 70Prosocial behavior 83 . . . . . . . . .Companionship 78Similarity 70Nurturance by sibling 28 - 77Nurturance of sibling 26 85Admiration by sibling 67 25 -29Admiration of sibling 69 - 28Affection 69 . . . - 36Dominance by sibling . . . . . . . . . — 65 55 . . .Dominance over sibling . . . . . . . 80 41Quarreling . . . . . . 88Antagonism . . . . . . 92Competition . . . . . . 63 36Parental partiality . . . . . . . . . 96

NOTE.—Scores are factor loadings on a principal components analysis with a general pro-max rotation. Factor loadings below .25 are not presented. Factors are minimally correlated( - .20 > r < .20), except Conflict and Rivalry (r = .35).

Family Constellation EffectsAnalyses of variance were conducted to

determine the influence of family constella-tion variables on the qualities of sibling rela-tionships. Five constellation variables wereexamined: (a) sex of subject, {b) sex of sibling,(c) relative age (i.e., older or younger sibling),{d) age difference, and (e) family size. Prelimi-nary regression analyses revealed that effectsdue to birth order could almost totally be ac-counted for by relative age of tbe child andfamily size. Consequently, birth order was notconsidered further because it was confoundedwitb these other factors that are germane tothe dyadic relationship.

To reduce the number of chance effects,five-way multivariate analyses of variance(MANOVAs) were first conducted on each ofthe four sets of the variables that corre-sponded to tbe factors presented in Table 2.Tbese MANOVAs included the scores for thescales that had high loadings (> .30) on thatfactor. Scales fhat had high loadings on morethan one factor were included in all of therelevant MANOVAs. If a multivariate effectwas significant {p < .05), univariate analysesof variance of the relevant factor and scaleswere conducted. Follow-up analyses weredone using Newman-Keuls tests. The resultsof the univariate analyses are reported inTable 3.^ A brief summary is presented in thefollowing section.

Warmth/Closeness.—The analysis ofvariance of Warmth/Closeness factor scoresrevealed a significant interaction of sex Xsibling-sex. As predicted, children felt greaterfeelings of closeness toward same-sex siblingstban opposite-sex ones. This effect wasqualified, however, by a significant interac-tion among sex, sibling-sex, and age differ-ence. Tbe difference in closeness betweensame-sex and opposite-sex siblings was onlysignificant for narrow-spaced dyads (narrowsame M = .36 vs. narrow opposite M = — .37,p < .05; wide same M = .07 vs. wide oppositeM = -.09, N.S.).

Similar sex x sibling-sex interactionswere found on the intimacy, companionship,prosocial bebavior, similarity, and nurturanceby sibling scales. For intimacy and compan-ionship, the difference between same- andopposite-sex siblings was only significant fornarrow-spaced dyads, but on the other scalesthe difference was significant regardless ofthe age spacing. Finally, greater companion-ship was reported to occur with younger sib-lings tban witb older siblings.

Relative Status/Power.—As expected,relative age had a strong effect on perceptionsof status and power. When subjects were theolder members of dyads, they reportedgreater nurturance of and dominance overtheir siblings than when they were younger

Results of the multivariate screening analyses are available upon request from the first author.

TABLE 3

SUMMARY OF SIGNIFICANT CONSTELLATION EFFECTS ON FACTOR AND SCALE SCORES

Variable and Effect F Description

Factor scores:Warmth/Closeness:

Sex X sib sex 9.79 Creater with same-sex sibs.Sex X sib sex x age

difF. 4.22 Creater with narrow-spaced, same-sex sibs than narrow-spaced, opposite-sex sibs.

Relative Status/Power:Rel. age 418.51 Creater influence over younger sibs.Rel. age X age diff. . . . 3.96 Createst influence over wide-spaced younger sibs; least

influence over narrow- or wide-spaced older sibs.Rel. age x fam. size .. 4.01 Less influence over older sib in 4+ child families than

in 2- or 3-child families. Regardless of family size,more influence over younger than older sibs.

Conflict:Age diff. 16.23 Creater when close in age.Age diff. X rel. age x

fam. size 3.86 Not interpretable.Rivalry:

Rel. age 8.34 Creater when sibs are younger.Rel. age x age diff. x

fam. size 3.38 Greatest when sibs are wide-spaced, younger in 4 +child family.

Scale scores:Intimacy:

Sex X sib sex 17.42 Creater with same-sex sibs.Sex X sib sex x age

diff. 4.25 Creater with narrow-spaced, same-sex sibs than nar-row-spaced, opposite-sex sibs.

Prosocial behavior:Sex X sib sex 4.84 Creater with same-sex sibs.

Companionship:Sex X sib sex 24.04 Creater with same-sex sibs.Sex X sib sex x age

diff. 6.01 Narrow-spaced, same-sex sibs greater than opposite-sex,narrow-spaced sibs.

Rel. age 15.25 Creater with younger sibs.Similarity:

Sex X sib sex 10.57 Creater with same-sex sibs.Nurturance by sibling:

Rel. age 179.67 Creater when sibs are older.Rel. age x age diff. . . . 4.61 Createst when sibs are wide- or narrow-spaced older.

Least when sibs are wide-spaced younger.Rel. age x fam. size . . 5.63 Createst by older sibs in 4 + child family; least by youn-

ger sibs in any family size.Sex X sib sex 7.36 No signiflcant post-hoc comparisons.

Nurturance of sibling:Rel. age 277.84 Creater when sibs are younger.Rel. age x age diff. . . . 17.22 Rank of scores is: wide-spaced younger sibs, narrow-

spaced younger sibs, narrow-spaced older sibs, wide-spaced older sibs. All four signiflcantly different fromone another.

Admiration by sibling:Rel. age 20.14 Creater when sibs are younger.Age diff. 5.12 Less when close in age.Rel. age x age diff. x

fam. size x sex 3.50 Not interpretable.Age diff. X sib sex x

sex 8.47 Not interpretable.Admiration of sibling:

Sex X sib sex x agediff. "4.43 Not interpretable.


TABLE 3 {Continued)

Variable and Effect Description

Affection No signiflcant effects.Dominance by sibling:

Rel. age 87.90 Creater when sibs are older.Sex X sib sex 4.85 No significant post-hoc comparisons.Sex X sib sex x rel.

age 4.76 Older same-sex sibs more dominant than older opposite-sex sibs. All older sibs more dominant than youngersibs.

Dominance over sibling:Rel. age 86.20 Creater when sibs are younger.Rel. age x fam. size .. 3.70 When target children are younger, less dominant in 4 +

child family than 2-child family. Regardless of familysize, less dominant when target children are younger.

Quarreling:Age diff. 9.31 Creater when close in age.Age diff. X rel. age . . . 4.38 Creater if close-spaced older sibs than wide-spaced

older sihs or younger sibs.Age diff. X rel. age X

fam. size 3.79 Not interpretable.Antagonism:

Age diff. 14.26 Creater when close in age.Competition:

Age diff. 7.04 Creater when close in age.Age diff. X rel. age x

fam. size 4.58 Not interpretable.Parental partiality for either:

Rel. age 6.42 Creater when sibs are younger.Rel. age x age diff. x

fam. size 3.75 Not inteipretable.Partiality for sibling (vs. subject):

Rel. age 8.80 Creater when sibs are younger.Satisfaction:

Rel. age X age diff. . . . 7.13 Creater with wide-spaced older sibs than nanow-spacedolder sibs.

Importance No significant effects.

NOTE. Rel. age = relative age of sibling and subject; age difE = age difference (less than 4 years vs. 4 or more);fam. size = family size {two, three or more than three children). Degi'ees of freedom for F's are 1 and 148, except foreffects involving family size, where they are 2 and 147. All F's and follow-up comparisons reported in the table aresignificant (p's < .05).

members of dyads. Inversely, nurturance anddominance by siblings were greater when thesubjects were the younger members thanwhen they were older members. Addition-ally, subjects were admired more by youngersiblings tban by older ones.

The size of the difference in age also in-fluenced perceptions of relative power andstatus as indicated by significant interactionsbetween relative age and age diifference forfactor scores and for the two nurturance scalescores. In particular, when they were 4 ormore years younger than their sibling, chil-dren reported that they had the least power orstatus and engaged in the least amount of nur-turant or caretaking behavior. When tbeywere 4 or more years older, they engaged inthe greatest amount of nurturant and caretak-ing behavior. Children also reported less ad-

miration of their sibling when the age spacingwas narrow than when it was wide.

Analyses also revealed several significantinteractions between relative age and familysize, indicating that the power differentiationwas also affected by family size. In particular,older siblings in large families of four or morechildren were perceived as more nurturantthan older ones in two- or three-childfamilies. Similarly, children in families withfour or more children perceived themselvesto be less dominant over older siblings tbanchildren in two- or three-child families did.

Conflict.—There was a significant effectof age difference on the Conflict factor scores,with children reporting more conflict withnarrow-spaced siblings than with wide-spaced siblings. Similar effects were ob-


served on the quarreling, antagonism, andcompetition scales. Additionally, children re-ported quarreling more with narrow-spacedolder siblings than wide-spaced older siblingsor younger siblings. Finally, a significant in-teraction occurred among sex, sibling-sex, andrelative age on dominance x sibling scores.Cbildren perceived older siblings of the samesex to be more dominant than older siblingsof the opposite sex.

Rivalry.—Children reported greaterrivalry and parental partiality when siblingswere younger than when tbey were older. Aninteraction among relative age, age differ-ence, and family size refiected the fact thatfeelings of rivalry with wide-spaced youngersiblings were particularly marked in familiesof four or more children.

The parental partiality scale assessed tbedegree of partiality. The data were also exam-ined to determine the effects on the directionof perceived partiality—that is, whether thechild or sibling was perceived as beingfavored. Children reported tbat their siblingswere favored more when the siblings wereyounger tban wben tbey were older.

Satisfaction and importance.—Finally,univariate analyses of variance were con-ducted on the satisfaction and importancescales. A significant interaction between rela-tive age and age difference was observed ontbe satisfaction scores. Cbildren were moresatisfied witb tbeir relationsbips witb wide-spaced older siblings than their relationshipswith narrow-spaced older siblings. No signifi-cant effects were found on tbe importancescores.

Size of EffectsTbe preceding findings indicate that fam-

ily constellation variables were related to sib-ling relationship qualities, but tbe questionremained as to the size of these effects. Multi-ple regression analyses were conducted inwhich each of the four relationship qualityfactors was predicted from equations com-prised of the five constellation variables andall interactions among the constellation vari-ables that were significant or approachedsignificance {p < .10) for the relevant factor inthe previous ANOVAs. For example, the re-gression equation for the affection factor con-tained the five constellation variables, and theinteractions among sex and sibling sex, sexand family size, sex and age spacing, siblingsex and age spacing, age spacing and relativeage, and sex, sibling sex, and age spacing.

Tbe Relative Status/Power scores couldbe predicted quite accurately from tbe equa-

tion of constellation variables {R = .84), prin-cipally because of tbe inclusion of relativeage, r = .81. In contrast, the equations of con-stellation variables did not account for morethan 20% of tbe variance on tbe three otherfactors, Warmth/Closeness R = .36, Confiict R= .43, and Rivalry R = .38. It should benoted that these equations contained aminimum of 11 variables, M = 14 variables;thus, if anything, these analyses provide over-estimates of the strength of relations becauseof capitalization of sample-specific variance.

DiseussionThese two investigations are first steps

toward the development of a comprehensiveframework for describing the qualities of sib-ling relationsbips. In the first study, cbildrenwere able to provide detailed descriptions ofbotb tbe positive and negative aspects of tbeirrelationships witb brothers and sisters. Therichness of tbeir comments underscores themultifaceted nature of their sibling relation-ships.

By beginning witb tbe insiders' perspec-tive on sibling relationsbips, we tried to in-sure that the qualities in the list are psycho-logically meaningful to the children and thatthe list incorporates the significant aspects ofsibling interactions as tbey occur in a widerange of naturalistic contexts. Altbough theywere responding to questions posed by us,the children provided the data for the sortingand categorization. Because the actual sortingwas conducted by adults, it is possible tbattheir conceptions may bave altered tbe list insome unknown manner. However, tbe threeindependent lists were very similar, and thecoders reported that the categories emergedquite readily.

Certainly, the qualities mentioned by thechildren are limited to those that they areaware of and are willing to discuss with anunfamiliar interviewer. However, we re-viewed previous studies of sibling relation-ships and were unable to identify any obviousomissions in the list.

In the second study, a self-report mea-sure of sibling relationsbips was developedfrom tbe list of relationship qualities. Onestrength of this measure is that children's rat-ings are likely to refiect the nature of theirinteractions in a wide range of social contexts.At the same time, the ratings are not objectiveones. They are affected by the children'smemories, their interpretations of events, andtheir willingness to report their actual percep-tions on a questionnaire. The high test-retest

reliabilities and tbe low correlations with so-cial desirability provide some encouragingevidence for the validity of the SRQ, but theSRQ still must be interpreted as a subjectiveself-report measure.

Despite these limitations, we believetbat the results of the second study providedvaluable information about sibling relation-ships. The features were found to form fourdistinct, interpretable factors: (a) Warmth/Closeness, {b) Relative Status/Power, (c) Con-fiict, and {d) Rivalry. The first three factorshave been found in most studies of adults'interpersonal bebavior and relationsbips(Wiggens, 1979; Wish et al., 1976). Appar-endy, children as young as 11—12 years olduse the same three dimensions as adults.

The fourth factor of Rivalry has notemerged in studies of other relationships andmay be particular to sibling relationships.One of tbe special characteristics of relation-ships between brotbers and sisters is that theyare based on shared biological and affectiveties with parents. The relative attention andtreatment by others outside the dyad may bemore salient in sibling relationships than inother kinds of relationships. Although theRivalry factor seems particularly interesting,its importance should not be overstated. It isonly one of four dimensions underlying chil-dren's perceptions.

Tbe pattem of relations among the fourfactors is interesting in several respects.Warmth/Closeness and Confiict were essen-tially uncorrelated with each other, support-ing the hypothesis that positive and negativequalities are not bipolar opposites (Shaver etal., in press). Perhaps many children have am-bivalent feelings about their siblings (Bubler,1939). Siblings may also vary in their style ofcoping with potential conflicts. For example,some siblings wbo are not close to eacb otbermay figbt regularly, whereas others may pre-fer to avoid each other. Finally, variation inthe intensity or frequency of contact in siblingrelationships may lead the Warmth/Closenessand Conflict factors to be more independenttban one might intuitively expect. Comparedto siblings who only interact with each otberinfrequently, tbose who frequently interactmay be likely to bave both more positive in-teractions and more negative interactions.

Warmth/Closeness and Conflict werealso unrelated to Relative Status/Power. Onemight have expected younger siblings to re-sent differential status, but it appears tbatsome children expect power differences be-cause of fhe difference in age. The only two


factors that were correlated with eacb otherwere Conflict and Rivalry. Perceptions of dif-ferential attention by parents could fosterfeelings of antagonism and conflict betweensiblings. Additionally, frequent conflict be-tween siblings could make tbe parents' taskof treating tbeir cbildren "differently butequally" particularly difficult.

In general, the present results are con-sistent with previous conceptualizations ofsibling relationships. For example, Dunn's(1983) distinguished between complementaryand reciprocal aspects of sibling relationsbips.Tbe relationship qualities tbat loaded on dieRelative Status/Power factor \vould seem tobe instances of complementary aspects of therelationship, whereas those that only loadedon the other factors would be instances of re-ciprocal features.

Behavior geneticists have argued that theenvironments of children from the same fam-ily may be quite different (Rowe & Plomin,1981; Scarr & Crajek, 1982). One such reasonis that siblings may create different environ-ments for each other. We believe tbat tbecomplementary qualities that loaded on theRelative Status/Power factor are promisingcandidates for such nonshared influences oneach other's development. In contrast, the de-gree of Warmth/Closeness or Conflict mayprove to be a source of shared influences, al-though mismatches on these factors could besources of nonsbared influences as well.

Family Constellation Variables andPerceived Relationship Qualities

Some qualities of sibling relationshipsseem to be influenced by family constellationvariables. For example, cbildren reported thatolder members of dyads have greater statusand power than younger members, particu-larly wben tbe difference in age is great. Thisfinding is consistent with past research (Big-ner, 1974a; Bragg et al., 1973; Sutton-Smith &Rosenberg, 1968), but it is an extension ofprevious results in that the asymmetry in therelationship is shown to exist in positive qual-ities, such as nurturance and admiration, aswell as in sheer power or dominance.

Feelings of warmth and closeness weregreater in same-sexed dyads than in opposite-sexed ones. Similar results have been foundin Bowerman and Dobasb's (1974) study ofadolescents' perceptions of sibling relation-ships and in observational studies of youngchildren's sibling interactions (Dunn & Ken-drick, 1981; Whiting & Pope-Edwards, 1977).Children with same-sex siblings wbo wereclose in age to them reported the strongest


feelings of warmth and closeness. Thus, simi-larity in age and gender appears to promoterelationships that resemble friendships insome respects.

At the same time, the highest level ofconflict is reported by children whose sib-lings are close in age, a finding consistentwith previous research on quarreling or nega-tive behavior among siblings (Koch, 1960;Minnett et al., 1983). This finding illustratesan important difference between friendsbipsand close-age, same-sex sibling relationships.Confiict may be avoided in friendships so tbatthe continuation of the relationship is notthreatened. On the other band, tbe institu-tional stmcture of tbe family guarantees tbesurvival of sibling relationships, thus allowingthe frequent expression of confiict.

Interestingly, children with opposite-sexsiblings close in age bad tbe lowest scoreson the Warmth/Closeness factor. Moreover,these children reported more conflict thanthose children with a wide-spaced, opposite-sex sibling. Although one might think that thesimilarity in age might foster affectivecloseness in opposite-sex sibling relation-ships, this is not the case. Rather, it appearsthat there is a taboo against intimacy andcloseness in tbese relationships just as thereis in opposite-sex peer relationships duringthis developmental period (Buhrmester,1983).

Although only the effects of constellationvariables on the factor scores have been high-

lighted here, corresponding effects were com-monly found on the scales that loaded on therelevant factors. At the same time, there werenoteworthy differences between the resultsfor the factor scores and tbose for the scalesmaking up the factors. In part, these differ-ences reflect the fact that some qualitiesloaded on more tban one factor. Even whentwo scales load on the same single factor,however, differences were present. The facttbat tbere are similarities and differences inthe effects found on the various qualities il-lustrates the value of a multitiered approachin which both specific relationship qualitiesand general dimensions are examined.

Although infiuenced by family constella-tion variables, the qualities of sibling relation-ships are by no means solely determined bytbem. A regression equation comprised of tberelevant constellation variables predictedscores on tbe Relative Status/Power factorwitb considerable accuracy, but tbe constella-tion variables accounted for less tban 20% oftbe variance on the other tbree factors.Clearly, studies of family constellation vari-ables cannot substitute for studies of tbe qual-ities of sibling relationships. Individual differ-ences within any particular type ofconstellation seem to be the rule ratber tbantbe exception.

Future DirectionsWe believe it is time to move away from

studies of family constellation variables andfocus our attention on the sibling relationsbip.Figure 1 presents a diagram of the variables

FAMILYCONSTELLATION

VARIABLES

Relative AgeAge DifferenceSex Sib-Sex

PatternFamily SizeBirth Order

IPARENT-CHILDRELATIONSHIPS

Qualities ofRelationships

Management ofSiblingRelationships

SIBLINGRELATIONSHIP

Warmth/ClosenessRelative Power/

StatusConflictRivalry

CHARACTERISTICSOF INDIVIDUAL

CHILDREN

CognitiveSocialPersonality

FIG. 1.—A diagram of the primary determinants of sibling relationship qualities


that appear to be most relevant to tbe study ofsibling relationships. The arrows among vari-able sets represent potential paths of causalinfluence that warrant investigation. At thecenter of the diagram are the qualities of therelationship—the focus of the present study.Although tbe Sibling Relationship Question-naire appears to be one promising method ofmeasuring these qualities, a comprehensivebattery of measures is still needed. In subse-quent research it would be important to deter-mine how parents perceive their children'srelationships with eacb otber. Similarly, ob-servational methods, such as those used instudies of young children's sibling relation-ships, can provide valuable information abouttbe qualities of such relationships (see Dunn,1983).

In the diagram, family constellation vari-ables, such as relative age, age spacing, andthe two children's sexes, are depicted as oneof the determinants of the qualities of siblingrelationsbip. The results of the present andprevious studies of sibling relationships sup-port tbis hypothesis (see Dunn, 1983; Koch,1960), but, as was demonstrated in this study,family constellation variables are not the soledeterminants of the qualities of sibling re-lationships.

The qualities of the relationships be-tween each child and each parent are alsolikely to infiuence the sibling relationsbip. Infact, in tbe present study perceptions of pa-rental partiality were associated with feelingsof competition and confiict. Similarly, Bryantand Crockenberg (1980) found that parentswho are responsive to their children's bebav-ior are likely to foster prosocial bebaviors be-tween their children.

One important, but neglected, factor istbe way parents manage tbe interactions be-tween their children. Parents may vary inwhat they expect their children's relation-ships with each other to be like and in tbediscipline or management techniques theyuse to promote such relationships. For ex-ample, Kendrick and Dunn (1983) found tbatmaternal intervention in tbeir children's quar-rels was associated witb frequent hostile be-havior among siblings 6 montbs later.

The parental management methods andthe qualities of parent-child relationships arelikely to be affected by family constellationvariables (Hilton, 1967; Lasko, 1954). Just asin the case of sibling relationsbips, however,constellation variables are likely to be onlyone of many determinants of parental manage-ment methods or parent-child relationships.

The social, personality, and cognitivecharacteristics of the two children are likelyalso to shape the nature of tbeir relationship.For example, temperament or sociability caninfluence the qualities of the relationship(Kendrick & Dunn, 1983). As yet, relativelyfew investigators have examined the impactof children's individual characteristics ontheir relationships with each other.

The qualities of sibling relationsbips canalso be expected to have an impact on theother variables depicted in the diagram. Forexample, we previously hypothesized tbatsibling conflict may strain parent-child rela-tionships. Similarly, one would expect siblingrelationsbips to affect the individual charac-teristics of the children (Furman & Buhrmes-ter, 1982).

In the present diagram, family constella-tion variables are not depicted as baving adirect effect on individual personality, social,or cognitive characteristics. Of course, the sexof children does affect tbese characteristics(Maccoby & Jacklin, 1974), but the infiuenceof the other constellation variables and tosome degree sex is expected to occur indi-rectly tbrougb their impact on sibling, parent-child, and perhaps marital relationships. Inprevious research, family constellation vari-ables have not been found to be very strong orconsistent predictors of personality or socialcharacteristics (see Wagner et al., 1979). If aconstellation variable is only moderately re-lated to the relationship qualities responsiblefor some effect on a cbild's personality, thenone can expect to find only small effects forconstellation variables. Even wben familyconstellation effects are observed, they can bedifficult to interpret without knowing themediating links.

In light of these considerations, weshould turn our attention to the study of thequalities of sibling relationsbips and theircauses and consequences. Many investigatorshave recognized that family relationshipsmust be conceptualized as a system, but thespecific processes interrelating tbe differentrelationsbips bave not been elucidated. It ishoped that the present study can contribute tothis effort by providing a framework for de-scribing and assessing sibling relationships.We believe tbat sibling relationships mayplay a larger role in development than hasbeen found in past family constellation re-search.

RefereneesAbramovitch, R., Pepler, D., & Corter, C. (1982).

Pattems of sibling interaction among pre-


school-age children. In M. E. Lamb & B. Sut-ton-Smith (Eds.), Sibling relationships: Theirnature and significance across the lifespan(pp. 61-86.) Hillsdale, NJ: Erlbaum.

Allport, G. W., & Odbert, H. S. (1936). Trait-names:A psycholexical study. Psychological Mono-graphs, 47(1, Whole No. 211).

Bigner, J. A. (1974). Second-borns' discriminationof sibling role concepts. Developmental Psy-chology, 10, 564-573. (a)

Bigner, J. A. (1974). A Wemerian developmentalanalysis of childrens' descriptions of siblings.Child Development, 45, 317-323. (b)

Bowerman, C. E., & Dobash, R. M. (1974). Struc-tural variations in inter-sibling affect. Journalof Marriage and the Family, 36, 48-54.

Bragg, B. W. E., Ostrowski, M. V., & Finley, G. E.(1973). The effects of birth order and age oftarget on use of persuasive techniques. ChildDevelopment, 44, 351—354.

Braiker, H. B., & Kelley, H. H. (1979). GonHict inthe development of close relationships. InR. L. Burgess & T. L. Huston (Eds.), Socialexchange in developing relationships (pp. 135—168). New York: Academic Press.

Bryant, B. K., & Grockenberg, S. B. (1980). Gorre-lates and dimensions of prosocial behavior: Astudy of female siblings with their mothers.Child Development, 51, 529-544.

Buhler, G. (1939). The child and his family. Lon-don: Harper & Bros.

Buhrmester, D. (1983). Toward a model ofsocioemotional development in preadoles-cence and adolescence. Unpublished doctoraldissertation, University of Denver.

Gicirelli, V. G. (1972). Goncept leaming of youngchildren as a function of sibling relationships tothe teacher. Child Development, 43, 282-287.

Gicirelli, V. G. (1973). Effects of sibling structureand interaction on children's categorizationstyle. Developmental Psychology, 9, 132-139.

Grandall, V. G., Grandall, V. J., & Katkovsky, W. A.(1965). Ghildren's Social Desirability Ques-tionnaire. Journal of Consulting Psychology,29, 27-36.

Dunn, J. (1983). Sibling relationships in early child-hood. Child Development, 54, 787-811.

Dunn, J., & Kendrick, S. (1981). Social behavior ofyoung siblings in the family context: Differ-ences between same-sex and difFerent-sex dy-ads. Child Development, 52, 1265-1273.

Dunn, J., & Kendrick, S. (1982). Siblings: Love,envy and understanding. Gambridge, MA:Harvard University Press.

Furman, W. (1984). Some observations on the studyof personal relationships. In J. G. Masters & K.Yarkin-Levin (Eds.), Interfaces between devel-opmental and social psychology (pp. 15—42).New York: Academic Press.

Funnan, W., & Bulmnester, D. (1982). The contri-

bution of peers and siblings to the parentingprocess. In M. Kostelnik (Ed.), Child nurtur-ance: Vol. 2. Pattems of supplementary par-enting (pp. 69-100). New York: Plenum.

Furman, W., & Burhmester, D. (in press). Ghil-dren's perceptions of their social relationships.Developmental Psychology.

Harter, S. (1982). The Perceived Gompetence Scalefor Ghildren. Child Development, 53, 87-97.

Hilton, I. (1967). Differences in the behavior ofmothers toward first and later bom children.Journal of Personality and Social Psychology,7, 282-290.

Irish, D. P. (1964). Sibling interaction: A neglectedaspect of family life research. Social Forces, 42,279-288.

Kaiser, H. F. (1962). Formula for component scores.Psychometrika, 1, 83-87.

Kendrick, G., & Dunn, J. (1983). Sibling quarrelsand matemal responses. Developmental Psy-chology, 19, 62-70.

Koch, H. L. (1960). The relation of certain formalattributes of siblings to attitudes held towardeach other and toward their parents. Mono-graphs of the Society for Research in ChildDevelopment, 25(4, Serial No. 78).

Lamb, M. E., & Sutton-Smith, B. (1982). Siblingrelationships: Their nature and significanceacross the lifespan. Hillsdale, NJ.: Erlbaum.

Lasko, J. K. (1954). Parent behavior towards flrstand second children. Genetic PsychologicalMonographics, 49, 96-137.

Maccoby, E. E., & Jacklin, G. N. (1974). The psy-chology of sex differences. Stanford, GA: Stan-ford University Press.

Minnett, A. M., Vandell, D. L., & Santrock, J. W.(1983). The effects of sibling status on siblinginteraction: Influence of sex of child, birth or-der, and age spacing. Child Development, 54,1064-1072.

Olson, P. H. (1977). Insiders' and outsiders' viewsof relationships: Research studies. In G. Levin-ger & H. L. Raush (Eds.), Close relationships:Perspectives on the meaning of intimacy.Amherst: University of Massachusetts Press.

Rowe, D. G., & Plomin, R. (1981). The importanceof nonshared (E ) environmental influences inbehavioral development. Developmental Psy-chology, 17, 517-531.

Scarr, S., & Grajek, S. (1982). Similarities and differ-ences among siblings. In M. E. Lamh & B. Sut-ton-Smith (Eds.), Sibling relationships: Theirnature and significance across the life span(pp. 357-381). Hillsdale NJ: Erlbaum.

Shaver, P., Furman, W., & Buhrmester, D. (inpress). Aspects of a life transition: Networkchanges, social skills, and loneliness. In S.Duck & D. Perlman (Eds.), The Sage series inpersonal relationships (Vol. 1). London: Sage.

Sutton-Smith, B., & Rosenberg, B. G. (1968). Sib-


ling consensus on power tactics./oMma/ ofGe- The effects of age, sex, and modernity on thenetic Psychology, 12, 63—72. behavior of mothers and children. Report to

Sutton-Smith, B., & Rosenberg, B. G. (1970). The the Ford Foundation.sibling. New York: Holt, Rinehart & Winston. Wiggens, J. S. (1979). A psychological taxonomy of

Wagner, M. E., Schubert, H. J. P., & Schubert, trait-descriptive terms: The interpersonal do-D. S. P. (1979). Sibship-constellation effects on main. Journal of Personality and Social Psy-psychosocial development, creativity, and chology, 37,395-412.health. In H. W. Reese & L. P. Lipsitt (Eds.), Wish, M., Deutsch, M., & Kaplan, S. J. (1976). Per-Advances in child development and behavior ceived dimensions of interpersonal relations.(Vol. 14, pp. 57-148). New York: Academic Press. Journal of Personality and Social Psychology,

Whiting, B. B., & Pope-Edwards, G. (Eds.). (1977). 33, 409-420.

Documents

Children's Perceptions of the Qualities of Sibling ... · Children's Perceptions of the Qualities of Sibling Relationships ... yielded a list of 15 salient qualities. In the second