See discussions, stats, and author profiles for this publication at:

https://www.researchgate.net/publication/227631322

Perceptions of Sibling
Relationships during Middle
Childhood and Adolescence

Article in Child Development · September 1990

Impact Factor: 4.92 · DOI: 10.1111/j.1467-8624.1990.tb02869.x

CITATIONS READS

176 34

2 authors, including:

Wyndol Furman
University of Denver
98 PUBLICATIONS 7,677
CITATIONS

SEE PROFILE

Available from: Wyndol Furman

Retrieved on: 18 May 2016
Perceptions of Sibling Relationships during
Middle Childhood and Adolescence

Eluane Buhrmester
University of Texas at Dallas

Wyndol Furman
University of Denver

BUHRMESTER, DUANE, and FURMAN, WYNDOL. Perceptions of Sibling Relationships during Middle
Childhood and Adolescence. GHILD DEVELOPMENT, 1990, 61,1387-1398. Ghildren in grades 3, 6, 9,
and 12 were administered the Sibling Relationship Questionnaire. Relationships were rated as
progressively more egalitarian across the 4 grade groups, with adolescents reporting less dominance
and nurturance by their older siblings than younger participants. Adolescents also reported less
companionship, intimacy, and affection with siblings than younger participants reported. Levels of
perceived conflict with younger siblings were moderately high across all 4 grades, whereas ratings
of conflict with older siblings were progressively lower across the 4 grades. The flndings suggested
that sibling relationships: (o) become more egalitarian and less asymmetrical with age, (fc) become
less intense with age, and (c) encompass experiences that are partially determined by the child's
standing in the family constellation.

A number of studies have examined over younger siblings than did 8-year-old par-
qualities of sibling relationships (i.e., warmth, ticipants; additionally, they found an increase
power, and conflict) and how they are related between ages 4 and 8 in the extent to which
to family constellation (Abramovitch, Gorter, younger siblings exercised power over their
& Lando, 1979; Bryant & Grockenberg, 1980; older siblings. The amount of instruction and
Dunn & Kendrick, 1982; Furman & Buhrmes- help that older siblings provided to younger
ter, 1985; Koch, 1960; Minnett, Vandell, & siblings also decreased as children got older.
Saintrock, 1983). Not as much is known, how-
ever, about how these features differ with Past studies also jdeld an unclear picture
age. Several investigators have examined how of age trends in warmth or closeness between
sibling relationships change during early siblings. Vandell et al. (1987) found an in-
childhood (see Dunn, 1983) and during adult- crease with age in the extent to which older
hood (see Gicirelli, 1982), but information siblings worked and played cooperatively
about age trends during middle childhood with younger siblings and a parallel increase
and adolescence is scarce (Bryant, 1982), with in the positive emotional tone of younger sib-
the few existing studies yielding discrepant lings' behavior toward older siblings. The au-
findings. thors interpreted these results as being con-
sistent with retrospective accounts and case
Bigner (1974), for example, found that studies suggesting that sibling relationships
children ascribed more power to older sib- become closer and more supportive during
lings than to younger siblings, and the adolescence and young adulthood (Gicirelli,
amount of power attributed to older siblings 1982; Ross & Milgram, 1982). On the other
increased with age. An opposite age trend, hand, Raffaelli and Larson (1987), in a study
however, was revealed in laboratory obser- of fifth and ninth graders, found no systematic
vaidons by Vandell, Minnett, and Santrock developmental trends in either the amount of
(1987). They found that 11-year-old partici- time spent with siblings or in the emotional
pants, especially girls, exercised less power closeness of their interactions. Similarly,

This research was supported by grant no. lROlHD 16142 from the National Institute of Ghild
Healdi and Human Development (Wyndol Furman, principal investigator). Portions of this research
were presented at the biennial meeting of the Society for Research in Ghild Development, Balti-
more, April 1987. Appreciation is expressed to Wendy Ritz for her assistance in the data collection.
We are also indebted to the feculty and students of the Cherry Greek School System. Reprints of this
paper can be obtained from Duane Buhrmester, School of Human Development, University of
Texas at Dallas, Richardson, TX 79083.
[Child Development, 1990, 61, 1387-1398. © 1990 by the Society for Research in Child Development, Inc.
All rights reserved. 0009-392O/9O/6105-O026$01.00]
1388 Child Development
Buhrmester and Furman (1987) found no dif- boys). Average ages were 8-4, 11-4, 14-4, and
ferences among second-, fifth-, and eighth- 17-5, respectively.
grade participants' ratings of companionship
with siblings. The participants attended suburban pub-
lic schools in Denver, Golorado, which
Finally, it is unclear whether sibling rela- served predominantly Gaucasian children of
tionships become more or less conflictual as middle- and upper-middle-class families. Not
children grow older. Although Vandell et al. surprisingly, the number of siblings varied
(1987) expected a decrease in conflict during somewhat across the four grades, F(3,387) =
middle childhood, they found that 11-year- 6.90, p < .001 (M = 1.54, 1.76, 1.66, 2.23,
olds engaged in more conflict with younger respectively). Mean ages of participants' older
siblings than 8-year-old children did. Raffaelli and younger siblings are presented in Ta-
and Larson (1987), on the other hand, ex- ble 2.
pected to find an increase in conflict during
adolescence, but found no definitive develop- Measures
mental trends. The Sibling Relationships Questionnaire
(SRQ) consists of 15 scales (see Table 1), each
The primary purpose of the present study containing three items. Each item asks how
was to clarify developmental trends in chil- characteristic of the relationship a particular
dren's perceptions of the qualities of sibling feature is. A five-point Likert-type format (1
relationships. We gathered ratings of chil- = hardly at all, 2 = not too much, 3 = some-
dren's perceptions of sibling relationships us- what, 4 = very much, 5 = extremely much) is
ing the Sibling Relationship Questionnaire used for all scales except the parental partial-
(SRQ) (Furman & Buhrmester, 1985). The ity scale. In that case, response choices range
SRQ assesses 15 qualitative features of re- from "1 = almost always him/her [favored]"
lationships that together capture the major to "5 = almost always me [favored]," with a
dimensions of sibling relationships (i.e., midpoint of "3 = about the same."
warmth/closeness, relative power/status, con-
flict, and rivalry). Intemal consistency coefficients (Gron-
A secondary purpose of the study was to bach alpha) computed separately for each of
attempt to replicate and extend previous find- the four grade levels were .71, .79, .77, and
ings concerning associations between constel- .81, respectively. Of the 60 alpha coefficients
lation status (i.e., age spacing, relative age, that were computed, all were greater than .60
sex, and sex of siblings) and relationship qual- except those for the following five scales:
ities. Specifically, several investigators have third graders' competition (.57) and admira-
found that the balance of power in relation- tion towEird sibling (.57); ninth graders' nur-
ships is strongly related to the relative age of turance by sibling (.54); and twelfth graders'
the sibling, with subjects whose siblings are nurturance toward sibling (.53) and nurtur-
older than themselves reporting less power ance by sibling (.55). In other research, chil-
than subjects whose siblings are younger dren's perceptions of these qualities have
(Bigner, 1974; Bragg, Ostrowski, & Finley, been found to be moderately to strongly cor-
1973; Furman & Buhrmester, 1985; Minnett related with reports by other family members
et al., 1983). The warmth or closeness of rela- (Furman, Jones, Buhrmester, & Adler, 1989).
tionships has also been found to vary with
gender composition, with same-sex siblings Procedures
reporting greater warmth/closeness than op- Questionnaires were administered to
posite-sex dyads (Bowerman & Dobash, 1974; groups of children at schools in a testing ses-
Dunn & Kendrick, 1982; Furman & Buhrmes- sion lasting 20—30 min. Siblings' names were
ter, 1985). Finally, the extent of confiict in the written on each child's SRQ so he or she
relationship has been found to be greater could easily keep track of who was being
when the age spacing between siblings is nar- rated. Questions were read aloud to third and
row rather than wide (Furman & Buhrmester, sixth graders. For the purposes of this report,
scores for only one sibling relationship were
1985; Koch, 1960; Minnett et al., 1983). selected for analysis. Selections were made so
as to achieve an approximately equal distribu-
Method tion of subjects in the different combinations
Subjects of the following variables: sex of subject, sex
Participants were 106 third graders (68 of sibling, relative age of sibling (older/
girls and 38 boys), 112 sixth graders (52 girls younger), and the age spacing between the
and 60 boys), 85 ninth graders (39 girls and 46 subject and sibling (less than 4 years vs. 4 or
boys), and 60 twelfth graders (33 girls and 27 more years). The criterion of 4 years differ-
 Buhrmester and Furman 1389
ence in age was selected because it permitted jects' reports of being nurtured by younger
the most equal division of subjects. siblings, as well as reports of nurturing older
siblings, were infrequent and varied litde
Results across the four grade levels.
In order to limit the number of chance At first glance, the different develop-
findings, three multivariate analyses of vari- mental trends for subjects' ratings of nurtur-
ance (MANOVA) were conducted on sets of ance by older siblings and nurturance of
conceptually related scales assessing the gen- younger siblings would seem to indicate a
eral dimensions of warmth/closeness, relative discrepancy between earlier- and later-bom
power, and conflict as identified by Furman subjects' perceptions of the age at which nur-
and Buhrmester (1985).^ The scales included turemce decreased. For earlier-borns' ratings
in each MANOVA are grouped together as of relationships with younger siblings, the de-
shown in Table 1. The independent variables cline occurred between grade 9 and grade 12,
in the analyses were grade, sex, sibling sex, whereas later-boms' ratings of relationships
relative age (i.e., the sibling is older or youn- with older siblings declined between grade 3
ger than the participant), and sibling spacing and grade 6. It should be noted, however, that
(4 or less years vs. more than 4 years differ- later-boms' ratings of nurturance by older sib-
ence). Univariate analyses were conducted lings described relationships with siblings
when the corresponding multivariate efifect who were on the average 4 years older than
was significant (see Table 1). Follow-up anal- themselves, whereas earlier-boms' ratings of
yses used Newman-Keuls tests with .05 al- nurturance of younger siblings described sib-
pha level. The significant MANOVA effects lings who were on the average 4 years youn-
are reported in the text, whereas Table 1 sum- ger than themselves. When the age of the sib-
marizes the significant univariate effects. ling being rated is considered, we found that
Means and standard deviations broken down there was general agreement in reports of
by grade level and by subjects' ratings of nurturance (see Fig. 1). Perceived nurturance
older and younger siblings are presented in declined most rapidly when younger mem-
Taible 2. bers of dyads were on average between 10
and 15 years old and older members of dyads
Relative Power/Status were on average between 14 and 19 years old.
Grade differences.—The MANOVA of
the four scales reflecting status/power re- Similar age trends were evident for domi-
vealed a main effect of grade, F(12,759) = nance ratings. At each successive grade level
3.39, p < .001, which was qualified by a grade later-boms reported being dominated less by
X relative age interaction, F( 12,759) = 6.02, their older siblings. Later-boms' ratings of
p < .001. Follow-up ANOVAs yielded main dominance over older siblings and earlier-
eflects of grade for all scales except domi- borns' ratings of dominance by younger sib-
nance over sibling; grade x relative age in- lings were low and varied modestly across the
teractions were also found for all four scales. four grades, with only third graders' ratings of
dominance over older siblings being signifi-
In general, the findings Were consistent cantly higher than all other grade levels.
with the results of Vandell et al. (1987), but Taken together, these findings indicate that as
contrary to those of Bigner (1974). That is, children grow older, they perceive that older
grade differences in nurturance by sibling brothers and sisters exercise less power/status
were found for subjects who had older sib- over younger ones, whereas the power/status
lings. Subjects in grade 3 reported being most of the younger sibling is perceived as remain-
nurtured by older siblings, whereas subjects ing relatively low.
in grades 9 and 12 reported being least nur-
tured; scores for grade 6 fell in between those Constellation effects.—As expected, the
for grades 3 and 9. Interestingly, subjects' rat- MANOVA revealed a large main effect of rel-
ings of nurturance of younger siblings were ative age for the scales assessing aspects of
moderately high across grade 3 to grade 9, power/status, F(4,251) = 132.50, p < .001.
wiith only twelfth graders reporting signifi- Consistent with past findings (Bigner, 1974;
cantly lower levels of nurturance directed to- Furman & Buhrmester, 1985; Sutton-Smith &
ward younger siblings. Not surprisingly, sub- Rosenberg, 1970), subjects perceived older

^ Although Furman and BuhrmestEr (1985) found that the Parental Partiahty Scale loaded on a
fotirth factor labeled "Rivalry," they noted that this was an underidentifled factor that was moder-
ately correlated with the conflict dimension. Therefore, Parental Partiality scores were grouped with
the scales reflecting conflict in the current analyses.
 TABLE 1
SUMMARY OF SIGNIFICANT ANOVA EFFECTS

Variable and Effect F Description

Relative status/power scales:
Nurturance by sib:
Grade 10.20 Grade 3 (2.60) greater than all other grades
(2.12).
Relative age 323.40 Subjects nurtured more by older (2.98) than by
younger sibs (1.55).
Grade x relative age 23.04 Decreases with grade when sibs are older:
grade 3 greater than grade 6, which is
greater than grades 9 and 12. No grade
differences when sibs are younger (see Ta-
ble 2).
Relative age x age spacing 17.59 Nurturance by wide-spaced older sibs (3.08)
greater than by narrow-spaced older sibs
(2.68), which was greater than nurturance by
younger sibs (1.55).
Nurturance of sib:
Grade 12 (2.26) less than all other grades
Grade 3.43 (2.76).
Subjects more nurturant of younger (3.31) than
Relative age 223.43 older sibs (1.98).
Grade 12 less than all other grades when sib
Grade x relative age 4.04 is younger; no grade difference when sib is
older (see Table 2).
Relative age x age spacing 19.73 Nurturance toward wide-spaced younger sibs
(3.58) greater than by narrow-spaced youn-
ger sibs (2.97), which was greater than nur-
turance toward older sibs (1.98).
Dominance by sib:
Grade 3 (2.60) greater than all other grades
Grade 5.40 (2.39).
Subjects dominated more by older (2.94) than
Relative age 122.77 by younger sibs (1.82).
Subjects dominated more by narrow-spaced
Age spacing 4.04 (2.53) than wide-spaced sibs (2.27).
When sib is older, significant decreases be-
Grade x relative age 2.75 tween each successive grade. When sib is
younger, only grade 3 greater than grade
6, with no other grade differences (see Ta-
ble 2).
Dominance over sib:
Subjects more dominant over younger (3.04)
Relative age 148.33 than older sibs (1.80).
Subjects more dominant over narrow-spaced
Age spacing 6.74 (2.57) than wide-spaced sibs (2.33).
Grade 3 is greater than all other grades when
Grade x relative age 4.64 sibs are older. No grade difference when
sibs are younger (see Table 2).
Warmth/closeness scales:
Intimacy:
Grade 3.95 Grades 3 (3.05) greater than all other grades
(2.58).
Relative age 19.96 Subjects more intimate with older (3.00) than
younger sibs (2.43).
Age spacing 4.00 Subjects more intimate with narrow..spaced
(2.88) than wide-spaced sibs (2.58).
Sib sex X relative age 10.26 Subjects more intimate with older sisters
(3.23) than all other sibs (2.59).
Sex X sib sex 6.16 Girls more intimate with sisters (3.03) than
brothers (2.60); no signiflcant difference be-
tween sisters (2.53) and brothers (2.71) for
boys.
1390
 TABLE 1 (CONTINUED)

Variable and Effect F Description

Affection:
Grade 3 (4.12) greater than all other grades
Grade 2.78 (3.77).
Subjects more affectionate with wide-spaced
Age spacing 10.70 (4.02) than narrow-spaced (3.73) sibs.
Subjects more affectionate with older sisters
Sib sex X relative age 2.95 (4.01) than all other sibs (3.80).
Prosocial behavior: Subjects more prosocial with wide-spaced
(3.42) than narrow-spaced (3.25) sibs.
Age spacing 3.77 Subjects more prosocial with older sisters
(3.54) than all other sibs (3.28).
Sib sex X relative age 3.96
Grade 3 (3.63) greater than grade 6 (3.31),
Gompanionship: which is greater than grade 9 (3.01), which
Grade 16.88 is greater than grade 12 (2.51).
Sex X sib sex 14.05 For girls, more companionship with sisters
(3.48) than brothers (3.12); for boys, no
signiflcant difference between sisters (2.90)
and brothers (3.26).
Similarity:
Sex X sib sex 10.82 Girls feel more similar to sisters (3.22) than
brothers (2.83); for boys, no significant dif-
ference between sisters (2.64) and brothers
(3.06).
Admiration by sib:
Sex X sib sex .. 3.74 Girls feel more admired by sisters (3.64) than
by brothers (3.37); for boys, no signiflcant
difference between sisters (3.42) and
brothers (3.55).
Admiration of sib:
Grade 2.68 Grade 3 (3.77) greater than grade 6 (3.39) and
grade 9 (3.45); grade 12 (3.68) not different
from grades 3, 6, and 9.
Relative age 21.64 Subjects admire older sibs (3.79) more than
younger sibs (3.33).
Age spacing 3.58 Subjects admire wide-spaced sibs (3.62) more
than narrow-spaced sibs (3.49).
Gonflict/rivalry scales:
Quarreling:
Relative age 13.86 Subjects quarreling more with younger (3.14)
than older sibs (2.69).
Age spacing 15.40 Subjects quarreling more with narrow-spaced
(3.15) than wide-spaced sibs (2.71).
Grade x relative age 2.64 When sibs are older, grade 3 is greater than
grade 12. No grade differences when sibs
are younger (see Table 2).
Antagonism:
Grade 2.65 Grade 3 (2.88) greater than grade 12 (2.44);
grade 6 (2.63) and grade 9 (2.60) not differ-
ent from grades 3 or 12.
Relative age 10.35 Subjects report more antagonism with younger
(2.84) than older sibs (2.48).
Age spacing 12.41 Subjects report more antagonism with narrow-
spaced (2.86) than wide-spaced sibs (2.50).
Grade x relative age 3.91 Decreased with grade when sibs are older:
grade 3 greater than grades 9 and 12, and
grade 6 greater than grade 12. No grade
difference when sibs are younger (see Ta-
ble 2).

1391
1392 Child Development
TABLE 1 (CONTINUED)

Variable and Effect Description

Gompetition:
Grade 10.04 Grade 3 (3.09) greater than grade 9 (2.48) and
grade 12 (2.44); grade 6 (2.63) greater than
grade 12 but not grades 3 or 9.
Age spacing 6.29 More competition with narrow-spaced (2.80)
than wide-spaced (2.55) sibs.
Grade x relative age 2.70 Effect of grade holds only when sibs are older
(see Table 2).
Parental partiality for sib:
Relative age 14.38 Subjects report greater partiality for sib when
sib is younger (2.69) than when sib is older
(3.00).
NOTE.—Relative age = relative age of sibling and subject; age difference = absolute difference in age between
siblings (less than 4 years vs. 4 or more). Values in parentheses are means of relevant cells. Degrees of freedom for F's
are 1 and 270, except for effects involving grade, where they are 3 and 270. All F's and follow-up comparisons reported
in the table are significant at the .05 level.

siblings as more domineering and nnrturing warm toward siblings than did younger sub-
than younger siblings. Similarly, subjects re- jects.
ported directing more dominance and nurtur-
ance toward younger siblings than toward Constellation effects.—Counter to ex-
older ones. pectations, the MANOVA did not yield a
significant sex X sibling sex interaction,
The MANOVA also revealed main ef- F(9,277) = 1.77, p > .13. In order to explore
fects for sibling spacing, F(4,251) = 2.36, p < whether the findings reported by other inves-
.001, with the ANOVAs showing that greater tigators would nonetheless replicate for indi-
dominance over siblings, as well as domi- vidual scales, univariate findings were exam-
nance by siblings, were reported in closely ined. Consistent with previous findings, the
spaced rather than widely spaced sibling ANOVAs revealed significant sex X sibling
pairs. The MANOVA also revealed a relative sex interactions for the intimacy, companion-
age X age spacing interaction, F(4,251) = ship, similarity, and admiration by sibling
7.73, p < .001, with greater nurturance of scales. In each case, girls rated sisters signifi-
younger siblings reported when pairs were cantly higher than brothers, whereas there
widely rather than narrowly spaced. was a nonsignificant trend in the direction of
boys rating brothers higher than sisters.
Warmth/C loseness These results partially replicate previous
Grade differences.—The MANOVA for findings (Furman & Buhrmester, 1985), show-
the set of warmth/closeness scales revealed a ing that there was a tendency for girls, at least,
main effect of grade, F(21,732) = 3.16, p < to feel more positive toward same-sex sibling.
.001. Univariate ANOVAs found grade effects
for four of seven of the scales. Companionship The MANOVA also revealed a main ef-
ratings showed the most noteworthy grade fect for the spacing between siblings among
differences: third graders reported the highest the scales reflecting closeness/warmth,
levels of companionship; scores for each sub- F(7,242) = 3.05, p < .01. Follow-up ANOVAs
sequent grade level were significantly lower showed that children reported greater affec-
than the previous one. Twelfth graders re- tion, prosocial behavior, and admiration of
ported less affection, intimacy, nurturance by siblings who were more than 4 years different
siblings, and admiration toward siblings than in age than more closely spaced siblings. The
younger students, with the differences be- interesting exception to this general trend
tween grade 3 and grade 6 accounting for was that subjects reported greater intimacy
most of the grade effect. Thus, contrary to the ^vith siblings who were less than 4 years dif-
views of several authors (Cicirelli, 1982; Ross ferent in age.
& Milgram, 1982), older subjects reported
feeling more distant from siblings than youn- Finally, the MANOVA revealed a main
ger ones. Adolescents reported spending con- effect of relative age, F(7,242) = 6.50, p <
siderably less time engaging in enjoyable ac- .001, which was qualified by a relative age x
tivities with siblings and felt slightly less sibling sex interaction, F(7,242) = 2.37, p <
 TABLE 2
AGES OF SIBLINGS AND MEANS AND STANDARD DEVIATIONS OF SRQ RATINGS

Grade
3 6 9 12
Average ages of siblings
(in years):
Younger sibling 4.51 (3.93) 7.30 (3.20) 10.21 (4.71) 14.20 (3.97)
Older sibling 13.46(2.38) 17.16(3.60) 19.11(3.23) 23.33(5.57)
Relative status/power
scales:
Nurturance by sibling:
Younger 1.4r(.5O) 1.48^ (.52) 1.69" (.59) 1.75" (.55)
Older 3.89" (1.09) 2.87'' (.91) 2.45" (.68) 2.45° (.66)
Nurturance of sibling:
Younger 3.36" (.99) 3.56" (.75) 3.31" (.84) 2.68''(.73)
Older 2.18" (1.02) 1.77" (.57) 2.06" (.65) 1.91" (.60)
Dominance by sibling:
Younger 2.03" (1.00) 1.60''(.68) 1.86"''(.76) 1.83"''(.63)
Older 3.40" (1.23) 3.00''(1.13) 2.77° (1.15) 2.34'' (.99)
Dominance over sibling:
Younger 2.92" (1.31) 3.17" (.99) 2.98" (.80) 3.12" (.81)
Older 2.24" (.90) 1.71'' (.90) 1.70'' (.67) 1.44'' (.47)
Warmth/closeness scales:
Intimacy:
Younger 2.87" (1.20) 2.33''(1.08) 2.11''(1.02) 2.31''(1.08)
Older 3.25" (1.22) 2.92'' (1.32) 2.86'' (1.52) 2.96'' (1.18)
Affection:
Younger 4.23" (.78) 3.90''(.78) 3.67''(.80) 3.44''(.80)
Older 3.99" (.97) 3.75''(1.02) 3.79''(1.14) 4.12''(.80)
Prosocial behavior:
Younger 3.49 (.90) 3.28 (.81) 3.18 (.72) 2.96 (.88)
Older 3.52 (1.19) 3.44 (1.01) 3.30 (1.09) 3.38 (.90)
Gompanionship:
Younger 3.60" (.96) 3.38''(.83) 2.87''(.85) 2.45''(.95)
Older 3.67" (.99) 3.24''(1.05) 3.17''(1.19) 2.55''(1.27)
Similarity:
Younger 3.04 (.91) 2.87 (.89) 2.78 (.95) 3.08(1.07)
Older 2.74 (1.02) 3.03 (.95) 2.87 (1.05) 3.24 (1.06)
Admiration by sibling:
Younger 3.56(1.09) 3.58 (.66) 3.60 (.80) 3.59 (.82)
Older 3.51(1.25) 3.32(1.06) 3.32(1.19) 3.49 (.88)
Admiration of sibling:
Younger 3.61" (1.00) 3.17" (.90) 3.21" (.80) 3.32"''(.86)
Older 3.93" (.93) 3.60''(.94) 3.71''(1.11) 3.98"''(.81)
Gonflict/rivalry scales:
Quarreling:
Younger 3.07" (1.18) 3.04" (1.12) 3.27" (.89) 3.24" (.99)
Older 2.99" (1.20) 2.73"'" (1.20) 2.74"'' (1.10) 2.10'' (.98)
Antagonism:
Younger 2.80" (1.19) 2.73" (1.04) 2.89" (.78) 3.03" (.96)
Older 2.96" (1.26) 2.51"'' (1.10) 2.26''° (1.10) 1.97° (1.02)
Competition:
Younger 2.87" (1.17) 2.75" (.99) 2.52" (.97) 2.40" (.88)
Older 3.31" (1.09) 2.51"''(1.10) 2.44''(.97) 2.10° (1.06)
Parental partiality for
sibling:
Younger 2.79 (.75) 2.72 (.66) 2.56 (.74) 2.64 (.85)
Older 3.07 (.81) 2.99 (.54) 2.91 (.65) 3.04 (.77)
NOTE.—Superscripts indicate results of post hoc comparisons made across the four grades; means with different
superscripts are significantly different.
1394 Child Development
4.5-1 LATER-BORN S's RATINGS OF
NURTURANCE BY OLDER SIBS
EARLIER BORN S's RATINGS OF
NURTURANCE OF YOUNGER SIBS
4.0- LATER-BORN S's RATINGS OF
NURTURANCE OF OLDER SIBS

UJ
EARLIER-BORN S's RATINGS OF
O NURTURANCE BY YOUNGER SIBS
Z 3.5-
CC
3

3.0-

O NURTURANCE GIVEN TO
CD YOUNGER SIBS
P 2.5

2.0-

.-lorr-—• NURTURANCE GIVEN TO

1.5-
OLDER SIBS

1.0
5 6 7 8 9 10 11 12 13 14 15 16 17 18

MEAN AGE OF YOUNGER SIBLING IN DYAD

FIG. 1.—Ratings of nurturance plotted according to the average ages of younger siblings in dyads

.05. The ANOVAs showed that greater admi- more than 4 years apart in age were seen as
ration of siblings was reported when siblings less conflictual than those with siblings less
were older rather than younger. Subjects also than 4 years apart in age. Subjects reported
reported greater intimacy, afifection, and pro- less quarreling, antagonism, and competition
social behavior with older sisters than with with wide-spaced siblings and more domi-
older brothers or younger siblings of either nance by and over narrow-spaced siblings.
sex.
Conflict/Rivalry The MANOVA also revealed a main ef-
Grade differences.—The MANOVA for fect of relative age, F(3,265) = 4.16, p < .01.
the scales reflecting conflict and rivalry re- Follow-up ANOVAs showed that greater pa-
vealed a main effect of grade, F(9,801) = 4.36, rental partiality, quarreling, and antagonism
p < .001, which was qualifled by a signiflcant were reported with younger than with older
grade x relative age interaction, F(9,801) = siblings. It was possible, however, that the
2.71, p < .05. Contrary to Vandell et al.'s effects of relative age for the latter two scales
(1987) flndings of increased conflict during were merely artifacts of the grade-related de-
middle childhood, the ANOVAs revealed cline in quarreling and antagonism described
that twelfth graders reported considerably previously. That is, we would expect subjects
less quarreling, antagonism, and competi- to report less conflict with older siblings than
tion with older siblings than third graders, younger siblings because their older siblings
and that these developmental trends were are further along the developmental con-
gradual. Reports of quarreling, antagonism, tinuum and therefore less likely to quarrel
and competition w^ith younger siblings were with them. Figure 2 reveals, however, that
moderately high and did not vary across the when the ages of the siblings are considered
four grades. in the ratings of antagonism with younger and
older siblings, a discrepancy still remains.
Constellation effects.—The MANOVA This discrepancy was most apparent when
revealed a main effect of the age spacing be- the older member of the dyad averaged 14
tween siblings, F(3,265) = 5.71, p < .001. years old and the younger member of the
Consistent with flndings from past studies dyad averaged 18 years old. At this point,
(Furman & Buhrmester, 1985; Koch, 1960; later-born subjects' ratings of antagonism with
Minnett et al., 1983), the ANOVAs showed older siblings are approximately three-quar-
that relationships with siblings who were ters of a scale point lower than earlier-bom
 Buhrmester and Furman 1395

^ 3.5-
ANTAGONISM WITH
YOUNGER SIBLING
I 3.o^
u.
O
CD
P 2-5-
CC
z
a 2.0-
ANTAGONISM WITH
OLDER SIBLING

6 7 8 9 10 11 12 13 14 15 16 17 18

MEAN AGE OF YOUNGER SIBLING IN DYAD

FIG. 2.—Ratings of antagonism plotted according to the average ages of younger siblings in dyads

subjects' ratings of antagonism with younger terms of physical, social, and cognitive capa-
siblings. These flndings indicate that relation- bilities), but, by the time these siblings are 21
ships with younger siblings are generally per- and 18 years old, respectively, they will be
ceived as less harmonious and warm than virtually equal in competence. In sum, as sib-
tliose with older siblings. lings grow more competent and their devel-
opmental statuses become similar, their rela-
Diseussion tionships become more symmetrical and
egalitcirian.
When the current flndings are considered
in conjunction with results from other studies
of sibling relationships, three tentative gener- This developmental trend parallels, to
alizations about the developmental cotirse of some degree, the change in authority struc-
sibling relationships emerge. First, significant ture that takes place in peirent-child relation-
transformations occur in the power/status ships during adolescence (Youniss, 1980;
structure of sibling relationships wherein re- Youniss & Smollar, 1985). If anything, the
lationships become more egalitarian and less changes in the distribution of power in sib-
asymmetrical with age. Across the four grades ling relationships may precede those in par-
studied, there were noteworthy decreases in ent-child relationships. Such would be ex-
thie extent of nurturance and dominance per- pected if these changes reflect a convergence
ceived to be directed toward younger sib- in the competence of the members of dyads.
lings. These trends can probably be traced to That is, even during adolescence, the differ-
age-related changes in the developmental ence in the competence and status of parent
status of the individual children in the dyads. and child would be expected to be greater
Tbe current flndings indicate that the bulk of than that between two siblings.
this transformation is complete by the time
younger siblings are roughly 12 years old, the The changes that occur in sibling rela-
age at which children typically show a rea- tionships, however, are unique in at least one
sonable amount of self-sufficiency and no important respect: children's experiences
longer need continuous supervision. Thus, as with siblings differ greatly depending on
later-born siblings grow older, they become whether they are older or younger siblings.
more competent and independent, thereby Older siblings inherit positions of authority
requiring and accepting less nurturance and and responsibility that they never hold in
direction from older siblings. At the same their relationships with parents and peers. As
time, the relative difference in developmental they grow up, older siblings are faced with
status between older and younger siblings di- relinquishing power/status, whereas younger
minishes as children get older. For example, a siblings acquire a more equal footing. Addi-
3-year-old child is vastly more competent tional research is needed to explore the dy-
than his or her newborn younger sibling (in namics of this transformation in power/status
1396 Child Development
and to determine whether it has signiflcant Third, the course of experiences with sib-
consequences for child development. lings is partially determined by the child's
status in the family constellation. The most
Second, as children grow older, their influential aspect of constellation status is the
sibling relationships typically become less child's position of being the older or younger
intense. Ratings of every major dimension member of the dyad. Younger siblings have
of sibling relationships dropped ofl^ to some experienced being nurtured and dominated,
degree with age, including the exercise of whereas older siblings are nurturant and
power, the warmth/closeness of relationships, dominating. The size of the effects associated
and the extent of reported conflict. This trend with relative age was very large for the domi-
can partially be traced to the decreasing nance and nurturance scales, indicating that
amount of interaction among siblings as they relative age of the two siblings accounted for
grow older. In fact, the most pronounced age the vast majority of variance in these scores.
trend among the warmth/closeness scales was Although the strength of these effects dimin-
found for companionship, with ratings for ished substantially with age (as indicated
twelfth graders being over a full standeurd de- above), it seems clear that being an older ver-
viation lower than those for third graders. sus younger sibling during early childhood is
Data from other studies corroborate this associated with markedly different experi-
flnding. During the preschool years, siblings ences.
spend the vast majority of their time in the
presence of each other (Ellis, Rogoff, & There also appears to be a basic asym-
Cromer, 1981), but by the time siblings are metry in the sentiments that older and youn-
adolescents, they spend a relatively small ger siblings feel for one another. For instance,
fraction of their time together (Raffaelli & Lar- there is a discrepancy between older and
son, 1987). This decreasing rate of interaction younger siblings' perceptions of conflict:
undoubtedly affects the frequency of both later-born subjects reported that conflict with
positive and negative interactions. older siblings dropped ofF steadily with age,
whereas earlier-bom subjects' ratings did not
Although the changing social ecologies evidence a parallel decline in conflict with
of childhood and adolescence may partially younger siblings (see Fig. 2). In addition to
account for decreasing rates of sibling interac- less conflict with older siblings, later-bom
tion, they may also reflect the psychological subjects reported greater admiration for and
transition from dependence on family to in- intimacy with older siblings than earlier-bom
vestment in peer relationships (Buhrmester & subjects report toward younger siblings. It ap-
Furman, 1987; Ftirman & Buhrmester, 1989). pears as if younger siblings look up to and
Adolescents may want to spend less time value interacting with older brothers and sis-
with siblings who are part of the family from ters, whereas older siblings view younger sib-
whom they want to develop some autonomy. lings as an annoyance. This asymmetry in
In addition, as adolescents become more in- sentiments may be part of a separation and
volved in intimate friendships and romantic individuation stmggle in which earlier-bom
relationships, they have less time and perhaps adolescents try to distance themselves from
less socioemotional need to invest in relation- the family, whereas the later-bom children try
ships with siblings. It is important, however, to be "more grown up" by identifying with
not to overstate the degree to which sibling the greater autonomy of older siblings. This
relationships become more distant with age. interpretation is speculative, of course, but
There were, in fact, relatively modest age- deserves exploration in future research.
related decrements in ratings of intimacy, af-
fection, and admiration (amounting to less Other constellation variables also affect
than one-half standard deviation difference the course of sibling relationships, but to a
over the age-range studied). Thus, the emo- lesser degree. In general, wider spacing be-
tional attachment between siblings remains tween siblings tends to foster more facilitative
moderately strong throughout adolescence, relationships than narrower spacing, with
despite the decline in companionship (Weiss, \vider spacing associated with more nurtur-
1974). Moreover, our examination of group ance, prosocial behavior, and affection and
averages undoubtedly masked important vari- narrower spacing associated with greater
ations in the developmental course of sibling quarreling, antagonism, and dominance. In
relationships. Some sibling relationships may addition, the gender composition of the dyad
become supportive egalitarian friendships affects the warmth/closeness of these relation-
during adolescence, whereas others may be- ships. Cenerally, same-sex siblings feel closer
come distant. than opposite-sex siblings. It is important to
 Buhrmester and Furman 1397
note, however, that although these constella- relationships than previously available, sev-
tion effects have now been replicated across eral cautionary notes are in order. First, we
several studies, they generally account for a assessed children's perceptions of their rela-
rather modest amount of the variance in sib- tionships. Although there is evidence that
ling relationships (with relative age being children's perceptions substantially agree
thie noteworthy exception). In fact, Stocker, with the perceptions of parents and siblings
Dunn, and Plomin (1989) recently demon- of their relationships (Furman et al., 1989),
strated that child temperament, matemal be- their perceptions are not always veridical
havior, and child age accounted for more of with accounts of observed behavior. This
the variance in the quality of sibling relation- does not imply, however, that children's per-
ships than did family structure variables. ceptions are less worthy of investigation (Ol-
son, 1977). To the contrary, self-perceptions
In addition to the three foregoing gener- of relationships may be psychologically im-
alizations, we also put forth the following portant variables determining development
hypothesis: The qualities of sibling rela- and psychosocial adjustment.
tionships during the preschool and early
elementary years may be more influential in Second, the flndings indicate that sibling
shaping the distinctiveness of children's per- relationships are affected in complex ways by
sonalities than relationships during adoles- family constellation. Although we assessed
cence. This admittedly speculative conten- one of the largest cross-sectional samples ex-
tion is a logical derivative of the conclusions amined to date, statistical power for sensi-
sp)ecifled above, but it is not as yet grounded tively detecting possible three-, four-, and
in empirical flndings. Two related considera- flve-way interactions is still lacking. Our re-
tions suggest this hypothesis. First, it seems sults may capture the most pronounced age
likely that the influence of any relationship on and constellation effects but undoubtedly
individual development is related to the miss more complex and subtler effects.
scope and intensity of that relationship. That
is, inasmuch as sibling relationships are more Finally, the current results provide a pic-
intense during early childhood than adoles- ture of age-related differences in sibling re-
cence, it follows that sibling relationships lationships but do not address issues con-
should have their greatest effects on per- cerning continuity and change over time.
sonalities and adjustment during early child- Longitudinal studies are now Ccdled for to de-
hood. termine how siblings resolve issues during
development concerning power, conflict, and
Second, behavior geneticists have argued dependency.
that siblings may create different environ-
ments for each odier, thereby contributing to References
the differences between them (Rowe & Plo-
min, 1981; Scarr & Grajek, 1982). It is our Abramovitch, R., Corter, C, & Lando, B. (1979).
view that "complementary" features of sib- Sibling interactions in the home. Child Devel-
ling relationships such as dominance and nur- opment, 50, 997-^1003.
turance, in contrast to "reciprocal" features Bigner, J. A. (1974). Second-boms' discrimination
like warmth and conflict, are likely to play of sibling role concepts. Developmental Psy-
the greatest role in fostering dissimilarities chology, 10, 564-573.
among siblings (see Dunn, 1983; Rowe & Plo- Bowerman, C. E., & Dobash, R. M. (1974). Struc-
min, 1981). Inasmuch as the current flndings tural variations in inter-sibling affect. Journal
indicate that the greatest asymmetry in these of Marriage and the Family, 36, 48-54.
complementary features occurs in early child- Bragg, B. W. E., Ostrowski, M. V., & Finley, G. E.
hood, it may be that sibling relationships con- (1973). The effects of birth order and age of
tribute more to the development of dissimilar- target on use of persuasive techniques. Child
it)' during that time than during adolescence. Development, 44, 351-354.
This possibility may help explain why Rowe Bryant, B. K. (1982). Sibling relationships in middle
and Plomin (1981) found litde evidence using childhood. In M. E. Lamb & B. Sutton-Smith
a sample of adolescents that the qualities of (Eds.), Sibling relationships: Their nature and
sibling relationships contributed to differ- significance across the Ufespan (pp. 87-121).
ences in siblings' personalities. Hillsdale, NJ: Erlbaum.
Bryant, B. K., & Crockenberg, S. B. (1980). Corre-
Limitations and Conclusions lates and dimensions of prosocial behavior: A
Although the current data provide a more study of female siblings and their mothers.
complete picture of age differences in sibling Child Development, 51, 529-544.
1398 Child Development
Buhrmester, D., & Furman, W. (1987). The devel- 115—135). Amherst: University of Massachu-
opment of companionship and intimacy. Child setts Press.
Development, 58, 1101-1113. Rafifaelli, M., & Larson, R. W. (1987). Sibling in-
Cicirelli, V. C. (1982). Sibling influence throughout teractions in late childhood and early adoles-
the lifespan. In M. E. Lamb & B. Sutton-Smith cence. Paper presented at the biennial meeting
(Eds.), Sibling relationships: Their nature and of the Society for Research in Child Develop-
significance across the lifespan (pp. 267-284). ment, Baltimore.
Hillsdale, NJ: Erlbaum. Ross, H. G., & Milgram, J. I. (1982). Important vari-
Dunn, J. (1983). Sibling relationships in early child- ables in adult sibling relationships. In M. E.
hood. Child Development, 54, 787-811. Lamb & B. Sutton-Smith (Eds.), Sibling rela-
Dunn, J., & Kendrick, C. (1982). Siblings: Love, tionships: Their nature and significance across
envy, and understanding. Cambridge, MA: the lifespan (pp. 123-152). Hillsdale, NJ: Erl-
Harvard University Press. baum.
Ellis, S., Rogoff, B., & Cromer, C. C. (1981). Age Rowe, D. C , & Plomin, T. (1981). The importance
segregation in children's social interactions. of nonshared E^ environmental influence in
Developmental Psychology, 17, 399-407. behavioral development. Developmental Psy-
Furman, W., & Buhrmester, D. (1985). Children's chology, 17, 517-531.
perceptions of the qualities of sibling relation- Scarr, S., & Grajek, S. (1982). Similarities and differ-
ships. Child Development, 56, 448-461. ences among siblings. In M. E. Lamb & B. Sut-
Furman, W., & Buhrmester, D. (1989). Age dif- ton-Smith (Eds.), Sibling relationships: Their
ferences in perceptions of social networks. nature and significance across the life span
Unpublished manuscript. University of Den- (pp. 357-381). Hillsdale, NJ: Erlbaum.
ver. Stocker, C, Dunn, J., & Plomin, R. (1989). Sibling
Furman, W., Jones, L., Buhrmester, D., & Adler, T. relationships: Links with child temperament,
(1989). Children's, parents', and observers' per- matemal behavior, and family structure. Child
ceptions of sibling relationships. In P. G. Zu- Development, 60, 715-727.
kow (Ed.), Sibling interaction across culture Sutton-Smith, B., & Rosenberg, B. G. (1970). The
(pp. 163-180). New York: Springer-Verlag. sibling. New York: Holt, Rinehart & Winston.
Koch, H. (1960). The relation of certain formal at- Vandell, D. L., Minnett, A. M., & Santrock, J. W.
tributes of siblings to attitudes held toward (1987). Age differences in sibling relationships
each other and toward their parents. Mono- during middle childhood. Journal of Applied
graphs of the Society for Research in Child Developmental Psychology, 8, 247-257.
Development, 25(4, Serial No. 78). Weiss, R. S. (1974). The provisions of social rela-
Minnett, A. M., Vandell, D. L., & Santrock, J. W. tionships. In Z. Rubin (Ed.), Doing unto others
(1983). The effects of sibling status on sibling (pp. 17-26). Englewood Cliffs, NJ: Prentice-
interaction: Influence of birth order, age spac- Hall.
ing, sex of child, and sex of sibling. Child De- Youniss, J. (1980). Parents and peers in social de-
velopment, 54, 1064-1072. velopment: A Sullivan-Piaget perspective. Chi-
Olson, D. H. (1977). Insiders' and outsiders' views cago: University of Chicago Press.
of relationships: Research studies. In G. Levin- Youniss, J., & Smollar, J. (1985). Adolescent rela-
ger & H. L. Raush (Eds.), Close relationships: tions with mothers, fathers, and friends. Chi-
Perspectives on the meaning of intimacy (pp. cago: University of Chicago Press.