Professional Documents
Culture Documents
1990 Intimacy of Friendship, Interpersonal Competence, and Adjustment During Preadolescence
1990 Intimacy of Friendship, Interpersonal Competence, and Adjustment During Preadolescence
1990 Intimacy of Friendship, Interpersonal Competence, and Adjustment During Preadolescence
and Adolescence
Author(s): Duane Buhrmester
Source: Child Development, Vol. 61, No. 4 (Aug., 1990), pp. 1101-1111
Published by: Blackwell Publishing on behalf of the Society for Research in Child
Development
Stable URL: http://www.jstor.org/stable/1130878
Accessed: 20/06/2010 06:35
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=black.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Blackwell Publishing and Society for Research in Child Development are collaborating with JSTOR to digitize,
preserve and extend access to Child Development.
http://www.jstor.org
Intimacy of Friendship, Interpersonal
Competence, and Adjustment during
Preadolescence and Adolescence
Duane Buhrinester
University of Texas at Dallas
Several authors have argued that there tions, intimacy has most often been equated
are important developmental changes in the with the depth and breadth of self-disclosure
nature and significance of friendship during (Jourard,1979), although Sullivan (1953), and
early adolescence (Berndt, 1982; Buhrmester more recently Reis and Shaver (1988), indi-
& Furman, 1986; Gottman & Mettetal, 1987; cated that the core process of intimate interac-
Sullivan, 1953). Friendships among preschool tions is not disclosure, per se, but ratherthe
and elementary school-aged children revolve experiences of feeling understood, validated,
primarily around playmate activities and and cared for that accompany self-disclosure.
group acceptance, whereas adolescent friend- As a feature of relationships, intimacy usually
ships become more intimate in nature. Al- describes a somewhat broader arrayof qual-
though the increased intimacy of adolescent ities. For example, Sullivan believed that
friendship is generally thought to have impor- truly intimate exchanges required a type of
tant implications for both short-and long-term relationship he described as a "collabora-
socioemotional functioning (Buhrmester & tion." Although exactly what features com-
Furman, 1986; Sullivan, 1953), few studies prise intimate collaborative friendships is
have empirically examined these implica- open to debate, at a minimum they involve
tions. The present study investigated whether engaging in mutual activities, self-disclosure,
there are age differences between preadoles- and reciprocalfeelings of satisfactionwith the
cence and adolescence in how importantinti- relationship (Furman & Robins, 1985; Man-
macy of friendship is to adjustment and narino, 1976; Sullivan, 1953). The current
growth of interpersonalcompetencies. study is primarilyconcerned with intimacy as
a feature of relationships.
The construct of intimacy has been used
both to describe types of interactions and Consistent with early accounts (Douvan
types of relationships. As a feature of interac- & Adelson, 1966; Sullivan, 1953),recent stud-
I would like to thank the staffand students of Seeds UniversityElementarySchool and of the
Los Angeles Unified Schools for participatingin this study. I would also like to thankDanny Kaye
for his helpful comments on this manuscript.This researchwas supportedby grant3975 fromthe
Academic Senate of the Universityof California,Los Angeles. Requests forreprintsand information
concerning the measures used in this study should be sent to the author at the School of Human
Development, the University of Texas at Dallas, Richardson, TX 75083.
[Child Development,1990, 61, 1101-1111. ? 1990 by the Society for Researchin Child Development,Inc.
All rights reserved.0009-3920/90/6104-0015$01.00]
1102 Child Development
ies have documented that, during early ado- (roughly 8 to 12 years) and adolescent (older
lescence, children's friendships become more than about 13 years) samples to determine
intimate as indicated by more frequent com- whether there is an increase with age in the
panionate exchanges, personal disclosure, strength of the association between friend-
and provision of emotional support(Buhrmes- ship intimacy and psychosocial adjustment
ter & Furman, 1987; Gottman & Mettetal, more broadly defined (e.g., anxiety, depres-
1987; Youniss & Smollar,1985; see Steinberg, sion, hostility, sociability, and self-esteem).
1989, for review). There are several theoreti- One purpose of the current study was to re-
cal reasons for believing that accompanying dress this situation.
the increase in the level of friendship inti- A second general implication of the in-
macy there is also an increase in the impor- creased friendship intimacy across the transi-
tance of friendship intimacy to socioemo- tion to adolescence concerns changes in the
tional functioning. First, not having intimate
friends may be a significant source of stress. types of social competencies that are impor-
tant in friendship relations. The demands of
Sullivan argued that the need for intimacy the interpersonal task faced by adolescents
intensifies during early adolescence, and if
left unsatisfied through friendship, leads to appear to be quite different from those faced
by younger children. Because childhood
heightened feelings of loneliness, alienation, friendships center around play activities and
and depression. Second, adolescents appear
to have an increased desire for self-disclosure group acceptance, being a competent friend
involves knowing how to enter ongoing
and self-explorationrooted in a need for "con-
sensual validation" of personal worth (Parker games, being a fun and "nice" play partner
& Gottman, 1989; Sullivan, 1953). Youths (i.e., sharingand helping), and refrainingfrom
who do not have intimate friendships may insulting or aggressing against one's friends
miss out on importantvalidating interactions, (Asher, 1983). In contrast, adolescent friend-
which can leave them feeling less secure, ship demands greater facility in a number of
close relationship competencies (here re-
more anxious, and less worthy. Finally, ferred to as "interpersonal competencies").
youths who lack intimate friendships may be To a greater extent than is true of younger
deprived of important sources of social sup- children, then, adolescents must be capable
port and coping assistance. The support of of initiating conversations and relationships
intimate friends may be particularly impor-
tant during adolescence as the young person outside of the classroom context. They must
also be skilled in appropriately disclosing
confronts a variety of uniquely adolescent
stressors (e.g., bodily changes, sexuality, dat- personal informationand tactfully providing
emotional support to friends (Youniss &
ing, and strained family relationships),many Smollar, 1985). In addition, adolescents are
of which cannot comfortably be discussed
with parents (Buhrmester & Furman, 1987). expected to express honestly their opinions
and dissatisfactions with each other while at
Thus, whereas relationship intimacy should the same time effectively managing conflicts
be associated with several aspects of adjust-
ment during both middle childhood and early (Bigelow & LaGaipa, 1980; Shantz, 1987).
These interpersonalcompetencies are similar
adolescence, several theoretical considera- to the skills called for in matureadult roman-
tions suggest that the association should be
tic relationships and only partially overlap
stronger during early adolescence. with the playmate skills demanded in early
childhood.
Relatively few studies have examined
developmental changes in the associationbe- According to this reasoning, adolescents
tween friendship intimacy and adjustment.A who lack these specific relationship com-
handful have focused on the links between petencies are more likely to have difficulty
self-esteem and friendship intimacy. Self- achieving intimacy in their friendships. They
reported closeness of friendship has been are likely to have fewer friendships that are
found to be positively related to self-esteem more superficial in nature, and thus these
among preadolescents (Bukowski & Hosa, youths are also likely to be at greaterrisk for
1989; Cauce, 1986; Mannarino, 1978), and adjustment problems. At the same time, be-
similar findings have been reported in some cause they are unable to establish intimate
studies of adolescents (O'Donnell, 1976). friendships, they probablymiss out on experi-
Other studies, however, have revealed weak ences that could further promote the growth
correlationsbetween friendship intimacy and of these competencies (Hartup & Sancilio,
self-esteem (Berndt, 1987; Blyth & Traeger, 1986). That is, many of the subtleties concern-
1987; Jourard, 1979). Almost no one, how- ing when and what to self-disclose and how
ever, has directly compared preadolescent to provide emotional support are probably
Duane Buhrmester 1103
learned and refined through hours of interac- issues concerning: (a) the extent of conver-
tions with close friends. Moreover, one would gence among self- and friend's ratings of
expect the association between interpersonal friendship quality and interpersonal compe-
competence and intimacy in friendship to be tence, and (b) the validity of the Interpersonal
stronger during adolescence than preadoles- Competence Questionnaire (Buhrmester et
cence because it is during adolescence that al., 1988) for use with preadolescents and ado-
the interactional processes that demand and lescents.
foster close relationship competencies be-
come central in friendship relations. Method
No studies have attempted to investigate Subjects
this developmental change directly. Man- The initial preadolescent sample con-
narino (1976, 1979) and McGuire and Weisz sisted of 133 fifth and sixth graders (68 girls),
(1982) found that children who scored high on ages 10-13 years (M = 11.3), whereas the ini-
measures of altruism and affective perspec- tial adolescent group consisted of 100 eighth
tive taking (which are constructs related to in- and ninth graders (44 girls), ages 13-16 years
terpersonal competence) were more likely to (M = 14.4), from ethnically and racially di-
be involved in stable close friendships than verse metropolitan Los Angeles schools. The
children who scored low on these measures. preadolescents represented 85% of the chil-
These investigators, however, studied only a dren from seven mixed-aged classrooms in
narrow age range of children (10-12 years) two elementary schools, whereas the adoles-
and consequently did not attempt to evaluate cents constituted 44% of the youths from nine
whether there were developmental changes classrooms in one junior high school. The
in the importance of these competencies. Fur- lower percentage of adolescents taking part in
thermore, they did not assess altruism and the study was due to a lower rate of return of
perspective taking as it was manifested parental consent forms. After the friendship
specifically in the context of intimate friend- pairing procedure (explained below), the
ship (see Price & Ladd, 1986). This is prob- sample consisted of 102 preadolescents and
lematic because we know that children be- 70 adolescents.
have differently toward friends and
nonfriends (Hartup, 1983). Assessment Design
A two-step procedure was used to assess
One general obstacle to research on ado- friendship intimacy. Initially, nomination pro-
lescent social competence has been the un- cedures were used to identify relationships in
availability of adequate measures. Although which pairs of students reciprocally reported
methods have been developed to assess the that they were friends. Students then rated
skills needed by younger children to gain the intimacy of their relationships with recip-
peer-group acceptance (Asher, 1983; Dodge, rocal friends. This method of assessment is
Pettit, McClaskey, & Brown, 1986; Putallaz & similar to that employed by Berndt (1981),
Gottman, 1981), little work has been done to and differs in important ways (which will be
develop measures that specifically assess the discussed later) from the more common prac-
interpersonal competencies called for in in- tice of having students rate their self-pro-
timate relationships (Ford, 1982). Recently, claimed "best friends." Students also rated
however, Buhrmester, Furman, Wittenberg, their own adjustment and competence, as
and Reis (1988) developed a measure for col- well as providing ratings of their preceptions
lege-aged adolescents that specifically as- of their friends' interpersonal competence.
sesses the competencies assumed to be im-
portant in both close friendships and romantic Procedures
relationships. This measure was modified Data were gathered in two testing ses-
slightly in the present study to investigate sions. During the first session, participants
these issues. were given rosters of names listing all same-
sex schoolmates participating in the study and
In the cross-sectional study reported instructed to indicate which peers they con-
here, groups of preadolescents and adoles- sidered close friends. Children were told that
cents were administered several question- "close friends are kids you know very well,
naires to investigate age differences in the spend a lot of time with in and out of school,
strengths of associations among friendship in- and who you talk to about things that happen
timacy, socioemotional adjustment, and inter- in your life." A tally of nominations revealed
personal competence. In addition to the two that preadolescents nominated about twice as
central substantive issues discussed previ- many peers as close friends (M = 4.1) than
ously, the data also addressed two secondary did adolescents (M = 1.9), t(231) = 7.62, p <
1104 Child Development
.001. This age difference is attributable to at Friendship intimacy scores were created by
least two factors: (a) Past studies have found averaging the nine items. Alpha coefficients
that adolescents report having significantly scores were satisfactory for both preadoles-
fewer close friends than preadolescents cents (.93) and adolescents (.93).
(Berndt & Hoyle, 1985; Epstein, 1986), and Adolescent Interpersonal Competence
(b) a lower percentage of the adolescent sam-
ple's grademates took part in the study
Questionnaire (AICQ).-This newly devel-
oped 40-item questionnaire assessed the fol-
(roughly 18% of the eighth/ninth graders at-
lowing five domains of competence that are
tending the school vs. 85% of the fifth/sixth
important in close relationships (sample items
graders' grademates), and thus it was rela- in parentheses): self-disclosure ("How good is
tively less likely that all of an adolescent's this person at opening up and letting friends
close friends took part in the study. Children
also indicated which peers they considered get to know everything about him/herself?"),
providing emotional support to friends ("How
"fairly good friends" (defined as "kids you
know and like, but who you don't feel quite as good is this person at making friends feel bet-
ter when they are unhappy or sad?"), manage-
close to as close friends").
ment of conflicts ("How good is this person at
Children's responses were examined to knowing how to disagree with friends without
identify instances of reciprocal friendship. getting into big arguments?"), negative asser-
A relationship was considered to be a close tion ("How good is this person at telling
reciprocal friendship if both children nomi- friends that they have been neglectful or in-
nated each other as close friends. If a student considerate?"), and initiation of friendships
had two or more close friendships, then one ("How good is this person at phoning friends
of the peers was randomly selected with the to set up a time to do things together?").
constraint that the selection did not interfere
with another peer being paired with a close The AICQ was developed by modifying
friend. It was possible to pair 76% of the the Interpersonal Competence Questionnaire
preadolescents and 70% of the adolescents (ICQ), a measure that was originally devel-
with reciprocal close friends. In cases where oped to assess college students' interpersonal
reciprocated friendships could not be found, competence in close friendships and romantic
schoolmates whom subjects had nominated as relationships (Buhrmester et al., 1988). Stud-
friends (but who had not reciprocated the ies of the college ICQ demonstrate that its
five scales are adequately reliable, they con-
nomination) were identified so every child
had two peers to rate on questionnaires (see form to the predicted five-factor simple struc-
below). These nonreciprocated friendships ture, and they correlate in predictable and
were not included in the analyses. Instances discriminant ways with theoretically related
of reciprocal fairly good friends were also variables. In rewording the ICQ items, an ef-
fort was made to make the vocabulary appro-
identified, and students rated one fairly good
friend on the friendship and competence priate for young adolescents without chang-
measures. The findings for fairly good friends ing the substantive content of questions.1
are not reported here, however, because of
Following each item were blank lines on
space limitations and because there were which to rate the competence of the identified
very few significant associations with these close friend (and the fairly close friend) and to
ratings. rate the respondent's self-perceptions of com-
Measures petence. The names of friends being rated ap-
Friendship Intimacy (FI).-This ques- peared at the top of the page. Respondents
tionnaire consisted of the items from the were instructed to use Levenson and Gott-
Companionship, Intimate Disclosure, and man's (1978) 5-point rating scale to indicate
Satisfaction scales from the Network of Rela- the level of competence and comfort that each
tionships Inventory (Furman & Buhrmester, target child would have in handling each type
1985). An illustrative intimate disclosure item of situation (e.g., "1 = Poor at this; would be
reads: "How often do you share secrets and so uncomfortable and unable to handle this
private feelings with this person?" Subjects situation that it would be avoided if possible"
rated the qualities of their relationships with to "5 = EXTREMELY good at this; would
identified friends using a 5-point Likert-type feel very comfortable and could handle this
scale (e.g., "1 = Never or hardly ever" to "5 situation very well"). Scores were created by
= VERY often or EXTREMELY much"). averaging all 40 AICQ items.2 Cronbach
1 Copies of this and all measures used in the study can be obtained from the author.
2
Owing to space limitations, data for the five individual AICQ scales are not presented in this
report. Information about the AICQ factor structure, scale reliabilities, and other descriptive data are
available from the author.
Duane Buhrmester 1105
TABLE 1
MEANS AND STANDARD DEVIATIONS OF FRIENDSHIP INTIMACYAND COMPETENCE SCORES
PREADOLESCENTS ADOLESCENTS
alpha coefficients were computed separately Hostility, r = -.06 and -.37 (among preado-
for the two age groups and were high: preado- lescents and adolescents, respectively); Socia-
lescents (.93) and adolescents (.92). bility-Anxiety/Depression, r = -.15 and
-.27; Hostility-Anxiety/Depression, r = .62
Self-Esteem Scale.-Rosenberg's (1965) and .50. Self-esteem was also moderately cor-
10-item scale was used to obtain global self-
related with Sociability, r = .43 and .47, Anx-
evaluations. Respondents rated their level of
iety/Depression, r = -.59 and -.53, and
agreement ("1 = Strongly agree" to "5 =
Hostility, r = -.34 and -.26, for the two age
Strongly disagree") with statements such as
"I feel I have a number of good qualities." groups, respectively. These associations are
The scale was satisfactorily reliable for pre- similar to those reported for other measures of
youths' adjustment (Achenbach, 1985) and in-
adolescents (.82) and adolescents (.82). dicate that the scales do not assess orthogonal
Socioemotional Adjustment (SA).-A dimensions of adjustment, although the mod-
40-item self-report questionnaire was devel- est to moderate size of the associations sug-
oped for this study as a relatively brief yet gests that the scales nonetheless assess dis-
broad measure of several spheres of self- tinguishable aspects of adjustment. Whereas
perceptions of socioemotional adjustment. It information about the validity of these scales
was necessary to develop this measure be- is limited, the available data indicate that
cause existing measures of child adjustment scale scores are correlated in theoretically
were either too narrow in scope (i.e., assess- predictable ways with other measures of so-
ing only one sphere of adjustment) or took cial functioning (Buhrmester, 1989).
more time to administer than was allotted by
school personnel. Items similar to those found Results
on the Child Behavior Check List, the Child
Manifest Anxiety Scale, the Child Depres- PreliminaryAnalyses
sion Inventory, and the Hopkins Symptoms Effects of sex, age, and rater.-Table 1
contains the means and standard deviations
Checklist were employed because of their es-
for the AICQ and FI scores for each age
tablished validity. Students responded on a 5-
point Likert-type scale (e.g., "1 = Never or group and sex. Consistent with findings from
not at all" to "5 = Very often or very much"). past studies, a two-way analysis (ANOVA) of
FI scores revealed a significant main effect of
Factor analyses revealed three orthogo- sex, F(1,168) = 4.02, p < .05, with girls (M =
nal dimensions labeled Sociability, Hostility, 3.22) rating their friendships as more intimate
and Anxiety/Depression (see Buhrmester, than boys (M = 3.06). Counter to expecta-
1989, for details). Scale scores were computed tions, adolescents did not rate their close
by averaging items that assess each dimen- friendships as more intimate than preadoles-
sion. Example items included: "How well do cents. This finding is surprising because past
you work with other people?" (Sociability; studies examining similar age groups using
10 items), "How often do you lose your similar questionnaires have found significant
temper?" (Hostility; 10 items), and "How of- age differences (e.g., Buhrmester & Furman,
ten do you feel unhappy or down?" (Anxiety/ 1987).
Depression; 20 items).
Next, an ANOVA was conducted on
Cronbach alpha coefficients computed AICQ scores to determine whether percep-
for the current samples were adequate, rang- tions of subjects' interpersonal competence
ing from .80 to .87 for preadolescents and .72 varied according to sex, age, and the type of
to .82 for adolescents. The intercorrelations rater (self vs. close friend). There was a sig-
among the three SA scores were: Sociability- nificant effect of type of rater, F(1,167) =
1106 Child Development
TABLE 2
CORRELATIONS
BETWEEN
FRIENDSHIP
INTIMACY
ANDADJUSTMENT
SELF-RATED
INTIMACY FRIEND-RATED
INTIMACY
Adjustment:
Sociability ............... .20*b .56**a .31** .53**
Hostility ................. .04b -.30**a -.07 -.24*
Anxiety/depression ....... .00b - .32**a - .24* -.30**
Self-esteem .............. - .01b .34**a .27** .29**
NOTE.-Coefficients differ significantly for the two age groups if their superscriptsdiffer.
* p < .05.
** p < .01.
12.53, p < .001. Follow-up comparisons using is more important to adjustmert during ado-
the Newman-Kuels procedure and a .05 alpha lescence than preadolescence, although there
level indicated that self-ratings (M = 3.82) were discrepancies between the results for
were higher than ratings by close friends (M the self-report and friend-report measures of
= 3.62). These differences are similar to friendship intimacy. Fischer's procedure
those found for college students' ICQ ratings (with alpha set at .05) revealed that self-
of self, and friends (Buhrmester et al., 1988). reported friendship intimacy was more
There was also a significant effect of grade, strongly associated with adjustment for ado-
F(1,167) = 16.25, p < .01, with adolescents lescents than preadolescents, with significant
(M = 3.57) rated as less competent than age differences found for all four dimensions
preadolescents (M = 3.85). This age differ- of adjustment. Significant age differences
ence may be attributable to a developmental were not found, however, for the correlations
trend observed in several studies where, as between friend-rated intimacy and adjust-
children get older, they lower their estimates ment, seemingly because of the relatively
of their own competence (Stipek & MacIver, stronger correlations for the preadolescent
1989).3 There were no other significant main sample.
effects or interactions.
The correlations contained in Table 3
Finally, a MANOVA was conducted on support the contention that interpersonal
the four adjustment scores. There were no competence is related to friendship intimacy
significant main effects or interactions of sex during adolescence. Both self- and friend-
or age. rated competence scores were moderately to
strongly related (r = .45 to .68) to self- and
Associations among Friendship Intimacy, friend-rated friendship intimacy scores for the
Adjustment, and Competence adolescent sample. It is noteworthy that self-
Table 2 provides clear evidence that ratings of competence were related just as
friendship intimacy is related to adjustment strongly to friend ratings of friendship inti-
for adolescents. All four dimensions of self- macy as to self-ratings of friendship intimacy,
reported adjustment were significantly and revealing that the associations held up across
moderately (r = -.24 to .56) correlated with (as well as within) independent sources of rat-
both self- and friend ratings of friendship inti- ings.
macy. For these and all other correlational
There was mixed support for the hypoth-
analyses reported here, separate analyses
were also conducted to determine whether esis that interpersonal competence is more
correlations differed for boys and girls within important for friendship intimacy during ado-
each age group. No significant sex differences lescence than preadolescence. There were
in coefficients were found, and therefore all marked age differences in the correlations be-
results are reported for the combined samples tween self-rated friendship intimacy and self-
of boys and girls. and friend-rated AICQ scores, with compe-
tence more strongly related to intimacy for
The results in Table 2 also provide some adolescents than preadolescents. No signifi-
support for the view that friendship intimacy cant age differences were found; however, for
3 An informal comparison between the AICQ means and the ICQ means for college students
reported by Buhrmester et al. (1988) revealed that the means for the eighth and ninth graders were
quite similar to college students' means (difference + .15 scale points) for all scales except emo-
tional support competence (where college students scored about .30 scale points higher).
Duane Buhrmester 1107
TABLE 3
CORRELATIONSOF INTERPERSONALCOMPETENCE WITH ADJUSTMENTAND FRIENDSHIP INTIMACY
Friendship intimacy:
Self-rated ................ .12b .45**a - .03b .63**a
Friend-rated ............. .29** .46** .56** .68**
Adjustment:
Sociability ............... .55** .63** .32**b .57**a
Hostility 00b - .26*a - .05 .14
................. - .30**a - .06 - .26*
Anxiety/depression ....... .07
Self-esteem .............. .27** .36** .22* .39**
NOTE.-AICQ = Adolescent Interpersonal Competence Questionnaire. Coefficients differ significantly for the
two age groups if their superscriptsdiffer.
* p < .05.
** p < .01.