Interparental Conflict and Family Cohesion : Predictors of Loneliness, Social Anxiety, and Social Avoidance in Late Adolescence

Journal of Adolescent Research
http://jar.sagepub.com
Interparental Conflict and Family Cohesion: Predictors of Loneliness, Social Anxiety, and Social
Avoidance in Late Adolescence
H. Durell Johnson, Joseph C. Lavoie and Molly Mahoney
2001; 16; 304
Journal of Adolescent Research
DOI: 10.1177/0743558401163004

The online version of this article can be found at:
http://jar.sagepub.com/cgi/content/abstract/16/3/304
Published by:
http://www.sagepublications.com

can be found at:
Journal of Adolescent Research
Additional services and information for

http://jar.sagepub.com/cgi/alerts

Email Alerts:

http://jar.sagepub.com/subscriptions

Subscriptions:

http://www.sagepub.com/journalsReprints.nav
Reprints:

http://www.sagepub.com/journalsPermissions.nav
Permissions:
http://jar.sagepub.com
Downloaded from
by Plamen Dimitrov on February 28, 2007
© 2001 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

---

JOURNAL OF ADOLESCENT RESEARCH / May 2001
Johnson et al. / INTERPARENTAL CONFLICT
Interparental Conflict and Family Cohesion
Predictors of Loneliness, Social Anxiety, and
Social Avoidance in Late Adolescence
H. Durell Johnson
Pennsylvania State University,
Worthington Scranton
Joseph C. LaVoie
Molly Mahoney
University of Nebraska at Omaha
Family environment is related to characteristics of adolescents’ personal development
and social interactions. Although potentially different for males and females, decreased
family cohesion and increased interparental conflict can inadvertently provide family
environments that are associated with increased feelings of loneliness, which may be
associated with problems in adolescents’ social interactions (i.e., social anxiety and
social avoidance). Analyses of responses from 124 late adolescents revealed that feel-
ings of loneliness were related to perceived levels of interparental conflict for males and
females and decreased family cohesion for females. Furthermore, late adolescents’ feel-
ings of social anxiety and social avoidance were related to their feelings of loneliness.
The findings in this study show how deteriorated family systems may provide contexts
that are associated with adolescents’ feelings of loneliness as well as their ability to
engage in social interactions outside of the family system.
Late adolescence has been characterized as one of the loneliest periods in
the life span (Steinberg, 1999, p. 321). If unresolved, adolescents’ feelings of
loneliness can become an obstacle in the formation of normative social rela-
tionships through the development of anxiety and social avoidance (Lau &
Kong, 1999; Vernberg, Ewell, Beery, & Abwender, 1994). Although experi-
ences of loneliness during late adolescence are attributed to the restructuring
We would like to thank Beck Stull for her assistance in data collection and the editor and anonymous reviewers
for their helpful comments on earlier versions of this article. Completion of this research was supported in part
by University of Nebraska at Omaha University Committee on Research Grant OGA 22-010-217. Correspon-
dence concerning this article should be addressed to H. Durell Johnson, Human Development and Family
Studies, Pennsylvania State University, Worthington Scranton, 120 Ridge View Drive, Dunmore, PA
18512-1699.
Journal of Adolescent Research, Vol. 16 No. 3, May 2001 304-318
© 2001 Sage Publications, Inc.
304
by Plamen Dimitrov on February 28, 2007
http://jar.sagepub.com
Downloaded from

© 2001 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

---

Johnson et al. / INTERPARENTAL CONFLICT
305
of peer relationships and social groups and the moving to new environments
(i.e., college), the association between disruptive family systems and the
loneliness that faces older adolescents has received little attention in the liter-
ature. Current research with children and high school adolescents (e.g.,
Stark, Humphrey, Laurent, Livingston, & Christopher, 1993) has suggested
that disruptive/hostile family environments are associated with interpersonal
deficits and difficulties during social interactions. Two such family charac-
teristics are interparental conflict and family cohesion. Both are equally
important when attempting to understand adolescents’ feelings of loneliness
and the social and personal difficulties that are associated with this feeling.
Interparental Conflict, Family Cohesion, and
Adolescents’ Social and Personal Adjustment
Interparental conflict. Although children and adolescents witness conflict
between their parents, research has shown that perceptions of intense
interparental conflict are associated with difficulties in social and personal
development for children and adolescents. Howes and Markman (1989)
report that high levels of marital conflict decrease the likelihood that parents
will establish strong relationships with their children. As a result, children
and adolescents may experience increased feelings of rejection due to the
emotional distance between themselves and their parents. In addition, Collins
and Laursen (1992) state that exposure to continuously high levels of family
conflict is associated with “psychosocial problems during adolescence and
later in life” (p. 217). Cummings, Davies, and Simpson (1994); Harold,
Fincham, Osborne, and Conger (1997); and Snyder (1998) suggest that per-
ceptions and interpretation of intense interparental conflict may be more
damaging to child and adolescent development than the act of witnessing
conflict. This research shows that adolescents’ reports of and awareness of
conflict were associated with adjustment difficulties and internalizing symp-
toms (i.e., anxiety, depression, worry, and withdrawal) over a 12-month
period (see also Neighbors, Forehand, & Bau, 1997). Despite these associa-
tions, Neighbors et al. and Cummings et al. were unable to find a relationship
between perceptions of interparental conflict and adolescent females’ reports
of internalizing symptoms and general psychopathology. This evidence sug-
gests that perceptions of interparental conflict are related to a wide range of
children’s and adolescents’ personality characteristics; however, these rela-
tionships are different for adolescent males and females. Therefore, disrup-
tions in the marital relationship may be associated with higher levels of dis-
tress for males than for females, and this sensitivity may be associated with
by Plamen Dimitrov on February 28, 2007
http://jar.sagepub.com
Downloaded from

© 2001 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

---

306
JOURNAL OF ADOLESCENT RESEARCH / May 2001
adolescent males’ reports of internalizing symptoms and general
psychopathology.
Family cohesion. Family cohesion (i.e., the emotional bonding among
family members and the feeling of closeness) is expressed by feelings of
belonging and acceptance within the family system (McKeown et al., 1997).
Wentzel and Feldman (1996) and McKeown et al. have found that adoles-
cents’ perceptions of low cohesion within their families were associated with
heightened feelings of depression and reduced social acceptance. Reinherz,
Stewart-Berghauer, Pakiz, Frost, and Moeykens (1989) suggest that low
cohesion, expressed by feelings of not belonging, is associated with chil-
dren’s and adolescents’ feelings and behaviors that are reflective of their
family environment (i.e., loneliness). However, Wentzel and Feldman report
that levels of cohesion reported by male and female adolescents have differ-
ent implications for their personal and social adjustment. Lower levels of
family cohesion were associated with female reports of feeling excluded and
depressed, whereas male reports were not associated with their perceptions
of family cohesion. Furthermore, Leadbeater, Blatt, and Quinlan (1995) and
Maccoby (1990) report that female adolescents are more likely to experience
feelings of exclusion and isolation in response to difficulties in maintaining
relationship closeness (i.e., cohesion). Although males report a desire for
closeness with their parents, the level of desired closeness with parents does
not approach the level of females’, and they are less sensitive than females to
disruptions in family closeness (Fischer, Munsch, & Greene, 1996). As a
result, disruptions in the closeness of parent-adolescent relationships may be
associated with higher levels of social and personal distress for females than
for males, and this sensitivity may be associated with adolescent females’
reports of depression and feelings of exclusion.
Despite the evidence that links perceptions of interparental conflict and
family cohesion with internalizing symptoms and social adjustment difficul-
ties, their association with loneliness and the gender differences in late ado-
lescents’ feelings of loneliness has not been established. Collins and Laursen
(1992) suggest that children’s and adolescents’ social behavioral styles
resulting from exposure to disruptive family environments are associated
with behaviors that make social interactions difficult (i.e., aggression and
externalizing problems). However, results concerning the association
between low cohesion and high levels of interparental conflict and late ado-
lescents’ functioning have not been consistently reported in the literature.
Although the dynamics appear to differ for male and female adolescents,
interparental conflict and low family cohesion are associated with various
personality and social difficulties including depression, loneliness, aggres-
by Plamen Dimitrov on February 28, 2007
http://jar.sagepub.com
Downloaded from

© 2001 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

---

Johnson et al. / INTERPARENTAL CONFLICT
307
sion, and subsequent poor social adjustment (Cummings et al., 1994; Har-
old & Conger, 1997; Perry, Perry, & Kennedy, 1992; Wentzel & Feldman,
1996). It is the relationship between marital and family disruption and ado-
lescents’ reports of internalizing symptoms and general psychopathology
that make the association between family environment and feelings of lone-
liness plausible.
Interparental Conflict, Family
Cohesion, and Loneliness
The resulting stress from exposure to disruptive family environment can
lead to child and adolescent adjustment problems (Fincham, Grych, &
Osborne, 1994). Examination of interparental conflict and decreased family
cohesion has shown a relationship with adolescent feelings of depression
and other internalizing symptoms. According to Leadbeater et al. (1995),
internalizing symptoms that result from the disruption of interpersonal rela-
tionships increases adolescents’ vulnerability to other “negative interper-
sonal events” (p. 12). Although research has predominately examined the
association between interparental conflict and family cohesion and depres-
sion, it is likely that loneliness is also associated with these family dynamics.
Koening, Isaacs, and Schwartz (1994) reveal that depression and loneliness
are highly correlated constructs. A study of adolescent depression and loneli-
ness by Koening et al. reveals that feelings of loneliness are associated with
male and female adolescents’ reports of depression. Because the relationship
between interparental conflict, family cohesion, and depression has been
established in the research, the association between loneliness and depres-
sion suggests that adolescents’ feeling of loneliness should also be associated
with marital conflict and family cohesion.
Loneliness and Social Adjustment
Researchers have proposed many different ideas to explain the association
between interparental conflict and low family cohesion and adolescent
adjustment. One idea not widely examined proposes that loneliness associ-
ated with disruptive family environment may be related to difficulty in estab-
lishing and maintaining relationships in other social situations (McKeown
et al., 1997; Wentzel & Feldman, 1996). Loneliness associated with adoles-
cent reports of interparental conflict and the perception of decreased family
closeness may be associated with anxiety and stress in other social situations
(cf. Leadbeater et al., 1995; Wentzel & Feldman, 1996). Evidence for this
relationship is seen in work by Lau and Kong (1999), who report that feelings
by Plamen Dimitrov on February 28, 2007
http://jar.sagepub.com
Downloaded from

© 2001 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

---

308
JOURNAL OF ADOLESCENT RESEARCH / May 2001
of loneliness are associated with self-reports of having an inadequate physi-
cal appearance, low self-concept, low emotional stability, and inadequate
social skills. Given this pattern, it follows that male and female adolescents’
feelings of loneliness may be associated with decreased effectiveness in
employing skills needed for social situations outside of the family environ-
ment (Asher, Parkhurst, Hymel, & Williams, 1990). Schwab, Scalise, Ginter,
and Whipple (1998) show that individuals who report feeling lonely are
increasingly aware of being embarrassed and/or rejected during social situa-
tions. This feeling may increase adolescents’ anxiety when faced with social
situations and their desire to avoid these situations all together. Newcomb
and Bagwell (1996) state that individuals who experience high levels of lone-
liness are more likely than others to distance themselves from social situa-
tions and feel increased apprehension when faced with social interactions. As
a result, feelings of loneliness may be related to adolescents’ feelings of anxi-
ety during social interactions as well as their increased social isolation from
others.
Current Study
The present study addresses the proposed notion that late adolescents’
perceptions of their family environment (i.e., perceptions of interparental
conflict and family cohesion) are associated with their reports of loneliness,
social anxiety, and social avoidance. It is proposed that interparental conflict
and family cohesion are positively associated with adolescents’ feeling of
loneliness, which, in turn, are positively related to their reports of social anxi-
ety and avoidance. However, gender is expected to moderate the association
between adolescents’ reports of loneliness and perceptions of interparental
conflict and family cohesion. Perceptions of interparental conflict are
expected to be associated with increased reports of loneliness for males, and
family cohesion is expected to be associated with increased reports of loneli-
ness for females.
METHOD
Participants
The nonrandom sample consisted of 124 first-year college undergradu-
ates (70 females, mean age = 19.69 years, range = 17 through 21 years of age;
and 54 males, mean age = 19.41 years, range = 17 through 21 years of age)
by Plamen Dimitrov on February 28, 2007
http://jar.sagepub.com
Downloaded from

© 2001 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

---

Johnson et al. / INTERPARENTAL CONFLICT
309
who were recruited from psychology classes at an urban commuter univer-
sity. The sample was predominantly White and from middle-class back-
grounds. Due to potential differences in dynamics of intact nonbiological
families (e.g., stepfamilies or foster families) or single-parent families, only
participants from intact biological families were included in the current
study.
Perceived Family Environment Measures
Perception of interparental conflict. The Children’s Perception of
Interparental Conflict Scale (CPIC) (Grych, Seid, & Fincham, 1992) was
used to assess participant interpretation and response to interparental con-
flict. This 48-item scale yields a total score of perceived interparental conflict
comprised from three subscale scores that measure perceived threats to self,
self-blame, and destructive parental conflict. The internal consistency of the
three scales has been reported as ? = .84 for Threats to Self, ? = .83 for
Self-Blame, and ? = .90 for Parental Conflict. Test-retest reliability estimates
over a 2-week period are r = .68, .76, and .70, respectively. The reliabilities of
the subscales for this study were alpha(s) = .93 for Threats to Self, .89 for
Self-Blame, and .91 for Parental Conflict. Scores on the CPIC have been cor-
related with parental reports of interpersonal conflict, thus supporting the
validity of the measure (see Grych et al., 1992). Furthermore, Bickham and
Fiese (1997) show support for the use of the CPIC in adolescent research. In a
study of late adolescents, the factor structure, reliability, and validity mea-
sures all resembled scores found with younger samples. For the present
study, the general score was used because it provides data on the participant’s
overall level of perceived conflict within the family. The reliability (alpha)
for the overall scale in the current study was .95.
Family cohesion. The Cohesion subscale of the Family Adaptability and
Cohesion Scale-III (FACES-III) (Olson, Portner, & LaVee, 1985) was used to
measure family cohesion. This measure is a 10-item subscale that assesses
the level of cohesion within the family environment (e.g., Family members
feel very close to each other). The test items were scored on a 5-point
Likert-type scale ranging from 1 = almost never to 5 = almost always. The
extremely low correlation between the adaptability and cohesion subscales
(r = .03) enables the use of individual scales for research. The reliability of
the cohesion scale for this study was ? = .90. Validation studies of the
FACES-III scale support its use with a variety of family structures (Olson et al.,
1985).
by Plamen Dimitrov on February 28, 2007
http://jar.sagepub.com
Downloaded from

© 2001 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

---

310
JOURNAL OF ADOLESCENT RESEARCH / May 2001
Loneliness and Social Anxiety/Avoidance Measures
Loneliness. The UCLA Loneliness Scale (Revised) (Russell, Peplau, &
Cutrona, 1980) consists of 20 items that assess feelings of loneliness. The
measure is composed of 10 items reflecting satisfaction with social relation-
ships (e.g., There are people I feel close to) and 10 items reflecting dissatis-
faction (e.g., I am unhappy being so withdrawn). Responses to these ques-
tions range from 1 (never) through 4 (often), with a possible range of 20 (low
loneliness) through 80 (extreme loneliness). The internal consistency of the
measure in the current study was ? = .92.
Social anxiety and social avoidance. The Franke and Hymel (1984) Social
Anxiety and Social Avoidance Scale was used to assess each participant’s
social adjustment. The measure consists of 12 items comprising two separate
subscales of 6 items that assess social anxiety (e.g., I worry a lot about what
others think of me) and social avoidance (e.g., I have the most fun when I’m
with others). Possible responses range from 1 (not at all true about me)
through 5 (always true about me). Responses yield two scores (Social Anxi-
ety and Social Avoidance) ranging from 6 (low anxiety/avoidance) to 30
(high anxiety/avoidance). The internal consistency of the two subscales has
been reported as ? = .85 for Social Anxiety and ? = .74 for Social Avoidance.
The reliability (alpha) of the two subscales in the current study was .88 for
Social Anxiety and Social Avoidance.
Procedure
Data collection was completed in 30- to 45-minute sessions using small
groups. The participants were told that the purpose of the study was to exam-
ine how perceptions of their family environment are associated with their
emotional reactions toward social interactions. The procedure used during
data collection was similar to procedures used by Laursen (1993, 1995) and
Fuligini and Eccles (1993) in which participants completed self-report ques-
tionnaires asking them to report on their perceptions of interpersonal con-
flicts and their outcomes. Each participant was given a randomly ordered
conflict questionnaire packet that asked them to report on their perceptions of
interparental conflict and family cohesion and their feelings of loneliness,
social anxiety, and social avoidance. They were then told to answer each
question as accurately as possible by circling or entering the response most
relevant to their experiences. When the questionnaires were completed and
collected from the group, the participants were debriefed about the study.
by Plamen Dimitrov on February 28, 2007
http://jar.sagepub.com
Downloaded from

© 2001 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

---

Johnson et al. / INTERPARENTAL CONFLICT
311
TABLE 1: Mean Scores for Late Adolescent Male and Female Reports of Family
Environment and Social Functioning
Females
Males
(n = 70)
(n = 54)
Interparental conflict
30.37 (18.51)
29.39 (18.42)
Family cohesion
35.18
(7.74)
31.25
(8.41)**
Loneliness
33.03 (10.39)
38.09 (10.85)**
Social anxiety
16.94
(5.50)
19.11
(4.16)*
Social avoidance
12.37
(3.75)
14.03
(4.49)*
NOTE: Standard deviations in parentheses. Row means significantly different.
*p < .05. **p < .01.
TABLE 2: Intercorrelations for Late Adolescent Male and Female Reports of
Family Environment and Social Functioning
1
2
3
4
5
1. Interparental conflict
—
–.46**
.39**
.26*
.09
2. Family cohesion
–.59**
—
–.45**
–.35**
–.11
3. Loneliness
.42**
–.35*
—
.58**
.39**
4. Social anxiety
.35*
–.38*
.50**
—
.33*
5. Social avoidance
.25
–.02
.56**
.59**
—
NOTE: Correlations for female and male participants above and below diagonal, re-
spectively; n = 70, females and n = 54, males.
*p < .05. **p < .01.
RESULTS
Descriptive and Correlational Analyses
Table 1 displays the means and standard deviations for the measures of
interparental conflict, family cohesion, loneliness, social anxiety, and social
avoidance. t tests revealed that females reported higher levels of family cohe-
sion, t(130) = 2.07, p < .01, and males reported higher levels of loneliness,
t(130) = –2.85, p < .01; social anxiety, t(130) = –3.39, p < .05; and social
avoidance, t(130) = –2.12, p < .05. No significant difference was found
between males’ and females’ reports of interparental conflict.
Table 2 presents the zero-order correlations between family environment
variables and reports of loneliness, social anxiety, and social avoidance. In
general, the correlational patterns for males and females are quite similar and
by Plamen Dimitrov on February 28, 2007
http://jar.sagepub.com
Downloaded from

© 2001 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

---

312
JOURNAL OF ADOLESCENT RESEARCH / May 2001
consistent with the hypothesized relationships between interparental con-
flict, family cohesion, loneliness, social anxiety, and social avoidance.
Data Analyses
The proposed relationship among family environment variables, loneli-
ness, social anxiety, and social avoidance was tested using multiple regres-
sion procedures. First, stepwise regression procedures recommended by
Pedhazur (1997) were used to determine the relationships among gender,
family environment variables, and adolescents’ reports of loneliness. The
stepwise method is useful because selection tests are calculated at each step
to determine the contribution of each predictor already in the equation as if it
were entered last (Pedhazur, 1997). Therefore, stepwise selection allows one
to identify predictors that are significant at early stages but lose predictive
significance when other variables are entered into the equation. Although this
process may result in entering more variables into the equation than are nec-
essary, it provides the option of backing up from the last step to one in which
the set of variables included in the equation is deemed most useful. Second,
multivariate regression procedures recommended by Pedhazur (1997) and
Stevens (1996) were used to determine the relationships between gender,
loneliness, and adolescents’ feelings of social anxiety and social avoidance.
Due to the high correlation between adolescents’ reports of social anxiety
and social avoidance, multivariate regression is necessary to account for the
relationship between the dependent variables. According to Stevens,
multivariate regression is useful because it allows one to predict several
related dependent variables from a set of predictors. Finally, significance of
the relationships between family environment and personality constructs was
measured by analyzing standardized path coefficients (i.e., beta weights) and
the total amount of variance explained in the dependent measures (R2).
Interparental conflict, family cohesion, and loneliness. Step 1 of the
regression analyses revealed a significant association between interparental
conflict and feelings of loneliness and accounted for a significant percentage
of the variance, ? = .15, p < .01, R2 = .24, p < .001. Family cohesion and gen-
der were not associated with feelings of loneliness. Step 2 revealed that the
Gender × Family Cohesion interaction was significantly associated with feel-
ings of loneliness, and the increase in explained variance was also signifi-
cant, ? = –.44, p < .001, R2(change) = .28, p < .001. Inspection of the Gender
× Family Cohesion interaction shows that as family cohesion decreased,
female feelings of loneliness increased; however, little change was evident in
male feelings of loneliness (see Figure 1). The Gender × Interparental Conflict
by Plamen Dimitrov on February 28, 2007
http://jar.sagepub.com
Downloaded from

© 2001 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

---