The Association Between Sibling Closeness Perceptions,

advertisement
The Association Between Sibling Closeness Perceptions,
Observed Family Communication Patterns, and Adoptive Status
Diana R. Samek & Martha A. Rueter
©2008 Regents of the University of Minnesota.
All rights reserved.
DEPARTMENT OF FAMILY SOCIAL SCIENCE, UNIVERSITY OF MINNESOTA, MINNEAPOLIS – ST. PAUL
Introduction
Methods
Adolescent externalizing behaviors have been negatively associated with family
closeness perceptions (Anderson & Henry, 1994; Roisman, 2002; Hamilton,
2005; Meadows, 2007; Branje, van Lieshout, van Aken, Haselager, 2004), as
well as observed family communication patterns (FCPs) (Rueter & Koerner,
2008), and adoptive status (Keyes, Sharma, Elkins, Iacono & McGue. 2008). In
fact, recent evidence suggests that FCPs interact with adoptive status to predict
externalizing (Rueter & Koerner, 2008).
Families participating in the Sibling Interaction and Behavior Study (SIBS)
(McGue et al., 2007) were included in this analysis. SIBS is a longitudinal
research study designed to examine gene x environment influences on adolescent
drug and alcohol use. Each family consisted of at least one parent and two
adolescent children (mean age = 14.9 years, SD = 1.9). Final study sample
included 1,232 adolescents (692 adopted adolescents, and 540 non-adopted
adolescents) from 616 families. Of the 1,232 adolescents (female: 54.8%),
53.7% were Caucasian, 37.5% were Asian, and 8.8% reported another ethnicity.
Figure 3. First-order factor means across Family
Communication Patterns
1.75
Control
Family Communication Patterns Theory (FCPT) (Koerner & Fitzpatrick,
2002, 2004, 2006) proposes that family communication is a combination of two
dimensions, conversation and conformity. These dimensions work together to
create a family’s shared social reality. In short, the conversation dimension is a
continuum of open communication amongst family members, whereas the
conformity dimension is a continuum of adhering to beliefs, values, and attitudes
within the family. The combination of these two dimensions produces four family
communication patterns (See Figure 1).
Utilizing FCPT on a sample of
adopted and non-adopted youth,
Rueter & Koerner (2008) found
that adopted adolescents were 5
times more likely, compared to
non-adopted adolescents, to have
externalizing problems in families
that placed little emphasis on
family communication and greater
emphasis on parent control
(Protective families), and were 3
times more likely in families that
emphasize
neither
family
communication
nor
control
(Laissez-Faire families).
Figure 1.
Koerner & Fitzpatrick’s (2006)
Family Communication Patterns
Model
High Conformity Orientation
Protective
Consensual
Low
Conversation
Orientation
High
Conversation
Orientation
Laissez-Faire
Pluralistic
Low Conformity Orientation
This study extends earlier research by testing five hypotheses which propose
that sibling closeness perceptions link FCPs to adolescent externalizing
problems, as well as propose a moderating effect of adoptive status on the
relationship between FCPs and sibling closeness perceptions. (See Figure 2 for
study model). Moreover, this study analyzed sibling emotional closeness and
sibling behavioral closeness as separate constructs in order to further understand
their similarities and differences.
Figure 2. Study model depicting study
hypotheses
Adoptive Status
Sibling
Behavioral
Closeness
Perceptions
Sibling
Emotional
Closeness
Perceptions
Communication
1.25
Warmth
0.75
Listening
Responsiveness
0.25
-0.25
Measures
Sibling emotional and behavioral closeness were assessed using 6 indicators
that made up two different scales from the Sibling Relationship Questionnaire
(Furman & Buhrmester, 1985). The emotional closeness factor consisted of 3
items that make up the Affect scale (e.g., “…How much do you and this sibling
care about each other?”). The behavioral closeness consisted of 3 items that make
up the Companionship scale (e.g., “…How much free time do you and this
sibling spend together?”). Indicator loadings ranged from .773 - .883.
The Family Communication Patterns variable was computed based on the
same measurement construction plan created by Rueter & Koerner (2008). FCPs
were rated by judges observing structured videotaped family interactions, and
assessed as two dimensions, conversation (warmth, listening responsiveness,
communication), and conformity (control), which became latent class indicators
of a 4-class variable. Reliability was assessed using intraclass correlations
(ICCs) (Shrout & Fleiss, 1979). ICCs ranged from .60 - .76.
Externalizing was assessed with Latent Class Analysis using 5 indicators: (1)
self-reports from the Delinquent Behavior Inventory (Gibson, 1967) (α = .894),
(2) teacher ratings on externalizing behavior (adapted from Conners, 1969 and
Rutter, 1967) (α’s ranged from .88 to .93; Spearman-Brown interteacher
reliability ranged from .74 to .77.), (3) externalizing symptom counts from the
DICA-R (Weiner et al., 1987) (range = 0 – 32), and trained observers ratings of
child hostility to (4) mother and (5) father) (ICC’s ranged from .58 to .78). This
variable has been used in a similar fashion in earlier research (Rueter & Koerner,
2008).
Results
Family Communications Patterns. The FCP latent class variable was created using
16 1st-order factors as indicators of the 2nd-order FCP variable. 1st-order factors
had 3 indicators each, where each indicator represented the family member’s
behavior towards the other family members (4 family members (mother, father,
elder adolescent, younger adolescent) x 4 measures (control, warmth, listening
responsiveness, communication). Due to Rueter & Koerner’s (2008) demonstration
of the presence of 4 classes using the present study’s sample, class comparisons
were not analyzed. To produce the 4-class FCP variable, first-order factor means
were set to the values reported by Rueter & Koerner (2008) (See Figure 3). Out of
the 616 families, 22.2% were classified as Protective, 6.8% as Consensual, 30.5%
as Pluralistic, and 40.5% as Laissez-Faire.
-0.75
Protective
Consensual
Pluralistic
Laissez-Faire
Note: First bar in every set is mother’s mean factor score. Second bar is father’s mean factor score. Third bar: Elder
sibling’s mean factor score. Fourth bar is younger sibling’s mean factor score. Bars above 0 represent behaviors
above the overall mean, bars below 0 represent behaviors below the overall mean.
Hypothesis Testing
Hypothesis 1: FCPs will be related to adolescent externalizing problems.
Proportions of adolescents placed in the high externalizing class for each FCP were:
Protective = 14.3%; Consensual = 8.3%; Pluralistic = 16.5%; Laissez-Faire =
20.8%. Results supported hypothesis 1a. Adolescents placed in the Laissez-Faire
class were more likely to be placed in the high externalizing class, compared to
adolescents in the Consensual class (Fisher Exact: p < .01; OR = 2.50, 95% CI =
1.12 – 5.56)
Family
Communication
Patterns
Sex
Age
Adolescent
Externalizing
Problems
Figure 5. Interaction of Adoptive Status on the
relationship between FCPs and Sibling Emotional
Closeness
0.55
0.45
0.35
0.25
0.15
0.05
-0.05
Hypothesis 2: Association of sibling closeness perceptions to FCPs and
externalizing. Figure 4 describes the pattern of closeness means across FCPs. It
was expected that for the full sample, adolescents in Consensual families would
report greater closeness, compared to adolescents in Laissez-Faire families. Results
supported hypothesis 2a for emotional closeness: The change in –2 x
Loglikelihood indicated that, with one exception, adolescents in all classes reported
significantly different emotional closeness (Chi-square difference test, all p’s < .03).
The exception was that adolescents in Pluralistic and Laissez-Faire families
reported similarly low levels of emotional closeness, relative to Consensual.
There was no significant difference in behavioral closeness when comparing
adolescents in Protective and Consensual families (relative to Laissez-Faire) or
between Pluralistic and Laissez-Faire families (relative to Consensual). It was
expected that adolescents who reported greater sibling closeness would be less
likely to be placed in the high externalizing class. Results supported hypothesis 2b
for emotional closeness, (CR = -1.859, p < .10; Estimate = -.383, 95% CI = -.787 .021), but not for behavioral closeness (CR = .039, ns).
Figure 4. Sibling emotional and behavioral
closeness factor means across FCPs
0.6
Full sample Emotional
0.5
Full sample Behavioral
0.4
0.3
Adopted Adolescents Emotional
0.2
Externalizing. By examining the adjusted Bayesian Information Criterion (BIC), as
well as the Lo-Mendell-Rubin adjusted LRT for the 1-,2-,3- and 4-class models of
externalizing, it was determined the 2-class model of externalizing was the best fit
due to a relatively substantial drop in the 2-class BIC compared to the other latent
class models, as well as the significant LRT p-value (p < .001). 82.4% of the
adolescents are in the low externalizing group, and 17.5% are in the high
externalizing group.
Hypothesis 3: Moderation of adoptive status between FCPs and sibling
closeness. Interaction results supported hypotheses 3a and 3b for emotional
closeness, but not for behavioral closeness. Adopted adolescents reported
significantly greater emotional closeness in Consensual families (M = .511),
compared to Protective families (M = .128), Chi-square difference test: p < .05. By
contrast, no significant differences were found non-adopted adolescents in these
classes (Consensual = .435; Protective = .204;). On the other hand, non-adopted
adolescents reported significantly greater emotional closeness in Protective (M =
.204) compared to Pluralistic (M = -.024) families, Chi-square difference test, p <
.05. Adopted adolescents reported no significant differences in emotional closeness
means for these classes (Protective = .128; Pluralistic = .03).
Adopted Adolescents Behavioral
0.1
1E-15
Consensual
Protective
Adopted
Pluralistic
Non-Adopted
Conclusions
This study’s findings suggest that the combination of clear, attentive family
communication and developmentally appropriate parent control are uniquely
associated with adolescents’ perceptions’ of sibling emotional closeness, as well as
with adolescent adjustment. Interaction results indicate the unique effect of
communication for adopted adolescents, compared to their non-adopted peers.
Family members who are genetically related may be able to rely on similar
cognitive processes or attitudes to influence feelings of similarity and closeness in
families that emphasize conformity and not conversation (due to the heritability of
these constructs). By contrast, open communication amongst family members
appears equally important in the development of feelings of sibling emotional
closeness in adoptive families.
Differences in emotional versus behavioral closeness were found indicating the
importance in operational definitions and conceptualizations of closeness. This
research indicates that spending time and having fun with a sibling (behavioral
closeness) is not associated with adolescent adjustment, however it is highly
associated with reporting greater feelings of love and care for a sibling (emotional
closeness). Moreover, being part of a family that actively emphasizes
communication and developmentally appropriate control may lead to increased
feelings of sibling emotional closeness. These results suggest that one way in which
adolescents may come to feel close to their siblings is through family level
communication behaviors.
Non-adopted Adolescents
Emotional
-0.1
-0.2
Protective
Consensual
Pluralistic
Non-adopted Adolescents
Behavioral
Note: Laissez-Faire is the reference class (i.e., M = 0). Bars above 0 represent closeness perceptions greater than
Laissez-Faire class, bars below 0 closeness represent perceptions below Laissez-Faire class.
Acknowledgements
This research was supported by grants from the National Institute on Alcohol Abuse and
Alcoholism (AA 11886) and the National Institute on Mental Health (MH066140).
Download