Adolescents in High Risk Families

Adolescents in High Risk Families:
Are There Different Dynamics at Home That Affect Behavior
Problems?
Karina R. Sokol, Priscila Diaz, William Fabricius, Melinda Baham
Arizona State University
Introduction
Method
Results
Figures
Figure 1. Internalizing Behavior Problems
Adolescents living in step-families are at risk for more internalizing
and externalizing behavior problems than adolescents in intact families.
Similarly, Mexican-American adolescents are at an increased risk for behavior
problems compared to Euro-American adolescents. Thus, this study examines
individuals in both of the aforementioned populations in order to assess what
dynamics at home contribute to or buffer behavior problems.
Previous research has shown multiple family and parent influences
can affect adolescent behavior problems. A substantial body of research
supports an association between interparental conflict and adolescent
adjustment problems. High levels of conflict between parents has been found
to relate to more internalizing problems and externalizing problems
(Cummings, Goeke- Morey, & Papp, 2004; Dumka, Roosa & Jackson, 1997;
Fainsilber-Katz & Gottman, 1993; Fincham, 1994).
On the other hand, it seems that the parent-child relationship may
mediate the relationship between interparental conflict and child outcomes
(Gonzales, Pitts, Hill & Roosa, 2000). Drawing from the parent-as-a-mediator
hypothesis, interparental conflict is assumed to affect child well-being by
altering the parent-child relationship. The changes in the parent-child
relationship, as an alternative to the conflict itself, ultimately lead to child
problems (Black & Pedro-Carroll, 1993).
Recently, researchers have found that time spent with a parent relates
to adolescent outcomes in divorced families independent of interparental
conflict (Luecken & Fabricius, 2003). In intact families, children who spend
more time with their fathers have better grades than children who spend less
time with their fathers. Time with mothers does not significantly predict
school grades on its own given its multicollinearity with time with fathers
(Cooksey & Fondell, 1996). When a parent takes the time to spend a portion of
his/her day with his/her child, the parent is conveying to the child that he is
meaningful. The child realizes that the relationship between himself and his
parent is important to the parent. Drawing from emotional security theory, the
child may infer that the relationship between himself and his parent is
emotionally secure regarding his parent’s love and the ability to care for him
(Davies & Cummings, 1994).
Given that prior literature (Amato & Rezac, 1994) has suggested that
the time a child spends with a parent might be harmful if the parents have high
levels of conflict, the time by conflict interaction was included in this study.
In the current study, the model we tested specified that time spent
with a parent and interparental conflict independently predict the parent-child
relationship which, in turn, predicts adolescent internalizing and externalizing
behaviors. Time spent with each parent, interparental conflict, and parent-child
relationship have not been explored in any sample, regardless of a multiethnic,
multi-family type sample. Additionally, specific analyses on high risk families
will test if the hypothesized model differs by gender, family type, or ethnicity.
Contact Information
For more information please contact:
Karina R. Sokol
Department of Psychology
Arizona State University
Tempe, AZ 85287-1104
(480) 727-7172 (office)
[email protected]
PAYS website: http://devpsych.sfsu.edu/PAYS/index.htm
Source: Sokol, K.R., Diaz, P., Fabricius, W. V., & Baham, M. (2006). Adolescents in
High Risk Families: Are There Different Dynamics at Home That Affect Behavior
Problems? Poster presented at 2006 Society for Prevention Research, San Antonio,
TX.
This research was supported in part by NIMH grant (1RO1MH064828-01).
Time X
Conflict
Mom
Participants
Our sample consisted of 367 families (105 Euro-American intact families, 83
Euro-American stepfather families, 102 Mexican-American intact families, and
77 Mexican-American stepfather families). The families were participants in
the first wave of the Parents and Youth Study (PAYS). PAYS is a two-site, 5year longitudinal project that examines how Mexican-American and EuroAmerican fathers and stepfathers influence adolescents’ mental health and
behavioral outcomes. Adolescents (195 female, 172 male) were in the 7th grade
for the results reported here and the age ranged from 11-13 years. The families
were recruited in schools from both the Phoenix and the Riverside areas.
Data Collection
Simultaneous in-home interviews were conducted in different rooms for the
three participating family members (mother, father, adolescent) by a team of
three trained interviewers. The interviews lasted approximately two hours and
were conducted in either Spanish or English, depending on the request of the
participant. Adolescent’s teachers were also asked to complete a questionnaire
regarding the adolescent’s behaviors at school.
Measures
Marital Conflict
A modified version of the Children’s Perception of Interparental Conflict
(CPIC; Grych, Seid & Fincham, 1992) was administered to the adolescent,
mother, and father/stepfather. Fifteen items were chosen from the original
CPIC and one new item was added, for a total of 16 items with a response scale
ranging from 1 to 3. All the items were summed within reporter and averaged
across reporters. The alphas were .80 for father/stepfather report, .80 for
mother report, and .82 for adolescent report.
Time Spent with Parent
Time was measured by the 5-item Parent Child Interaction (PCI) Scale
(Coltrane, Parke & Adams, 2004) with a response scale ranging from 1 to 5.
Both the adolescent and the mother reported on the PCI for a global measure of
time the adolescent and mother spent together (adolescent α= .68, mother α=
.59). Similarly, both the adolescent and the father/stepfather reported on the
PCI for time the adolescent spent with father (adolescent α= .68,
father/stepfather α= .57). The items were summed within reporter and then
averaged across reporters.
Parent-Child Relationship
Shortened versions of the Child Report of Parent Behavior Inventory (CRPBI;
Schaefer, 1965) scales of acceptance and rejection were used to assess parentchild relationship (10 items for each scale) on a 1 to 3 rating scale. To obtain a
global measure of mother-child relationship for the acceptance scale, the
adolescent and father/stepfather reports were summed for each reporter and
averaged across the two reporters. The rejection scale was also summed for
each reporter, and the adolescent and father/stepfather report were averaged.
Similarly, for father-child relationship, the mother and adolescent reports were
summed for each reporter and the two reporters were averaged (separately for
acceptance and rejection). All scales had reliability levels above α = .70.
Internalizing and Externalizing behavior
Measured by the teacher report of the Behavior Problems Index (BPI; National
Longitudinal Study of Youth, 1979). This scale contains two subscales:
internalizing behaviors (10 items) and externalizing behaviors (18 items), each
rated on a 1 to 5 scale. Two teachers reported on each adolescent and the items
were summed for each teacher. The reports of these teachers were averaged
separately for internalizing and externalizing. If only one teacher report was
obtained, that report was solely used. The alpha for internalizing was .89 and
.96 for externalizing behavior.
Overall Analyses
Structural equations modeling with observed variables (Bollen,
1991), often referred to as path analysis, was used. It was hypothesized that the
amount of marital conflict observed by the child and the amount of time the
child spent with the parent would each independently influence the relationship
the parent had with the child, and that relationship, in turn, would impact the
child’s behavior. Finally, because it was of interest to examine the simultaneous
impact of mother and father, the time spent with each parent and the
relationship with each parent was tested in the same model, shown in Figures 1
(internalizing) and 2 (externalizing). Time and conflict variables were allowed
to correlate with each other and the four relationship variables were allowed to
correlate.
The results of the model for internalizing behavior problems indicate
that the hypothesized model fits the data well, χ2(13) = 23.54, p = .04, with a
CFI value of 0.98, an RMSEA value of 0.047, and an SRMR value of 0.03.
The hypothesized model for externalizing behavior problems was also a good
fit to the data, χ 2(13) = 23.14, p = .04, with a CFI value of 0.98, an RMSEA
value of 0.046, and an SRMR value of 0.03.
Specific Analyses
Time
Mom
.26
-.20
-.18
Marital
Conflict
Family Type. First, the model was fit separately in intact families and stepfamilies to ensure the appropriateness of a direct comparison. For internalizing
behavior problems, the model did not fit in intact families but was a good fit in
step-families. In intact families, the previously discussed model fit the data
once a direct path from time spent with dad to acceptance from mom was
added. A new significant path emerged between the interaction between time
with dad and conflict and acceptance from dad in this model such that more
time with dad was positively related to acceptance from dad, regardless of the
level of conflict, but that the relationship between low amounts of time with
dad and acceptance from dad depended on the level marital conflict.
For externalizing behavior problems, the model did not fit in intact
families but was a good fit in step-families. In intact families, the previously
discussed model fit the data once a direct path from time spent with dad to
acceptance from mom was added. The original model fit the data well in stepfamilies.
Ethnicity. The original model was fit separately in Euro-Americans (EA) and
Mexican-Americans (MA). For internalizing behavior problems, the models fit
similarly to those for intact and step-families with EAs similar to step-families
and MAs similar to intact families. Specifically, the model fit for EAs but only
fits for MAs when a path is added from time with dad to acceptance with mom.
Additionally, in EAs there were significant paths from the interaction
terms between time with dad and conflict. The time by conflict interaction
significantly predicted acceptance with dad such that high amounts time with
dad was related to high levels of acceptance from dad, regardless of the level of
conflict, but that the relationship between low amounts of time with dad and
acceptance from dad depended on the level marital conflict. The interaction
between time with dad and conflict for rejection from dad was such that high
amounts time with dad was related to lower levels of rejection from dad,
regardless of the level of conflict, but that the relationship between low
amounts of time with dad and rejection from dad depended on the level marital
conflict
For externalizing behavior problems, the original model fit EA
families well, with similar significant path coefficients as the overall model. In
MA families, the previously discussed model fit the data once a direct path
from time spent with dad to acceptance from mom was added.
Rejection
Mom
.28
.37
Time
Dad
.24
Internalizing
-.22
Acceptance
Dad
.39
-.23
Rejection
Dad
Time X
Conflict
Dad
Figure 2. Externalizing Behavior Problems
Time X
Conflict
Mom
Time
Mom
Acceptance
Mom
.26
-.20
-.18
Gender. The model fit equally well in both males and females.
Acceptance
Mom
Marital
Conflict
Time
Dad
Time X
Conflict
Dad
.28
-.22
.37
.39
-.23
Rejection
Mom
-.21
-.17
Externalizing
Acceptance
Dad
Rejection
Dad
*Blue dotted lines indicate paths that were tested but were not significant
Conclusion
Given our statistical analyses, it seems that generally, both time with each
parent and marital conflict independently predict the parent-child
relationship, which in turn predicts adolescent behavior problems. The
models fit well in both female and male adolescents. The model also fits
well in step-families and EA families, but not in intact families and MA
families. In the latter, it may be that the family acts as more of a unit and
time with fathers is related to acceptance from mothers. In other words,
spending time with fathers indicates that mother is more accepting. Time
may spill over to affect the entire family dynamic relationship within these
families.
Representative References**
Black, A.E. & Pedro-Carroll, J. (1993). Role of parent-child relationships in mediating the
effects of marital disruption. Journal of the American Academy of Child Psychiatry, 32,
1019-1027.
Davies, P.T. & Cummings, E.M. (1994). Marital conflict and child adjustment: an
emotional security hypothesis. Psychological Bulletin, 116, 387-411.
Dumka, L.E., Roosa, M.W., & Jackson K.M. (1997). Risk, conflict, mother’s parenting,
and children’s adjustment in low income, Mexican immigrant, and Mexican American
families. Journal of Marriage and the Family, 59, 309-323.
Gonzalez, N.A., Pitts, S.C., Hill, N.E., & Roosa, M.W. (2000). A mediational model of the
impact of interparental conflict on child adjustment in a multiethnic, low-income
sample. Journal of Family Psychology, 14, 365-379.
Luecken, L.J, & Fabricius, W.V. (2003). Physical health vulnerability in adult children
from divorced and intact families. Journal of Psychosomatic Research, 55, 221-228.
**Note: A complete list of references can be found on the PAYS website