model) and a model of how mother-daughter interactions cause daughters'
symptoms (the causal model). Of course, proof of reactive or causal connec-
tions between variables requires an experimental design and, although our
study had the advantage of yielding longitudinal data, the design was nonex-
perimental. When research findings from a variety of studies yield consistent
empirical relations between family behaviors and internalizing symptoms, it
will be crucial to conduct experimental intervention research with families of
depressed adolescents to more adequately examine these reactive and causal
models. In addition, our findings raise questions regarding the sequencing of
mother-daughter interactions. For example, the submissive behavior of inter-
nalizing daughters may be a response to mothers' high levels of conflictual
behaviors or, daughters' submissive behaviors may be an immediate response
of daughters to their own perceived expressions of conflict -- self-monitoring
attempts to dilute the impact of their assertive interactions. These questions
are best addressed with conditional probability analyses in which the sequen-
tial nature of mother-daughter interactions could be examined.

Variables Not Included in the Model. Our analyses were restricted to
examining the relation of participants' subjective understanding of interper-
sonal behaviors to internalizing symptoms. In an earlier article (Powers et
al., 1994), we proposed that subjective understanding would be moderated
by a wide range of individual differences variables (such as level of social-
cognitive reasoning or temperament) and environmental variables (such as
social class or parental symptomatology). We expect to examine the role of
these moderating variables in subsequent analyses. Although we expect mod-
erating variables to strengthen the models presented in this chapter, it is
important to note that our findings for the simple models examined here
were quite strong, with results for the reactive model explaining from 14%
to 37% of the variance in behavior and results for the causal model explaining
from 28% to 50% of the variance in internalizing symptoms.

This research was funded by the National Science Foundation. We thank
Kristen Pollack and Virginia Wright for their coordination of the data col-
lection and management for this project and for their participation in dis-
cussions regarding the ideas presented in this chapter.

Abramson L. Y., Seligman M. E. P., & Teasdale J. (1978). "Learned helplessness in humans:
Critique and reformulation". Journal of Abnormal Psychology, 87, 49-74.

Achenbach T. M. (1991). Manual for the Youth Self-Report and 1991 profile. Burlington, VT:
University of Vermont.

-275-