ICPR Abstracts: Session 18

---

Session 18: Symposium

The Use of Individual Versus Couple Data in Close 
Relationships Research: Theoretical and 
Methodological Issues


Assessing Power Over Behaviors Using Individual and 
Couple- Derived Data

Christopher R. Agnew
Purdue University
  
Traditional social psychological models of behavioral 
prediction (e.g., the theories of reasoned action and 
planned behavior) hold that the prediction of behavior is 
maximized by obtaining predictor variables that match the 
criterion in terms of action, target, time, and context.  This 
paper argues that the nature of the behavior itself must 
also be considered when attempting to maximize the 
prediction of behavior.  Specifically, behaviors that require 
the coaction of another person (interdependent behaviors) 
should be best predicted by data obtained from the co-
acting dyad.  In contrast, behaviors that require action 
from only one person (independent behaviors) are 
sufficiently predicted from individual data.  In presenting 
the results from a longitudinal study of college couples, 
this paper highlights the use of a simple methodology 
which allows the researcher to assess relative power over 
behaviors and behavioral domains as well as determine 
whose data (male, female, or both couple members in 
some combination) provide the most complete prediction 
of different behaviors. 


Estimating Mutual Influence in Couples Research:
An Extension of the Kraemer-Jacklin Method

David A. Kenny, University of Connecticut
William L. Cook, Maine Medical Center

One model that elaborates how members of a couple 
become interdependent is the "partner effect model."  
Consider the effect of satisfaction on commitment in 
relationships.  A woman's commitment is a function of 
both her and his satisfaction, and a man's commitment is a 
function of his and her satisfaction. Because each person's 
commitment is influenced by characteristics of the partner, 
their commitment levels will be correlated or 
nonindependent.  Kraemer and Jacklin (1979) proposed a 
data analysis strategy for such data.  Recently, Kenny 
(1996) has developed an extension of the Kraemer-Jacklin 
method.  We apply these modifications to several data 
sets.  For instance, we examine influence tactics in 
heterosexual and homosexual couples.  We can test 
whether the differences between the two types of couples 
can be explained by gender.  We hope to show the utility 
of the estimation of partner effects in couples research.



Relationship Health?: The Investment Model and 
Predictions of Self-Report Versus Couple Interaction 
Relationship Health Measures

Stephen M. Drigotas, Cinzia Pacione, 
and Robert B. Hampson, 
Southern Methodist University

The question of what constitutes relationship health is 
investigated in a cross sectional study of dating 
relationships. Specifically, the individual-level dyadic 
adjustment scale (Spanier, 1976) and the couple-level 
Relationship Competence Model (Beavers, 1985) are 
predicted to tap different elements of relationship health 
as predicted by Investment Model (Rusbult, 1983) 
constructs.  Results indicate that while the investment 
model powerfully predicts both relationship health 
measures, the two health measures are not related to each 
other.  The findings suggest the need for a comprehensive 
relationship health measure that incorporates both 
individual and couple interaction aspects of health. 


The Convergence of Husbands' and Wives' 
Childbearing Motivations During Marriage

Warren Miller and David Pasta
Transnational Family Research Institute

There is relatively little systematic empirical knowledge 
about how husbands and wives affect each others' 
childbearing motivations in the course of marriage. We 
collected data from 324 married couples at five points in 
time during a five year period.  Using a hierarchical linear 
model analysis of several childbearing motivational 
measures, we examine first the amount of couple 
convergence over time and second how both the initial 
(time- invariant) characteristics of the husband, wife, and 
couple and the follow-up (time-varying) events 
experienced by them affect this convergence. The results 
are discussed in terms of couple influence processes.

Discussant
Mark Attridge
United HealthCare Corporation
Golden Valley, MN