Full text

This article is a primer on issues in designing, testing, and interpreting interaction or moderator effects in research on family psychology. The first section focuses on procedures for testing and interpreting simple effects and interactions, as well as common errors in testing moderators (e.g., testing differences among subgroup correlations, omitting components of products, and using median splits). The second section, devoted to difficulties in detecting interactions, covers such topics as statistical power, measurement error, distribution of variables, and mathematical constraints of ordinal interactions. The third section, devoted to design issues, focuses on recommendations such as including reliable measures, enhancing statistical power, and oversampling extreme scores. The topics covered should aid understanding of existing moderator research as well as improve future research on interaction effects.

There are many examples of theories in family psychology in which the association between two variables is hypothesized to be dependent on some other variable. Studies that evaluate these theories have evaluated interactions and moderator effects.  ^{[  1  ]}  For example, a perusal of articles published in the Journal of Family Psychology in 2003 (Volume 17) indicates that regression analyses have been used to study interaction effects in basic research evaluating outcome variables such as marital violence, children's health and adjustment, attitudes about physical punishment, adolescent mothers' psychological adjustment, and quality of parenting. In addition to basic research on family functioning, moderator effects have also been used to study predictors of differential responses to therapy interventions. The importance of moderator or interaction research was underscored by Cohen, Cohen, West, and Aiken (2003), according to whom “it is safe to say that the testing of interactions is at the very heart of theory testing in the social sciences” (p. 255). However, others (e.g., Judd & McClelland, 1989, 1998; Luce, 1995) warn that interactions are often signs of trouble, indicating scaling problems, model misspecifications, and other...