Moderator Variables in Social Work Research

Moderator variables are "third variables" that affect the size or nature of the relationship between an independent and dependent variable. Because of this role, such expressions as specification, contingency, conditional, and qualification have been used in reference to moderator variables. In classical experimental research in psychology, moderator variables have been referred to as involved in "interaction effects" with the independent variable. Considering their important role in specifying under what conditions and for what subgroups variables will have significant effects, surprisingly little attention has been awarded moderator variables in social work research. This paper reviews the history of moderator variable applications in the related disciplines and exemplifies recent applications in the study of gender discrimination and research on stress, strain, and social support. Special attention is given to the comparison of correlations and slopes, factorial analysis of variance, and moderated regression analysis, since these are the dominant analytic strategies for detection and valuation of moderated effects. Less prominent procedures and refinements are briefly presented, as is the important role of moderator variable in theory explication and causal inference.