A guide for the estimation of gender and sexual orientation effects in dyadic data: An actor-partner interdependence model approach

The study of gender differences is a pervasive topic in relationship science. However, there are several neglected issues in this area that require special care and attention. First, there is not just one gender effect but rather three gender effects: gender of the respondent, gender of the partner, and the gender of respondent by gender of the partner interaction. To separate these three effects, the dyadic research design should ideally have three different types of dyads: male-female, male-male, and female-female. Second, the analysis of gender differences in relational studies could benefit from the application of recent advances in the analysis of dyadic data, most notably the Actor-Partner Interdependence Model. Third, relationship researchers need to consider the confounding, mediating, and moderating effects of demographic variables. We use the American Couples (Blumstein & Schwartz, 1983) data set to illustrate these points.