Dyadic data analysis: New methods for social psychology

New methods for the analysis of dyadic data are presented by using data of exchangeable and distinguishable dyads. Based on a correlational approach, the pairwise dyadic model and the pairwise latent variable model by Griffin and Gonzalez (1995) provide estimates of dyadic non-independence and the u nderlying dyadic-level and individual-level correlations between attributes of dyad members. In contrast, the partner-effect model developed by Kenny (1996) is based on a regression model and provides estimates of actor, part ner, and reciprocity effects. These effects can be interpreted as independe nt influences between dyad members. For the exchangeable case, dyads of 83 older MZ and DZ twins were used to demonstrate the interdependence of attac hment styles. For the distinguishable case, dyads of 214 younger heterosexu al couples were used to show interdependence between attachment styles of r omantic partners.