A General Framework for the Statistical Analysis of Sequential Dyadic Interaction Data

Abstract

Recent interest in sequential dyadic interactions has motivated researchers to develop methods appropriate for the analysis of such data. After briefly reviewing a series of methodological papers focusing on the analysis of discrete-valued observations, we present a general framework for studying many substantive effects, including dominance and autodependencies, in social interactions measured on dyads. We show how this framework allows a researcher to study dyadic interactions measured at two or more time points on one or more relations. The methods described here are general enough to permit the simultaneous analysis of the sequential relational variables and attribute variables (such as sex of actors or emotional status of the dyad) recorded on either the dyad or the actors.