The Design and Analysis of Data from Dyads and Groups

Abstract

In this chapter we begin with terminology and definitions that are essential for understanding dyadic and group data. We then discuss a statistical technique that has become indispensable for the analysis of such data—multilevel modeling (MLM). Next we provide a detailed discussion of dyadic data structures in which both members are measured on the same set of variables. For dyadic data, the discussion centers on the very popular Actor-Partner Interdependence Model (APIM). We then turn our attention to group data. We discuss first the extension of the APIM to the study of groups. We then discuss the analysis of data from groups in which each person rates or interacts with other members of the group, and finally we discuss the analysis of intergroup data. The final data structure, one likely unfamiliar to most readers but important, represents a blend of dyadic and group structures, the one-with-many design. In each section we emphasize that a person's response depends not only on that person but also on the partners with whom he or she interacts. We avoid presenting extensive computational and computer syntax details, but we do cite the sources that present these details. (PsycINFO Database Record (c) 2015 APA, all rights reserved). (chapter)