A Method and Metric for Assessing Similarity among Dyads

Abstract

Ascertaining substantive similarities between members of dyads is often difficult to accomplish. Assessments of similarity between dyad members are frequently made with statistical tools that are inappropriate for that purpose. In this article, problems in defining and measuring similarity with dyadic data are reviewed and a simple alternative is proposed. The method of planned comparisons between related and random dyads is illustrated with the use of sociopolitical attitude data from three-generation families. Results show both greater and less intergenerational attitude similarity in certain domains than would have been discerned on the basis of correlations or difference scores alone.