| Sumario: | The writer reviews three analytic approaches for treating the dyad as the unit of analysis. She argues that these approaches are useful for the specific, but quite common, situation in which researchers have information from or about two members of a dyad. She outlines the following three approaches: intraclass correlations as a measure of similarity, repeated measures analysis of variance, and hierarchical linear modeling. She then uses all three approaches to analyze the same data taken from the first wave of a short-term longitudinal study of 197 families with adolescent children. Her analyses reveal that although the three analytical approaches were designed to answer slightly different questions, they yielded a similar pattern of results, with a number of important differences. Details of this pattern of results are provided.
|
|---|