DYADS

This program was developed for the analysis of dyadic data or, to be more precise, data that are a mix of individual scores and scores of dyads between individuals. An example could be an analysis of the difference in extraversion between friends as compared to that of non-friends. One could perform an F-test, where the cases are all possible dyads in a class. The dyadic variable friends / non-friends would act as a fixed factor and the difference in extraversion as the dependent variable. However, there is a problem, usually indicated by the term multiple membership: the cases are not independent, since one single individual (with his specific score on the extraversion scale) can be a member of several dyads, and even worse: not all individuals feature in the same number of dyads. As a consequence, the distribution of the F-statistic may deviate from the F-distribution as it is assumed by a regular F-test.

Downloadable files:

Documentation: dyads.pdf (pdf, 213 kB) (213 Kb)
Program: dyads.zip (zip, 694 kB) (695 Kb)
Example: dyadsexample.zip (zip, 4 kB) (4 Kb)