Key points are not available for this paper at this time.
This article deals with the application of regression models to observations on relations in dyads. I address three problems that have troubled past investigations of this sort: (1) how to incorporate measures of nodal characteristics into models explaining ties in dyads, (2) how to specify and estimate the effects of reciprocity, and (3) how to measure and control auto-correlation arising from dyads having common nodes. A model is developed that is designed to overcome these problems. An illustration is then provided via an analysis of data on interorganizational relations. The analysis yields empirical results that suggest the potential of these methods for testing certain social network theories.
James R. Lincoln (Wed,) studied this question.