The Big Tent: Integrating Macro Models for Intergroup Association with Experimental Data on Exchange Relations in a Minimal Group Setting

Purpose: This chapter aims: (a) to extend biased net models of homophily to complete networks; (b) to extend the scope of application of these models to processes of social exchange in a small-group laboratory setting; and (c) to link the theoretical model of attraction and repulsion with a standard statistical model of logistic regression as a way of estimating and evaluating the model.

Methods: We discuss the logic of biased net theory and show how it leads to formal mathematical models of tie formation and tie renewal under mechanisms of attraction and repulsion. We then estimate key theoretical parameters in the models by means of logistic regression.

Findings: The estimated effects of homophily in our models are moderate in strength, weaker than corresponding reciprocity effect, and processes of tie formation and tie renewal are driven more by considerations of direct reciprocity than group membership. Under attraction, homophily effects are stronger for tie renewal than tie formation. Under repulsion, the opposite holds.

Limitations: Participants in our study are divided into two groups based on a criterion that is likely to have been too weak to induce strong group identity. Measures that enhance the sense of group identity need to be introduced in future studies.