The best-known model of r ational, repetitive choice in sociology is the Camilleri-Berger Model. The Discrimination Model is a recent competitor to the Camilleri-Berger Model and a thorough comparison is called for. In this paper, the two models are presented and compared in some detail with regard to the way they handle utility and marginal utility; the way they explain (or fail to explain) matching; their cognitive assumptions; their heuristic capabilities for application to social situations; and their predictive accuracy. The Discrimination Model comes out favorably on all counts. In addition, it is a deepet theory because it can specify the conditions under which the Camilleri-Berger Model will furnish accurate predictions.