Although there have been numerous empirical tests of various theories of coalition formation, many of these data may be equivocal because the relative validity of a given theory may vary with the procedure used to test it. Accordingly, 3 different procedures, varying in communication restrictions and information availability, were used to test 3 theories of coalition formation: minimum resource theory, bargaining theory, and the weighted probability model. Group size (3, 4, and 5 members) was also varied. Based on data from 372 male undergraduates, the results indicate that group size and experimental procedure have significant effects on the validity of the 3 theories. The predictions of minimum resource theory were least accurate, whereas the predictions of the bargaining theory were most accurate. (21 ref) (PsycINFO Database Record (c) 2006 APA, all rights reserved). © 1978 American Psychological Association.