Type of Group and the Probability of Initiating Acts

Three hypotheses about the initiation of acts in group interaction are substantiated. The hypotheses specify that (1) individuals have constant tendencies to initiate acts in a given type of group; and (2) the proportion of acts initiated by an individual in any particular group equals his constant normalized by the sum of the constants of all group participants. Separate session estimates of individual constants for two sets of experiments on stranger triads show agreement of .894 and .965. For any given subject, (3) predictions from interaction with strangers will not hold for interaction with family members. Further analysis suggests that family solidarity underlies the observed changes, setting families apart from the typical ad hoc laboratory groups.