Research on mutual-aid groups has begun to examine reasons for joining and outcomes, but few investigations have focused on the processes of group development or interaction. The applicability of a therapy-group development model to student mutual-aid groups was examined to determine the extent to which specified formal group structure and behavioral expectations affect the ability of group members to perform constructively and cohesively. To test this model, participants were randomly assigned to groups of high or low formal structure for four weekly sessions to assess the performance of group behaviors, "ownership" of group functioning, and group cohesion. The results show that the processes of group development in mutual-aid groups parallel those in therapy groups. Formal structuring by the group leader can have both positive and negative effects, and must be used judiciously. Future research may be able to identify generic processes characteristic of different types of change-induction groups. (Author/CS)