Three seminars for nonpsychiatric physicians were compared. Attendance in the first was discontinuous and participation uneven. The second had better attendance and participation; however, it was characterized by a marked tendency to restrict open disagreement. The third seminar had regular attendance and optimal participation; its members disagreed freely without being judged disagreeable. The content, structure, processes, and pluralistic tolerance of the seminars were analyzed. Differences were found in membership, cohesiveness of practice groups, patterns of conflict resolution, intergroup relations, and community forces. Additional determinants included the influence of strangers, frequency of meetings, and seminar settings. Lessons are distilled that relate to the organization, processes, and goals of such seminars. A "process model" is emphasized.