Four hundred and seventy-seven professionals attended thirteen group relations conferences. Conferences varied across three dimensions: context, including sponsorship and history: design, involving duration, intensity (residential setting) and complexity; and linkages, the social and authority ties between members and staff. Three month follow-up questionnaires were collected from sixty percent of participants. Significantly more self-assessed learning was reported by those who attended the residential than the non-residential conferences. The results, from a large diversified sample, suggest that a combination of training in a residential setting, strong institutional sponsorship and pre-existing authority and social linkages between members and staff resulted in the most reported learning. Group relations conferences provide unique learning opportunities for mental health professionals, (Correa et al. 1981) and have been increasingly used in the United States and Europe during the last twenty years. Despite this, there is little research evaluating the outcomes of such training in terms of member learning or the differential effectiveness of the alternative forms of conferences currently available. Conferences vary along three major dimensions: a) context, including the institutional sponsorship as well as the history of previous conferences held at the same site, b) design, including the duration, intensity and number of events which make up the conference, and c) linkages, the social and authority relations among members and staff. A review of the first decade of group relations work in the United States concluded that the characteristics of a conference, including the setting have an important impact on member learning, (Klein 1978).(ABSTRACT TRUNCATED AT 250 WORDS)