The quality of participation in the performance of clinical trials of university members and community affiliates of the Northern California Oncology Group is evaluated and compared. The data, based on 738 patients on 33 protocols, were collected during a one year period, July 1, 1980--June 30, 1981. The comparisons are made on three types of criteria: accrual distribution, with respect to study phase and modality multiplicity; data quality, generally reflecting protocol adherence; and data completeness. The performance of the community affiliates was found to equal or surpass that of the university members in most measures. Therefore, it is concluded that the community affiliates are functioning as full and valuable participants in the Northern California Oncology Group.