This paper reports the results of a survey of data monitoring practices of the 12 clinical co-operative groups in cancer sponsored by the National Cancer Institute in the United States. Most of these co-operative groups conduct a large number of active clinical trials in a wide variety of cancers. The number of active phase III trials in 1991 was 244 with nearly 24,000 patients entered on these trials. The survey instrument used in this paper was modified from one originally developed by Geller and Stylianou, supplemented by questions on the responsibilities of data monitoring committees based on Hawkins. Topics covered included general policies and procedures on the monitoring of clinical trials in these groups as well as specific questions on the responsibilities and operational procedures of formal data monitoring committees. Although formal statistical stopping rules are almost always employed in cancer trials in these co-operative groups, use of a formal data monitoring committee is a relatively new, but increasing practice. For a variety of economic and practical reasons, the members of data monitoring committees for cancer trials are rarely independent of the study participants.