Patients who agree and those who refuse clinical trial entry may differ in attitudes towards decision control and the benefits associated with the trial arms. These differences, if they exist, have implications for the process of obtaining informed consent and for the generalization of the results of a clinical trial. This paper describes the development and initial application of methods designed to detect such differences. Developmental work involved creating an inventory of instruments designed to determine patients' attitudes towards participating in treatment decision making, permitting random selection of treatment, and undertaking the risks and benefits associated with the various treatments in a trial. Initial application involved modifying these instruments in terms of an actual chemotherapeutic trial for colonic adenocarcinoma, seeking responses to these measures from 60 non-eligible colorectal cancer patients, then determining whether those who would agree to trial entry differed systematically on these measures from those who indicated that they would refuse such a trial. Twenty-five of the respondents reported that, if faced with the actual decision, they would agree to trial entry: 35 would refuse. Refusers demanded more participation in decision making (Chi-square; P = 0.01) and a greater increment in treatment benefit (t-test; P = 0.0001). Twenty-two of the 35 refusers reported aversion to randomization as their primary reason for trial refusal. Since their particular content can be modified, these measures may be applicable to all clinical trials. They could be used to study the reasons patients accept or refuse trial entry and to determine if agreer-refuser attitude differences undermine the generalizability of a trials results.