Objective-To ascertain attitudes to different methods of obtaining informed consent for randomised clinical trials (RCTs). Design Structured interviews with members of the public, medical secretaries and medical students. Setting The public were approached in a variety of public places. Medical secretaries and students were approached in their place of work. Subjects-Fifty members of the public, 25 secretaries and 25 students. Main outcome measures Views on RCTs were elicited, with particular emphasis on how subjects thought the concept of randomisation should be explained. Each participant was presented with descriptions ofproposed clinical trials and asked to select his or her preference from a range of options. Results-Written information was preferred over verbal information in 91 % of replies. Most respondents (860o) would prefer to sign a consent form. Of the seven statements explaining randomisation, a significant difference was found in favour ofexplanations that were less explicit about the play of chance (ANOVA; p=0 0004). Eighty-three per cent of participants thought that randomised trials were morally acceptable when there was no prior medical preference between treatments. However, over half (55%) thought they would find it upsetting to be offered entry in such a trial and a quarter thought the outcome of treatment might be adversely affected. Conclusions Our results offer some support for the idea that "economy with truth" is less unsettling than a frank description of the stark reality of what randomisation means. It is a matter of debate as to whether, if we are correct, autonomy should have precedence over beneficence.