In answer to a questionnaire administered in 1961, 90% of responding physicians indicated a preference for not telling a cancer patient his diagnosis. To assess attitudinal changes, the same questionnaire was submitted to 699 university-hospital medical staff. Of 264 respondents, 97% indicated a preference for telling a cancer patient his diagnosis--a complete reversal of attitude. As in 1961, clinical experience was the major policy determinant, but the 1977 population emphasized the influence of medical school and hospital training. Our respondents indicated less likelihood that they would change their present policy or be swayed by research. Clinical experience was the determining factor in shaping two opposite policies. Physicians are still basing their policies on emotion-laden personal conviction rather than the outcome of properly designed scientific studies.