The study sought to modify comparative optimism about colorectal cancer in a community sample using a method of providing risk information found to be effective in a laboratory setting. The 3185 adults from General Practice lists were randomized to three groups: (1) control--no information; (2) risk information leaflet; (3) risk and screening information leaflet. Significant comparative optimism and high numeric estimates of absolute risk were found. Risk factor information did not reduce optimistic beliefs nor modify estimates of risk. Interest in screening was high overall and not influenced by the information. Comparatively optimistic risk perceptions appear resistant to change in community settings.