This study aimed to determine factors that influence fecal occult blood test performance in colorectal cancer screening. A random sample was selected of men and women ages 50 to 74 years of age who had been mailed a fecal occult blood testing kit in a screening program in fall 1986. One year after initial test mailing, sample group members (n = 504) were surveyed by telephone. Four months later, the survey sample received a second fecal occult blood test mailing. Multivariable analysis for subjects with validated past fecal occult blood test status (n = 322) revealed the past testing was positively associated with physician encouragement of screening, age, the belief that cancer is curable, perceived test efficacy, and strong intention to do testing. It also was discovered that persons who felt that they had little control over their health were more likely to have done past testing. Preliminary analysis of prospective adherence showed that the strongest statistically significant independent predictor was past test performance. Prospective adherence among past nontesters (n = 121) was associated with expressed commitment to do fecal occult blood testing and reported presence of colorectal cancer risk factors. Analysis of adherence among past testers (n = 201) revealed that belief in colorectal cancer curability and age were significant predictors. The findings reported here indicate that factors influencing adherence among past nontesters differ from those for past testers. Overall, these results suggest that to increase participation in colorectal cancer screening, physicians and other health professionals should (a) deliver educational messages that increase awareness of risk factors for colorectal cancer and curability of the disease, and (b) elicit from potential screenees a commitment to engage in recommended preventive behaviors. It may also be well to consider "tailoring" messages for past nontesters and past testers, respectively, by emphasizing colorectal cancer risk factors and highlighting curability.