Background. Few current data are available regarding factors associated with participation in cancer screening examinations in the general population. Methods. To identify factors associated with participation in cancer screening examinations, random population samples of 25- to 74-year-old men and women in six various-sized communities in three upper-Midwestern states (n = 4,915) were surveyed in 1987-1989. Multivariate-adjusted means were calculated and compared using analysis of covariance. Results. Statistically significant (P < 0.05) strong predictors (other than age and sex) of ever having had a specific cancer screening test were as follows (the numbers in parentheses following each listed association are the absolute maximum differences in mean proportions among the levels of the predictors): (1) rectal examination: higher education (14%); (2) fecal occult blood testing: higher education (6%) and never smoker (5%); (3) sigmoidoscopy: higher income (7%) and higher education (6%); (4) clinical breast exam: higher education (6%); and (5) mammography: higher income (25%), higher education (8%), and a positive family history of breast cancer (7%). There were no strong predictors (out of nine) of ever having had a Papanicolaou smear or a breast self-examination. Conclusions. The largest differences among the population for participation in cancer screening examinations involves income and the two most expensive cancer screening tests: higher income is a strong predictor of having a mammogram and, to a lesser extent, of having a sigmoidoscopy. The most consistent predictor of participation in cancer screening examinations across all cancer screening tests is education: higher education is a predictor of having each kind of cancer screening test.