The purpose of this study was to identify the relationship of selected predisposing and enabling characteristics of women > or = 50 years of age to mammography utilization. Andersen and Aday's theoretical model for health services utilization guided data collection. Data were collected from a convenience sample of 161 women members of four urban churches, using a mailed survey. Results showed that 81% reported at least one mammogram and 24% had followed mammography guidelines for the preceding 3 years. Results of logistic regression analyses with variables having a bivariate significance of p < or = 0.01 showed that higher income was associated with both ever having a mammogram and adherence. Willingness to pay > $50 out of pocket for a mammogram was significant for 3 year adherence. Additionally, the sociodemographic variables of age and religion were associated with adherence, whereas a college education was highly significant (odds ratio = 13.78) for ever having a mammogram. Having a regular place for health care and having yearly Papanicolaou tests were associated with ever having a mammogram, but not adherence. Finally, intending to get a mammogram was associated with ever having a mammogram. In this study, belief and knowledge variables showed no association with utilization, and social influence had bivariate significance only for ever having a mammogram. This study suggests the importance of addressing economic and health-care delivery system factors to promote increased mammography utilization, particularly for older women.