As neighborhood context is increasingly recognized as an important predictor of health outcomes and health behaviors, this analysis sought to determine the relationship between neighborhood-level socioeconomic status (SES) and regular mammography screening behavior. One thousand four hundred fifty-one women ages 40 to 79 years who obtained an "index" screening mammogram at one of five urban hospitals in Connecticut between October 1996 and January 1998 were enrolled in this prospective study. The logistic regression analysis includes the 1,229 women [484 African-American (39%) and 745 White (61%)] who completed telephone interviews at baseline and follow-up (average 29.4 months later) and for whom the study outcome, nonadherence to age-specific mammography screening guidelines, was ascertained. Neighborhood-level SES was determined using 1990 census tract information. Neighborhood-level SES variables (quartiles) were associated with nonadherence for African-American women [neighborhood-level education and composite socioeconomic position index (SEP Index)] and White women (neighborhood-level crowding and neighborhood-level assets). Using race-specific categorizations reflective of individual-level SES distributions, the SEP Index and neighborhood-level education were associated with nonadherence to mammography screening guidelines for African-American women (marginally significant for White women), independent of individual-level SES and other known predictors of mammography screening use [African-American women: SEP Index odds ratio (OR), 3.55; 95% confidence interval (95% CI), 1.33-9.51; neighborhood-level education OR, 3.21; 95% CI, 1.25-8.26; White women: SEP Index OR, 2.13; 95% CI, 0.97-4.67; neighborhood-level education OR, 2.31; 95% CI, 0.93-5.76]. The results of this analysis underscore the importance of examining neighborhood social context as well as individual factors in the study of mammography screening behavior.