The relation of education and income to the estrogen receptor status of primary breast tumors was studied using two patient groups. The first consisted of 887 women from northeastern Ohio who were diagnosed between late 1974 and 1985 and entered into either of two prospective breast cancer clinical trials. All estrogen receptor values were determined by one laboratory. Through loglinear regression, patterns of association were studied among patient characteristics such as age, race/ethnicity, menopausal status, census tract indices of poverty and education, tumor diameter, stage of disease, obesity, and height and weight at the time of diagnosis. In this group of patients, estrogen receptor status was directly related to age, poverty, educational level, and tumor size. Younger women (odds ratio (OR) = 1.57, 95% confidence interval (Cl) 1.14–2.17), women from census tracts with greater poverty (OR = 1.77, 95% Cl 1.28–2.44) or less education (OR = 1.98, 95% Cl 1.43–2.73), and women with larger tumors (OR = 0.67, 95% Cl 0.48–0.92) were more likely to have estrogen receptor-negative tumors at the time of diagnosis of primary breast cancer. The second group consisted of 604 patients from northeastern Ohio whose tumors were diagnosed between 1986 and mid-1992 at University Hospitals of Cleveland. All estrogen receptor values were determined by one laboratory. Results from this second group of patients confirmed those from the first. This association of estrogen receptornegative tumors with low economic and educational levels provides a potential explanation for the poor prognosis of these women