Objectives. To examine the effect of a patient's educational level on treatment selection for patients with newly diagnosed prostate cancer. A patient's educational level may impact a patient's understanding and perception of the risks and benefits of the treatment options for prostate cancer. Methods. We examined 3484 patients in CaPSURE with prostate cancer between 1992 and 2001. Chi-square and multinomial logistic regression analyses were performed to determine the role of education level in primary treatment received relative to other pretreatment predictors (age, race, insurance status, prostate cancer risk, comorbidity). Prostate cancer risk stratification was determined by serum prostate-specific antigen level and tumor stage and grade. Results. The mean patient age was 67.7 ± 8.3 years, and the mean prostate-specific antigen level was 13.0 ± 18.7 ng/mL. Of the 3484 patients, 16.7% had less than a high school education, 27.0% had completed high school or technical school, 19.5% had had some college, 18.0% had graduated from college, and 18.6% had had some graduate education. In bivariate analysis, the factors predictive of treatment selection were patient age, race, education, insurance status, risk group, and patient comorbidity (all P <0.001). In multinomial regression analysis, the factors predicting treatment received were age, race, cancer risk group, and comorbidity. For patients older than 75 years, those with a higher education level received more aggressive treatment (radiotherapy versus hormonal therapy) than did those with less education. Conclusions. Patient age, race, cancer risk group, comorbidity, and, for men older than 75 years, education level are the factors predictive of the primary treatment received by men with newly diagnosed prostate cancer. © 2003 Elsevier Inc.