Associations between level of education and cancer risk is well supported by scientific evidence, but previous studies could only partly adjust for relevant confounding factors. In this article, we examined how risk of cancer varies with level of education and identified factors that explain this variation using data from a prospective cohort study, including 93,638 Norwegian women who responded to an extensive questionnaire in 1991/1992 or 1996/1997. A total of 3,259 incident primary invasive cancer cases were diagnosed during follow-up, which ended in December 2001. We used Cox proportional hazards model to calculate relative risks (RR) with 95% confidence intervals (95% CI). Besides a similar overall risk of female cancers by level of education, we observed differing risks between educational groups for cancers of the lung, breast, cervix, kidney, and skin melanoma. Women with >16 years of education had an increased risk of breast cancer (RR, 1.46; 95% CI, 1.19-1.79) and a decreased risk of lung cancer (RR, 0.30; 95% CI, 0.13-0.70) and cervical cancer (RR, 0.38; 95% CI, 0.17-0.85) compared with the lowest educated women (7-9 years). The middle educated (13-16 years) had the lowest risk of kidney cancer (RR, 0.24; 95% CI, 0.08-0.71), whereas the risk of skin melanoma was highest among women with 10 to 12 years of education (RR, 1.53; 95% CI, 1.05-2.24) compared with the lowest educated women. After multivariate adjustment for potential confounders related to level of education, the variation in cancer risk according to educational levels declined into nonsignificance for all these sites.