The National Cancer Institute has recently changed its approach and has substituted summary-of-evidence statements for specific recommendations for breast cancer screening in women 40-50 years of age, leaving these women with a greater share of decision-making responsibility. To make an informed decision about breast cancer screening, women need accurate information about their breast cancer risk and the benefit of screening. Although it has been suggested that women younger than 50 years of age overestimate this risk and benefit, their estimates have not been quantified. The purpose of this study was to determine how women 40-50 years of age perceive their risk of breast cancer and the effectiveness of screening and how these perceptions compare with estimates derived from epidemiologic studies of breast cancer incidence and randomized clinical trials of screening. We mailed a questionnaire to 200 women, identified through the computerized medical records of Dartmouth-Hitchock Medical Center, who were between 40 and 50 years of age and had no history of breast cancer. Each woman was asked about her risk factors for breast cancer and asked to estimate her probabilities of developing breast cancer and dying of it within 10 years, with and without screening. The women's answers were compared with individual probabilities derived from the Gail et al. model, age-specific probabilities of developing and dying of breast cancer in the United States, and the results of randomized clinical trials of screening. The mailed questionnaire was completed and returned by 145 (73%) of the 200 women. Respondents over-estimated their probability of dying of breast cancer within 10 years by more than 20-fold (median, 22.3; interquartile range, 11.1-74.2). Assuming a 10% relative risk reduction from screening, respondents overestimated the relative risk reduction by sixfold (median, 6.0; interquartile range, 5.0-7.5) and the absolute risk reduction more than 100-fold (median, 127.5; interquartile range, 47.1-399.6). The median perceived estimate of absolute risk reduction was 6.0 breast cancer deaths per 100 women; the median calculated estimate was only 0.04 per 100 women. These findings suggest that many women younger than 50 years of age substantially overestimate their breast cancer risk and the effectiveness of screening. A balanced presentation of information about breast cancer risk and screening effectiveness may improve decision making for women younger than 50 years of age and reduce their anxiety about breast cancer, regardless of whether they choose to be screened.