The Women's Health Trial (WHT) was designed as a multicenter randomized trial to evaluate whether a low-fat diet could reduce the incidence of breast cancer. The dietary goal of the WHT was to reduce the percent of daily calories consumed as fat from a mean of 39% to 20%. Unlike other models of dietary change, the methods of meeting each subject's goal were chosen by her. Self-monitoring and self-correction guided dietary changes. The nutritional component of the WHT intervention focused on two major content areas: the knowledge base for decreasing dietary fat and nutritional skills needed to make the changes. Three behavioral models were used in the WHT intervention: social learning theory, cognitive-behavioral theory, and self-control theory. The dietary changes made by intervention women were dramatic and long lasting. Over 80% of intervention women had met their fat goal by the 3-month follow-up. These changes were maintained for the 3-year follow-up period. There were no significant differences between numbers of intervention (16%) and control women (14%) who reported difficulties caused by current dietary habits at the 1-year follow-up, suggesting that consuming a low-fat diet was not problematic. At the 1-year follow-up, intervention subjects reported significantly more positive reactions of others to their eating habits than did control subjects. A description of the design and results of the WHT dietary intervention will be useful for scientists involved in many different aspects of health behavior change.