Some methodological issues in epidemiological studies of fat and cancer.

Three types of approach to the study of the possible relationship between fat and cancer are discussed, namely, correlational, case-control, and cohort studies. The advantages and limitations of each method are analyzed, and it is emphasized that a necessary prerequisite to direct observational studies in humans is the development of a satisfactory and uniform method of a diet assessment. Despite the problems, epidemiological research in this area offers considerable potential for cancer prevention.