Because life-style patterns affect many cancer risks, research on health-risk behavior and behavior change is critical to cancer prevention. This report recommends priorities for the next decade of psychosocial research on cancer prevention and detection. The leading priority for future research is to fill gaps in basic knowledge left by the rush to intervention and outcome studies. Such research must be theoretically driven and should aim to develop broad principles applicable to diverse health behaviors. Studies that include relevant process data on various stages of behavior change are considered more desirable than simple outcome studies. Epidemiologic investigations should be expanded to include measures of relevant behaviors, so that their impact on clinical outcomes might be established. More research is needed on lay perception of health risks and on individual and health-system barriers to effective cancer prevention and detection. Studies that address the needs of minority and underprivileged populations are crucial. Funding agencies' narrow categorical mandates impede interdisciplinary research on multiple risk factors and their interactions; these boundaries must be relaxed to promote such approaches. Funding agencies should also consider basic research as a long-term investment towards the development of effective interventions.