Purpose: To determine the research productivity of faculty in family medicine departments at U.S. medical schools, as well as the individual and environmental characteristics and prior socializing experiences predictive of research productivity. Method: In 2000, a 43-item questionnaire was mailed to 796 faculty to obtain descriptive data toward formulating a conceptual model of the research productivity of family medicine faculty. Prior to model construction and testing through full-model regression, the model's factors were reduced through factor analysis. Results: A total of 474 questionnaires (63%) were returned. Eighty-percent of respondents spent a half-day or less per week on research; on average they produced less than one scholarly product per year. Few had research experience, nor could identify a research agenda or current research project. Mixed messages were perceived related to research, both at institutional and disciplinary levels. In testing a conceptual model, psychological and cognitive characteristics were most predictive of research productivity, along with time spent on research. Psychological and cognitive factors included enhancing research skills, establishing a definable research agenda, fostering research networks, having multiple research projects underway, maintaining in-depth knowledge of a research area, and clearly understanding research expectations for promotion and tenure. Conclusion: The clinical and academic demands on family medicine faculty reduce the likelihood that they will engage in research. These demands prevent the development of a critical mass to provide mentorship and networking necessary for research productivity. Resources are needed to recruit faculty with an interest in research and to provide faculty development in research skills, mentorship, and networking.