A theoretical model of publication productivity was studied to see if it was applicable to professors in various types of institutions. The impact of several correlates of publications within doctoral universities and comprehensive colleges and universities were evaluated, thus addressing several questions about the extent to which organizational factors and individual differences in preparation, prior activities, and perceptions influence faculty publication. Data for the study were drawn from "Faculty at Work," a national survey conducted by the National Center for Research to Improve Postsecondary Teaching and Learning in 1987 and 1988. In addition to standard socio-demographic and career indicators, the survey instrument measured faculty perceptions of their colleges and universities as well as their own competence and efficacy, values and beliefs, and psychological dispositions. Measures of career, self-knowledge, social knowledge, behavior, and environmental response were created for the study. Results indicate that the proposed model for understanding faculty publication differences is a useful one. Other findings include: past role performance does influence current productivity; and consistent grant and publication involvement, larger commitment of time to research, and commitment to teaching are important correlates of publication rate and should be included in any future development of a causal model. Tables and figures are appended. Contains 67 references. (SM)