Although it is commonly held that environmental, social, and psychological factors influence health, specific causal models of these influences are rarely tested directly. Methods of structural analysis were applied to the problem of the relationships among variables thought to influence the health status of college students. Data were collected from 365 students on personal characteristics, physical features of their college residences, the perceived social climate of the residences, psychological distress, and health status. Canonical correlation analyses reduced the data to five indicators representing four factors: dormitory size, the degree of perceived influence and social support in the dormitory, and reports of academic pressure and physical symptoms. Four sets of structured equations, representing alternative causal models of the interrelationships among those variables, were tested for goodness-of-fit to the observed data. The first three models were based on the assumption of unidirectional causation. The best-fitting model was Model 4 which included a direct link from perceived influence to physical symptoms and an added reciprocal causal link between psychological distress and physical symptoms. Conclusions drawn about the relationships among the variables examined in Model 4 emphasized the important role played by perceptions of personal control in self-evaluations of health status. (Author/NRB)