Testing the health belief model: LISREL analysis of alternative models of causal relationships between health beliefs and preventive dental behaviour

This study tests the causal relationships postulated in the Health Belief Model by using a causal modeling technique, the LISREL system. It analyzes causal relationships between American women's health beliefs and their preventive dental visits. Data are from a 1980 nationalfamily dental survey. Various causal models consistent with previous theory and research are specified and tested. A model that best fits the data shows that, in contrast to the one-way causal relationships from health beliefs to dental visits postulated in the originalformulation of the Health Belief Model, four types of causal relationships exist. The findings are compared with previous studies. The importance of causal analysis is discussed. We conclude that a number of requirements must be completed before our final model can be accepted: (I) replication of this model on a new set of data; (2) longitudinal-prospective analyses exhibiting the hypothesized relationships through time; and (3) experimental testing by manipulation of one or more independent variables in experimental-clinical trials.