A causal model based on Pender's (1987) Health Promotion Model was tested to evaluate Pender's hypothesis that demographic and biological characteristics affect health-promoting behaviors indirectly through three mediating cognitive-perceptual variables. A sample of 3,025 noninstitutionalized adults completed a telephone survey from which indicators of the conceptual variables were selected. Initial tests of the causal model using the LISREL 7 program indicated that the basic model did not fit the data. Therefore, the model was modified so that the exogenous variables—sex, age, income, marital status, education, and body mass index—had direct effects on select health-promoting behaviors. Further, the variables of self-actualization and interpersonal support were required to share common indicators as were health responsibility and interpersonal support. Though the modified model fit the data, little of the variance in health-promoting behaviors was explained, since all significant effects were weak.