Data from 197 randomly selected male adults were analyzed in order to examine the dimensionality of two of the concepts in the Health Promotion Model-perceived control of health and definition of health. The model, tested with LISREL, examines the impact of age on exercise with the key concepts modeled as intervening variables. The addition of multiple indicators resulted in a model that failed to meet the data constraints, thereby calling into question the putative dimensionality of perceived control of health and definition of health. The proportionality constraints within multiple-indicator structural equation models seem to provide a more stringent test of whether several indicators measure the same concept than the test provided by the internal consistency constraints implicit in factor analysis.