It has been hypothesized that there is more demand creation by doctors in high physician density areas. Empirical tests of actual utilization have had difficulties accounting for potential confounders, particularly possible differences in the health status of diverse populations. To eliminate that problem, physicians were presented with identical, hypothetical medical cases. Actual physician density was significantly and positively correlated with the aggressiveness of proposed treatment, holding constant other physician characteristics. This result lends support for the hypothesis of demand-inducement.