This paper presents the results of a revealed-choice experiment testing the theoretical predictions of a model of a mixed system of public and private finance. In the context of a mixed system of health care finance, we investigate behavioural responses to changing the public sector allocation rule (needs-based vs. random), the supply of health care resources, and the size of the public health care budget on the following outcomes: the equilibrium market price for health care resources, the number of individuals who purchase private insurance, the probability of health treatment in the public system for those without private insurance, the health status of individuals left untreated, and the incomes of individuals who receive treatment. Our findings are generally consistent with the predictions of the theoretical model, although individuals consistently exhibit greater willingnessesto-pay for private insurance than predicted resulting in a larger than predicted amount of private insurance being purchased. A commonly used risk-aversion measure only partially explains this observed deviation. We also find that, relative to a system of public financing only, a mixed system of health care finance results in higher health care prices and sicker, poorer people being left untreated.