We develop a simple methodology to estimate the returns to education despite heterogeneous labor/leisure preferences. The labor supply behavior of doctors and physician assistants is consistent with people choosing between the two careers based on differing tastes for leisure.