This paper investigates how the Fresh Fruits and Vegetable Program (FFVP), a national program which provides funding for the distribution of free fresh fruits and vegetables to students in participating schools, affects childhood obesity. Using a panel data set, we combine matching methodology and difference-in-differences analysis to estimate the effect of the FFVP on childhood obesity outcomes. The results suggest that estimates of the FFVP effect are very sensitive to use of different matching methods. With the use of a stricter matching method, the estimate of the FFVP effect is negative and significant, indicating that the program reduces children’s body mass index. Less strict matching methods yield opposite results.