Although the majority of school districts in the United States assign students to schools on the basis of geographic location, there is increasing interest from parents and policy makers in school choice programs. These programs allow parents (and children) to choose their school. In many cases, when students opt to attend a nonlocal school, the parent is responsible for transportation. This creates a trade-off between school quality and travel expenditure. This paper estimates the value of school quality in a random utility model framework, using data on rank-ordered preferences submitted in a school choice program in Garland, Texas. I find that a standard deviation of high school quality is valued between $495 and $783, substantially lower than hedonic estimates that include noneducational benefits associated with good schools. In addition to estimating the average value of school quality, this approach can also be used to estimate the value of school quality for demographic groups or for individual students.