Preferences for schools are important determinants of equitable access to high-quality education, effects of expanded choice on school improvement and school choice mechanism design. Standard methods for estimating consumer preferences are not applicable in education markets because students do not always get their first choice school. This review describes recently developed methods for using rich data from a school choice mechanism to estimate student preferences. Our objectives are to present a unifying framework for these methods and to help applied researchers decide which techniques to use. After laying out methodological issues, we provide an overview of empirical results obtained using these models and discuss some open questions.