The adoption of market theory as a guiding principle of education policy increased the need for assessments of school performance that families could use to compare academic benefits of attending one school to another. Prominent among measures used by states are the school proficiency and growth indicators resulting from high-stakes tests. Using a longitudinal dataset of college-bound public high school students in North Carolina (N = 17,565), we test the usefulness of proficiency and growth scores of high schools in predicting students' performance in college. We find both indicators to be useful and have the strongest associations with outcomes for students taking a mix of honors and non-honors classes. We also find that academic track placement has a stronger association with outcomes than either measure of school performance. Therefore, this study shows that reliance on school quality indicators can be inadequate because they exaggerate differences between schools and ignore differences within schools.