We use the score reports of 375,000 US-based GMAT® test takers to document patterns of mismatch between student ability and school quality. We find substantial levels of mismatch in candidates' selection of MBA programs for score-sending, and by extension, applications. The evidence suggests that the high levels of mismatch in the selection of schools by candidates can explain some of the academic mismatch attributed to school policies. Using discrete choice models, we find candidates are discouraged by distance and motivated by in-state programs. We also find that much of the application overmatch is explained by candidates' preference for high-quality programs.