How Much Should We Rely on Test Scores to Measure School Quality?

While schools may influence academic and non-academic outcomes, it is not clear whether schools that improve test scores are the same schools that causally improve longer-run outcomes. This policy brief uses rich administrative data covering the full population of Trinidad and Tobago to show that (1) School causal effects are multidimensional. Effects on test scores are weakly related to effects on crime, teen births, and adult employment; and (2) Parents of lower-achievers value effects on non-test outcomes relatively more than on tests while the opposite is true for parents of high-achievers. These findings suggest that policy evaluations based solely on test scores may be misleading about the welfare effects of school choice.