This article explores the use of standardized tests to hold schools accountable. The history of testing for accountability is reviewed, and it is shown that currently between-school differences account for less than 10% of the variance in student scores, in part because the progress of individuals is small compared to the spread of achievement within an age cohort, and, possibly, due to lack of alignment between instruction and assessment. A review of the literature on the effects of the introduction of such tests in high-stakes accountability regimes suggests that the effects can be positive and the size of the effects is substantial. Therefore, although the validity of such tests may be problematic in terms of the intended inferences, their introduction may nevertheless be justified by their impact. The article concludes with a number of suggestions on improving tests for high-stakes accountability. (Contains 4 figures and 3 footnotes.)