Tests are a natural part of education, from the quizzes, essays, and classroom tests that teachers have traditionally administered to the high-stakes tests that states use to make decisions about graduation, promotion, and school funding and governance. In this article, the author stresses the need to learn the unintended consequences of high-stakes testing, including reliability of the actual tests, nonrandom human error, and fudging dropout figures. The pros and cons of the state approach of using high-stakes testing to increase student achievement are examined.