The past decade has seen a tremendous amount of research on the use of value-added modeling to assess individual teachers, and a significant number of states and districts are now using, or plan to use, value added as a component of a teacher's summative performance evaluation. In this article, I explore the various mechanisms through which the use of value added might affect teacher quality and describe what we know empirically about the potential of each mechanism. Given that many of these mechanisms work through the behavioral response of teachers to the high-stakes uses of evaluation, I argue that the jury is still out on how the use of value added will affect the quality of the teacher workforce.