A common criticism of merit-pay plans is that they fail to systematically target rewards to the most effective teachers. This study presents new evidence on this issue by evaluating data from Tennessee's Career Ladder Evaluation System and the Project STAR class-size experiment. Because the students and teachers participating in the experiment were randomly assigned, inferences about the relative quality of teachers certified by the career ladder should be unbiased. The results indicate that Tennessee's career ladder had mixed success in rewarding teachers who increased student achievement. Assignment to career-ladder teachers increased mathematics scores by roughly 3 percentile points but generally had smaller and statistically insignificant effects on reading scores. © 2004 by the Association for Public Policy Analysis and Management.