Teach For America (TFA), a prominent alternative teacher certification program, has evolved substantially since it began in 1989. However, most studies ignore this potential variation when evaluating TFA's effects on student outcomes. Using 12 years of administrative data, I test whether TFA's effects on achievement differ over time, focusing on a major reform introduced in 2005 and testing for changes in the qualifications of TFA corps members recruited before and after this reform. I find that TFA's effects on achievement increased over time in some subjects and grade levels. However, these changes do not occur immediately following the reforms, nor are they explained by most observable teacher quality indicators. Instead, results indicate gradual improvement over time in many subjects. These results help explain divergent findings in previous research from different periods and underscore the importance of understanding the continuing evolution of alternative teacher certification programs.