Professional development — formal in-service training to upgrade the content knowledge and pedagogical skills of teachers — is widely viewed as an important means of improving teaching and learning. While many interventions include professional development, professional development was the central intervention of the two recent research and demonstration projects — the Professional Development in Reading Study (the "Reading PD study," for short) and the Middle School Mathematics Professional Development Impact Study (the "Math PD study") — whose findings are synthesized in this report. The studies were carried out by the American Institutes for Research and MDRC for the U.S. Department of Education. The professional development that was provided went far beyond the "one-shot" workshop approach that has been widely criticized; it instead included intensive summer institutes, follow-up group sessions, and coaching of individual teachers. The evaluations of the interventions employed random assignment design, and, as a result, they supply unusually rigorous evidence about the effects of the professional development that was offered both on instruction and on student achievement.

The impacts of both interventions were substantially less positive than had been hoped. The Reading PD study increased teachers’ content knowledge; the Math PD study did not. In both studies, the professional development had positive effects on some targeted instructional practices but not on others. Most critically, students of teachers who received the training scored no higher on subject-matter achievement tests than students of teachers who did not receive the training. Moreover, in the reading study, professional development that included one-on-one coaching as well as group workshops did not lead to significantly larger impacts than professional development involving just the workshops; in the mathematics study, receiving two years of professional development did not lead to better results than receiving just one year.

A number of factors likely reduced the effectiveness of the professional development and the researchers’ ability to measure that effectiveness. For example, teacher turnover in the Math PD study meant that many teachers did not receive the full dose of professional development that had been planned. And the two-year time frames of the two studies may not have allowed enough time for major changes in teaching and learning to take hold.

Nonexperimental analyses that were conducted as part of these two studies, along with other research, suggest that the theory of change underlying the studies is correct: professional development of the type that was delivered is associated with increased teacher knowledge and that teacher knowledge and improved instruction is associated with higher student test scores. But changes in teacher-related variables must be substantial — considerably larger than they were in these studies — to move the needle on student achievement even a small amount.

By themselves, the findings of the two studies do not mean that professional development efforts cannot work. New thinking emphasizes a broader conception of teacher learning that involves all teachers in a school in a professional learning community that is engaged in a continuous and collegial cycle of learning, practice, reflection, and improvement. Randomized trials to test professional development that is reinforced within professional learning communities are in order. At the same time, in-service training should not be the only vehicle for improving student achievement.