What really makes teacher professional development effective? And more importantly, why does it work?
A new meta-analysis by Lynch, Gonzalez, Hill, and Merritt (2025) offers a clear and nuanced answer. The authors looked at 46 experimental studies on professional learning programs in mathematics and science. But unlike many others, they didn’t just ask whether these programs improved student performance. They also examined what changed in teachers and how those changes related to student learning.
The good news? On average, the effects on teacher outcomes were strong (+0.52 SD) regarding subject knowledge and teaching practices. But the real crux lies in the next step: which teacher outcomes predict student achievement?
The answer is straightforward: instructional quality.
Programs that succeeded in improving how teachers teach also significantly positively affected student learning outcomes (+0.24 SD for each standard deviation improvement in instructional quality). Simply put: better teaching leads to better learning—but only when professional development directly impacts classroom practice.
This finding has major implications for designing, funding, and evaluating teacher training. Too often, we assume a direct line from professional development to student success. But this study reminds us that such a shortcut doesn’t exist. There’s always an intermediate step: what do teachers do differently after the training?
That makes this study both timely and relevant. It doesn’t just show that professional development can work—it shows how it works. And it reminds us: between input (the training) and output (student results), there is a middle step that we cannot afford to skip.
The one that truly counts.
Abstract of the working paper:
Despite evidence that teacher professional development interventions in mathematics and science can increase student achievement, our understanding of the mechanisms by which this occurs – particularly how these interventions affect teachers themselves, and the extent to which teacherlevel changes predict student learning – remains limited. The current meta-analysis synthesizes 46 experimental studies of PK-12 mathematics and science professional development interventions to investigate how these interventions affect teacher-level outcomes, including knowledge and classroom instruction, and how these impacts relate to intervention effects on student achievement. Compared with controls, treatment group teachers in PK-12 mathematics and science PD intervention studies demonstrated stronger performance on teacher-level outcomes (pooled average impact estimate: +0.52 SD). Programs with larger impacts on teacherlevel outcomes also tended to have significantly larger mean impacts on student achievement. We discuss implications for future research and practice.