A Meta-Analysis of the Experimental Evidence Linking Mathematics and Science Professional Development Interventions to Teacher Knowledge, Classroom Instruction, and Student Achievement

THE BOTTOM LINE

Professional development (PD) for teachers is one of the most common strategies schools use to improve teaching. But with so many approaches out there, how do we know which ones actually make a difference, and why?

With limited time for PD, it's important that the training teachers receive is grounded in the features most likely to improve their practice and, ultimately, student learning.This meta-analysis synthesizes 46 experimental studies of PD programs in math and science to explore why some PD programs are more effective than others at changing teacher knowledge and instructional practice, something earlier reviews haven’t been able to do. 

The study has three key findings:

1. The math and science PD programs included in the meta-analysis significantly improved both teacher knowledge and instructional practice.
2. PD programs that focused on strengthening teacher knowledge (both content and pedagogical knowledge) and content-specific formative assessment were associated with stronger improvements in classroom instruction than programs that did not contain these features, while PD duration, a focus on content-general instructional strategies, and inclusion of curriculum materials did not boost the size of PD impacts on teacher knowledge or classroom instruction.
3. PD programs that improved classroom instruction were more likely to also see significant gains in student achievement, while improvements in teachers’ content or pedagogical content knowledge were not associated with improved student outcomes. 

THE STUDY

The study asks three research questions:

Research Question 1: What is the causal impact of math and science PD programs on teacher outcomes?

The question seeks to establish a causal relationship between math and science PD programs and improvements in two teacher-level outcomes: 

• (1) teacher content knowledge- a teacher’s understanding of subject matter content
• (2) classroom instruction- observable teaching practices teachers use, such as questioning techniques, discussion routines, and instructional formats, that shape how students engage with content

Research Question 2: Do specific design features, focus areas, or contexts of PD programs predict stronger impacts on teacher outcomes?

PD design features the authors tested:

• Does combining professional development with new curriculum materials lead to stronger improvements in teacher outcomes than providing PD alone?
• Do teachers in longer PD programs have better outcomes than those in shorter ones?

PD focus areas the authors tested:

• Does a focus on improving teachers’ knowledge (content knowledge and pedagogical content knowledge) lead to stronger impacts on teacher outcomes compared to PD programs without this focus?
• Does a focus on content-general instructional strategies (e.g., strategies to improve classroom climate)  influence the effectiveness of PD on teacher outcomes?
• Does a focus on content-specific formative assessment (in which teachers learn techniques to adjust instruction based on student learning data) lead to greater improvements in teacher outcomes compared to programs without this focus?

Contextual features the authors tested:

• Teacher sample size
• Demographic and socioeconomic characteristics of the student population
• Geographic setting (urban, suburban, or rural districts)
• Grade level (PK-12)

Note: Unlike the causal analysis used to measure the overall impact of PD (RQ1), the specific features, focus areas, and contexts of PD programs were not causally tested. Instead, researchers compared the differences that already existed across programs, so the links show associations, not cause-and-effect relationships.

Research Question 3: Are changes in teacher outcomes associated with improvements in student achievement? If so, which teacher outcome has a stronger impact on student achievement: improvements in teacher knowledge or improvements in instructional practice?

While many studies assume this link, few have actually tested the full chain, from what PD actually changes about teacher practice to how changes in teachers’ practices affect student achievement.

KEY FINDINGS

1. PD in math and science improves both teacher knowledge (+0.52 SD) and instructional practice (+0.49 SD).  In other words, on average, participating in the studied PD caused teachers to score about 19 percentile points higher on measures of instructional quality compared to those who didn’t participate. 

2. PD that improved classroom instruction tended to have positive impacts on student learning, while PD that mainly improved teacher knowledge did not. When teachers improved how they taught in the classroom, students learned more. But just knowing more about math or science didn’t make a difference for student achievement. Specifically, 1 SD improvement in classroom instruction was associated with a 0.24 SD increase in student achievement.

   Figure 1: PD programs improve both teacher knowledge and classroom instruction. However, only instruction improvements are associated with student learning gains.
   

   PD program characteristics that show a positive impact on teacher outcomes:

3. PD that included a focus on strengthening teacher knowledge had larger average impacts on classroom instruction than those that did not. 

4. PD that included a focus on content-specific formative assessment showed larger impacts on teacher outcomes, especially classroom instruction, compared to programs without this focus.

   PD program characteristics that show no impact on teacher outcomes:

5. When comparing across PD programs, quality matters more than quantity. The number of hours of PD was not associated with its effectiveness. This reinforces prior research that the quality and focus of PD matter more than time spent.
6. Including new curriculum materials in PD programs was not associated with stronger teacher outcomes.

7. High-quality PD can be effective across diverse environments, including high-poverty schools where support is often most needed. PD program impacts did not significantly vary by context, including setting type, student income level, or grade level.

   Figure 2: PD programs that include teacher knowledge-building or content-specific formative assessment are associated with improvements in classroom instruction.
   

IMPLICATIONS FOR POLICY AND PRACTICE

1. PD can be effective in improving teacher knowledge and student outcomes. Overall, studies of PD programs produced moderate-sized average impacts on these intermediate outcomes.
2. Invest in PD that strengthens instructional practice, not just content knowledge. This study finds that changes in classroom instruction are likely the primary driver of student learning. Districts should look for PD that offers strategies teachers can immediately use to improve their instructional practice.
3. More or longer PD is not necessarily better. Even when PD reached or exceeded a 49-hour threshold, it was not associated with stronger impacts on teacher knowledge or instruction compared to programs shorter than 49 hours.

FULL WORKING PAPER

This report is based on the EdWorkingPaper “A Meta-Analysis of the Experimental Evidence Linking Mathematics and Science Professional Development Interventions to Teacher Knowledge, Classroom Instruction, and Student Achievement,” published in May 2025. The full research paper can be found here: https://edworkingpapers.com/ai24-1023