A Randomized Controlled Trial (RCT) of...
Networks for School Improvement (NSI) to Increase the Proportion of Students On Track for High School Graduation
Reviewed
Mathematica (July 2026) published RCT findings for Networks for School Improvement (NSI), a $300M Gates-funded initiative in which schools collaborate to use continuous improvement methods to test and refine strategies addressing a common challenge. This large, high-quality RCT found small positive impacts on academic and behavioral outcomes in schools focused on 9th grade improvement (e.g., moving the average student from the 50th to 53rd percentile), and no positive impacts in schools focused on 8th grade improvement, after 2-3 years of implementation.
Study Report’s Summary of the RCT results:
- The results from the random assignment study were generally similar to the matched comparison results for the 8th-grade on-track NSI. The random assignment study found that the 8th-grade on-track NSI did not have a consistent impact on the foundation’s prioritized outcomes (Appendix Exhibits B.8–B.10). However, in contrast to the matched comparison results, there were no impacts on math test scores in the random assignment study. This could be because the math test score impacts were driven primarily by NSI focused on improving that outcome, and only one of the five 8th-grade on-track NSI in the random assignment analysis were focused on improving math test scores.
- The random assignment and matched comparison results were also similar for the 9th-grade on-track NSI. The random assignment study found that 9th-grade on-track NSI had medium-sized, positive impacts on course-related outcomes (Appendix Exhibit B.11). However, the random assignment analysis also found a medium-sized positive impact on attendance rates in schools’ second and third years of participation (Appendix Exhibit B.12). The impact in Year 3 is equivalent to a 2-percentage-point increase in the attendance rate, from 83 to 85 percent (0.11 standard deviations). Both of the 9th-grade on-track NSI that focused on improving attendance were in the random assignment study, which may have contributed to the positive impacts in the random assignment analysis.
We have no suggested revisions to the study report’s summary of the RCT results.
No-Spin’s Study Overview
Large, high-quality RCT of Networks for School Improvement (NSI) aimed at increasing the proportion of students on track for high school graduation finds small positive impacts on academic and behavioral outcomes in schools focused on 9th grade improvement, and no positive impacts in schools focused on 8th grade improvement, after 2-3 years of implementation.
Program:
- In an NSI network, approximately 20 schools iteratively test and refine strategies to address a common challenge – such as increasing course pass rates or school attendance – while learning from one another. The study report describes this approach as “network-based continuous improvement.”
- From 2018 to 2025, the Gates Foundation awarded more than $300 million to intermediary organizations that led 34 NSI networks across the country. The intermediaries supported teams of school staff in conducting continuous improvement.
Study Design:
- The study randomly assigned 125 schools located in six U.S. school districts to NSI participation focused on improving 8th grade outcomes, versus a control group. The study randomly assigned an additional 91 schools located in four U.S. school districts to NSI participation focused on improving 9th grade outcomes, versus a control group.
- Across both samples, the schools’ percentage of economically disadvantaged students averaged 80-90%, and the percentage of Black or Latino students averaged 85-90%.
- Because the first year of NSI implementation was affected by COVID, we summarize the findings for years 2 and 3 of implementation (below). These findings apply to the 8th or 9th graders in the study schools in those years (i.e., to the “cross-sectional” sample of 8th or 9th graders).
- Based on our review criteria, this was a high-quality RCT.
Findings in years 2 and 3 of NSI implementation:
- In the 8th grade NSI, the study found no discernible positive impacts on any of the targeted academic or behavioral outcomes.
- In the 9th grade NSI, the study found a pattern of small positive impacts on the targeted academic outcomes, including GPA for all courses, share of core courses passed, and share of 9th graders earning 5+ credits. The standardized effect sizes averaged 0.07 (the equivalent of moving the average student from the 50th to 53rd percentile) and were mostly statistically significant.
- In the 9th grade NSI, the study also found a statistically significant impact of approximately 2 percentage points on attendance rates (85% in the treatment group versus 83% in the control group). The study found no discernible impact on the share of students with no suspensions.
- The study report also presents findings of a matched comparison-group study of NSI, which were generally similar to the RCT results.
Click or tap a highlight to see No-Spin’s comment