Recent controversy has surrounded the effectiveness of school turnaround efforts–programs designed to significantly reform schools or districts through changes to organization, staffing, and governance–to address student outcomes.
A recent meta-analysis by Redding & Nguyen seeks to shed light on the issue by examining the association between different forms of school turnaround and attendance, standardized test scores, and graduation rates. They highlight four school turnaround mechanisms—transformation and turnaround (including comprehensive instructional reform or teacher and principal effectiveness reforms), changes in human capital (through replacing the principal or teachers), changes in governance or management of low-performing schools (such as through state takeover of schools or school districts, closures of schools, or transitioning schools to charter status), and high-stakes accountability policies.
Thirty-five of 13,800 studies screened were eligible for the analysis, showing that:
- There is evidence of positive association between school turnaround, transformation, and school restart and improvements in student attendance, math and ELA test scores, and graduation.
- Though transformation and restart models appeared to have the largest standard deviation unit increases on math test scores (0.177 and 0.214 SDUs, respectively), the authors could not find significant evidence that one method was more successful than the others.
- However, accountability-driven school closures and state turnaround efforts (based on No Child Left Behind waivers and student test scores) did not appear to have any effect, positive or negative.
An important qualification to these results is that these programs take time to show effects—few programs showed much success in the first year of implementation, while stronger positive impacts were seen in years 2-4 (increasing 0.127 SDUs in math and 0.084 in ELA). In addition, the authors note that schools and districts engaged in school turnaround efforts may revert to pre-reform performance levels after some improvement, though the extent of this was not possible to be examined in this study.