Reducing COVID-19 Learning Loss through Tutoring: A Meta-Analytic and Cost-Effectiveness Analysis

Which tutoring programs and practices are most effective at supporting students of color and students from low-income families? v What hypothetical combination of tutoring program design features would produce the most benefits to students relative to their costs? Evidence suggests that the COVID-19 pandemic has resulted in dramatic learning losses and widened education gaps across racial and economic lines. Researchers, policymakers, and pundits alike have identified tutoring as a promising approach to address COVID-related learning loss, support struggling students, and curb growing educational inequalities. However, researchers know little about what programs are most cost effective, limiting their ability to provide credible guidance to leaders managing tight budgets. Thus, there is an acute need for a comprehensive meta-analytic and cost-effectiveness analysis of the causal research on tutoring. This study will identify high-impact and cost-effective tutoring models to inform schools' efforts to create tutoring programs that are financially viable for the long-term. Kraft and colleagues will review the research literature on tutoring by: 1) compiling existing reviews of the tutoring literature; 2) searching electronic databases; and 3) reviewing working paper series organized by the National Bureau of Economic Research, The Annenberg Institute at Brown University, and other academic centers and private research firms to capture white papers and research reports. They will also review the references in their preliminary list of papers to cross-check their search process and contact scholars in the field to inquire about additional studies. Data from each study will be systematically recorded to estimate standardized effect sizes. Analyses will include modeling pooled effect sizes and heterogeneity across program characteristics, actual and simulated cost-effectiveness analyses, and robustness tests.