Personalizing Math Instruction at Scale: A Meta-Analytic and Cost-Effectiveness Analysis of Math Tutoring Programs

This project will identify the most cost-effective and scalable methods of delivering mathematics tutoring to K-12 public school students for the purpose of accelerating learning. While there is a notably large body of rigorous research documenting that high-dosage tutoring produces large academic gains for children who have fallen behind academically, less is known about the features of effective programs needed to guide program design. Additionally, the widespread adoption of tutoring programs has been hampered by the high upfront costs of intensive individualized instructional programs. This project will therefore isolate the key features of the most effective tutoring programs and the costs associated with these programs to identify program models with the potential to deliver the greatest benefits at the lowest cost. Program features could relate to tutor characteristics, the dosage of tutoring delivered, the timing of tutoring sessions, the training and support provided to tutors, tutor-student ratios, the content of tutoring sessions, and more. The project is particularly timely given policymakers and educators are currently looking to invest in affordable strategies to support students' learning in mathematics given early evidence that Covid-19-related educational disruptions have negatively affected mathematics achievement. This project is funded by the EHR Core Research (ECR) program, which supports work that advances fundamental research on STEM learning and learning environments, broadening participation in STEM, and STEM workforce development.

The project will include a thorough meta-analysis of all existing studies of mathematics tutoring interventions that use quasi-experimental or experimental methods designed to isolate the causal effects of the tutoring program in question. More specifically, the research team will search for, collect, and code all studies based on the features of the intervention under examination, the size of the effects the intervention had on student outcomes, as well as information that will allow them to calculate costs for each intervention. Researchers will then pool estimates across studies to identify the average effects of mathematics tutoring on student achievement and average per-pupil costs and to examine how the effects and costs of tutoring programs vary by design and program features. Finally, the researchers will examine which features of interventions predict the largest academic gains for students and calculate cost-benefit ratios for each intervention. Meta-analytic and cost-effectiveness reviews are a critical step in developing new scientific knowledge by generating conclusions with greater precision and external validity and accounting for publication bias which can skew perceptions about the overall effectiveness of an intervention when only well-known or successful programs are highlighted in the literature. This project will provide a comprehensive review across all existing studies, ultimately allowing the research team to accurately identify those mathematics tutoring models with the greatest promise for both impact and scalability.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.