To Address Pandemic Learning Loss, Evidence Points to Tutoring

Individualized instruction consistently outperforms all other educational interventions

A student works with a tutor

In the aftermath of the pandemic, schools across the country face an urgent crisis of student achievement. Most students will require a minimum of three school years to recover the academic learning that was lost. The deleterious effects of the pandemic were even more pronounced among our nation’s most vulnerable students, including those from the most economically disadvantaged circumstances.

In response to the Covid crisis, the federal government injected $190 billion of stimulus aid into the K–12 public education system via the Elementary and Secondary School Emergency Relief Fund. ESSER funds gave district leaders wide latitude to select educational interventions to address pandemic learning loss, including high-dosage tutoring, small-group interventions, afterschool programs, and extended schoolyears. Many of these leaders have placed their bets on tutoring.

Recently, however, questions have arisen about whether the increasing investment in high-dosage tutoring has been motivated by suspect science. Paul T. von Hippel profiled an influential 1984 article by Benjamin Bloom which claimed that tutoring could raise student achievement by two standard deviations (see “Two-Sigma Tutoring: Separating Science Fiction from Science Fact,” features, Spring 2024), akin to moving an average student (at the 50th percentile) to the 98th percentile of the achievement distribution. “The idea that tutoring consistently raises achievement by two standard deviations is exaggerated and oversimplified,” von Hippel concludes. Frederick Hess writes in Forbes that gains of that size would be “astonishing, representing more than a year’s worth of learning in early elementary school and roughly five years’ worth for high schoolers.” And while Hess concedes that “Tutoring can be a powerful tool,” he suggests that “the same can be said about many other strategies to improve schooling.”

I agree with Hess that tutoring can be a powerful tool. And certainly, Bloom’s oft-cited claim that tutoring reliably improves student achievement by two standard deviations is a vast overstatement. At the same time, we shouldn’t let this myth-busting distract us from the fact that tutoring has the largest impact on student learning from among a wide range of educational interventions.

Here’s what we know:

In a recent peer-reviewed meta-analysis of 89 randomized control trials of tutoring interventions, Andre Nickow and coauthors show that the average impact of tutoring on student achievement is 0.29 SD, or approximately four additional months of learning for the typical elementary aged student. This result, which was based on studies published between 1985 and 2019 of tutoring providers that served students in different grade levels and subject areas with different tutor types (e.g., teachers, paraprofessionals, and parents) shows that tutoring significantly outperforms other educational interventions, including class size reduction (0.13–0.20 SD); vacation academies (0.06–0.16 SD); summer school (.08–.09 SD); and extended school day/year (0.05 SD). Tutoring is not an “education boomlet … triggered by research findings which later proved to be exaggerated, incomplete, or colored by wishful thinking,” as Hess suggests, but rather an intervention with a host of evidence to support not just its effectiveness for improving student achievement but its relative dominance over other strategies to support post-pandemic academic recovery.

Overstating what the research shows is a mistake. Tutoring is not a panacea, and it won’t close achievement gaps overnight. There is wide variability in the impact that tutoring has on student achievement. And while tutoring has the potential, for example, to double the annual growth in high school math achievement, much more work is necessary to identify—and scale—the tutoring providers that meaningfully improve student achievement for students in different grade levels, with different educational needs (e.g., students with dyslexia), and in different schooling environments. More evidence is also needed to determine which features of tutoring program design are associated with improvements in student achievement, including the tutor type (e.g., teachers or college students), modality (in-person, virtual, or hybrid), and student-to-tutor ratio.

At Accelerate, we are developing an extensive research portfolio with a focus on the design, implementation, and, ultimately, the impact of tutoring on student achievement to support ongoing knowledge generation for the field. Our goal is to identify effective tutoring models that meaningfully improve student achievement and can be scaled so that significantly more than the 10 percent of students nationally who currently receive high-dosage tutoring can do so. We believe that evidence and return on investment should be the primary factors that inform how education leaders select interventions. Given the challenge of student learning loss that our nation currently faces, tutoring remains one of our best bets for accelerating learning recovery and closing persistent achievement gaps.

Matthew P. Steinberg is the Managing Director of Research and Evaluation at Accelerate.

Last Updated


Notify Me When Education Next

Posts a Big Story

Business + Editorial Office

Program on Education Policy and Governance
Harvard Kennedy School
79 JFK Street, Cambridge, MA 02138
Phone (617) 496-5488
Fax (617) 496-4428

For subscription service to the printed journal
Phone (617) 496-5488

Copyright © 2024 President & Fellows of Harvard College