What works in small scale studies doesn’t always translate to district-wide success. When Metro Nashville Public Schools tried to scale up high-dosage tutoring—a strategy backed by compelling research—the 125,000 hours of instruction that were delivered by trained educators produced minimal gains.
This was puzzling to the district and to researchers because studies have traditionally shown that tutoring is a powerful way to boost student outcomes. A 2024 meta-analysis of 89 studies of tutoring in literacy and math, for example, found academic gains equal to about three to four months of extra learning for early elementary students who received high-dosage tutoring.
It would make sense, then, that when nearly one-fifth of U.S. districts directed a portion of their COVID relief dollars to tutoring programs, their efforts would have been a boon for students. However, in districts like Nashville, the impact of these efforts was largely disappointing.
In the Nashville study, the tutoring delivered to 6,888 students resulted in limited effects in reading (about 1 month of academic gains) and no discernible effects in mathematics. And yet Metro Nashville Public Schools followed research guidance closely:
- It relied heavily on experienced professionals (73 percent of tutors were teachers, paraprofessionals, or district staff) who received extensive training.
- It carefully aligned tutoring materials with existing reading and math curricula.
- It delivered high doses of tutoring—90 minutes per week over a full semester.
At the same time, researchers noted that non-participating students received alternative supports—personalized instruction through computer-adaptive learning programs and small-group in-class support—that proved nearly as effective as the tutoring itself.
Given this, researchers identified one possible cause for the meager returns: Tutoring materials were too aligned with classroom instruction; rather than targeting skill gaps, they often repeated classroom content.
The Nashville experience illustrates why research doesn’t always scale: Many studies included in the 2024 meta-analysis were conducted with fewer than 50 students and examined resource-intensive 1:1 tutoring. As student-to-tutor ratios and sample sizes increased in these studies, the effect sizes of tutoring fell.
But scale isn’t the only issue. These studies didn’t always examine what tutors were teaching. The Nashville results suggest that content matters most: When tutoring simply repeated classroom instruction rather than targeting specific skill gaps, even well-trained tutors delivering high doses of instruction produced minimal gains.
The takeaway for district leaders: Start small, ensure tutoring targets genuine skill gaps rather than duplicating classroom content, and carefully evaluate costs and benefits before scaling up.
