Forty-two percent of U.S. public schools offered high-dosage tutoring as of June 2025, according to the National Center for Education Statistics (NCES). School leaders rate it well above the other types of tutoring the agency tracks. In the same survey, 56% of schools using it called it very or extremely effective, compared with 27% for standard tutoring.
NCES defines high-dosage tutoring as sessions three or more times per week, at least 30 minutes each, one-on-one or in small groups. Tutors must be educators or well-trained tutors with specific training, the sessions must align with an evidence-based curriculum, and drop-in homework help does not count. The National Student Support Accelerator uses the term high-impact tutoring and says it "is best scheduled during the school day to ensure all students have access."
What the strongest studies found
Much of the evidence comes from randomized trials, which assign students to tutoring or not at random so the groups can be fairly compared. A 2020 review by Andre Nickow, Philip Oreopoulos, and Vincent Quan, published by the National Bureau of Economic Research (NBER), pooled 96 such studies from preschool through grade 12. It found an average effect of 0.37 standard deviations on learning.
A standard deviation is a statistical yardstick for how spread out scores are. Brown University's Matthew A. Kraft and two coauthors wrote in 2024 that effects of 0.3 to 0.4 standard deviations are roughly equal to an 11 to 15 percentile point gain, or about the reading growth upper elementary students make in a full school year.
The Nickow review found stronger results, on average, for programs run by teachers and paraprofessionals (tutors who do the work as a job but are not certified teachers) than for programs using volunteers or parents. Its authors also wrote that "Tutoring programs conducted during school tend to have larger impacts than those conducted after school."
In Chicago, the University of Chicago Education Lab studied Saga Education, a nonprofit whose model the lab describes as "daily 50-minute, two-on-one instruction to students during the school day." In a first trial with 2,633 ninth and 10th graders in Chicago public schools, math test scores rose 0.16 standard deviations and grades improved in math and other courses, according to a 2021 NBER working paper by Jonathan Guryan, Jens Ludwig, and colleagues. A second trial with 2,710 students found a larger math gain of 0.37 standard deviations.
That paper put Saga's cost at $3,500 to $4,300 per participant per year. A later trial of about 4,000 students tested groups of four, with two students working with a tutor while two used math software. Per-pupil costs were about 30% lower, and math scores rose 0.23 standard deviations, almost as much as under the two-on-one model, according to a 2024 NBER paper by Monica P. Bhatt and coauthors.
Bigger programs tend to post smaller gains. Kraft, Beth E. Schueler, and Grace Falken pooled hundreds of randomized trials in a working paper from the Annenberg Institute at Brown University and found an average effect of 0.42 standard deviations. When they limited the sample to U.S. programs measured with standardized tests, the estimate fell to 0.21 for programs serving 400 to 999 students and 0.16 for programs serving 1,000 or more. Only nine studies in their sample looked at programs that large.
Programs that followed a set of recommended practices lost less of their effect as they grew, the authors found. Those practices were in-person tutoring at school during the school day, no more than three students per tutor, a curriculum provided to tutors, at least three sessions a week, and at least 15 hours of planned tutoring.
What school leaders report
NCES tracks tutoring through its School Pulse Panel, a survey of a national sample of public schools that the agency labels experimental data. The effectiveness ratings below are school leaders' own judgments of their programs.
| Tutoring type | Schools offering, May 2024 | Rated very or extremely effective, May 2024 | Schools offering, June 2025 | Rated very or extremely effective, June 2025 |
|---|---|---|---|---|
| High-dosage | 46% | 52% | 42% | 56% |
| Standard | 66% | 27% | 60% | 27% |
| Self-paced | 24% | 17% | 22% | 15% |
| On-demand online | 13% | 18% | 11% | 14% |
Source: National Center for Education Statistics, School Pulse Panel, May 2024 and June 2025 collections. Effectiveness ratings are among schools offering each type. NCES notes questionnaire changes between collections, so compare years with caution.
In June 2025, high-dosage tutoring's rating beat standard tutoring's by 29 percentage points (56% minus 27%), about twice as high. Overall, 85% of public schools reported providing some type of tutoring during the 2024-25 school year, NCES said.
Money and staffing limit access. In October 2024, 22% of public schools said they provided no tutoring. Among those schools, 48% cited a lack of, or reductions in, funding, and 36% said they could not find staff.
Questions parents can ask
A June 2024 brief from EdResearch for Action, a partnership of the Annenberg Institute and Results for America, sets out design principles drawn from tutoring research. Its authors, Carly D. Robinson, Kraft, Susanna Loeb, and Schueler, describe traits a parent can check.
- How often, and for how long? Effective programs meet three or more times a week for at least 10 weeks, the brief says. The NCES definition sets a 30-minute minimum per session.
- How many students per tutor? "The most effective programs have no more than three students per tutor," the authors wrote.
- The same tutor each time? The brief calls for a consistent tutor, meaning "each student always meeting with the same tutor for the duration of the program."
- During the school day? Programs held during the school day get better attendance and results than those held after school or in the summer, according to the brief.
- What materials? The authors favor high-quality materials that complement grade-level class work and note that simpler, previous-grade material has been found to leave students further behind.
- How is progress tracked? Programs that use data and ongoing informal assessments are better able to understand student progress, the brief says. State results are one outside check, and allk12 has a guide to reading your child's state test scores.
- Who are the tutors? Paraprofessionals, community members, college students, and classroom teachers can all succeed if they get training and ongoing support, the authors wrote.
A separate allk12 parent guide to academic recovery looks at when a content gap calls for a tutor and when the problem is organizational or emotional.
The EdResearch brief also cites a recent study showing that on-demand programs offering 24/7 homework help may not reach the students who need them most. In June 2025, 14% of schools offering on-demand online tutoring rated it very or extremely effective.
Sources
NCES: About One-Quarter of Public Schools Reported That Lack of Focus or Inattention From Students Had a Severe Negative Impact on Learning in 2023-24 (July 18, 2024)
NCES: Public School Leaders Reports Suggest No Overall Change in the Number of Teaching Positions Since Last School Year (December 12, 2024)
NCES: School Pulse Panel Interactive Results
NCES: School Pulse Panel Tutoring Data File, 2022-23 to 2024-25
NBER: The Impressive Effects of Tutoring on PreK-12 Learning: A Systematic Review and Meta-Analysis of the Experimental Evidence
NBER: Not Too Late: Improving Academic Outcomes Among Adolescents
NBER: Can Technology Facilitate Scale? Evidence from a Randomized Evaluation of High Dosage Tutoring
University of Chicago Education Lab: High-Dosage Tutoring Research Partnership With Saga Education
Annenberg Institute at Brown University: What Impacts Should We Expect from Tutoring at Scale? Exploring Meta-Analytic Generalizability
EdResearch for Action: Design Principles for Accelerating Student Learning With High-Impact Tutoring (June 2024)
National Student Support Accelerator: High-Impact Tutoring FAQs
