Rigorous research provides strong evidence that high-dosage tutoring can produce large learning gains for a wide range of students, including those who have fallen behind academically.
Tutoring is not a silver bullet. Students are unlikely to experience learning gains if schools do not commit time and resources to implement a high-quality tutoring program well.
EdResearch for Recovery
Design Principles Series
Carly D. Robinson, Matthew A. Kraft, & Susanna Loeb | Annenberg Institute at Brown University
Beth E. Schueler | University of Virginia
DESIGN PRINCIPLES FOR EFFECTIVE TUTORING
AT A GLANCE
FREQUENCY GROUP SIZE
Tutoring is most likely to be effective when Tutors can effectively instruct up to three or four
delivered in high doses through tutoring programs students at a time. However, moving beyond this
with three or more sessions per week or intensive, number can quickly become small group
week-long, small-group programs taught by instruction, which is less personalized and requires
talented teachers. a higher degree of skill to do well. One-to-one
tutoring is likely most effective but also more costly.
Because the skills required for tutoring are different Researchers have found tutoring to be effective at
from the skills required for effective classroom all grade levels—even for high school students who
teaching, a wide variety of tutors (including have fallen quite far behind. The evidence is
volunteers and college students) can successfully strongest, with the most research available, for
improve student outcomes, if they receive adequate reading-focused tutoring for students in early
training and ongoing support. grades (particularly grades K-2) and for math-
focused tutoring for older students.
Tutoring programs that support data use and Ensuring students have a consistent tutor over time
ongoing informal assessments allow tutors to more may facilitate positive tutor-student relationships
effectively tailor their instruction for individual and a stronger understanding of students’ learning
Using high-quality instructional materials that are Tutoring interventions that are conducted during
aligned with classroom content allows tutors to the school day tend to result in greater learning
reinforce and support teachers’ classroom gains than those that are after-school or during the
DELIVERY MODE PRIORITIZATION
Most research has focused on in-person tutoring, Programs that target lower-performing students
but there is emerging evidence that tutoring can be can support those students who most need
effective when delivered at a distance. personalized instruction but can also create a
negative stigma where tutoring is perceived as a
punishment. Programs that target all students in a
lower-performing grade level or school benefit from
broader organizational commitment and the
perception that tutoring is for everyone but are
EdResearch for Recovery Design Principles - February 2021 1
THE EVIDENCE BASE
Rigorous research provides strong evidence that high-dosage tutoring can produce large
learning gains for a wide range of students, including those who have fallen behind
Tutoring can meaningfully increase learning for a wide variety of K-12 students.
A recent meta-analysis reviewed studies of tutoring interventions that have been evaluated by randomized
controlled trials in the past few decades and found that, on average, tutoring increased achievement by
roughly an additional three to 15 months of learning across grade levels.
Another review of almost 200 rigorous studies found that high-dosage tutoring—defined as more than three
days per week or at a rate of at least 50 hours over 36 weeks—is one of the few school-based interventions
with demonstrated large positive effects on both math and reading achievement.
While effective tutoring programs can be expensive, their large average effects make them highly cost effective
relative to many other educational interventions.
Tutoring is one of the most effective ways to increase achievement for students from lower income
A 2017 study examined interventions that aimed to improve educational achievement for elementary and
middle school students from low socioeconomic backgrounds. Of all the interventions examined, including
feedback and progress monitoring, cooperative learning, computer-assisted instruction, and mentoring of
students, tutoring was most effective.
High-dosage tutoring can be scaled and still improve student learning outcomes.
Many educational programs that show effects in smaller trials appear less effective when implemented for large
groups of students. Large-scale tutoring will not likely replicate the gains found in small-scale studies
evaluating tutoring programs under ideal circumstances. However, studies of 15 larger-scale tutoring
programs serving between 500 and 7,000 students still found that these programs generated meaningful
gains (an average effect size of 0.25 standard deviations).
Although high-dosage tutoring is an excellent strategy for addressing COVID-19 learning
loss, students most likely to benefit from high-dosage tutoring are the least likely to have
adequate access without direct school or district action.
COVID-19 is projected to widen inequality in educational outcomes.
New research consistently finds evidence that the pandemic and initial school closures increased inequality in
educational outcomes across racial and socioeconomic lines, creating an urgency to identify programs and
policies—such as high-dosage tutoring—to mitigate COVID learning loss.
COVID-19 is likely to widen gaps in access to individualized instruction.
The pandemic substantially widened socioeconomic gaps in parents’ searches for online learning resources
during spring 2020. Wealthy families are also more likely to hire professional tutors to combat learning loss,
exacerbating gaps. Students from lower-income families and schools will likely require additional attention and
tutoring resources given lower levels of access to and engagement with supplemental learning tools.
DESIGN PRINCIPLES: ACCELERATING STUDENT LEARNING WITH HIGH-DOSAGE TUTORING 2
DESIGN PRINCIPLES FOR EFFECTIVE TUTORING
Tutoring is most likely to be effective when delivered in high doses through tutoring
programs with three or more sessions per week or intensive, week-long, small-
group programs taught by talented teachers.
Tutoring interventions appear to be more effective as the number of tutoring sessions per week and
the overall number of weeks increase.
The most effective tutoring interventions involve three or more sessions per week with sessions that last for
about 30-60 minutes per day. Studies have found little evidence that once-a-week tutoring is sufficient to
generate meaningful gains.
Most effective tutoring programs last for at least 10 weeks, but many last for the entire school year.
However, elementary students may benefit from shorter but more frequent sessions (i.e. 20 minutes, five times
Intensive “vacation academy” programs where small groups of struggling students focus on a single
subject over week-long vacation breaks have also generated positive results.
With roughly ten-to-one student-teacher ratios, vacation academy programs are more affordable than one-to-
one tutoring models. Although they produce meaningful positive effects, the gains are smaller than those
observed for state-of-the-art high-dosage tutoring.
Vacation academies must be carefully designed to ensure impact. Successful programs have drawn on
talented teachers who are likely better able to effectively serve small groups than inexperienced volunteers or
even average-performing certified teachers. These programs also target students falling below proficiency
thresholds but without significant attendance or behavioral issues.
Tutors can effectively instruct up to three or four students at a time. However,
moving beyond this number can quickly become small group instruction, which is
less personalized and requires a higher degree of skill to do well. One-to-one
tutoring is likely most effective but also more costly.
Tutoring does not need to be one-on-one.
The Match Corps/Saga Education tutoring model pairs one tutor with two students at a time and has shown
large positive student learning outcomes across numerous studies.
Grouping students by skill level or language learner status may make for a more effective tutoring session.
DESIGN PRINCIPLES: ACCELERATING STUDENT LEARNING WITH HIGH-DOSAGE TUTORING 3
Because the skills required for tutoring are different from the skills required for
effective classroom teaching, a wide variety of tutors (including volunteers and
college students) can successfully improve student outcomes, if they receive
adequate training and ongoing support.
Although teachers tend to be the most consistently effective tutors, recent studies have found that
AmeriCorps members and paraprofessionals (teaching assistants) can be just as effective when
tutoring one-to-one or small groups.
One of the most promising studies in recent years used AmeriCorps fellows to conduct daily two-to-one
tutoring sessions with low-performing students in 12 Chicago Public Schools. In addition to improved math
test scores, students' math grades improved by 0.58-GPA points, moving students from a C- average to a
C+. Students were 50% less likely to fail their math course and 28% less likely to fail a non-math course.
Several paraprofessional tutoring programs have proven to effectively increase their scale, including those that
deploy service fellows as year-long tutors through AmeriCorps and Saga Education.
Successful tutoring programs that rely on volunteers tend to involve intensive training—often at least
a couple of weeks—and ongoing support throughout the program, including structured materials and
Tutoring programs that employ service fellows often require tutors to pass a subject-related exam, as well as
undergo trainings focused on instructional techniques, social-emotional learning, and cultural competency.
While completely unpaid volunteers have not historically performed well as tutors, “paid volunteers” like paid
employees of local businesses whose time is donated and college students on work-study programs show
Researchers have found tutoring to be effective at all grade levels—even for high
school students who have fallen quite far behind. The evidence is strongest, with
the most research available, for reading-focused tutoring for students in early
grades (particularly grades K-2) and for math-focused tutoring for older students.
Reading-focused tutoring interventions for kindergarten and first graders have been evaluated far
more often than other tutoring programs.
Of 203 studies about the effects of tutoring programs on student learning, 148 assessed the effects on literacy
development, the vast majority of which focused on elementary school grades.
Fewer studies evaluate tutoring initiatives in middle and high school. At these grade levels, there is
greater evidence of success in math outcomes compared to reading outcomes.
A study of the Match tutoring program, which provided 9th and 10th grade males in 12 Chicago Public
Schools with intensive two-to-one math tutoring as a for-credit class during the school day, found the program
reduced math course failures by over 50%, in addition to reducing overall course failures and increasing math
DESIGN PRINCIPLES: ACCELERATING STUDENT LEARNING WITH HIGH-DOSAGE TUTORING 4
Tutoring programs that support data use and ongoing informal assessments allow
tutors to more effectively tailor their instruction for individual students.
Tutoring programs that effectively use data are more likely to be successful.
Many of the most successful tutoring interventions to date collected ongoing implementation data and used it
to continuously improve their programs.
Successful tutoring organizations often have clear learning objectives and engage in rigorous evaluations to
assess the effectiveness of their program.
Frequent assessments of learning allow tutors to personalize instruction based on individual
Formative assessments, whether conducted in the classroom or in a tutoring session, provide the tutor with
timely feedback on each student and allow for personalized instruction.
The Reading Partners program monitors student progress using mid-year and end-of-year assessments and
updates each student’s reading plan after the mid-year assessment.
For formative assessments to result in more student learning, tutors need time and support to review the
assessment, as well as knowledge of how to address each student’s needs.
Minnesota Math Corps tutors are provided with regular professional development to improve their data-based
Ensuring students have a consistent tutor over time may facilitate positive tutor-
student relationships and a stronger understanding of students’ learning needs.
Tutoring programs that pair a student with a consistent tutor for the duration of the program may
provide better learning environments.
Many tutoring interventions keep one tutor with a set student or group of students for the duration of the
Tutors that meet regularly with students have the opportunity to get to know individual students’ strengths and
weaknesses, while developing caring tutor-student relationships.
Like positive teacher-student relationships, positive tutor-student relationships may lead to greater
academic, social, and motivational outcomes.
Some successful tutoring interventions are relationship-driven programs that focus on building trusting
relationships between tutors and students, in addition to enhancing academic achievement.
When students have positive, caring relationships with their tutors, they may become more motivated and
have better attitudes toward school. Research on youth mentoring programs suggest positive mentoring
relationships can have a wide range of benefits for students’ social-emotional well-being.
Small-group vacation academy participants who were assigned to the same teacher for the week-long
program saw the largest reductions in suspensions after the program.
DESIGN PRINCIPLES: ACCELERATING STUDENT LEARNING WITH HIGH-DOSAGE TUTORING 5
Using high-quality instructional materials that are aligned with classroom content
allows tutors to reinforce and support teachers’ classroom instruction.
Students likely learn more when their tutoring sessions complement and are responsive to their
classroom grade-level instruction.
High-impact programs tend to use high-quality instructional materials that align with their core classroom
content or grade-level standards.
Remediation, or giving students simpler, previous grade-level materials, had been found to result only in
students falling further behind the grade-level material.
Rather than focusing on items that students have failed to master previously, tutors can address missed
concepts and skills that are most critical to accessing the upcoming content.
Tutors can take on the role of catching struggling students up, while taking some of the pressure off classroom
teachers and allowing them to maintain the pace of their instruction. However, tutors need to ensure they build
foundational skills while making connections to the content students are learning in class.
There is some suggestive evidence that tutoring programs that employ school-based coordinators who
connect tutors to teachers may be more successful.
Tutoring interventions that are conducted during the school day tend to result in
greater learning gains than those that are after school or during the summer.
The recent meta-analysis of tutoring studies found that the effects of programs conducted during the
school day are roughly twice as large as those conducted outside of school.
Providing tutoring during school or immediately after school increases the likelihood that students will attend
their tutoring sessions and facilitates a more academic-focused culture.
Tutoring that was a feature of the No Child Left Behind supplementary education services was delivered by
independent providers outside of the school day. Poor student attendance and a disconnect from students’
school experiences often led to disappointing results.
Most research has focused on in-person tutoring, but there is emerging evidence
that tutoring can be effective when delivered at a distance.
Virtual learning offers the potential to lower tutoring costs and potentially provide stronger matches
between students’ needs and tutors’ skill sets by expanding the potential supply of tutors.
Preliminary results from a recent evaluation of the Saga Education tutoring program in Chicago and New York
City public schools found that a blended model was equally effective at increasing student learning as the
costlier in-person only tutoring.
A recent small-scale evaluation of an elementary math online tutoring program found promising results.
Students who received online one-to-one tutoring showed greater gains on a math assessment than those
who did not, which compares favorably to effects found by other, in-person elementary math tutoring
DESIGN PRINCIPLES: ACCELERATING STUDENT LEARNING WITH HIGH-DOSAGE TUTORING 6
Programs that target lower-performing students can support those students who
most need personalized instruction but can also create a negative stigma where
tutoring is perceived as a punishment. Programs that target all students in a lower-
performing grade level or school benefit from broader organizational commitment
and the perception that tutoring is for everyone but are more costly.
Decisions about which students to target are central; however, several different models have led to
Need-driven: Most tutoring interventions that have undergone evaluation have been need-driven and targeted
to students who perform below particular thresholds. In 19 schools across three states, Reading Partners
identified more than 1,100 students who were half a year to two and a half years behind grade level in reading
to receive intensive reading tutoring weekly for 28 weeks on average.
Curriculum-driven: Tutoring can also be curriculum-driven and provided at critical stages of students’ learning
development or school transitions, points at which the disruption can negatively affect student achievement.
For instance, Reading Recovery tutoring programs occur during first grade because it is a crucial point for
Universal tutoring: The Match tutorial model, where all students in a grade or school receive tutoring, may
make tutored students feel less stigmatized, could address the needs of the mid-performing students, and
may provide a vehicle for high-performing students in underserved communities to excel. Districts or states
can also target low-performing schools but provide tutoring schoolwide to address equity concerns without
generating stigma for individual students.
Framing matters for avoiding negative stigma.
Vacation academies in Lawrence, Massachusetts were pitched as a “special opportunity” rather than
remediation and achieved high student engagement with student prizes and interspersed recreation. These
programs were one part of broader “turnaround” policy efforts and therefore were implemented in some of the
state’s lowest-performing schools.
CAVEATS & CONSIDERATIONS
Tutoring programs may exacerbate educational inequalities if schools do not take an
active role in connecting students who could benefit from tutoring with tutors.
Many students who would benefit from tutoring are unlikely to sign up for tutoring on their own.
Schools can reduce barriers for students by actively making tutoring a core part of students’ educational
It is much harder for tutors to deliver effective individualized instruction if they have to
tutor more than three or four students at a time, especially if tutors are not experienced
DESIGN PRINCIPLES: ACCELERATING STUDENT LEARNING WITH HIGH-DOSAGE TUTORING 7
While tutoring can involve up to four students with one tutor, teachers might be better able to tutor up
to four students whereas paraprofessionals and volunteers may do better with one or two students at
Paraprofessionals and volunteers may be better suited to one-to-one tutoring because they are less likely to
have developed the skills in behavior management and group instruction that are needed for working with
Tutoring is not a silver bullet. Students are unlikely to experience learning gains if schools
do not commit time and resources to implementing a high-quality tutoring program well.
Studies have found little evidence that once-a-week tutoring is sufficient to generate meaningful
One meta-analysis found that high-dosage tutoring was 20 times more effective than low-dosage tutoring in
math. In reading, high-dosage tutoring was 15 times more effective than low-dosage tutoring.
Simply providing students with access to tutoring is unlikely to be effective for all students.
Tutoring interventions often are not successful when there are no minimum dosage requirements, little
oversight, and minimal connections with the students’ schools. A key element of successful tutoring programs
is being able to establish a rigorous and caring culture.
FOR MORE INFORMATION
This brief is one in a series aimed at providing K-12 education decision-makers and advocates with an evidence
base to ground discussions about how to best serve students during and following the novel coronavirus pandemic.
Click here to learn more about the EdResearch for Recovery Project and view the set of COVID-19 response-and-
recovery topic areas and practitioner-generated questions. To receive updates and the latest briefs, sign up here.
Briefs in this series will address a broad range of COVID-19 challenges across five categories:
Supporting All Students
Finances and Operations
This EdResearch for Recovery Project brief is a collaboration among:
Funding for this research was provided by the Bill & Melinda Gates Foundation. The findings and conclusions contained
within are those of the authors and do not necessarily reflect positions or policies of the foundation.
DESIGN PRINCIPLES: ACCELERATING STUDENT LEARNING WITH HIGH-DOSAGE TUTORING 8
APPENDIX: EXAMPLES OF EFFECTIVE TUTORING MODELS
NAME & DESCRIPTION FREQUENCY GROUP SIZE PERSONNEL FOCUS MEASUREMENT CURRICULUM
Minnesota 20 minutes each 1:1 (K-3) + AmeriCorps members Reading Tutors use research- The K-3 curriculum is
or year-long Fellows, focused on the National
Reading day (K-3) Small group
PreK-3 based assessments
Reading Panel’s “Big
(PreK) to monitor their
Corps Embedded in college grads students’ progress Five Ideas in Literacy”
classroom and work with and teaches those
Consistently structured Three-day training prior through direct instruction
lessons with clear/detailed (PreK) coaches to make
to the start of each with heavy modeling by
directives to tutors school year and sure they’re tutor, practice by
significant, multi-layered providing the right student, and immediate
training and support instruction to each feedback and clear
throughout the year. child. rationale from the tutor.
Tutors are given The PreK curriculum is
detailed literacy focused on the
manuals and online emergent literacy skills
resources to identified by the National
supplement the Early Literacy Panel.
contents of the manual.
Saga One class 2:1 AmeriCorps members or 6-12 (focus Tutors use formative The first half hour of
Education period of about year-long Fellows, on 9th assessments such each tutoring session is
50 minutes daily generally recently grade) as exit tickets which skill-building curriculum
Recent college graduates during the college grads allow them to revise for remediation. The
lead small-group tutoring school day for Math (focus the next day’s second half of each
during designated class the full school Two weeks of training on Algebra I) lesson based on the session is tied to what
period. prior day’s learning.
year prior to starting the students are learning in
school year Pre- and post-tests their classrooms.
Corps members have
caseload of 6-10 are used for each
Continued professional unit, which show
students. development throughout tutors how much
the school year review time to allow
for the first 2-3
weeks of the next
Reading Five months to 1:1 Volunteers (high school K-4 Monitor student Research-based,
Partners over a year through retirees) progress using the scripted curriculum
Primarily midyear and end-of- aligned with Common
from the community Twice-weekly, Training involves serves year assessments. Core State Standards
provide one-on-one 60-minute participation an students in and other state
literacy tutoring to sessions orientation called a federally Update each standards
struggling readers in “shadow session” designated student’s individual
kindergarten to 4th grade. low-income reading plan after
Continued support and schools. the midyear
Tutoring takes place in a coaching provided assessment.
Center” at each school or
online through their
Minnesota Either two 45- 2:1 AmeriCorps members 4-8 After receiving Curriculum is built
minute sessions and local community intervention in all around research that
Math Corps or three 30- members. subskills within a suggests skills with
minute sessions given unit, students whole and rational
Initial four-day training complete a mastery numbers in late
Highly trained tutors work each week for a led by trainers with
with pairs of students in total of 90 assessment covering elementary and middle
doctoral degrees and
grades 4-8, using minutes per experience conducting each of the subskills. school are foundational
evidence-based math week math intervention If any subskills are for future math
interventions created by research. not mastered, a development.
experts. remedial lesson is
Tutors required to provided to facilitate
Students are paired based attend ongoing trainings mastery.
on similar performance on during the school year,
a placement test. which focus on data-
based decision making,
and goal setting.
Schools identify an
“Internal Coach” who
and support to tutors.
DESIGN PRINCIPLES: ACCELERATING STUDENT LEARNING WITH HIGH-DOSAGE TUTORING 9