Your student is struggling in math.
Maybe their test grades have started slipping.
Maybe homework that should take 30 minutes is taking two hours.
Maybe they understand everything while the teacher is explaining it, but can't reproduce it on their own.
So you start considering tutoring.
Then comes the obvious question:
Does tutoring actually work?
Or is it simply another educational expense that makes parents feel like they're doing something?
As someone who runs a tutoring company, I obviously have a perspective on this question. But fortunately, we don't have to rely on the opinion of someone who sells tutoring.
Tutoring has been studied extensively, including through randomized controlled trials and large research reviews.
And overall, the evidence is remarkably positive.
But there is an important catch.
Not all tutoring is equally effective.
Research suggests that who provides the tutoring, how often it happens, how personalized it is, and what actually occurs during those sessions can make an enormous difference.
So the better question isn't simply:
“Does tutoring work?”
It's:
“What kind of tutoring works, and why?”
What Does the Research Actually Say?
One of the strongest recent pieces of evidence comes from a large systematic review and meta analysis published in the American Educational Research Journal in 2024.
Researchers Andre Nickow, Philip Oreopoulos, and Vincent Quan analyzed 89 randomized controlled studies of PreK through 12th grade tutoring programs.
Randomized studies are particularly valuable because students are assigned to treatment and comparison groups, making it easier to determine whether tutoring itself contributed to differences in academic outcomes.
Across the studies, tutoring produced an average improvement of 0.288 standard deviations in student learning.
The researchers found positive effects in both mathematics and literacy.
That number probably means very little to most parents.
So here's the important part:
In educational research, an effect of that size is meaningful.
And this wasn't one unusually successful tutoring program.
The researchers were combining evidence from dozens of randomized experiments.
J-PAL, the Abdul Latif Jameel Poverty Action Lab, has similarly reviewed a large body of randomized tutoring research and concluded that tutoring consistently produces significant improvements in learning outcomes.
So if the question is simply whether tutoring can improve academic learning, the evidence is strong.
Yes, it can.
But that's only the beginning of the story.
Tutoring Isn't One Thing
Imagine two students both tell you:
“I have a tutor.”
Student A meets with the same experienced math tutor several times each week.
The tutor identifies specific gaps in the student's knowledge, teaches concepts, asks the student to explain their reasoning, assigns targeted practice, monitors progress, and adjusts instruction based on mistakes.
Student B logs into a homework help website once every few weeks when they're stuck.
Someone shows them how to complete that night's assignment.
Technically, both students received tutoring.
But these are completely different educational experiences.
This is one reason research about tutoring needs to be interpreted carefully.
“Tutoring” can describe everything from intensive individualized instruction to occasional homework assistance.
And unsurprisingly, those approaches don't necessarily produce the same results.
Some Research on Private Tutoring Is Much Less Impressive
This is an important point, particularly coming from a tutoring company.
Not every study finds dramatic benefits from tutoring.
A 2024 meta analysis examining 21 studies of private tutoring and academic performance found only a small association overall. After the researchers adjusted for possible publication bias using a statistical method known as trim and fill, the estimated relationship was no longer statistically significant.
Meanwhile, another meta analysis focused on experimental studies of private tutoring found moderate positive effects on academic achievement across the experimental designs it examined.
At first glance, that can sound contradictory.
But these studies are examining different bodies of evidence using different inclusion criteria and research methods.
There is also a major problem with simply comparing students who receive private tutoring with students who don't.
Students are not randomly assigned by life to get tutors.
Students who receive tutoring may already be struggling.
Their families may have greater financial resources.
Their parents may be more involved academically.
They may attend different schools.
They may be preparing for particularly difficult courses or exams.
That makes ordinary observational comparisons difficult to interpret.
This is one reason randomized evaluations of tutoring programs are so useful.
And across those experiments, the evidence for well designed tutoring is considerably stronger.
The takeaway isn't:
“Every tutor works.”
It's:
“Certain forms of tutoring have very strong evidence behind them.”
Why Can Tutoring Be So Effective?
Think about a typical classroom.
One teacher may be responsible for 25 or 30 students.
Those students have different strengths.
Different weaknesses.
Different background knowledge.
Different learning speeds.
Different questions.
Different levels of confidence.
The teacher has to keep the entire class moving.
Now compare that with one on one tutoring.
The student solves a problem incorrectly.
The tutor can immediately stop.
“Walk me through what you did here.”
The student explains.
The tutor realizes the student understands the new concept but is making an older mistake with fractions.
Now the lesson changes.
Instead of continuing through 15 more problems, the tutor can spend ten minutes repairing the fraction issue.
That level of responsiveness is difficult to reproduce in a large classroom.
The major advantage of tutoring isn't simply that the student receives more instruction. It's that the instruction can respond to the individual student.
Good Tutoring Finds the Actual Problem
Suppose a student is struggling in Algebra II.
The obvious response is:
Teach them more Algebra II.
But that may not solve the problem.
Perhaps the student doesn't understand factoring.
Or exponent rules.
Or fractions.
Or solving basic equations.
The student experiences the problem as:
“I don't understand Algebra II.”
But an effective tutor may discover:
“You actually understand today's concept. You keep getting stuck because of a skill you learned two years ago.”
That distinction matters enormously.
Repeating the teacher's explanation of the current lesson won't necessarily fix a missing prerequisite.
Good tutoring should be diagnostic.
The tutor is constantly asking:
What does this student understand?
Where does the reasoning break down?
Is this a current concept or an older gap?
What should we work on next?
That is very different from simply helping a student finish tonight's worksheet.
The Tutor Should Not Be Doing Most of the Talking
This is one of the easiest mistakes to make in tutoring.
A knowledgeable tutor sees a difficult problem.
They explain it beautifully.
The student nods.
The tutor asks:
“Does that make sense?”
The student says:
“Yeah.”
Then they move on.
It feels like a successful lesson.
But there's a problem.
The tutor demonstrated that the tutor understands the material.
We still don't know whether the student does.
Effective tutoring should repeatedly transfer the cognitive work back to the student.
Instead of explaining every step, the tutor might ask:
“What do you think we should do first?”
“Why?”
“What information does the problem give us?”
“What would happen if we tried this?”
“Can you explain that step back to me?”
“Now try the next one without me.”
The tutor should provide support when needed.
But gradually, that support should disappear.
The goal is not for the student to solve problems successfully with the tutor.
The goal is for the student to eventually solve them successfully without the tutor.
Frequency Matters
One of the clearest findings from tutoring research is that intensity matters.
The 2024 meta analysis found that effects tended to be larger in programs where tutoring occurred at least three days per week.
A 2024 research brief from EdResearch for Action similarly identifies frequent sessions as an important characteristic of high impact tutoring. Their research synthesis describes effective models as typically occurring at least three times per week for a sustained period, often in one on one or very small group settings.
That does not mean every student needs tutoring three times a week.
This distinction is important.
Much of the strongest research examines structured interventions designed to accelerate learning among students who are significantly behind.
A student who generally earns A's in AP Chemistry but needs help preparing for one difficult unit has a very different need.
But the broader principle makes sense:
A student who has substantial academic gaps is unlikely to repair them through occasional emergency tutoring.
Consistency creates opportunities to teach, practice, assess, revisit, and build.
Consistency With the Same Tutor Can Matter Too
Imagine having a different teacher every time you walked into class.
Every new teacher would need to figure out:
What you know.
What you don't know.
How quickly you learn.
Where you tend to make mistakes.
What explanations work for you.
What you're currently studying.
A consistent tutor gradually develops that knowledge.
Over time, they may recognize patterns.
“You understand the calculus concept. It's the algebra manipulation that's getting you again.”
Or:
“You tend to rush through the last step.”
Or:
“You know this material better than you think. You're changing correct answers when you become nervous.”
EdResearch for Action identifies sustained tutor student relationships as one of the characteristics associated with strong tutoring programs.
Tutoring is partly about content.
But it is also about knowing the learner.
Professional Tutors Tend to Produce Stronger Results
Who provides the tutoring matters.
The large research review by Nickow and colleagues found that effects tended to be largest when tutoring was provided by teachers or paraprofessionals.
J-PAL's evidence review similarly reports that tutoring led by teachers and professional tutors generally produced stronger effects than tutoring provided by nonprofessionals or parents.
That doesn't mean a volunteer, older sibling, parent, or peer cannot help a student.
Of course they can.
But knowing a subject and teaching a subject are different skills.
A strong tutor needs to recognize misconceptions, ask useful questions, explain ideas in multiple ways, adjust difficulty, determine when to intervene, and know when to stop helping.
Sometimes the best tutoring move is actually not giving the student the answer.
Small Groups Work, but Smaller Is Usually Better for Personalization
Tutoring doesn't necessarily need to be one on one.
Research supports small group tutoring as well.
EdResearch for Action describes high impact tutoring models as generally using one on one instruction or small groups of up to approximately three students per tutor.
Small groups can have advantages.
Students can hear other students' reasoning.
They can solve problems together.
Programs can serve more students at lower cost.
But as groups become larger, one of tutoring's greatest advantages begins to disappear:
individualization.
If one student needs help with fractions, another needs help with factoring, and another already understands both, it becomes increasingly difficult to personalize instruction.
Tutoring Appears to Work Online Too
Parents sometimes assume tutoring must happen in person to be effective.
That isn't necessarily supported by the evidence.
A randomized controlled trial published in the Journal of Public Economics in 2024 evaluated a completely online mathematics tutoring program for secondary school students from disadvantaged neighborhoods.
Tutoring was delivered primarily by qualified math teachers after school.
Researchers found positive effects on mathematics test scores and grades. The program also reduced grade repetition and increased the likelihood that students passed the subject.
This doesn't prove that online and in person tutoring are always equally effective.
The quality of the tutor, student engagement, program design, technology, subject, and individual learner can all matter.
But it does challenge the assumption that a tutor needs to be physically sitting beside a student for meaningful learning to occur.
Personalized instruction can still be personalized through a screen.
Tutoring Is Especially Useful When a Student Has Fallen Behind
One of tutoring's biggest advantages is its ability to move backward.
Classrooms generally cannot do this indefinitely.
Suppose the class is learning quadratic equations.
Your student doesn't understand factoring.
The teacher may review factoring briefly.
But eventually, the class has to continue.
A tutor doesn't face exactly the same constraint.
They can say:
“Before we keep working on quadratics, we're going to spend some time fixing factoring.”
That can be enormously valuable.
Research reviews find tutoring particularly useful as an intervention for students who have fallen behind academically. J-PAL describes supplemental one on one and small group tutoring as an effective method for helping students learn, particularly those who are behind.
Sometimes moving forward requires going backward first.
Tutoring Can Work for Strong Students Too
Tutoring is often treated as remediation.
The student is failing.
Call a tutor.
But that is only one use.
A student might already be doing well and want to:
Prepare for an AP exam.
Improve an SAT or ACT score.
Move into an advanced mathematics course.
Strengthen academic writing.
Prepare for a difficult cumulative final.
Work through more challenging problems.
Develop better study habits.
In those situations, the tutor isn't necessarily repairing a major deficit.
They're providing targeted instruction and practice.
The research base is much larger for tutoring as an academic intervention than for every possible form of enrichment, so we should be careful not to generalize every research finding to every tutoring situation.
But the underlying instructional advantage remains:
One student. Specific goals. Immediate feedback. Instruction adjusted to performance.
Tutoring Shouldn't Become Permanent Academic Life Support
This is where even well intentioned tutoring can go wrong.
Imagine a student who has a tutor beside them every time they complete math homework.
The student reaches a difficult problem.
They immediately ask for help.
The tutor provides the next step.
Another difficult problem appears.
More help.
Homework gets finished.
Grades improve.
Everyone is happy.
Except the student may be becoming more dependent.
A useful question for parents is:
“Is my student becoming more independent?”
Good tutoring should gradually reduce how much assistance a student needs.
Perhaps initially the tutor models a problem.
Then they solve one together.
Then the student attempts one with prompts.
Then the student solves one independently.
Eventually, the tutor asks:
“Explain to me why your solution works.”
That's progress.
Homework Help and Tutoring Aren't Necessarily the Same Thing
There is nothing wrong with helping students complete homework.
Sometimes that's exactly what they need.
But parents should recognize the distinction.
Homework help asks:
“How do we finish this assignment?”
Tutoring asks:
“What does the student need to understand so they can eventually do this themselves?”
The best sessions may involve both.
A tutor can use tonight's homework as a window into the student's understanding.
But if every session consists entirely of completing assignments, there is a risk that the student becomes better at finishing homework without becoming substantially better at the subject.
How Can Parents Tell Whether Tutoring Is Working?
Don't evaluate tutoring exclusively by whether your student likes the tutor.
Rapport matters.
But liking someone doesn't necessarily mean learning from them.
Don't evaluate it exclusively by grades either.
Grades can change slowly and are affected by many factors.
Instead, look for multiple forms of progress.
Is homework becoming easier?
Does the student need fewer hints?
Can they explain concepts more clearly?
Are they making fewer repeated mistakes?
Can they solve problems without looking at examples?
Are they becoming more willing to attempt difficult problems?
Do they know what they don't understand?
Are test scores improving over time?
Is confidence increasing alongside competence?
One of my favorite questions is simple:
“Can you do this without the tutor sitting next to you?”
Ultimately, that's the test that matters.
What Should Parents Look for in a Tutor?
The research suggests several useful principles.
Look for someone who knows the subject well but also knows how to teach it.
Look for consistency.
Look for individualized instruction rather than a generic lesson.
Look for a tutor who asks the student questions rather than simply lectures.
Look for someone who identifies foundational gaps.
Look for regular opportunities for the student to work independently.
Look for progress monitoring.
And perhaps most importantly, look for a tutor whose goal is eventually to become less necessary.
The best tutor isn't the person who makes a difficult assignment look easy.
It's the person who helps the student become capable of handling difficult assignments independently.
An Important Limitation in the Research
There is one more point worth making.
Some of the strongest tutoring research comes from structured programs operating during the school day, often multiple times per week.
That isn't identical to the tutoring most families purchase privately after school.
Researchers have also found that results can become smaller when tutoring programs are expanded to much larger scales. A recent analysis of 282 randomized controlled trials found that estimates more closely resembling large scale U.S. tutoring programs remained meaningful, but were substantially smaller than estimates from the full collection of studies.
That's an important reminder not to turn:
“Tutoring works.”
into:
“Any tutoring program will produce enormous gains.”
Research doesn't support that conclusion.
Implementation matters.
Tutor quality matters.
Frequency matters.
Attendance matters.
Student engagement matters.
Instructional quality matters.
And the student's specific needs matter.
The Bottom Line
So, does tutoring actually work?
Yes. There is strong experimental evidence that well designed tutoring can meaningfully improve student learning.
A 2024 systematic review of 89 randomized tutoring studies found substantial positive effects overall, with benefits across mathematics and literacy.
But that's not the same as saying every student who hires a tutor will suddenly earn an A.
Tutoring appears to be most powerful when it is consistent, individualized, focused on actual learning gaps, delivered by capable tutors, and structured so that students do the thinking themselves.
And that's an important distinction for parents.
If tutoring becomes someone sitting beside your student and helping them finish every assignment, you may see short term improvement without solving the underlying problem.
But when tutoring identifies what a student doesn't understand, provides targeted instruction and practice, gives immediate feedback, and gradually builds independence, it can do something much more valuable.
It can help a student become a better learner.
And ultimately, that should be the goal.
Not creating a student who always needs a tutor.
Creating a student who increasingly doesn't.
