Tutoring
What does a tutor cost, and what are you paying for
Federal wage data puts the median tutor at about $21 an hour, while private tutoring is commonly advertised far above that. The gap explains more than the price does.
6 min read
Key takeaways
- US federal wage data for May 2025 puts the median tutor at $20.84 an hour, with the middle half earning between about $16.83 and $26.38.
- What a family pays is usually higher than what a tutor earns, because agencies, platforms and centres take a margin.
- Price tracks credentials, subject scarcity and local cost of living far more than it tracks results.
- Frequency matters more than any other purchasing decision: the research finds the largest effects where tutoring happens at least three times a week.
- Occasional expensive tutoring is usually worse value than regular cheaper tutoring, which is the opposite of how most families buy it.
US federal wage data for May 2025 puts the median tutor at $20.84 an hour, with most tutors earning between roughly $17 and $26. What families pay is typically higher than what tutors earn, and the most important decision when buying tutoring is not the hourly rate at all. It is how often the sessions happen.
What the data says
The Bureau of Labor Statistics tracks tutors as their own occupation. From the May 2025 Occupational Employment and Wage Statistics:
| Percentile | Hourly wage |
|---|---|
| 10th | $14.15 |
| 25th | $16.83 |
| Median | $20.84 |
| 75th | $26.38 |
| 90th | $36.53 |
About 175,000 people are employed as tutors nationally, at a mean hourly wage of $23.10.
An important caveat: these figures cover employed tutors reported by employers. Independent tutors setting their own rates, and the agency margins a family pays on top, are not what this measures. It is a floor and a reference point rather than a price list.
Why what you pay is higher
A rate quoted to a parent usually includes more than the tutor's time:
- Recruitment and vetting, including background checks
- Scheduling, rescheduling and payment handling
- A replacement if the tutor is ill or leaves
- The company's margin
So a family paying an agency may find the tutor in the room receives roughly half the rate. That is not automatically poor value. Vetting someone who will be alone with your child has real worth, and so does having a replacement when the relationship does not work. But the premium is buying administration and insurance, not teaching quality.
What actually moves the price
In rough order of influence:
- Local cost of living. The same qualifications cost very different amounts in different metropolitan areas.
- Subject scarcity. Advanced maths, physics, chemistry and standardised test preparation command more because fewer people can teach them.
- Credentials. Qualified teachers charge more than undergraduates, and in this case the premium has some evidence behind it.
- Format. One-to-one costs more than small group, in person more than online.
- Agency or independent. Discussed above.
Note what is missing from that list: demonstrated results. Almost no tutor or agency can show evidence of their effect, so price is set by what the market will bear for a given set of credentials.
The decision that matters more than price
A 2024 meta-analysis of 89 randomised tutoring experiments found a pooled effect of 0.288 standard deviations, which is large for an education intervention. More useful than the headline number is what the authors found predicts it:
- Programmes using teachers or paraprofessionals as tutors beat volunteer and parent tutoring.
- Effects are strongest in earlier grades.
- Programmes running at least three days a week do better than less frequent ones.
- Tutoring during the school day beats after-school tutoring.
That third point has direct financial consequences. Most families buy one session a week, because that is what a weekly budget supports at $40 to $80 an hour. The evidence favours three. If a budget forces a choice between one session a week with an expensive tutor and three with a cheaper one, the research points at the cheaper one.
A worked comparison
For a student needing help across one school year:
| Option | Rough annual cost | Sessions per week | What the evidence says |
|---|---|---|---|
| Weekly private tutor at $60 | ~$2,200 | 1 | Below the frequency where effects are largest |
| Twice weekly at $40 (group of 2) | ~$2,900 | 2 | Closer, and often better value |
| Tutoring centre package | $2,000 to $5,000 | 1 to 2 | Varies enormously; ask about group size |
| School-provided intervention | Free where offered | Often 3+ | The configuration the research likes best |
| AI tutor subscription | $100 to $500 | Unlimited | Cheap and frequent; no comparable evidence base |
These are planning figures, not quotations. Rates vary widely by area and subject.
What each format costs, and what you give up
The hourly rate is only one of four things that change the real price of tutoring. The others are group size, travel and commitment length.
One to one, in your home. The most expensive format, because you are also paying for the tutor's travel and for a slot they cannot fill with anybody else. Typically the top of whatever the local range is.
One to one, online. Usually cheaper than the same tutor in person, because the travel disappears and a cancelled session costs the tutor less. For secondary and university subjects where the work is reading, writing and problem-solving, the research gives no reason to prefer the room, provided the tutor can write and draw in a shared space.
Small group, two to four students. Often half the per-student cost of one to one. The meta-analysis found positive effects for group ratios as well as one to one, so this is not obviously worse value - but it stops being tutoring and starts being a small class once the group grows past about four, and the student who most needs attention is the one who gets least of it.
A centre or franchise package. Usually sold as a block of hours with an assessment up front. The assessment has real value; the block has a real cost, and the group size during those hours is the question to ask before signing. Packages are also where the largest gap tends to sit between the price a family pays and the wage the tutor receives.
University tutoring centres and school intervention. Free where they exist, and usually staffed by people who know the specific course. The research likes during-school delivery most of all. Almost nobody asks whether these exist before paying privately, and they very often do.
Why frequency beats rate, in numbers
Take a household with $2,500 for the year, which is a realistic figure for a family that has decided tutoring matters.
| How it is spent | Sessions per year | Pattern |
|---|---|---|
| $75 an hour, weekly in term | ~33 | One session a week |
| $45 an hour, twice weekly | ~55 | Two a week |
| $25 an hour in a group of three, three times weekly | ~100 | Three a week |
The evidence favours the bottom row and most families buy the top one, because the top row is how tutoring is sold. If a cheaper tutor or a small group is the difference between one session and three, the research says take the three.
The exception worth naming: a genuinely specialist need - a diagnosed learning difficulty, an unusual exam, a subject at the very top end - is a case where one expensive hour with the right person beats three with the wrong one.
Questions worth asking before you pay
- How many students are in a session?
- Who is the specific tutor, and what are their qualifications in this subject?
- What happens if the match does not work?
- How will we know in six weeks whether this is working?
That last one is the question almost nobody asks, and the one most likely to change the outcome. A tutor who answers it with something checkable is a better bet than one who charges more.
Where TruLearn fits
TruLearn is free during its private beta, and the intention afterwards is a monthly subscription costing well under one hour of human tutoring. The point of that pricing is frequency: the evidence favours three or more sessions a week, and an hourly model prices most families out of that.
The honest limitation, which this article's own standard demands we state: human tutoring has decades of randomised evidence behind it and we have none of our own yet. We say so on our research page. Our argument is that the mechanics we have built - a plan, teaching out loud, explain-back graded against a human-written rubric, scheduled review - are the ones the evidence favours, not that we have proved an outcome.
More: AI tutor or human tutor, and what makes tutoring work.
Frequently asked questions
- How much does a private tutor cost per hour?
- Federal wage statistics for May 2025 put the median tutor wage at $20.84 an hour, with the 25th to 75th percentile running from about $16.83 to $26.38 and the 90th percentile at $36.53. Families often pay more than this, because the rate a tutoring company charges includes its own margin on top of what the tutor is paid.
- Why are agency rates so much higher than those wages?
- An agency carries recruitment, vetting, scheduling, payment handling and its own profit. A rate quoted to a parent typically covers all of that, so the tutor in the room may be receiving roughly half. This is not necessarily bad value, since the vetting and the replacement guarantee have real worth, but it is worth knowing what you are buying.
- Is a more expensive tutor better?
- Price correlates with credentials, subject scarcity and local costs, not reliably with outcomes. The research finds that the tutor's training does matter - programmes using teachers and paraprofessionals outperform volunteer tutoring - but within the professional market, a higher rate is weak evidence of a better result.
- How often should tutoring happen?
- The evidence favours frequency. A 2024 meta-analysis of 89 randomised studies found larger effects for programmes running at least three days a week. A weekly session is the most commonly purchased pattern and sits below the threshold where the evidence is strongest.
- Is AI tutoring cheaper?
- Substantially, because the marginal cost of a session is a model call rather than an hour of a person's time. The honest caveat is that AI tutoring has nothing like the evidence base that human tutoring has, so the comparison is between a well-studied expensive option and a cheap option whose effects are not yet established.
Keep reading
Tutoring
What the research says actually makes tutoring work
Tutoring is one of the best-evidenced interventions in education, but the effect depends on who tutors, how often, and when. Most purchased tutoring misses on frequency.
Oct 3, 20265 min read
Tutoring
Is ChatGPT a good tutor?
It is an excellent explainer and a poor teacher, and the difference is not about model quality. A tutor decides what to teach, checks it landed, and refuses to do the work.
Oct 3, 20265 min read
Tutoring
How to choose a tutor
Most families choose on price and availability, then hope. Five questions, asked before you pay, that predict whether tutoring will work.
Oct 3, 20265 min read