Why 20 Minutes A Day Beats A 90-Minute Tutoring Session
- How often should your child practice math? The answer might surprise you.
- Research on distributed practice consistently finds that spreading study across sessions beats massing it into one block, for the same total time.
- Tutoring studies point the same way: three or more sessions a week produces real gains, once a week mostly doesn’t.
- One review found high-dosage tutoring roughly twenty times more effective in math than low-dosage tutoring.
- The reason weekly sessions dominate anyway is scheduling and hourly pricing, not evidence.
- Four twenty-minute sessions beat one ninety-minute session, and they’re also easier to get a kid to do.
Ask a parent how their child’s tutoring is going and you’ll usually hear about a Saturday morning session. Ninety minutes, once a week, at the kitchen table or a learning center.
It’s the worst possible schedule, and almost nobody chooses it for pedagogical reasons.
What the research says
The spacing effect is one of the oldest and most replicated findings in the study of learning. Give two groups the same total practice time. Give one group all of it at once and the other group the same amount spread across several sessions. The spread group retains substantially more, and the advantage grows the longer you wait to test.
This holds across ages, subjects and skill types. It holds for math specifically. It’s not a marginal effect.
The tutoring research lands in the same place. Reviews of high-dosage tutoring, the kind that actually moves achievement, converge on three or more sessions per week of about thirty to sixty minutes. Large programs serving between 500 and 7,000 students produced average effect sizes around 0.25 standard deviations, with meta-analysis estimating gains of roughly three to fifteen months of learning. One review found high-dosage tutoring about twenty times more effective in math than low-dosage tutoring.
The same reviews find little evidence that once-weekly tutoring produces meaningful improvement.
Why the massed session fails
Three things happen in a ninety-minute block that don’t happen in four twenty-minute ones.
Fatigue. The last thirty minutes of a long math session are worth much less than the first thirty. Anyone who’s watched a ten-year-old at minute seventy already knows this.
No forgetting, so no retrieval. This is counterintuitive but it’s the core of the spacing effect. When practice is bunched together, the material is still sitting in working memory and you never actually have to retrieve it from long-term storage. Retrieving something you’d partly forgotten is the act that strengthens the memory. Spacing forces that. Massing prevents it.
Six days of drift. Whatever happened on Saturday has six days to decay before anyone touches it again. By the next session you’re re-teaching rather than building.
Why weekly sessions dominate anyway
Not because anyone believes in them. Because of two constraints.
Scheduling. Getting a tutor and a student in the same place at the same time is hard, so you do it as few times as possible and make each one count.
Price. At $50 an hour, three sessions a week is $600 a month. Very few families do that. So they buy one longer session, which feels like getting more for the money and isn’t. The full cost picture is here.
Both constraints come from a human tutor’s time being the product. Neither has anything to do with how children learn.
What a better week looks like
Four sessions, fifteen to twenty-five minutes each. Tuesday, Wednesday, Friday, Sunday. Not consecutive if you can help it, since a day off between sessions is part of what makes spacing work.
Same skill across sessions, not four different topics. The point is repeated retrieval of the thing that’s broken.
Stop while it’s still going well. Ending a session at a good moment matters more than finishing the worksheet. A kid who ends on a win comes back willingly, and willingness is the whole game over a school year.
Mix in old material. Two or three problems from something they learned three weeks ago, every session. This is interleaving, and it’s the other half of the spacing research.
The scheduling problem is a pricing problem
You can’t run this schedule with a $60-an-hour tutor. Not because it wouldn’t work, but because it costs $960 a month and nobody’s doing that.
That’s why Math Anon prices monthly with unlimited sessions: $79 for one student, $129 for up to three, $199 for up to five. Twenty minutes on a Wednesday costs the same as ninety minutes on Saturday, which removes the reason to ration.
Removing the meter changes behavior in a way that’s easy to underestimate. When each session has a price, you save up questions for it. When it doesn’t, your kid asks the question the night it comes up, which is the night they can still remember what confused them.
Short sessions also work better for a kid who’s tense about math, since twenty minutes is a much easier thing to agree to than ninety. More on that here.
Try the schedule. Start with the assessment so you know what to practice, then run four short sessions a week for a month and see what happens.
Related reading
