The argument

It pays for effort, not for grades

Written 10 September 2026 by Daniel, who built this

Most of the case against paying kids for schoolwork is really a case against paying for grades. That is a narrower claim than it first appears, and the difference between rewarding a report card and rewarding an hour of work turns out to be the part the research keeps landing on.

The objection, at full strength

It goes roughly like this. Attach money to a grade and you have told your child that the grade is the point. So they optimize for the grade. They take the class they can coast in rather than the one that would stretch them. They do the minimum that scores and stop. They negotiate. And on the day you stop paying, the reason to work disappears with the payment, because you spent a year teaching them that the reason to work was the payment.

I do not think that is a weak argument. I think it is basically correct, and any parent who feels uneasy handing a child five dollars for an A is responding to something real.

What the objection is actually about

Read it again and notice what every step depends on. Chasing the number. Taking the easy class. Doing the minimum that scores. Every one of those is a story about a grade, which is an outcome the kid only partly controls, arrives weeks after the work, and can be reached by several routes, not all of them learning.

That distinction is not something I invented to be convenient. The best known research on this is a set of large randomized experiments the economist Roland Fryer ran across several US school districts, and the result people usually skip past is that the design of the incentive mattered more than whether there was one. Paying children for outputs, for test scores and grades, produced very little. Paying them for inputs, the concrete things they could actually do that day, produced meaningfully more. Dallas paid second graders a small amount per book read, and reading scores moved.

One reading of that is unflattering to the whole idea and worth sitting with: a young child told to go get better grades often does not know what to actually do next, so the money floats free of any action and buys nothing. Told to read a book, they know exactly what to do. The incentive works when it is attached to a behavior rather than to a result, which is a fairly narrow thing to conclude and not a license to pay kids for whatever you like.

The paper is Financial Incentives and Student Achievement: Evidence from Randomized Trials, and I have written up what it found, what a second Fryer study in Houston found that is considerably less flattering, and where the whole literature actually disagrees with itself, on a separate page about the research.

So what this pays for

There is no grade anywhere in PiggyBankMath. No report card to upload, no test score to hit, and nothing that talks to a school. A kid gets paid for finishing a round of timed math practice. That is the whole trigger.

A round is somewhere between ten and twenty questions depending on age, it takes a few minutes, and it pays whatever rate you set. Finishing it is a thing a nine year old can decide to do on a Tuesday, which is the entire point. It does not require the week to have gone well.

Getting the answer wrong still counts as work

This is the part I would want to know as a parent. A kid who finishes a hard round and gets a chunk of it wrong has still done the work, and the app still pays for the session. What accuracy changes is mastery, not money: an answer that is correct but slow does not count toward a fact being learned, because a slow correct answer usually means fingers were involved.

The result is that grinding easy problems does not pay better than attempting hard ones. A kid cannot farm 2 + 2 for forty rounds and come out ahead, so there is no version of this where the profitable move is to avoid the difficult thing. That was the failure mode the original objection warned about, and it is the one I most wanted closed.

The cap says the work still counts after the money stops

Every kid has a daily maximum you set. When the day's earnings reach it, the money stops. The practice does not: mastery still records, the streak still counts, the range still widens. A kid who wants to keep going after the cap is never told the work stopped mattering, only that the budget did.

That is a deliberate answer to the sharpest version of the worry, which is that paid work teaches a child that unpaid work is worthless. I do not know that a daily cap fixes that. I know that stopping the money while visibly continuing to count the work is the most honest thing I could build.

What I am still unsure about

I would rather write this down than leave it out.

Fryer's input experiments paid for reading books, not for math drills, and I am extending that finding to a neighboring activity rather than citing a study of this exact thing. I think the extension is reasonable, since a completed practice session is an input in the same sense that a finished book is. It is still an extension.

Two specifics cut against me and belong on this page rather than buried somewhere kinder. The first is that when Fryer looked at math test scores across all four cities, he found nothing: the input designs trended slightly positive, the output designs slightly negative, and every estimate was too small and too noisy to call anything but a null. The reading result is the one that held up, and this is a math app.

The second is what happened after the money stopped. A year on from the Dallas experiment the reading effect had fallen to roughly half its size and was no longer statistically significant. That is one follow-up, in one city, on one subject, so it does not settle the question either way, but it is the closest thing to an answer anybody has and it is not a flattering one. My own view is that this works best as a way to start a habit during the stretch where math feels like nothing but a wall, and that the sensible end state is a kid who has forgotten the money is part of it. That is a hope with a mechanism behind it, not a finding. The full read of the research is here, including the parts that argue against this whole idea.

And a family that decides none of this belongs in their house has made a completely reasonable call. I built it for my own two kids because I wanted the record of what was earned and what was paid, and could not find that anywhere. It might not be for yours.