School punishes kids for learning at the wrong speed

The public school system is failing children.

In a study of more than 280,000 American fifth-graders tested in fall 2021, researchers used reading scores to estimate which grade’s lessons each student needed. Only 13% needed fifth-grade reading lessons. About 35% were ready for sixth-grade lessons or higher, while 52% needed lessons from fourth grade or below. They were all in fifth grade, but they didn’t all need to be taught the same thing. Reading study

When school sets one pace for a whole class, kids can be punished for learning too slowly and for learning too quickly. If you need more time, the teacher can start the next lesson while you’re still trying to understand the last one. If you already understand, you can spend the rest of the week doing more of the same work. And the same kid can be bored in one math lesson and lost in the next.

Being bored with schoolwork doesn’t necessarily mean a student is doing fine. A review of 29 studies involving middle-school, high-school, and college students found that students who were more bored tended to feel less like studying and get lower grades or test scores. Practice matters, but students need to be challenged, too. Boredom research

What happened to me

That lack of challenge became a real problem for me in ninth grade. Near the end of my first semester in 2025, I was nearly failing every class. I’d been in online school since seventh grade, but I still couldn’t choose my pace. Often, I understood the unit on day one and had a week’s worth of assignments left. Once I felt I’d learned it, I struggled to make myself finish the remaining assignments. The assignments piled up, and my grades fell.

That December, before the semester ended, I switched to homeschooling. A week later, I’d completed the credits for ninth grade with a perfect 4.0 unweighted GPA. I could take the unit tests and move past material I already knew without having to complete every assignment first.

Passing isn’t mastery

I was happy with that result, but being able to choose my pace didn’t fix everything about the system. The program only required 60% on a unit test to move on. A student could miss 40% of the points and finish the unit without ever having to go back and learn what they’d missed. If the next unit depended on those skills, they’d be starting it with something important missing. Even when the next lesson didn’t depend on it, why should passing mean we stop helping them learn the parts they got wrong?

This isn’t just a homeschooling problem. A 2018 study found that 36% of North Carolina students who earned a B in Algebra I scored below the level the state expected on its Algebra I exam. Their report cards said they were doing well, but they couldn’t show the same understanding on the state test. North Carolina study

A national study followed students who started ninth grade in 2009 and later went to college. Among those with high-school GPAs between 3.0 and 3.49, about 35% took at least one catch-up course covering material below college level. Even among students with GPAs of 3.5 or higher, the figure was about 18%. Good grades hadn’t guaranteed that they were ready for the next step. National study

School can feel like a conveyor belt: you’re expected to move with the group, even when you need something different. Teachers already try to account for this through extra help, small groups, and harder work for students who need it. But one person can only give so many children individual attention at once. Asking teachers to work harder doesn’t fix that.

What better pacing could look like

My experience showed me two things I’d want from a better system: the freedom to move past work I already understand, and help with anything I don’t. Those should come together. A child who can already add fractions should be able to show that with a few new problems and explain how they got the answers. They shouldn’t need another week’s worth of the same worksheet. A child who keeps getting stuck should get help with the part that’s confusing them: another explanation, a worked example, and a chance to try without help. Then check again later. They can keep moving forward in other areas while working on that skill.

Researchers have tested this kind of change in classrooms. In one study, children were followed from first through third grade. Each year, some teachers were randomly chosen to receive software and training that helped them choose reading activities based on each child’s skills. Children who got those lessons in first, second, and third grade finished with stronger reading skills than children who got them for only one or two of those years. Everyone still had a classroom teacher; the software helped the teacher decide what each child needed to work on. Individualized reading study

I built it for German

Before I went looking for these studies, I’d already built a German-learning app for myself. I just wanted a way to learn that worked for me. It’s a Duolingo-style app where algorithms decide what I practice next, and large language models, a type of AI, help adapt the practice and explanations around those decisions. If I know what a word means but keep misspelling it, the app focuses on the spelling. It cuts back on exercises I’ve already shown I can do, makes room for new words, and brings back things I struggled with. If I need a hint, that answer doesn’t count as proof I can do it myself.

It also brings older material back as I learn new things. Instead of saving a year’s worth of review for the month before a final exam, the app mixes words I’ve already learned into new lessons, so I keep practicing them alongside new vocabulary. Words I remember need less frequent review; words I forget come back sooner. The goal is long-term memory: knowing a word months later and being able to bring it into working memory when I’m putting a sentence together. Research on spreading practice over time supports this approach to remembering what we learn. Spaced-review research

There’s also evidence for mixing different kinds of practice, so students have to work out how to solve each problem instead of repeating the same method. A study tested this in 54 seventh-grade math classes. Over four months, some classes practiced different kinds of problems mixed together; others practiced one kind at a time. Both groups then completed the same mixed review. On a surprise test a month later, the students who had practiced the mixed problems averaged 61%, compared with 38% for the other group. Math-practice study

In another study, eighth-graders reviewed U.S. history in two ways. For some facts, they read the information again. For others, they tried to answer questions from memory, then checked the correct answers. Nine months later, they remembered more of the facts they’d practiced answering questions about. Trying to remember an answer had helped them remember it later. These studies took place in classrooms, with teachers still teaching. Long-term memory study

What I’d want schools to do

Using this approach in other subjects wouldn’t mean throwing away everything schools already teach. A school’s curriculum could supply the topics and lessons. We’d need to break them into clear skills and decide what a student should be able to do to show they’ve learned each one. Some existing school standards already ask students to explain their reasoning, not just produce an answer. I’d want a system that follows through on that for each child, changing the practice and support instead of accepting an overall passing grade. Example: elementary math standards

I don’t want every kid to finish ninth grade in a week. I want every kid to have the chance to master what they’re being taught: to do it on their own, explain it, and still be able to use it later. That means giving a struggling student the time and help to learn it, and letting a student who already understands take on something harder. A child shouldn’t have to leave their school to get that chance. I did.