Six students did not get it. So on Thursday they come to the back table, and you teach the lesson again. Same example, same steps, same words, just slower and with more encouragement. Some of them nod. On Friday's exit ticket, four of them make the same mistake they made on Tuesday.
This is the most common shape of reteaching in elementary math, and it is the least effective. Not because the teaching was bad. Because a student who did not understand an explanation the first time is unlikely to understand the same explanation the second time. If the route did not work, repeating the route is not a plan.
Reteaching is not repetition. It is diagnosis followed by a different route. Find the specific understanding that is missing, go back to a model that makes it visible, and rebuild from there. The procedure the student got wrong is almost never the thing that needs fixing.
Diagnose Before You Teach Anything
The first move in reteaching is not teaching. It is finding out what the student actually did.
Take a real error. A fourth grader computes 43 minus 27 and writes 24. There are at least three different students who produce that answer for three different reasons. One subtracted the smaller digit from the larger in each column because that is what felt legal. One tried to exchange and lost track of what the 3 became. One does not have a stable sense of what 43 or 27 are as quantities, so nothing about the answer looked wrong to them.
Those three students need three different lessons. Reteaching the algorithm helps at most one of them. So ask, before anything else: "Walk me through what you did." Then: "Is 24 more or less than 27? Does that make sense?" Two questions, and you usually know which student you have.
Try this tomorrow:
Before your next reteach group, sit with each student for ninety seconds and have them narrate one wrong answer out loud. Do not correct anything yet. Write down what they say. You will usually find that your six students split into two or three genuinely different groups, and the group you thought was one problem is actually two.
Go Back to a Model, Not to the Steps
Once you know the gap, reteach it with something the student can see. For 43 minus 27, the fastest route back to meaning is usually the open number line, because it turns subtraction into a distance instead of a column procedure.
Ask how far it is from 27 to 43. That question has nothing to do with exchanging and everything to do with quantity. Students jump to a benchmark, then to another, and count the jumps.
Open Number Line: 43 - 27 as a Distance
Three hops: 3, then 10, then 3. Sixteen. A student who counts that distance has answered the question without exchanging anything, and more importantly, they now have a result they believe. That belief is what you build the algorithm back onto later.
Notice the hop widths. The 10 is drawn much wider than the 3s because it is much larger. When a reteach drawing ignores proportion, it stops carrying information about size, which is exactly the thing the struggling student is missing.
Then Show the Same Problem a Second Way
One model is a rescue. Two models is understanding. After the number line, put the same numbers into a bar and ask a different question: if 43 is the total and 27 is one known amount, what is the amount we do not know?
Bar Model: Known Total, One Unknown Amount
Same numbers, same answer, completely different question. The number line asked how far. The bar asks how much is left over. A student who can move between those two readings has understanding rather than a trick, and you can hear it when they explain which one they would use for a new problem.
A single model can be memorized as its own procedure. Two models of the same problem force the student to attend to what the numbers mean rather than where they go. This is also the fastest way to find out whether the reteach worked: give a new problem and ask which model fits. A student who picks correctly understood something. A student who picks whichever one you used last did not.
Reteach the Understanding, Not the Grade
There is a strong pull, especially in the week before a unit test, to reteach whatever will be assessed. It is the wrong instinct and it costs more time than it saves.
Most persistent errors in fourth grade trace back to something from second. A student who cannot exchange in subtraction usually does not have a secure sense that ten units of one compose one unit of ten. Reteaching the exchange step will not fix that. Spending twenty minutes composing and decomposing tens with a bar model will, and it will also fix the three other things that gap was quietly breaking.
This feels like going backward. It is not. Going back one layer and rebuilding it properly is faster than patching the same surface error every unit for a year.
What Makes It Stick
A few things separate reteaching that holds from reteaching that has to happen again in three weeks.
Spacing beats massing. Three ten minute sessions across two weeks outperform one thirty minute session, reliably and by a wide margin. Split the reteach group time rather than extending it.
Vary the representation. Practice the same idea across a context, a visual model, and symbols. A student who has only practiced one form has learned one form. Mixing representations is uncomfortable and it is where durable learning comes from.
Make them explain. End every session with a claim: "I claim the answer is 16 because..." If the student cannot finish that sentence without pointing at a procedure, the understanding is not there yet, whatever the answer sheet says.
Let the wrong idea surface first. Ask for a prediction before you show anything. A student who predicts 24, then counts 16 on the line, has felt the contradiction. A student who is told 16 has only received information.
Try this tomorrow:
Take one problem your reteach group got wrong. Have them predict the answer, then build it on an open number line, then build the same problem as a bar. Ask which drawing matched the story better and why. Ten minutes, one problem, two models, and you will learn more about what they understand than a page of practice would tell you.
What to Stop Doing
Stop assigning more of the same practice. A student who is practicing a misunderstanding gets better at the misunderstanding.
Stop making the reteach group permanent. Groups should be formed around a specific gap and dissolved when that gap closes. A student who sits in the same seat all year learns something about themselves that is much harder to reteach than subtraction.
Stop leading with the answer. The instinct to relieve a struggling student by showing them the right way immediately is kind and it removes the exact moment where the learning would have happened.
Keep the language steady through all of it: unit, compose, decompose, partition, iterate, equal. When the reteach uses the same words as the original lesson, the student is being handed a second route to the same place rather than a second thing to remember.
Diagnose first: have the student narrate the error before you teach anything. Reteach with a model that makes the missing idea visible, then show the same problem a second way so the understanding cannot collapse back into a procedure. Space the sessions, vary the representation, and make students state a claim. Reteaching the same lesson slower is the one approach that reliably does not work.
Ready for Math Success?
Hundreds of research-backed resources, live teacher support, professional development courses, all for $21/month per teacher. No curriculum required. Just real help from math learning experts who get it.
Sign Up for Full Access ($21/mo)Also try our free Foundations masterclass → Start Free Course
Keep reading: Subtraction With Regrouping and Build the Number Line