You ran the activity. It was bumpy. Some students went quiet. A few handed you the answer they thought you wanted and waited to be told they were right. The discussion didn't catch the way you'd pictured it. You walk out wondering whether the whole thing was a misfire.
Here's the better read, and it's worth having before the next rough day, not after: the bumpiness is usually the skill forming, not the routine failing. The activity asks students to do something math class often hasn't asked of them, and the first time anyone uses a new skill, it's rough. That roughness is the thing working, not the thing breaking.
The bumpy first run and the smooth fourth run are the same activity
Start with what's actually new for students. Most of these routines ask them to talk about the thinking, not just produce the answer. To say why one method beats another, why an argument holds, what they'd need to see to change their mind. For a lot of students, that's a genuinely unfamiliar request. They've gotten good at producing answers. Producing reasoning out loud is a different skill, and they haven't built it yet.
So the first run is bumpy the way the first day of any new skill is bumpy. Nothing's wrong with the routine. What's missing is the fluency students develop by doing it more than once.
This is the part worth holding onto: the bumpy first run and the smooth fourth run are the same activity. The worksheet didn't change. The questions didn't change. What changed is that students learned how to do the thing it was asking for. A teacher who reads the first run as failure stops before that happens. A teacher who reads it as the beginning keeps going, and gets the fourth run.
Struggle is the mechanism, not the obstacle
There's a temptation, watching students strain, to step in and smooth the path: show the method, give the example, get them unstuck. Sometimes that's right. But a body of research suggests that the strain itself, kept in the right range, is doing more work than it looks like.
Manu Kapur's studies of productive failure found that students who wrestled with a hard problem before being taught the method ended up with deeper understanding and better transfer than students who were taught the method first and then practiced it (Kapur, 2008, 2014). The struggling group often got worse answers in the moment, and learned more in the end. The wrestling wasn't wasted time on the way to the lesson. It was part of the lesson.
The reason has to do with what struggle does to a learner's attention. Grappling with a problem you can't yet solve makes you aware of exactly what you don't know, and that awareness is what makes the eventual explanation land and stick (Loibl, Roll & Rummel, 2017). A student handed the method before they've felt the need for it has nowhere to put it. A student who's been stuck has a question the method answers.
Underneath all of it is a finding that holds across very different teaching traditions: the two features that consistently predict student learning are explicit attention to meaning and students genuinely struggling with important mathematics (Hiebert & Grouws, 2007). Not one or the other. Both, together. Which means a routine that has students working at something hard, then surfacing and discussing the thinking, isn't a departure from good instruction. It's a fairly precise description of it.
Not all struggle is the good kind
This is the honest caveat, and it matters, because "let them struggle" is easy to hear as "leave them stranded." It isn't.
Productive struggle is students grappling with mathematics that's within reach but not yet automatic. Unproductive struggle is students stuck with no way in at all, spending the period frustrated and learning that math is something that happens to other people. The first builds understanding. The second builds avoidance.
The teacher's job isn't to remove the struggle. It's to keep it in the productive zone. That's mostly about the problem and the entry: a low floor so everyone can start, a high ceiling so no one runs out of room, and more than one way in so a student who's blocked on one path can find another (Sullivan et al., 2020). When a student is genuinely stuck with no entry point, that's the signal to step in, not with the answer, but with a way to begin. When a student is straining but moving, that's the work. Letting that student sit in it a little longer is often the most useful thing you can do.
Part of what's hard is what students believe math is
There's one more layer under the bumpiness, and it's the slowest to shift.
Students read their own classroom for cues about what math is. In a room where the answer ends the conversation, they come to believe math is fast procedure-recall: you either know it or you don't, and if you don't get it quickly, you're not a math person. Alan Schoenfeld documented how that belief quietly governs behavior. Students who think math problems should yield in a couple of minutes give up the moment one doesn't, because persistence feels like evidence they're failing (Schoenfeld, 1989).
So when these routines ask students to stay with a hard problem, they're pushing against a belief, not just a skill gap. Part of what the early runs are doing is teaching students that being stuck for a while is normal, expected, and not a verdict on them. That belief changes slowly, by being lived in a classroom over and over, not by being announced. The bumpy runs are where it starts changing.
The question to ask after a rough run
So, the move for the day after a hard one. When a run doesn't go the way you hoped, the instinct is to ask "did this work?" That question doesn't lead anywhere useful, because it points at the routine, and the routine is rarely the problem.
Ask a different one: what were students being asked to do that they haven't had to do before?
That question relocates the difficulty from the activity to the skill, which is where the actual fix lives. Maybe it was the first time they'd been asked to justify a choice out loud. The first time a "wrong" answer was treated as worth discussing. The first time the answer was the start of the conversation instead of the end. Name that, and the next run has a job: not "make the activity work," but "give students another rep at the thing that was new." That's a problem you can solve. "Did this work?" isn't.
The bumpy first run was the skill forming. The fourth run is what it looks like once it has.
References
Hiebert, J., & Grouws, D. A. (2007). The effects of classroom mathematics teaching on students' learning. In F. K. Lester (Ed.), Second handbook of research on mathematics teaching and learning (pp. 371–404). Information Age Publishing.
Kapur, M. (2008). Productive failure. Cognition and Instruction, 26(3), 379–424.
Kapur, M. (2014). Productive failure in learning math. Cognitive Science, 38(5), 1008–1022.
Loibl, K., Roll, I., & Rummel, N. (2017). Towards a theory of when and how problem solving followed by instruction supports learning. Educational Psychology Review, 29(4), 693–715.
Schoenfeld, A. H. (1989). Explorations of students' mathematical beliefs and behavior. Journal for Research in Mathematics Education, 20(4), 338–355.
Sullivan, P., Bobis, J., Downton, A., Feng, M., Hughes, S., Livy, S., McCormick, M., & Russo, J. (2020). Exploring a framework for teaching mathematics: Focusing on student engagement. In J. Way, C. Attard, J. Anderson, J. Bobis, H. McMaster, & K. Cartwright (Eds.), Research in mathematics education in Australasia 2016–2019 (pp. 241–266). Springer.