Let them struggle first, then teach
Ten minutes of failing at a problem they haven't been taught makes the teaching that follows land twice as hard.
Aug 15, 2026 · 1 min read
Give students a problem just beyond what they've been taught, let them generate their own imperfect attempts, and only then teach the canonical method. In Kapur's studies this "productive failure" sequence beat instruction-first on conceptual understanding and transfer, even though the struggling students looked worse during the lesson (Kapur, 2014). The failed attempts do real work: they make students notice exactly what the proper method solves. (Manning's education-agent-skills library packages this as the productive-failure-protocol skill.)
In practice
- Pick a problem where intuition gets you partway: rich enough for several wrong approaches, close enough that the real method feels like relief when it arrives.
- Collect two or three student attempts on the board before teaching, and teach by comparing the method against them: "here's what your version couldn't handle."
- Set the frame out loud: "I don't expect anyone to solve this yet. I want good attempts." Struggle without that frame reads as unfairness.
- Not for everything: it needs prior knowledge to build on, and a debrief that always happens. Struggle without the teach is just failure.
Quick takeaway: the struggle isn't in the way of the lesson. It's the part that makes the lesson stick.