The Review Discussion Nobody Actually Reviews Anything With

Almost every lesson plan I've ever opened ends the same way: a short block labeled "Review" with three or four questions that ask kids to repeat the thing they heard eleven minutes ago. "What did Jesus say to the paralyzed man?" "Where did the storm happen?" And the teacher reads them off, a couple of confident kids answer, everyone nods, and the lesson closes. That's not review. That's a verbal attendance check.

I spent five years building curriculum before I joined the design side here, and I wrote those exact questions. Dozens of them. I thought I was reinforcing the content. What I was actually doing was testing whether the loudest kid in the room was paying attention in the last ten minutes, which, congratulations, he was.

The problem is that recall at the end of a lesson tells you almost nothing about retention. Cognitive science has a boring, well-replicated finding here: information you just heard sits in working memory for a short window and then mostly evaporates unless something forces you to retrieve it later, after you've had a chance to forget it. The uncomfortable part for lesson planners is that the forgetting is the mechanism. You need a little bit of struggle to pull the memory back for it to actually stick. Asking a kid to repeat a fact ninety seconds after you said it skips the forgetting entirely. No struggle, no retrieval, no durable memory. Just an echo.

So the end-of-lesson review that most curriculum defaults to is doing the one thing least likely to help anyone remember anything.

What spaced retrieval actually looks like in a 30-minute room

Here's the frustrating catch. The research on retrieval practice and spacing is some of the most solid stuff in all of learning science, and it's also genuinely hard to apply in a ministry setting, because you see most of these kids once a week for thirty minutes and roughly forty percent of them weren't there last Sunday. You can't space something across sessions when attendance is a coin flip.

What you can do is move the retrieval instead of adding it. Don't review at the end. Open the next lesson with the retrieval from the previous one. Not "here's a recap," but an actual question the kids have to dig for. "Two weeks ago we talked about a guy who got lowered through a roof. Nobody tell me the story yet. Why would four grown men tear a hole in somebody's house?" Then wait. Let it be awkward. The kid who has to reconstruct that from partial memory is doing the exact cognitive work that builds retention, and the kid who missed that week gets pulled into the story by his peers instead of by you.

I've watched this land and I've watched it die, and the deciding factor is almost always whether the teacher can tolerate the silence. That's it. The technique is simple. The teacher's nervous system is the obstacle.

Why I stopped trusting the questions a tool hands me

This is where I'll be honest about the thing I'm supposed to be enthusiastic about. An AI lesson plan generator for churches will happily produce a tidy review section in half a second, and most of them generate exactly the recall questions I've been complaining about, because that's what's in the training data. That's what review sections have looked like for forty years. The tool isn't wrong. It's just reflecting the convention back at you, confidently.

So the design decision that actually matters isn't whether you use a generator. It's what you ask it for.

• Ask for retrieval prompts, not recap questions.
• Ask it to write the opening of next week's lesson as a retrieval of this week's, which most teachers never think to request and which solves the spacing problem without adding a single minute to the lesson.
• Ask for questions a kid can't answer by parroting the last sentence you said, which forces a different kind of prompt and, in my experience, noticeably better output.

I don't have a clean answer for the attendance problem. Spacing assumes continuity, and ministry rarely gives you continuity, so some of the cleanest findings in cognitive science are partially broken the moment they hit a real Sunday school hallway. I'm not going to pretend otherwise. What I'm fairly confident about is narrower: the review block at the end of your lesson, the one every template includes, is the weakest tool in the plan. Move the work to the front of the next session and let the kids actually forget a little first.

That's the whole adjustment. It costs nothing. It just requires giving up the satisfying sound of a kid repeating the right answer back to you on your way out the door.