The Review Question That Tells You Nothing

"So what did we learn today?"

Every kids lesson ends with some version of that question, and almost every version of it is worthless as a measure of whether anything stuck. I want to explain why, because once you see the mechanism you can't unsee it, and it changes how you'd actually write the back third of a lesson.

Here's what's happening cognitively. When you ask a class of eight-year-olds "what did we learn," the kid who raises a hand is pulling the answer out of working memory, which is the short-term buffer that holds whatever you said in the last few minutes. That buffer is loud and recent. It feels like recall. It isn't. Recall, the kind that predicts whether a kid will still hold the idea next Sunday, means pulling something out of long-term memory after the context that put it there is gone. Asking the question 90 seconds after you said the answer just measures echo.

I spent five years building curriculum for church and nonprofit programs before this, and I wrote those exact review questions into lesson plans. Confidently. "Recap: what was today's big idea?" I thought I was reinforcing. What I was actually doing was rewarding the three kids fastest at repeating the last thing they heard and teaching the rest of the room that the answer is always available in the air if you wait a beat.

The research term for the thing that actually works is the testing effect, sometimes called retrieval practice. The short version: trying to pull information out of your own head, with effort, with a little struggle, strengthens the memory far more than hearing it again or seeing it again. The struggle is the point. A review question that's easy to answer is doing almost no work. A review question that makes a kid go quiet for three seconds and frown is doing real work.

That's the part most children's ministry curriculum gets backwards. The review section is written to feel successful. Everybody answers, everybody's right, the teacher feels like the lesson landed, the kids feel smart, parents at pickup hear "we learned about Jacob today." Smooth. And mostly hollow, because nothing was retrieved against any resistance.

So what would you change. A few specific things.

Stop reviewing at the end of the lesson and start reviewing at the start of the next one. The gap is the mechanism. When you open this Sunday by asking what happened last week, before you've re-cued anything, you're forcing actual retrieval from long-term memory. That's the moment that tells you something true. If the room goes silent, that's not a failed opener. That's accurate data. It means last week didn't stick, which is information you needed and the end-of-lesson recap was hiding from you.

I'll be honest about the friction here. Teachers hate the silence. A cold-open review question that nobody can answer feels like you bombed, and your instinct is to rescue it: "Remember, it was the story about the coat?" The second you give that cue, you've turned retrieval back into recognition and killed the benefit. The discipline is to wait. Count to five in your head. Let one kid drag it up the hard way. That single effortful pull is worth more than ten smooth end-of-lesson recaps.

The other change is to ask for application instead of summary. "What did we learn" invites a label. "When would you actually use this" invites retrieval plus transfer, which is harder and more durable. A nine-year-old saying "I guess if my brother took my stuff I shouldn't get him back" is doing more cognitive work than the same kid reciting "the lesson was about forgiveness." One is a stored phrase. The other is the kid building a bridge from the idea to a situation, which is most of what retention even is.

Where does this break down. Two places. Preschool, mostly. Under five, working memory and the whole retrieval machinery isn't developed enough for the silence-and-struggle approach to do what it does in older kids, and a confused four-year-old just disengages. With that age you're aiming for repetition and recognition on purpose, and that's fine, that's developmentally right. Don't make a preschooler retrieve. The other place it breaks is when a kid's quietness isn't "can't retrieve" but "won't risk being wrong in front of the room," and those two look identical from the front. I don't have a clean way to tell them apart in the moment. If a kid is anxious, the silence isn't telling you about memory at all, it's telling you about safety, and pushing the wait time just makes it worse.

Which is the honest limit of all this. Retrieval practice is one of the better-supported findings in learning science, and it still depends on a kid feeling safe enough to try and fail out loud. The cognitive principle is solid. The room it lands in is not controlled, and a lot of what determines whether the effortful question works is stuff that has nothing to do with memory and everything to do with whether your eight kids trust you not to make a face when they get it wrong.

So if you change one thing this week, don't change how you teach the story. Change where you put the question. Move it from the end of today to the start of next time, and then sit in the quiet long enough to find out what your class actually kept.