On the tension between teaching to the test and teaching for keeps — and what cognitive science says you can do about both

Right about now you're somewhere in that particular stretch of the school year where the pacing guide is a work of fiction, state testing is either already done or breathing down your neck, and you're doing the quiet math in your head: how much can we actually cover before the end? What do they actually remember from October? What counts as review vs. just hoping for the best?

And underneath all of it, there's a tension most teachers carry but rarely name out loud: you want your students to do well on this test, and you know that a lot of what gets reviewed the week before a test is going to evaporate within two weeks of taking it. You're trying to prepare them for Friday and trying to give them something that actually sticks. Those aren't always the same goal.

This post is about both. It won't pretend the tension doesn't exist. But it will give you a sharper lens for making review decisions — and some structures you can use this week that serve both goals at once.

– The Shift

Here's the experience almost every teacher has had. You run a review session. Students are engaged — maybe there's a Kahoot going (this is me, all the time, even while giving PD sessions), maybe it's a whole-class Q&A, maybe you've got a review game that took you an hour to make. The energy is good. They know this stuff. You feel cautiously optimistic. Then Friday's test comes back and half the class tanks it, including students who were confidently shouting answers two days ago.

What happened? They weren't faking it in the review. They genuinely recognized the answers when they saw them. The problem is that recognition and retrieval are different cognitive skills — and tests demand retrieval.

Recognition is easy: given options, pick the right one. Retrieval is harder: from nothing, produce what you know. A quiz game is fundamentally a recognition exercise. Multiple choice, context provided, answer options visible. A test strips all of that away and asks students to produce from an empty page. If your review only practiced recognition, you've inadvertently built a confidence illusion — students feel prepared because they could identify the right answer when they saw it. The scaffold disappears on test day.

– Five Structures Worth Using This Week

Cold Recall at the Start of Class (5-7 minutes)

Write three questions on the board from content covered two to three weeks ago. Students answer on paper, no notes, before anything else happens. Collect nothing. Correct it together quickly. Repeat every class for the days you have left. This is distributed retrieval practice built into the class structure with essentially zero additional planning.

Blank-Page Retrieval + Gap Check (10-12 minutes)

Students get a blank piece of paper and write everything they can remember about a topic — no notes, no prompts. Eight minutes. Then they open their notes and circle what they missed. The circle is the critical step: students are now looking directly at their own gaps, in real time, before the test. This is genuinely uncomfortable for students who thought they knew it. That discomfort is diagnostic information they can actually act on.

Intentionally Mixed Problem Sets

Take a review set and scramble it: questions from Unit 1, Unit 3, and Unit 5 in the same assignment rather than separated by unit. Tell students explicitly why — "I'm mixing this on purpose. On the test you won't be told which unit a question belongs to. This is practice for that." Students who understand why something is hard will push through it. Students who just think the assignment is confusing will give up (and that is a great data point for you to leverage).

Two-Column Error Analysis (after any practice test)

Students divide a page: what I wrote versus what was correct. For each error: one sentence on why they got it wrong (didn't know it, misconception, misread the question), one sentence on what they'd do differently. This turns test review from passive grade-checking into active metacognitive work. Students who can name their own errors are significantly better positioned than students who just see a score.

Teach It Back (15 minutes)

Students write a 90-second explanation of one key concept as if explaining it to a student who missed that unit entirely. No notes. Then they compare with a partner: what did each explanation include, what did each miss? The act of constructing the explanation forces retrieval and organization simultaneously. The comparison surfaces gaps that neither student noticed alone. And when a peer explains something differently than you did, it sometimes clicks in a way your explanation never quite managed.

– The Honest Part

Some of this is going to feel boring to students compared to a competitive game, and that's a genuine tradeoff. You know your students and you know what they need to stay in the room emotionally, not just physically. Engagement matters — a checked-out student isn't retrieving anything regardless of how theoretically sound the activity is.

But here's something worth saying directly to your students, and it tends to land: The reason this feels harder than Kahoot isn't that you don't know it. It's that your brain is doing something different. It's not just recognizing an answer — it's actually building the memory. Most students have never had anyone explain to them why effortful practice works better than easy practice. When you name it, many of them find it genuinely interesting, and some of them will carry that understanding into how they study for things long after they've left your class.

Which brings us back to the tension from the opening: preparing students for Friday vs. giving them something that lasts. The structures above serve both, more than most review approaches do. Retrieval practice produces better test performance and better long-term retention. Spaced repetition works for Friday and for the week after.

– Why It Works

There are a few key findings supporting this framework, but always remember — you know your context. These research items provide tested perspectives, so use these to run your own local research on what works best in your classroom.

Being Tested Is Better Preparation Than Restudying

Roediger and Karpicke's 2006 research on the testing effect showed something that still surprises people: students who took repeated retrieval tests on material remembered significantly more on delayed tests than students who spent the same time rereading their notes. Even imperfect retrieval — even getting it wrong — strengthened memory more than re-exposure. The act of pulling information from memory is itself a learning event, not just a measurement of learning.

Translation: any review activity that requires students to produce information from nothing is doing more cognitive work than one that asks them to select from options. The effort isn't a sign the activity isn't working. The effort is the activity working.

Memory Is the Residue of Thought

Cognitive psychologist Daniel Willingham puts it plainly in Why Don't Students Like School?: students remember what they think about. Not what they encounter. Not what they recognize. What they genuinely have to think hard about.

This one question — what will students actually be thinking about while they do this? — is worth asking about every review activity you design. If the answer is "how to beat their friend's score," that's what gets encoded. If the answer is "what caused the Civil War" or "how to solve this type of equation," that's what gets encoded. Same activity. Completely different cognitive experience depending on how it's designed.

Cramming Is a Short Game

Cepeda and colleagues' meta-analysis of 317 experiments is unambiguous: study distributed across multiple sessions significantly outperforms the same total time concentrated in one session. Cramming produces a real short-term gain (which is why students keep doing it and why it sometimes seems to work) but it collapses fast. Information packed in the night before fades within days.

The honest implication for right now: if you're a week out from a test, spacing out studying is not really an option. What you can do is run short retrieval touchpoints across the days you have left — five minutes at the start of class, not one ninety-minute cram session the day before — and you'll outperform the cram with less total time. Five sessions of ten minutes beats one session of fifty.

Mixing It Up Makes It Stick

Most review sessions are blocked: we go back through Unit 1, then Unit 2, then Unit 3 in the order we taught them. This feels organized. It's also, according to Rohrer and Taylor's interleaving research, significantly less effective than mixing topics together in a single session.

When students work through a mixed set (where they don't know in advance which unit's tools apply) they have to do two things instead of one: remember the content and identify which concept or strategy applies to this particular problem. That extra step is harder. It's also why interleaved practice produces a significantly larger benefit to performance on delayed tests. The difficulty is the mechanism. A review session that feels frustrating and effortful is often working better than one that feels smooth.

Teaching It = Learning It

When students have to explain something to someone else, they learn it more deeply than when they prepare to be tested on it themselves. Research on what's sometimes called the protégé effect consistently shows that preparing to teach material pushes students toward the kind of organized, gap-aware thinking that passive review never quite gets to. They notice what they don't know when they have to explain it to someone else. They organize information differently when they're responsible for someone else understanding it. That reorganization is the learning.

– The Traps

The confidence illusion trap. Review that produces visible engagement and correct answers in the room can mask shallow learning. Students who correctly identify answers in a group review context often can't produce those same answers alone under test conditions. The visible confidence isn't lying exactly — it's just recognition, not retrieval.

The sequence trap. Reviewing in the same order you taught means students rebuild fluency within each unit but can't identify which unit's tools apply when a test mixes them. Real tests mix them. Review should too.

The one-big-session trap. A two-hour review the day before a test feels like serious preparation and produces genuinely worse results than six fifteen-minute sessions across the preceding weeks with the same total time. This is one of the most counterintuitive findings in memory research and one of the most practically important. If you're looking for permission to stop treating review day as a single high-stakes event: here it is.

The guilt trap. This one isn't in the research. It's just real. Many teachers feel vaguely guilty about using games and engagement tools for review, like they should be doing something more rigorous. And many feel guilty about not doing more review, period. But remember, the most important thing is whether students are retrieving, producing, and thinking about the actual content.

– Your Move This Week

1. Tomorrow: Write three questions on the board from content you haven't explicitly reviewed in two weeks. Let students answer cold before class starts. Don't grade it. Correct it together. Do it again the next day with three different questions.
2. This week: Take one existing review activity (whatever you'd normally do) and ask Willingham's question: what will students actually be thinking about while they do this? If the honest answer is "the game mechanics" or "copying from the board," add one element that makes retrieval unavoidable. Close notes. Remove the multiple-choice options. Ask them to produce before they check.
3. Before the next test: Build one mixed review set that pulls from at least three different units or topics without labeling which is which. Tell students it's intentional. Watch what they do with the ambiguity. That's exactly the situation the test puts them in.

What's the review approach that actually shows up on your students' tests — not the one that feels best in the room, but the one that transfers? Drop it in the comments.

---

This is part of Accingo's Pedagogical Playground, where we explore teaching practices that honor both research and classroom reality.