How far in advance should test review start?

About a week out for a unit test is a practical target. Three shorter sessions spread across the week let students retrieve the material several times with gaps in between, rather than once in a single sitting, and the total class time spent is the same.

How long should a test review game be?

Fifteen to twenty-five minutes per session works well for this structure. Shorter sessions are easier to keep tight and leave room in the period for going over what the class missed, which is where the useful information is.

Should review questions be harder than the test?

No. Review should reach the hardest level the test will reach, but not beyond it. Questions harder than anything on the assessment tend to create anxiety without preparing anyone for what they'll actually face.

Is it better to review one topic at a time or mix them?

Both, in that order. Early sessions can be topic-by-topic, since that's how gaps get identified and addressed. The final session before the test is a good place to mix topics in random order, because recognizing what a question is about is part of what a real test demands.

What if the review shows the class isn't ready?

That's the review doing its job, and it's the argument for starting a week out rather than the day before. With several days left you can reteach, adjust pacing, or move the test. On test eve you can only find out.

Test Review Games: How to Use the Week Before an Exam

The standard approach to test review is one long session on the day before, covering everything. It's convenient, and it has one structural drawback worth weighing: everything reviewed on Thursday gets recalled once, in a single sitting, immediately before being tested.

Splitting the same total minutes across three shorter sessions is a practical way to apply [spaced retrieval](/articles/games-memory-recall) in a classroom timetable. Students pull the material from memory several times with gaps in between rather than once. Whether it works better than a single long review in your particular class is worth trying, but it costs no extra instructional time to find out.

This is about sequencing that week. For building the board itself against your unit objectives, see [end-of-unit review games](/articles/end-of-unit-review-games).

A three-session structure

Assume a test on Friday and roughly 60 minutes of review time to spend across the week. Rather than spending all 60 on Thursday:

Monday, 20 minutes: breadth sweep. Quick coverage across every topic on the test, mostly at recall level. The purpose isn't depth. It's to make students confront the full scope early, while there's still time to act on it, and to give you an early read on where the class is weak.

Wednesday, 25 minutes: weak areas only. Built from what Monday exposed. Two or three topics, at application and multi-step level. This is the session doing the most work, because it's the only one targeted at actual gaps.

Thursday, 15 minutes: mixed cumulative. Questions from all topics in random order, at test-level difficulty. Short and fast. The purpose is to practice the switching that a real test demands, not to cover new ground.

The total time is the same. The distribution is what changes.

Why the mixed session goes last

Students who review one topic at a time get good at answering questions when they already know which topic the question is about. A test doesn't work that way. Question 14 might be about the first week of the unit and question 15 about the last.

That switching is its own demand, and the Thursday session is where it gets practiced. Keep the questions in deliberately random order. If your board tool lets you shuffle or pick clues out of sequence, use it.

Match the question format to the test format

This is one of the cheapest improvements available and it gets skipped constantly.

If the test is multiple choice, review questions should be multiple choice with plausible distractors. If it's short answer, students should be writing short answers during review, not selecting from four options. If it involves showing work, the review questions should require showing work.

A review that's all quick recall is poor preparation for a test that's all multi-step problems, no matter how thoroughly it covers the content.

What belongs in a test review and what doesn't

Include:

Everything on the test, at roughly the proportion it appears on the test
Questions students missed on earlier exit tickets and quizzes
Anything you retaught mid-unit, since that content had the shakiest foundation
At least a few questions at the hardest level the test will reach

Leave out:

Content that isn't being assessed, however interesting
Questions that are harder than anything on the test, which tend to produce anxiety without preparing anyone
Trick questions and edge cases, unless the test has them

A sample week

A high-school algebra unit on quadratics, with a Friday test:

| Session | Focus | Level | |---|---|---| | Monday, 20 min | Factoring, graphing, the quadratic formula, word problems, discriminant | Recall and one-step | | Wednesday, 25 min | Word problems and discriminant (weakest from Monday) | Application and multi-step | | Thursday, 15 min | All five topics, shuffled | Test-level |

If Monday shows the class is solid on factoring and shaky on word problems, Wednesday spends its entire 25 minutes on word problems. Reviewing factoring again on Wednesday would feel productive and add little, because the class already has it.

Running a shorter session well

A 15- or 20-minute review is a different exercise from a full-period board and needs a tighter structure:

Skip elaborate team formation. Use rows or table groups.
Cut the board down. Three categories with four clues is plenty for 20 minutes.
Shorten the timer. Ten to fifteen seconds for recall-level questions.
Keep every student writing answers. In a short session this matters more, not less, because there's no time for anyone to be a spectator.
Reserve the last three minutes to go over misses. Non-negotiable, even at 15 minutes.

The trap of the fun review

There's real pressure to make the day-before session enjoyable, especially with an anxious class. The risk is a session where the mechanics take over: elaborate scoring, bonus rounds, team theatrics, and eight actual questions in forty minutes.

Students will report that it was a great review. It wasn't one.

A rough test: count the questions. If a 40-minute review session covered fewer than 20 questions, most of the period went somewhere other than content.

After the test

The board you built is worth keeping. Two uses:

Reteaching. If the test results show a topic went badly across the class, the questions from that category are already written.
Next year. The sequence works again with updated clues, and you'll know which questions were ambiguous.

For building the board, Kleveroo's [Jeopardy board maker](/jeopardy-board-maker) gives the host control of the timer and scoring during play, and you can build one board for the unit and reuse it across all three sessions. Format options beyond a board are covered in [classroom review games for any subject](/articles/best-classroom-review-games), and the mechanics of running one with a full class are in [how to play a Jeopardy-style game in the classroom](/articles/how-to-play-jeopardy-in-the-classroom).

Test Review Games for the Week Before the Exam