Reading Week Is Not a Break: How to Build a 20-Card Midterm Deck

At a Canadian university, reading week can look like a holiday. There are no classes, the library is quieter, and everyone says they will “catch up.”

For many students, midterms arrive soon after.

The students who do well during reading week rarely spend it making a fresh 200-card deck for every course. They spend it choosing what to retrieve. The break is not a break from thinking; it is a change in the kind of thinking.

Dates and names vary. McGill, for example, lists a Fall Reading Break in October and a Winter Reading Break in March in its 2026–27 key dates. Other universities call it Reading Week, Reading Break or something else. The label is local; the problem is not. You have a short window, several midterms and more slides than you can review.

This guide is for Canadian university students who want to use reading week to build a small, usable deck, not a second copy of the lecture.

A student finishes a two-hour lecture, opens a flashcard app, and makes 200 cards.

Two weeks later, the deck is untouched. The student remembers making the cards, not what was on them.

This is the most common flashcard failure, and it is not a discipline problem. It is a design problem. A lecture is not a list of facts. It has a structure: a problem, a definition, an example, a qualification, a comparison. The structure is what you need to reconstruct in an exam. A deck that copies the slides preserves the surface; a deck that asks about the structure preserves the thinking.

A deck that tries to capture the whole lecture is a second copy of the lecture. A deck that captures the twenty decisions the lecture actually tested is a study tool.

This guide is for Canadian university students who want to use flashcards for retrieval, not transcription. The learning-science evidence is strong for practice testing; the practical problem is selection.

Reading Week Is Not a Break

Reading week is a study window with a misleading name. It is not a holiday, and it is not a second semester. It is a short block of time before midterms, and the worst way to use it is to start from zero.

A better reading-week sequence:

1. Day 1: inventory. List every midterm, its date and its format. A multiple-choice exam and a written exam need different cards.
2. Day 2: select. For each course, choose 15–25 cards. Use the lecture learning outcomes, not every slide.
3. Days 3–5: retrieve. Review the cards, then do practice problems or past papers.
4. Day 6: repair. Rebuild only the cards you keep missing; leave the rest.

The sequence is not a rigid timetable. It is a way to avoid the two failure modes: making cards for everything and reviewing nothing.

Reading week is also a rest week. If you are exhausted, sleep is not a moral failure. A smaller deck built after rest is better than a 200-card deck built in a panic. The point is not to turn every hour into output; it is to make the hours you do study count.

Why 200 Flashcards Fail

A 200-card deck fails for three reasons.

It contains too many facts. A lecture has examples, asides, definitions, qualifications and stories. Most of them are not worth a card. If everything is a card, nothing is a priority.

It contains the wrong kind of card. “What is the definition of X?” can be answered by recognition. “Given this example, which concept applies?” requires retrieval. The second is harder and more useful.

It is reviewed once. A deck made the night before the exam is a compression exercise, not a retrieval practice. Spacing matters; a deck you revisit three times beats a deck you read once.

There is also a measurement problem. A deck of 200 cards feels like progress because it is large. A deck of 20 cards feels insufficient because it is small. But the goal is not to own the lecture; it is to retrieve the parts that will be tested. A small deck you actually review beats a large deck you avoid.

The fix is not “make fewer cards because you are lazy.” It is “make cards that deserve to exist.”

What Deserves a Card

A card deserves to exist if it tests something you need to retrieve without the page in front of you.

Use four filters:

1. It is examinable. The lecturer said it matters, it appears in the learning outcomes, or it explains a problem set.
2. It is retrievable. The answer is a concept, a relationship or a procedure, not a paragraph.
3. It is atomic. One card, one idea. If the answer has three parts, make three cards or ask a question that requires the relationship between them.
4. It has a cue. The question should give enough context to retrieve the answer, but not so much that the answer is in the question.

A worked example from economics. The lecture defines price elasticity, shows a formula, gives three examples and mentions a policy debate. The three big ideas might be: what elasticity measures, why the midpoint method matters, and why elasticity changes along a demand curve. The policy debate is context, not a card. The formula is a card only if you can use it; otherwise it is a reference.

A bad card:

What is the definition of opportunity cost?

A better card:

A student chooses a paid internship over a summer course. What is the opportunity cost, and why is it not the internship salary?

Three card types are worth having. Concept cards ask you to explain a term in your own words. Application cards give a short scenario and ask which concept applies. Comparison cards ask how two concepts differ and why the difference matters. Definition-only cards are the weakest of the three, because recognition can carry you further than understanding.

The second card asks you to apply the concept. It is harder to guess and closer to what an exam asks.

The 20-Card Rule

Twenty is a heuristic, not a law. The point is to force selection.

After a lecture, ask:

• What are the three big ideas?
• What are the five terms I could not explain without notes?
• What are the two comparisons the lecturer made?
• What is one example I could reproduce?
• What is one thing I still do not understand?

That usually produces 15–25 cards. After you make the deck, audit it: how many cards ask for a definition, how many ask for an application, how many ask for a comparison? If 18 of 20 are definitions, the deck is a glossary, not a study tool.

If it produces 60, you are probably carding the slides, not the lecture. If it produces five, you may be carding only the bold terms and missing the relationships. The target is a deck that covers the lecture’s decisions, not its word count.

After you make the deck, audit it: how many cards ask for a definition, how many ask for an application, how many ask for a comparison? If 18 of 20 are definitions, the deck is a glossary, not a study tool.

If it produces 60, you are probably carding the slides, not the lecture.

Do not card the lecturer’s biography, the course logistics or the joke that got a laugh. Do not card a number unless you need to use it. Do not card a slide title unless the title is the concept. These are not examinable retrieval targets; they are the texture of the lecture.

A useful rule: if a card does not change what you can do on a problem, it is probably a definition you can look up. That does not mean definitions are unimportant. It means a definition card should ask you to use the definition, not recite it.

Do not card the lecturer’s biography, the course logistics or the joke that got a laugh. Do not card a number unless you need to use it. Do not card a slide title unless the title is the concept. These are not examinable retrieval targets; they are the texture of the lecture.

A useful rule: if a card does not change what you can do on a problem, it is probably a definition you can look up. Definitions are not useless; they are just not all worth retrieval practice.

How to Review Without Re-reading

A flashcard is a question, not a highlight. The review loop is:

1. Read the question.
2. Answer aloud or on paper before turning the card.
3. Check the answer.
4. If wrong, write the correction in your own words.
5. Return to the card after a gap.

The evidence for retrieval practice is strong. Roediger and Karpicke’s test-enhanced learning study showed that students who tested themselves remembered more than students who reread, even when the rereaders felt more confident. Dunlosky et al.’s review of effective learning techniques rated practice testing as one of the most effective strategies. The caveat is that a bad card tests the wrong thing; retrieval practice only helps if the question is worth answering.

The gap is the part students skip. Reviewing a card five times in one night feels productive; it mostly tests short-term familiarity. Spacing the same card over several days is harder and more useful.

A simple schedule:

• Day 1: make the cards, review once.
• Day 2: review the missed cards.
• Day 4: review all cards.
• Day 7: review the missed cards again.
• Before the exam: review the cards you still miss, not the whole deck.

A review session should be short enough to repeat. Twenty minutes of retrieval is better than ninety minutes of rereading the lecture. If you have a long commute, review on the phone; if you have a study room, use paper and say the answer aloud. The medium matters less than the retrieval.

The schedule is not magic. The principle is: retrieve, check, space, repeat what you miss.

Where the App Helps, and Where It Cannot

The Sovi app’s knowledge cards, flashcards and practice quiz can hold the deck and give you a review loop. That is useful. It cannot decide which twenty cards matter.

Use the app to:

• keep the cards in one place;
• review on a phone in a queue or on a bus;
• take a practice quiz;
• separate cards you have mastered from cards you keep missing.

Do not use it to:

• card every slide;
• copy a definition without an example;
• replace understanding with recognition;
• assume a quiz score is the same as exam performance.

The app is a delivery system for retrieval. The intellectual work is choosing the question. If you cannot explain why a card is in the deck, delete it.

The app is a delivery system for retrieval. The intellectual work is choosing the question.

A useful AI prompt is restrictive: “Give me one application question per concept, not a definition.” Then check the question against the lecture. If the question can be answered from the slide title, it is too easy. If it requires information the lecture did not cover, it is unfair.

AI can help you draft cards, but it tends to produce too many definition cards and too few application cards. If you use it, ask for “one application question per concept,” then delete the cards that only ask for a definition.

The kit below turns a lecture into a 20-card deck and a review schedule.

• Flashcard selection kit (PDF) to print and fill in by hand.
• Flashcard selection kit (Word) to type into.

Frequently asked questions

1. How many flashcards should I make per lecture?

A useful target is 15–25. The number is not sacred; the selection is. If you have 60 cards after one lecture, you are probably carding examples and definitions that do not need retrieval.

2. Is making flashcards enough to study?

No. Flashcards support retrieval, but you still need to understand the concept, solve problems and connect ideas. A deck is a review tool, not a substitute for the lecture or the reading.

3. Should I use pre-made decks?

Use them carefully. A pre-made deck may use a different textbook, a different emphasis or a different definition. If you use one, check it against your course and delete cards that do not match.

4. How often should I review?

Space the reviews. A practical pattern is day 1, day 2, day 4, day 7, then before the exam. Review the cards you miss more often than the cards you know.

5. What makes a good flashcard question?

A good question requires retrieval and gives a clear cue. It should not contain the answer, and it should not ask for a paragraph. Application questions are usually better than definition questions.

6. Can AI make my flashcards?

It can draft them, but it often overproduces definitions. Ask for application questions, check every card against your notes, and delete anything you cannot explain without the card.