Flashcards turn a fact into a retrieval prompt: you attempt the answer before checking it. Laboratory and educational research describes this as retrieval practice or the testing effect; a useful overview is Roediger and Karpicke's review. The size of the benefit depends on the material, feedback, delay, and test, so a card is a study aid rather than a guarantee.
The second ingredient is spacing reviews across time. A meta-analysis of 317 experiments found that the useful interval depends on how long the material must be retained. The schedules below are simple starting rules, not universal optimal intervals.
The Leitner box: spaced repetition with a shoebox
Sebastian Leitner's system from the 1970s needs one box with three to five compartments (or three envelopes — nobody's checking).
- All new cards start in Box 1, which you review daily.
- Answer a card correctly and it moves up one box. Box 2 gets reviewed every 3 days, Box 3 weekly, Box 4 every two or three weeks.
- Answer wrong — the card goes back to Box 1, no matter where it was.
That's the whole algorithm. Hard cards automatically show up often; easy cards get out of your way. A card that survives to the last box is essentially learned. The daily session stays short — usually 10–15 minutes — because most of your cards are resting in the higher boxes at any given moment. The one rule that makes or breaks it: demotion is non-negotiable. A "close enough" answer that stays in Box 3 is a fact you'll be missing on exam day.
How to write cards that actually teach
- One fact per card. "The three zones of a Cornell page" is three cards, not one. If a card keeps failing, that's usually the symptom — split it.
- Ask a question, don't state a topic. "Photosynthesis" on the front tests nothing. "What are the two inputs of photosynthesis?" tests something specific.
- Write both directions when it matters. For vocabulary, make one card for word → translation and another for translation → word. They're different retrievals, and you're usually weaker in one.
- Keep answers short enough to verify at a glance. If you can't tell in two seconds whether you got it right, the answer side is doing too much.
- Make your own cards. Formulating the question is half the studying. A borrowed deck skips that half.
If your source material is lecture notes, cards practically write themselves from a well-kept cue column — the Cornell method and flashcards feed each other nicely: cues become fronts, notes become backs.
Picking a card size
Size is a real decision, not a detail.
- Pocket, 2.5 × 3.5 inches — playing-card sized. Best for vocabulary, dates, and one-word answers. A rubber-banded stack lives in a jacket pocket, which means you'll actually review on the bus.
- Index, 3 × 5 inches — the classic. Enough room for a definition or a short formula with a usage note. The default if you're unsure.
- Large, 4 × 6 inches — for worked examples, diagrams, chemical structures, or anything where the answer needs breathing room. Better for desk review than commutes.
Whatever you pick, stay consistent within a subject — uniform cards shuffle cleanly and fit the same box.
Printing and cutting a deck
The flashcards generator lays out 4, 6, or 8 blank card structures per page, with plain, ruled, or prompt-and-answer zones, and can print cut lines. It does not store your study content or generate paired front and back pages. If you want the same ruling on both sides, duplex-print the blank layout and test one sheet before cutting because reverse-side alignment varies by printer. Regular paper works for light use; cardstock or a heavier backing is more durable for repeated shuffling.
When paper beats apps
Apps can automate review schedules, search large decks, and synchronise them across devices. Paper supports a different workflow: cards can be sorted physically, shared around a table, and reviewed without opening another screen. Claims that handwriting by itself improves memory are contested, so choose between paper and an app for practical reasons rather than assuming that either medium guarantees better learning.
A small paper deck can start with three labelled envelopes and a manageable first batch. If the deck becomes too large to sort or schedule reliably by hand, an app may be the more practical option.