Use flashcards as retrieval tools, not recognition tools

See the question, produce the full answer from memory before flipping — every time.

Why it works

The memory benefit of a flashcard comes entirely from the retrieval attempt, not from flipping the card. Flipping immediately (or hovering), then reading the answer, is recognition — and recognition produces minimal strengthening of the memory trace. Forcing a full production attempt before flipping, even when it is difficult and slow, is what generates the retrieval-practice benefit.

How to do it

1. When you see a flashcard front, say or write the full answer before flipping — with no hints.
2. If you cannot produce the answer, let the effort continue for at least 30 seconds before giving up.
3. Rate your recall honestly: (1) completely forgot, (2) partially knew it, (3) knew it well.
4. Items rated 1 or 2 return to the deck sooner; items rated 3 wait longer.

Evidence

Production testing (generating an answer) produces larger testing effects than recognition testing (choosing among answers). The generation effect compounds the retrieval practice benefit when you produce rather than recognize. (rct)

Karpicke & Roediger used cued-recall formats in lab conditions; real-world flashcard use with self-grading is more variable, and self-assessment accuracy affects which cards are retired correctly.

Sources

• Karpicke & Roediger (2008), the critical importance of retrieval for learning, Science

Common mistake

Marking a card as "known" after recognizing the answer on the back — you recognized the answer, you didn’t retrieve it. Only retire a card when you can produce the answer reliably before flipping.