How to make flashcards that actually work

Many cards look like a tiny lecture summary squeezed onto a small piece of paper. You reread them, you feel good, and nothing has moved in your head. A useful card does something else. It asks a precise question on the front and waits for you to bring out the answer before you flip it. For example, for AP US History, “In what year was the Treaty of Versailles signed?” on the front and “1919” on the back.

The principle comes from the generation effect. Slamecka and Graf (1978) showed that you remember a word you produced yourself better than the same word you only read (Slamecka and Graf, 1978). A card that makes you produce the answer uses that mechanism, while a card that shows everything lets you just read. Keep one idea per card. If you cannot say it in one sentence, split it into two simpler cards.

When people prepare cards, the temptation is to offer choices, because they are quicker to write and more reassuring to get through. Rowland (2014) gathered the work comparing testing with rereading in a meta-analysis. Recall tests, where you must find the answer, give a bigger benefit than recognition tests such as multiple choice (Rowland, 2014). The direction is clear.

In practice, hide the answer and say it out loud or write it down before checking. If you prefer to keep options in front of your eyes, you are mostly doing recognition, and your brain does not have to search. This applies to an AP Calculus exam, a finals paper or a pre-med course. Multiple choice keeps its place for practising the format of a test like the SAT, but for learning, the open question keeps the advantage in this meta-analysis. You can read more in Active recall, why quizzing yourself beats rereading.

Kornell (2009) worked with vocabulary flashcards of the kind people prepare for the GRE. One large deck studied with spacing beat four small decks, each studied in one block, and spacing also beat cramming the night before. Spacing was more effective for 90% of the participants (Kornell, 2009).

Here is the most interesting part. After the first session, 72% of the participants thought that massing worked better (Kornell, 2009). So your feeling can fool you. Studying everything at once feels productive, because everything is fresh, while leaving time between two passes feels more laborious. In practice, instead of finishing your chemistry cards in one evening, put them back into play over several days. The guide Spaced repetition, how long to wait between two reviews explains how to choose the gaps, and Does cramming the night before an exam work? covers what the night before really gives you.

Rawson and Dunlosky (2011) followed 533 students across three experiments, with more than 100,000 responses scored by hand. Their advice is precise. Recall an item correctly three times in the first session, then relearn it three times at widely spaced intervals (Rawson and Dunlosky, 2011). So a card is not finished because you got it right once.

Karpicke and Roediger (2008) worked with foreign-language vocabulary, in Swahili. Once a word was known, studying it again had no effect on delayed recall, while retesting it had a large positive effect (Karpicke and Roediger, 2008). They also saw that students’ predictions of their own success were not related to how well they actually did. Keep two habits. Do not reread a card you got right just to feel reassured, take it out of today’s pile and test it again later. And do not trust your feeling of knowing, trust your answer.