Raw notes make poor study material. Turn lecture notes and slides into short question-and-answer cards, and see what the research does and does not show.
By Bleart Emini, founder · Published · Last reviewed · 7 min read
Notes are what you wrote down so you would not lose information. They are not a good way to learn it. Reading them again feels productive and mostly is not. The fix is to turn each fact in your notes into a short question with a short answer, then practice answering from memory. This article shows how to do that, what the research says about why it works, and where the research stops.
Lecture notes are built for capture. You write fast, you abbreviate, you copy a slide because the professor said it mattered. The result is a record, and a record is organized by the order things were said, not by what you will be asked.
Three problems follow. Notes contain answers but no questions, so there is nothing to retrieve. They bundle many ideas into one paragraph, so you cannot tell which idea you actually know. And they invite rereading, which gives a strong feeling of familiarity without testing whether you can produce the answer on your own.
Slides are worse in one way. A slide is a prompt for a speaker, a few words that only make sense with the talk attached. Three weeks later the bullet "Operant vs. classical" tells you nothing.
The method is simple and a little tedious by hand.
The size of the card matters most. Piotr Wozniak, who built the SuperMemo spaced repetition software, put it as a rule in his essay "Twenty rules of formulating knowledge": stick to the minimum information principle. Each item should be as simple as possible, so one repetition retrieves only a small piece of information. He also notes that simple items can be reviewed at their own pace, and that minimum information does not mean minimum characters. He recommends cloze deletion (a sentence with a gap to fill) as a quick way to start, especially if you struggle to keep items small.
Here is why that helps in practice. If a card asks you to list five features and you know four, you will mark it wrong or right inconsistently, and the one you miss never gets isolated. Five small cards let the scheduler find the weak one.
Here is an example, not my story. Suppose a calculus slide reads: "Derivative at a point: slope of the tangent line; limit of the difference quotient as h approaches 0; power rule."
As a single card, that is useless. Split it:
Four cards, each answerable in under ten seconds. The last one is a practice problem, not a definition. Notes from a quantitative course should produce a few of those, because knowing a rule and applying it are different skills.
The same idea works for vocabulary-heavy subjects. For a psychology term, one slide on negative reinforcement becomes a card for the definition and a second card that asks you to classify a scenario ("A student takes a pill and the headache stops, so they take it again next time: positive reinforcement, negative reinforcement or punishment?" Answer: negative reinforcement, because something unpleasant was removed and the behavior increased).
Be skeptical of any claim that flashcards are magic. Here is what has actually been tested.
Practice testing beats rereading and highlighting. Dunlosky, Rawson, Marsh, Nathan and Willingham (2013) reviewed ten common study techniques in Psychological Science in the Public Interest. They gave practice testing and distributed practice high utility ratings because the benefits held across students of different ages and abilities and across many kinds of tasks. Rereading and highlighting got low ratings. This is the closest thing to a verdict for students, and it is a review of other people's studies, not one experiment.
Testing wins at longer delays, not right away. Roediger and Karpicke (2006, Psychological Science) had students study prose passages, then either restudy them or take free-recall tests. After five minutes, restudying produced better recall. After two days and after one week, the tested group remembered substantially more, even though the restudy group felt more confident. Two things to keep in mind: the materials were short passages, and the longest delay was one week. That is useful for a finals schedule, less so for a semester-long memory question.
Retrieval also helps understanding, not just facts. Karpicke and Blunt (2011, Science) compared retrieval practice with building concept maps from a text. Retrieval practice produced more learning, including on inference questions, and even when the final test asked students to make concept maps. Note that these were science texts and the comparison was against one specific alternative.
Generating answers helps, in lab tasks. The generation effect is the finding that people remember information they produce themselves better than information they read. A meta-analysis by Bertsch, Pesta, Wiscott and McDaniel (2007, Memory & Cognition) pooled 445 effect sizes from 86 studies and found an average effect of 0.40, close to half a standard deviation. The abstract describes tasks such as producing synonyms, so most of this work is about short items, and it does not directly test whether writing flashcards from lecture notes beats receiving ready-made ones.
That last gap matters, because it creates a real tradeoff. Writing your own cards forces you to process the material, and that likely has value. But no study was found that cleanly isolates that benefit for flashcard creation from notes. What the evidence supports firmly is the retrieval part: once the cards exist, answering them from memory is what does the work. So whoever makes the cards, make sure you read each one, fix the vague ones, and delete the ones that test trivia.
Also worth saying: most of this research uses college undergraduates, short retention intervals and structured materials. It says little about, say, a high school student learning to write essays.
Before Cardvoo, I typed my cards by hand, and I tried two other ways of getting notes into cards. One was copy-pasting my notes in. The other was writing term and definition cards. Both took too long, and copy-pasting always missed some important notes I wanted included. What I wanted was to describe a deck, say what I was studying, and have the cards generated, so the time goes to studying and not typing.
The honest caveat from the section above applies to that wish. Skipping the writing step may skip some of the learning from writing. The retrieval evidence is on firmer ground than the writing evidence, which is why reviewing generated cards, rather than accepting them blindly, is the part you should not skip.
All study articles · Features · Sign up
Features · Pricing · Explore public decks · Student discount · Study guides · Study articles · Help · Sign up · Log in · Privacy Policy · Terms of Service
Cardvoo runs in your browser and needs JavaScript turned on.