How to Study Without Notes: Straight From the Source

Yes, you can study without notes, and for a lot of material you should. Notes are storage, and what builds the memory is retrieval, the act of producing an answer with the source closed. If a chapter, a slide deck or a printed handout is already in front of you, skip the note and test yourself on it.

The evidence for that is stronger than most students expect, and it arrives with a real limit attached: at very short delays the ordering reverses and a second read wins. What follows is what the two papers found, the loop that replaces note-making, what you give up by dropping the artifact, and the cases where notes still earn the hour.

Do you need notes to study?

No, and the clearest way to see why is to separate the two jobs a set of notes does. The first is capture: getting a source that will not be available later into a form you can reopen. That job has a deadline, which is why a live lecture is worth writing down. The second is study: working the material into memory well enough to produce it under exam conditions. Only the first of those needs an artifact to exist.

The second job has been tested head to head against exactly the kind of artifact students build. In 2011 Jeffrey Karpicke and Janell Blunt published Retrieval Practice Produces More Learning than Elaborative Studying with Concept Mapping (PDF) in Science, and the finding sits in their opening lines: “we show that practicing retrieval produces greater gains in meaningful learning than elaborative studying with concept mapping”. Concept mapping there is an elaborative study activity done with the text open, in which students draw a diagram whose nodes are concepts and whose links are the relations between them. It is a tidier artifact than most people’s notes and it is built the same way, by reorganizing a source while looking at it.

Eighty undergraduates took part in the first experiment. They studied a science text once, or across four consecutive study periods, or by building a concept map from it, or by studying it once and then practicing retrieval, recalling as much of it as they could on a free recall test before restudying and recalling again. Learning time was exactly matched between the mapping and retrieval conditions. One week later, on short-answer questions covering both verbatim detail and inferences, retrieval practice scored 0.67 against 0.45 for concept mapping, which the authors describe as about a 50 percent improvement in long-term retention scores.

The second experiment removed the obvious objection, which is that a recall test and a short-answer test are much the same activity, so of course practicing one helps with the other. There, 120 students each built a concept map of one science text and practiced retrieval on another, and half were then assessed by building a concept map. “The advantage of retrieval practice occurred even when the criterial test involved creating concept maps.” Retrieval beat the artifact even when the exam was building the artifact.

The part worth carrying furthest is what the students expected. Asked during the learning session how much of the text they would still have in a week, they put repeated reading at the top of the four conditions and retrieval practice at the bottom, which is the reverse of the order the week-later scores came out in. In the second experiment 101 of the 120 students actually did better after retrieval practice than after mapping, while only 30 of them had predicted they would. Making an artifact feels like the productive option because the material stays in front of you the whole time, and that feeling carries no information about next week.

Hold the result to the size of its design. Undergraduates, science texts, one week between the session and the test. That settles nothing about a term-long revision plan, and it is silent on the case where there is no source to work from in the first place, where the first job is getting hold of one. That case has its own guide, how to catch up on schoolwork after falling behind.

Testing loses at five minutes and wins at a week

The result that makes this usable, and the one that keeps it honest, came five years earlier. In Test-enhanced learning: taking memory tests improves long-term retention (PDF), Henry Roediger and Jeffrey Karpicke ran Washington University undergraduates through prose passages taken from a TOEFL reading-comprehension book. Experiment 2 used 180 of them, aged 18 to 24. Each student got four consecutive periods, five minutes each for a study period and ten for a test, filled one of three ways: four study periods and no test at all, three study periods followed by one free recall test, or one study period followed by three consecutive tests. The final test came either five minutes later or one week later.

The authors state both halves in one passage: “On the 5-min test, recall was correlated with repeated studying: The SSSS group recalled more than the SSST group (83% vs. 78%), who in turn recalled more than the STTT group (71%). However, on the 1-week test, recall was correlated with the number of tests given earlier: The STTT group recalled more than the SSST group (61% vs. 56%), who in turn recalled more than the SSSS group (40%).” S is a study period and T is a test period, so SSSS is the group that only ever read.

Read the first half as carefully as the second. Five minutes after the session, the group that had only ever studied was at 83 percent against 71 for the group that had mostly tested itself. If your measurement is immediate, more reading really is the better use of the time, and every student who has felt that is right about what they felt. The reversal at one week does not erase the five-minute result. It dates it.

The awkward part sits in the abstract itself: “prior testing produced substantially greater retention than studying, even though repeated studying increased students’ confidence in their ability to remember the material”. The arrangement that produced the least durable memory produced the most confidence about having one.

The bounds here are narrow and worth keeping in view. Prose passages, undergraduates, free recall as the test, and delays of five minutes and one week. Nothing in that design speaks to a delay of a single day, to problem-solving subjects, or to a whole syllabus. What it does support is the shape of the decision: the further away the thing you are studying for, the worse an extra pass over the source looks against a pass in which you produce the material yourself. An earlier experiment in the same paper followed the same crossover across a middle delay, and how to review notes after class works through what it means for the hour straight after a lecture.

How do you study without making notes?

Five steps, and the fourth is where this usually falls apart.

1. Cut the source into pieces you can hold at once. A textbook subheading, a run of six or eight slides, one worked example. The size matters because step three has to be possible from memory, and a whole chapter is not.
2. Read the piece once, at normal speed. Read it with the explicit intention of being asked about it in two minutes. Do not annotate, do not highlight, do not stop to transcribe anything. One pass.
3. Close it and produce. Say the answers out loud, or write them on a blank page. Writing out everything you can remember in one uninterrupted go is the blurting method, which is the fullest version of this step and the one to read if the blank page is where you stall.
4. Open the source and check. This is what converts a memory test into studying, and it is the one people drop, because checking is where you find out you were wrong. Mark every miss.
5. Keep the question, drop the note. What survives the session is a list of the questions you missed. It is short, and it is a schedule of what to retrieve next. When those questions should come back is a subject of its own, worked through in active recall versus spaced repetition.

One source format needs its own handling. A slide deck is a speaker outline with the explanation deleted, so the gaps in it are structural and step two will not close them on its own. There is a specific procedure for that, and how to study from PowerPoint slides sets it out.

What skipping the notes costs you

Three things, and each is a reason the answer to the title question is conditional.

The first is the record. Retrieval builds the memory and leaves nothing behind on the desk. If the source goes away, a library book returned, a shared drive closed at the end of term, a paper you read once in a seminar, then whatever you failed to learn is not recoverable. Notes are insurance against your own forgetting, and the insurance is cheap only while the source is still open in front of you.

The second is selection. Deciding what matters in a chapter is real intellectual work, and note-making forces it in a way that reading does not. A question list does part of that job, since you cannot write a question without deciding what is worth asking, and it does less of it than a full set of notes would. That is a trade, and it is better made knowingly. If you do write notes and want the writing to pull its weight, the same principle applies one level up: write them with the book shut, which is what taking notes from a textbook without copying it out works through.

The third is that this whole method assumes a stable, re-openable source in front of you. Everything above depends on step four, and step four depends on the source still being there to check against. Material that passed once and was never recorded is a different problem with a different answer, and no amount of self-testing substitutes for having the content at all.

Where a generated question set fits, and where it does not

Step five above, keep the question and drop the note, is the mechanic GeniusPal runs, so it is worth being exact about what it does and what it leaves to you. On every plan, the free account included, it takes one source and returns a set of questions written on that source, and there are exactly two ways to hand it that source, a file you upload or a link you paste to a web page. Your own typing is not a third way, so material that lives only in your head has to become a file before anything can read it.

GeniusPal runs two separate meters and confusing them is how people get caught out. Building a set spends a generation, and a Free account gets two of those for the lifetime of the account with nothing on the calendar adding to them, while the Student plan is 14.99 dollars a month and meters 100 generations inside each of those months, and Genius is 59.99 dollars billed annually with the word unlimited on the card, which is display copy laid over a fair-use ceiling that has a real number behind it. Sitting a set you already own spends no generation at all, and on that second meter a Free account can work through each of its own sets twice in full, while on the paid plans that meter is not applied.

Every GeniusPal set comes back as a scored quiz, which is what a Free account can sit. The Student plan adds flashcards and written active recall over those same questions, with no further generation spent to switch between them. Two things stay uncounted on every plan, the free account included, and they are the daily review, which pulls a round together out of whatever has beaten you so far, and the coach page, which names the pattern your wrong answers keep falling into.

Two limits belong in the same paragraph as all that. The questions arrive written by a model, and step five above put the writing of the questions inside the work: choosing what is worth asking is the selection job that note-making forces on you, and handing it over means you have not done it. A set that arrives complete also removes the pause where you would have discovered that a paragraph made no sense to you. The honest use is as the second pass over material you have already read yourself, and as the thing that makes questions exist on the evenings you would otherwise have spent rereading.

The underlying point is smaller than any of the techniques. A note is a means and the memory is the end, and the two get confused because one of them is visible on the desk at the end of an evening and the other is not. Read a section, shut it, produce what you can, check against the source, and keep the questions you missed. An empty page at the end of the hour is not the same as an empty hour.