The Feynman technique

Richard Feynman was a physicist at the California Institute of Technology, and the technique is named in his honor. The myth is that he designed these four steps and taught them as a study method. We did not find the steps set out in anything he wrote or said on record. The label was attached later, by people writing about how to learn.

What is documented is the standard behind the method. David Goodstein, a Caltech physicist who worked alongside him, wrote a memoir of Feynman as a teacher (Goodstein, 1989). Goodstein describes the challenge Feynman set himself of reshaping physics so that it could be presented to young students, and adds:

“In fact, it was the standard by which he measured whether something was really understood.”

Goodstein then tells of asking Feynman to explain why one family of particles behaves as it does. Feynman answered that he would prepare a freshman lecture on it. A few days later he came back and, in Goodstein’s account, said:

“You know, I couldn’t do it. I couldn’t reduce it to the freshman level. That means we really don’t understand it.”

That is one colleague’s memory of a conversation, not a line from Feynman’s own pen, and it should be read as such. Still, it holds the whole idea: if you cannot explain something to a beginner, take that as news about your own understanding.

The same memoir is honest about how hard the standard is. Feynman gave two years to teaching introductory physics, and the lectures became a famous set of books. Goodstein’s verdict is that they often went over the heads of the freshmen and helped his fellow physicists more. Plain explanation is hard even for the person the method is named after, which is a good reason to practice it and a poor reason to expect it to be quick.

Popular retellings add scenes and quotations. Where we could not trace one to a source we had read, we left it out.

Research has not, as far as we found, tested the four steps as one package under this name. What has been tested, many times, is each of its parts: explaining to yourself, preparing to teach, teaching, and finding out how much less you know than you thought.

Explaining to yourself

Psychologists call this habit self-explanation: saying, as you study, what each line means and how it connects to what came before. In a study by Chi and colleagues (1994), 14 eighth-grade students were asked to explain each line of a text on the human circulatory system to themselves as they read it. Another 10 read the same text twice.

Both groups took a test before the reading and another after it, and the students who explained improved more. Among them, the ones who produced the most explanations understood the system in the most depth, and all of those arrived at a correct picture of how blood moves through it.

One small study proves little, so later researchers pooled many. Bisra and colleagues (2018) gathered 69 comparisons from 64 reports in which learners were prompted to explain while they studied or solved problems. The average effect size was 0.55. An effect size is a standard way of saying how far apart two groups ended up, and this one means the explaining groups scored about half a standard deviation higher. In plain terms, the gain was easy to see in the scores without being dramatic.

The broad review of study techniques by Dunlosky and colleagues (2013) sorted the methods into a high, a middle and a low group by how useful they are. Self-explanation is one of the 3 in the middle group: promising, with less classroom evidence than practice testing, which means quizzing yourself.

Preparing to teach, and teaching

The second step asks you to explain as if to someone else, and that framing helps even before anyone listens. Nestojko and colleagues (2014) ran 2 experiments in which people studied a passage. Some were told they would be tested on it, and others that they would teach it to another student. Nobody taught anyone, and everyone was tested.

The people who had expected to teach recalled the passage more completely and in better order, and they did better on questions about its main points. The authors take that as a sign that expecting to teach leads people to study in more effective ways. Listing the points an explanation has to reach, the first step here, borrows that mindset.

Explaining for real adds to that. Fiorella and Kuhlmann (2020) had 120 college students study a science text. Three groups then taught it on video to a made-up classmate, by talking, by drawing, or by both, while a fourth group spent the same time studying the text again. A week later all three teaching groups scored higher than the group that restudied, with effect sizes from 0.80 to 1.46, which are large gaps by the same measure. Talking while drawing did better than either alone.

Finding the gaps

The third step rests on an uncomfortable finding: people think they understand how things work far better than they do. Rozenblit and Keil (2002) named it the illusion of explanatory depth. In their first study, 16 graduate students rated how well they understood 48 everyday things, such as a zipper, a flush toilet and a helicopter, on a scale from one to 7.

Each student then had to write a step-by-step explanation of 4 of them and rate their understanding again. Nearly all of them lowered their ratings. The authors report that many were surprised, and humbled, by how much less they knew than they had thought. A written explanation did what a feeling of confidence could not: it showed where the knowledge stopped.

Four mistakes account for most sessions that feel busy and teach little. Each one removes the part of the method that the research supports.

• Explaining with the notes open. Then you are copying, and the gaps never show. The closed book is what makes the second step a form of active recall, the practice of producing an answer from memory instead of reading it again.
• Swapping plain words for neat ones. A sentence built from the textbook’s terms can sound right and explain nothing. If a term appears, its meaning has to appear with it.
• Stopping after the first explanation. The first try finds the gaps. The learning is in the trip back and the second try.
• Picking an idea that is too big. “Thermodynamics” is a course, not an idea. If the list of points runs past seven or eight, split it.

The Feynman technique for learning an idea is also slow, on purpose. Use it for the ideas a course is built on and the ones you keep getting wrong. For plain facts and vocabulary, quick recall is enough. The guide on how to study effectively ranks the methods side by side, and the study planner spreads your hours across the days before an exam.

A good way to run this test counts the hours from the line you already write or say about your day, such as an hour explaining chemistry after dinner. No stopwatch to carry around. Most journaling apps stop at how the session felt. They keep the frustration of a gap you could not close, which is worth keeping, but not the hours you gave to closing it.

One app built for this is Think in Numbers, a smart journaling app with time tracking built in.

You write or say your day in your own words, as it happens or all at once. The app reads what you wrote and counts the hours that went to each part of it: your study, your phone, and the time you never mentioned (shown as unaccounted). Then it shows those hours over days and weeks. A target for your study hours holds each week against your real ones.

Write my first day

Download on the App StoreGet it on Google Play

Or scan with your phone's camera.

For iPhone and Android. Try it on your own day before you pay.

The first half of the test is hours, and it comes first because the method is easy to admire and easy to skip. The second half is a result you score yourself: cover your list of points, explain, and count the ones you reached. Four of every five on a second try is a fair bar. All of them on a first try means the idea was too easy to be worth a session.

Here is the hours half for an example person: a made-up university student with finals on Monday, September 28. Each day they wrote a few lines about what they did, and the app counted the study hours from those lines. Their words do not say “Feynman technique.” They say what the student did, and one Saturday shows the method’s shape: a timed past paper, then a separate sitting for the mistakes.

Week one came to 5h and week three to 15h, on 6 of its days. Week three is the useful one to look at. It holds a day with 0m of study, and most of its other days hold a sitting of an hour and a half. Going back over mistakes needs a sitting of about that length.

The count has limits, and they are worth knowing. It shows hours, not how many points an explanation reached, so that half of the test stays on your own page. One Friday in week four also includes a lecture the app filed as study, which puts that week an hour above the student’s own sessions.

The sources for this page are one memoir and six research papers, each linked at its publisher.

1. Bisra, K., Liu, Q., Nesbit, J. C., Salimi, F. and Winne, P. H. (2018). Inducing Self-Explanation: a Meta-Analysis. Educational Psychology Review, 30(3), 703 to 725.
2. Goodstein, D. L. (1989). Richard P. Feynman, Teacher. Physics Today, 42(2), 70 to 75.
3. Chi, M. T. H., de Leeuw, N., Chiu, M.-H. and LaVancher, C. (1994). Eliciting Self-Explanations Improves Understanding. Cognitive Science, 18(3), 439 to 477.
4. Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J. and Willingham, D. T. (2013). Improving Students' Learning With Effective Learning Techniques: Promising Directions From Cognitive and Educational Psychology. Psychological Science in the Public Interest, 14(1), 4 to 58.
5. Nestojko, J. F., Bui, D. C., Kornell, N. and Bjork, E. L. (2014). Expecting to teach enhances learning and organization of knowledge in free recall of text passages. Memory & Cognition, 42(7), 1038 to 1048.
6. Fiorella, L. and Kuhlmann, S. (2020). Creating Drawings Enhances Learning by Teaching. Journal of Educational Psychology, 112(4), 811 to 822.
7. Rozenblit, L. and Keil, F. (2002). The misunderstood limits of folk science: an illusion of explanatory depth. Cognitive Science, 26(5), 521 to 562.