Study skills

Feynman technique: 4 steps, a worked example, and the evidence

By Studycamp · September 20, 2026 · 11 min read

Studycamp turns your slides, PDFs, and notes into flashcards, quizzes, and summaries, and keeps your deadlines in one place. Try Studycamp free

The Feynman technique is a four-step self-test. Pick one concept, explain it in plain words from memory, mark every spot where you get stuck or wave your hands, then go back to your material and fix those spots. It works because explaining shows you what you don't know.

One note up front: the name is a modern label. The research behind it is about self-explanation and learning by teaching, and that evidence is decent but not magic.

Key takeaways

The four steps: choose a concept, explain it simply from memory, find the gaps, then go back to the source and explain it again.
Do step 2 with your notes closed. One study found that teaching from memory beat teaching with notes on a delayed test.
Across 64 research reports, learners who were prompted to explain as they studied did moderately better than those who weren't. A major review rated self-explanation "moderate utility."
Studycamp covers each step: write your explanation in Notes, check the gaps with AI chat over your own files, then test yourself with a quiz. Do step 2 on your own first.
We found no record of Richard Feynman writing this routine down. The name appears to date from around 2011, though a real Feynman story does sit behind the idea.
It's slow, it suits concepts better than raw facts, and it can't catch an error you can't see. Pair it with active recall.

The four steps

Step 1: Pick one concept

Pick something small enough to explain in a couple of minutes. "Opportunity cost" works. "Microeconomics" doesn't. Write the concept at the top of a blank page.

Step 2: Explain it from memory, in plain words

Close your notes. Write or say the explanation as if to a friend who hasn't taken the course. Three rules make this step work:

No notes, no slides, no textbook.
Every technical term you use, you must define in ordinary words.
Include one concrete example with real numbers or a real situation.

The "twelve-year-old" framing is popular, but it's a device. What you want is an explanation with no borrowed vocabulary.

Step 3: Mark the gaps

Read it back. Circle every place where you:

paused or said "basically" or "kind of",
used a term without defining it,
couldn't give an example,
said something you now suspect is wrong.

These circles are the main point. You now have a specific list of what to study, instead of a vague feeling that you "mostly get it."

Step 4: Go back, fix, and explain again

Return to your slides or textbook, but only for the circled spots. Then explain the concept again from scratch, without looking. If you get stuck in the same place, try a second source.

A worked example: opportunity cost

Here's the technique on a real first-year economics topic. First, the explanation a student might write after reading the chapter twice.

The first attempt (the bad one)

Opportunity cost is the value of the next best alternative that you give up when you make a decision. So when you choose something, you have to consider all the alternatives you're giving up, which represents the true cost of the decision. It's an important concept in economics because resources are scarce.

It reads fine. It sounds like the textbook. Now run step 3 and ask what a curious friend would ask:

"All the alternatives, or just one?" (The first sentence says one. The second says all. That's a real error, and the student didn't notice it while reading.)
"The value of the alternative to whom?"
"How is this different from just the price I pay?"
"What does 'represents the true cost' actually mean? Give me a number."
"What does scarcity have to do with it?"

The student can't answer most of these. That's the useful result. The words "represents" and "important" were covering for missing understanding.

Going back to the material

Step 4 means checking the chapter for exactly those questions. The student finds that the cost is only the single best alternative given up, that it's measured in what it's worth to the chooser, and that it includes non-money things like time.

The second attempt (the fixed one)

Every choice costs you the best thing you didn't pick. Say you have a free Saturday afternoon. You could work a shift that pays $60, hang out with a friend, or study. If you study, the opportunity cost is the $60 shift, because that's the best option you gave up. Hanging out isn't added on top. If hanging out is worth more to you than $60, then that's the opportunity cost instead. It's about what you give up, not what you pay. That's why a "free" concert still costs you the evening, and why the cost of college is more than tuition: it also includes what you could have earned or done with that time.

What changed:

The one-versus-all error is gone.
There's a number and a scenario, so the idea can be checked.
It separates opportunity cost from price, which is the confusion exams love to test.
Every term is defined. Nothing is borrowed.

What the research says

Popular guides such as the Farnam Street article and the University of York study guide give the steps without citing research. So here is what the evidence does support.

Where the name comes from

We found no record of Feynman publishing a study method. The earliest use of the name we could trace is a September 2011 post by learning writer Scott Young, announcing a video called "The Feynman Technique." A widely shared four-step version followed on Farnam Street.

There is a real Feynman story behind the spirit of it. Physicist David Goodstein recalled asking Feynman to explain why spin-1/2 particles obey Fermi-Dirac statistics. Feynman offered to prepare a freshman lecture on it, then came back days later and said: "I couldn't reduce it to the freshman level. That means we really don't understand it." That's the idea: if you can't simplify it, you don't yet understand it. It's an anecdote about how Feynman judged his own understanding, not a documented study system.

Self-explanation has real support

Self-explanation means generating your own reasons for how or why something works. A 2018 meta-analysis in Educational Psychology Review combined 69 effect sizes from 64 research reports. Learners who were prompted to self-explain while studying or solving problems did better. Researchers measure that gap with an effect size, where about 0.2 is small, 0.5 is medium and 0.8 is large. The average here was g = 0.55, a medium boost: prompted learners scored clearly higher, though not by a huge margin. The British Psychological Society's summary notes that open-ended prompts did better than multiple-choice ones, which is a good argument for blank-page explaining over picking answers.

Teaching adds something on top

The other pillar is learning by teaching. In a chapter reviewing this research, Logan Fiorella reports a meta-analysis by Kobayashi (2019) covering 28 studies. Preparing to teach beat preparing for a test by a small-to-medium margin (d of about .30 to .40). Preparing to teach and then actually teaching came in higher, a medium effect (d of about .5 to .6).

That chapter also has details most Feynman guides miss:

In Fiorella and Mayer's studies, students who only prepared to teach did about as well on an immediate test as students who taught. Students who taught, by recording a video lecture for an imaginary student, did best on a delayed test. Planning to explain may not be enough. Doing it seems to matter.
In one study, students who taught from memory beat students who taught with notes on a delayed test. So the "no notes" rule in step 2 isn't a nicety. Explaining works partly as retrieval practice.
Results for written explanations are mixed. One study found that writing explanations for yourself worked better than writing them for an imagined audience.

Why the gaps stay hidden

There's also a reason step 3 exists. In a classic set of experiments, Rozenblit and Keil (2002) found that people feel they understand complex things with far more depth than they really do, and that this illusion is stronger for explanatory knowledge than for facts or procedures. Rereading your notes feeds that feeling. Trying to explain punctures it.

How strong is the evidence for the branded routine?

Weaker than the internet suggests. We found few direct tests of the Feynman technique by name, and none strong enough to build a study plan on.

The larger picture is the 2013 review by Dunlosky and colleagues, which rated self-explanation "moderate utility." Practice testing and distributed practice got the top rating. The authors noted that self-explanation hadn't been adequately tested in real classrooms, and that the durability of its effects was "woefully underexplored." That review is over a decade old, but the ranking is a useful sanity check: this is a good tool, not the best one.

Where the Feynman technique falls short

It's slow. One concept can take 10 to 20 minutes. That's fine for the ten ideas your exam hinges on. It's not a plan for 300 pages. The York guide also flags that the method suits complex ideas better than simple or memorization-heavy topics.
It's weak for isolated facts. Fiorella's review says learning by teaching suits conceptual material such as how a system works, and may be less effective for things like foreign-language vocabulary. Use flashcards for those.
You can be fluent and wrong. Nobody checks your explanation unless you do. Always compare against the source in step 4. In Studycamp, the AI chat can check a gap against your uploaded files, but it can't tell you what those files get wrong. Explaining to a real person who asks questions can help too. In one study, teaching followed by a question-and-answer period beat teaching alone.
It doesn't schedule itself. Explaining a concept once won't keep it in memory for the exam. Come back a few days later and try again, which is where active recall takes over. For how this fits with other methods, see how to study effectively.

A routine you can copy

After a lecture or reading, list the 3 to 5 concepts that would show up on an exam.
For each, set a 5-minute timer and explain it on a blank page, notes closed.
Circle the gaps. Check only those against your material.
Explain again, out loud if you can. If a friend is around, use them.
Two or three days later, do it once more, without looking at your last attempt.

In Studycamp, Notes support folders, so you can keep one page per concept and add each new attempt below the last one.

Run the Feynman technique in Studycamp

If your slides and PDFs already live in Studycamp, each step of the technique maps onto a feature. The AI checks your explanation against your uploaded files, not against the whole internet, so it only knows what you gave it.

Pick the concept. Open your course workspace and choose one idea from that material.
Explain it on your own, first. Write the explanation in your own words in Notes, or say it out loud and record it with voice notes, which transcribes the recording. Do this before you ask the AI anything. If the AI explains it first, you've skipped the part that works.
Mark the gaps and check them. Circle the vague spots, then ask the AI chat about each one. Its answers cite the passages from your files, so you can open the source and confirm.
Explain it again and test yourself. Rewrite the explanation without looking, then take a quiz built from the same files. Each answer comes with an explanation.

The chat and the quiz work from your files. If your files are wrong or incomplete, the checks will be too, so treat the AI as a second reader, not a judge.

Start studying free on the free plan (3 workspaces, 1 GB of storage and 1,500 AI credits a month), or see pricing for the paid plans.

FAQ

What are the 4 steps of the Feynman technique?

Pick one concept. Explain it from memory in plain words. Mark every place you got stuck or stayed vague. Go back to your material for only those gaps, then explain it again.

Who invented the Feynman technique?

We found no record of Richard Feynman describing it. The name seems to date from around 2011, more than 20 years after his death. It's built on a real Feynman idea, that if you can't explain something simply you don't understand it yet, plus later research on self-explanation and teaching.

Does the Feynman technique work?

The parts of it that are well studied, explaining ideas in your own words and preparing to teach, help learning on average. Meta-analyses find small-to-medium and medium effects. We found few direct tests of the branded four-step routine, so treat it as a sensible use of good principles, not a proven system.

How is the Feynman technique different from active recall?

Active recall means pulling information out of memory, usually with questions or flashcards. The Feynman technique is a form of recall that also asks you to build an explanation, so it tests understanding of how things connect, not just recall of a fact. Use both. See the active recall guide.

Can I use the Feynman technique for memorizing?

Not really. It's built for concepts, processes, and "why" questions. For dates, vocabulary, or formulas you need to memorize, flashcards and spaced practice tend to work better.