What the Feynman technique is, where the name really comes from, and what research on self-explanation and teaching says about explaining ideas to learn them.
By Bleart Emini, founder · Published · Last reviewed · 7 min read
You read a chapter, it all made sense, and then a practice question asks you to say it in your own words and you have nothing. The Feynman technique is a name for the fix: explain the idea in plain language, notice where you stall, and go back to the source for exactly those gaps. Research supports the core move, explaining to yourself or to someone else, but not under that name, and the evidence is narrower than the internet makes it sound. This article covers what to do, what is known, and where it breaks down.
The version you will find online has four steps.
The point is not a pretty explanation. The point is that a vague sense of "I get it" becomes testable the moment you have to produce the words.
The technique is popularly attributed to the physicist Richard Feynman, and the attribution is thin. One write-up that went looking for the four-step method, or the phrase itself, in anything attributed to Feynman reports that it could not find either. The earliest use of the name it found is a 2011 blog post by the writer Scott Young, with the four-step version circulating in 2012 (templacity.com, listed in Sources). That is a secondary source, not a scholarly history, so treat the dates as approximate.
What Feynman is credited with is a related habit: not trusting your own understanding until you can reduce an idea to something simple. That is a fair thing to admire. It is not the same as a documented study method with his name on it. This article uses the name because it is the term people search for, not because Feynman is known to have invented a protocol.
Here is an example, not my story. Say you are studying the psychology term negative reinforcement.
First attempt, written fast: "Negative reinforcement is when something bad happens so you do the behavior less." You reread it and the problem is obvious. That sentence describes punishment, and you have just confused the two.
Go back to the textbook for that gap only. Negative reinforcement means removing something unpleasant after a behavior, which makes the behavior more likely. Second attempt: "Your car chimes until you buckle your seatbelt. The chime stops when you buckle up. Because the annoying thing went away, you are more likely to buckle up next time. 'Negative' means something was taken away, and 'reinforcement' means the behavior goes up."
The second version has an everyday example, a stated meaning for each word, and a built-in way to tell it apart from punishment. The useful part was not writing it out. It was the moment the first draft exposed the mix-up.
My own experience here is small and I will keep it to what is true. One real example for me is a teacher breaking a psychology concept down into examples. I do not remember which concept, so I will not name one.
The harder part is studying without a tutor. When nobody is there to break a concept down for you, you have to do it alone, and that can take longer than the studying itself. This is the real cost of the technique. It works, but it is slow, and it is hardest on exactly the topics you understand least, which is where you most need it.
Two bodies of research sit under this technique. Neither tested "the Feynman technique" as such.
Self-explanation. Chi, de Leeuw, Chiu and LaVancher (1994, Cognitive Science) asked 14 eighth-grade students to explain to themselves after reading each line of a passage on the human circulatory system. A comparison group of 10 students read the same text twice with no prompts. The prompted group improved more from pretest to posttest, and within that group the students who produced more self-explanations understood more. The authors also reported that all the high explainers ended up with a correct mental model of the circulatory system, while many control students and low explainers did not. The limits matter: it was a small sample, one passage, and eighth graders rather than the college students many people reading this are. It also does not tell you that explaining out loud to another person beats explaining silently to yourself.
Learning by teaching. Chase, Chin, Oppezzo and Schwartz (2009, Journal of Science Education and Technology) studied what they call the protege effect. Students used a program in which they taught a character, called a teachable agent, about biology. Others used the same software but believed they were learning for themselves. In the first study, with eighth graders, the students who thought they were teaching spent more time on learning activities and learned more, with the benefit strongest for lower-achieving students. The authors suggest that feeling responsible for the agent drove the extra effort. Again the caveats: students were teaching a piece of software, not a person, and the findings are from school-age students in one kind of environment. The main thing it supports is that expecting to teach can raise effort, which is a motivational effect as much as a memory one.
The wider picture. A well-known review by Dunlosky and colleagues (2013, Psychological Science in the Public Interest) examined self-explanation among ten study techniques. Self-explanation is one of the techniques covered; this article does not quote the review's rating for it.
Taken together, the honest summary is this. Generating explanations has research behind it, mostly short lab and classroom studies. Whether the specific four-step routine beats other ways of studying was not found to have been tested directly in the sources checked. Explaining is also not a substitute for checking yourself: a smooth explanation can still be wrong, and the technique offers no built-in way to know. Pair it with testing yourself on the material.
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.