In a 2006 experiment, students who re-read a passage roughly fourteen times recalled 40% of it a week later. Students who read it three times and then tried to remember it recalled 61%. The re-readers were also the most confident they would do well.
You already know you should be using active recall — testing yourself instead of re-reading. Most people who study anything have heard it. And most people still re-read, because re-reading feels like it is working and self-testing feels like failing.
That feeling is the whole problem, and there is an experiment that measures it precisely. Understanding what it actually found will tell you why the wrong method feels right, and what to do at the moment studying starts to feel unproductive.
The popular version is “active recall beats re-reading.” It is repeated everywhere, and it is incomplete in a way that matters.
Henry Roediger and Jeffrey Karpicke ran the study people are referring to, published in Psychological Science in 2006. In their second experiment, 180 students learned a prose passage under one of three schedules. One group read it four times. One read it three times and then took a single recall test. One read it once, then tried three times to write down everything they could remember, with no feedback and no looking back.
Then they took a final test — some after five minutes, some after a week.
Here is the part that gets left out:
| Final test | Read four times | Read once, recalled three times |
|---|---|---|
| After 5 minutes | 83% | 71% |
| After 1 week | 40% | 61% |
After five minutes, re-reading won. Comfortably. If the test had ended there, the conclusion would have been that re-reading is the better technique.
A week later the ranking had completely inverted.
This is why re-reading is so hard to give up. Someone who crams the night before a morning exam is not being irrational or lazy. They are running the five-minute condition on themselves, and in the five-minute condition they are right. They get real evidence that their method works, and the evidence is accurate — for that test, at that delay.
What it does not do is leave anything behind. Re-reading is not useless. It is optimised for the wrong test.
The comparison is more lopsided than the percentages suggest, because the two groups did not put in the same work.
The researchers tracked how many times each group actually got through the whole passage. The re-readers made it through about 14.2 times. The recall group read it about 3.4 times — once, at the start, and never again.
So the group that did four times the reading remembered about two-thirds as much a week later. The extra twelve passes through the text did not just fail to help. The time spent on them would have been better spent trying to remember.
Before the final test, the researchers asked students to predict how well they would remember the passage in a week, on a scale of 1 to 7.
| Predicted | Actually recalled | |
|---|---|---|
| Read four times | 4.8 | 40% |
| Read once, recalled three times | 4.0 | 61% |
The predictions ran exactly backwards. The group that would go on to remember the least was the most confident it would remember the most, and the gap was statistically real rather than noise.
This is the finding with the most practical weight, and it is almost never the headline. It means you cannot use how studying feels as evidence that it is working. The sensation of fluency that comes from re-reading — the text getting easier, the sense that you know this — is produced by familiarity with the page, not by the ability to retrieve anything once the page is gone.
Retrieval feels worse because it is harder. That difficulty is not a sign you picked a bad method. It is closer to the point.
A large review of study techniques by Dunlosky and colleagues reached the same practical conclusion from a wider base of evidence: practice testing and spacing your sessions were rated high utility, while rereading and highlighting — the two things students report doing most — came out low.
The mechanics are unglamorous, which is part of why people skip them.
If you want one rule: the moment it starts feeling uncomfortable is the moment it starts working. Comfort is the signal you have drifted back to re-reading.
Consistency matters more than intensity here, the same way it does with any practice that compounds — the reset habits that work are the ones you actually repeat, and the same principle underlies learning how to learn as a skill in itself.
Not never. If you have a test in twenty minutes, re-reading is the better use of those twenty minutes — that is precisely what the five-minute result shows. If you are reading something for the first time and do not yet understand it, retrieval has nothing to pull on. Understand first, then retrieve.
The mistake is not re-reading. It is re-reading when what you actually want is to still know this next month, and designing your study around how memory really behaves rather than around what feels productive.
Worth being straight about the limits.
Re-reading is not a bad habit because it does nothing. It is a bad habit because it works on a timescale you do not care about, while feeling better than the thing that works on the timescale you do.
If you want to still know something next month, close the book and try to produce it from memory — and treat the discomfort as confirmation rather than a reason to stop.
Level Up Smarter publishes practical, science-informed strategies for learning faster, focusing longer, and building the skills that matter as the world keeps changing.
This article is educational and is not medical advice. It references a breathing practice covered elsewhere on this site; breathing practices are not a substitute for professional medical or mental-health care.
