Readwyn 0.1.0

Does highlighting help you remember what you read?

A little for memory, and hardly at all for understanding. A large 2013 review of study techniques rated highlighting low utility. In a 1992 study by Sarah Peterson, students who underlined as they read did worse on the comprehension questions that needed an inference linking ideas from different parts of the text. John Dunlosky, who described the study in 2013, suggests that the underliners may have spent their attention on individual ideas and less of it on how those ideas connected.

Dunlosky and four colleagues' 2013 paper in Psychological Science in the Public Interest rated ten techniques by how well the evidence for each held up across different students, materials and kinds of test. Practice testing and spreading study out over time came top. Highlighting landed in the bottom group with rereading, the keyword mnemonic, mental imagery and summarising. Their complaint about highlighting and rereading was that most students use them and neither reliably improves results. In the version he wrote for teachers, Dunlosky lists Air Force trainees, children and undergraduates among the groups that did no better with a highlighter than with plain reading. He also admits that he owns a favourite highlighter himself, and says he wouldn't take students' away from them.

Writing a summary in your own words sits in the same low group. The review found that it helps some students on some tasks, undergraduates among them, and Dunlosky notes that younger students who can't yet write a good summary need a lot of teaching before it does them any good.

A larger pool of studies gives highlighting slightly more credit. Héctor Ponce, Richard Mayer and Ester Méndez pooled 36 published articles from 1938 to 2019, 85 effect sizes in all. Highlighting your own text improved scores on memory tests (an average effect size of 0.36) but not on comprehension tests (0.20). It helped college students and not school students. The odd result is that readers given text with the important material already highlighted did better than readers who marked their own: an average effect of 0.44 on memory and on comprehension alike, and a gain for both age groups (0.41 for college students, 0.48 for school pupils).

The gap between the two suggests that the choosing is where readers go wrong. A 2025 analysis of the digital PIRLS 2021 reading test looked at how fourth-grade pupils in five countries used the highlighter built into the test. Marking the text that a question needed went with a correct answer in only one of the five countries. The same paper's review of earlier work notes that school pupils pick less relevant text than adults, and cites a 2018 study in which highlighting helped adults reading on paper but not adults reading on a screen.

Timing changes the picture too. Carole Yue, Benjamin Storm, Nate Kornell and Elizabeth Bjork had students read a passage twice, with or without a highlighter, and tested them a week later. Those who highlighted did better, mostly when their two readings came back to back rather than spread apart. Highlighting didn't cost them anything on the sentences they left unmarked. And the students who doubted that highlighting worked gained more from it than the ones who believed in it.

Dunlosky's advice on the same page is that highlighting is only the beginning: once students have read and marked a text, they should go back over it with one of the techniques higher up his list, and practice testing sits at the top. Peterson's underliners lost track of the links between ideas, and a highlight read on its own, months later, hides those links as well. Readwyn's Notebook has a "Show context" button for that: it swaps the highlight for the paragraph before it, its own paragraph and the paragraph after. Read with its surroundings, a highlight works as a prompt. Cover the line, recall what it said and why it mattered, then check; the post on why you forget what you read covers that kind of retrieval.