Short answer: barely, and not in the way most students use it.
What the evidence says
The most thorough review of study techniques we have is Dunlosky, Rawson, Marsh, Nathan and Willingham (2013), published in Psychological Science in the Public Interest. The authors assessed ten common techniques against decades of experimental work and rated each for utility. Five came out low: summarisation, highlighting and underlining, the keyword mnemonic, using imagery for text learning, and rereading.
Two came out high. Practice testing and distributed practice — retrieving material from memory, and spreading study out over time. Those are the two this site keeps returning to, and the gap between them and highlighting is not small.
Low utility does not mean zero. It means the benefit is small, inconsistent, and dependent on conditions most students do not meet. Across the studies reviewed, highlighting a text and then rereading it tends to leave readers roughly where reading it once without a pen would have left them.
The wrinkle worth knowing
Highlighting done by the reader sometimes outperforms reading a text someone else has already highlighted. That is a useful clue about mechanism. Whatever small benefit exists appears to come from deciding what matters, not from the mark that records the decision. The judgement is the active ingredient. The colour is a by-product.
Which suggests an obvious problem with how the technique is usually practised. A student highlighting their way through a first read has not yet got enough of the material in view to judge what is important, so the marking becomes near-continuous, and the discrimination that would have made it useful never happens.
When it can actively hurt
There is evidence that heavy marking impairs performance on questions requiring inference across a passage, apparently because attention concentrates on the marked fragments and away from the relationships between them. On a test asking you to recognise isolated facts, that costs little. On a test asking you to reason from one idea to another — which is most university assessment, and nearly all licensure examination — it costs.
This is the part students find hardest to believe, and it is worth sitting with. The technique is not merely neutral. Used the way it is usually used, on the material that matters most, it can leave you worse off than a plain reading would have.
Why it feels like it works
Because it makes the page familiar, and familiarity feels like knowledge.
You finish a chapter with a visible record of effort and a text that now reads more smoothly than it did an hour ago. Both of those are real. Neither is evidence that you could produce any of it on a blank page tomorrow morning. The sensation of fluency is generated by the ease of processing a text you have just processed, and it tracks how recently you read something far better than it tracks whether you have learned it.
This is the fluency trap, and highlighting is its most popular vehicle — partly because it is the only study technique that leaves physical evidence of work behind. Four hours of retrieval practice looks like nothing. Four hours of highlighting looks like four hours.
What to do instead
You do not have to put the pen down. You have to change what it is for.
- Write questions in the margin rather than marking answers in the text. A page of questions can be used to test yourself next week. A page of yellow cannot be used for anything.
- If you do highlight, do it on a second pass. Deciding what matters is only possible once you know what the passage contains, and the second pass is the only one where that is true.
- Cap it. A hard limit of one or two marks per page forces the discrimination that produces whatever benefit exists.
- Then close the book and write down what you remember. That single step does more than any marking system, and it takes less time than highlighting the same chapter.
The last one is the whole argument in miniature. If you have twenty minutes and a chapter, twenty minutes of attempted recall will leave you with more than twenty minutes of careful marking. Not slightly more. Substantially more, and the gap widens as the exam gets further away.
In the book
Chapter 1 covers the fluency trap and why it is so persistent. Chapter 4 covers what to do with the time instead. Appendix B lists the studies in full.