passive marking vs generative processing; what to do instead
Why does highlighting do almost nothing? Because highlighting is a form of passive review that does not require your brain to retrieve the material, manipulate it, or generate an explanation; it only asks you to recognize what looks important and drag a marker across it. The act of marking leaves a trace on the page, not in memory.
The mechanism behind the failure is a distinction psychologists draw between shallow and deep processing. Recognizing a sentence as "this seems like a key point" is shallow. Restating that point in your own words, connecting it to something you already know, or imagining a test question about it is deep. Shallow processing produces a fragile, easily disturbed memory; deep processing produces one that survives a delay. Highlighting is, by design, shallow, and shallow traces decay quickly once the book is closed.
The bigger danger is the feeling of competence. Highlighting makes the page familiar on a second pass. Research on judgments of learning shows that people consistently mistake how easily something comes to mind for how well they know it. A page covered in yellow feels like a page you have mastered. It is actually a page you have seen twice. The feeling is generated by perceptual fluency, not by memory, and it grows stronger the more you highlight.
There is also a more basic reason test scores barely move: highlighting boosts recognition more than recall. If you reread a highlighted passage, you can pick out the marked sentences. But a test usually asks you to produce the idea without the text in front of you. Recognition and recall are different tasks, and highlighting trains only the easier one. This is why students who highlighted extensively can sail through an open-book review and then fall apart on a closed-book exam.
Suppose you read a paragraph arguing that sleep deprivation impairs working memory by reducing activity in the prefrontal cortex. With highlighting: you mark three sentences, spend a few minutes, and next week recognize those sentences but cannot paraphrase the argument or recall the mechanism. With the closed-book method: you write "sleep loss hits working memory because the prefrontal cortex stops sustaining attention," check against the text, and a week later can produce the argument from scratch and connect it to other material. Time spent is similar; outcome is not.
Highlighting does help in two narrow cases. First, when you return to the text later for review, the marks guide your eye to what was flagged as important, which is a small efficiency gain. Second, when highlighting is paired with an active step, such as writing a margin question, summarizing the marked chunk, or making a flashcard, the marks function as a cue rather than as the learning event itself. And outside studying entirely, if your goal is to find a passage again rather than learn it, highlighting is fine: book-marking, navigation, annotation for collaboration. The failure is specific to using marks as a substitute for study.