What the data actually shows
The standout finding is the testing effect, demonstrated repeatedly by Roediger and Karpicke. In their work, students who studied material and then practised recalling it remembered far more on a later test than students who simply restudied the same material for the same time. The act of retrieving information — pulling it out of memory rather than re-reading it in — strengthens it, even though it feels harder and people often predict, wrongly, that rereading will serve them better.
Spacing and interleaving add to this. Distributing study sessions across days, rather than cramming them together, produces better long-term retention for the same total time. Interleaving — mixing problem types instead of doing many of one kind in a block — improves the ability to tell problems apart and apply the right approach, though it feels more confusing in the moment and tends to lower performance during practice while raising it on later tests.
A widely cited review by Dunlosky and colleagues (2013) evaluated common study techniques and rated their evidence. Practice testing and distributed practice came out as the most broadly effective and generalisable. By contrast, the techniques students use most — rereading and highlighting or underlining — were rated low in usefulness, because they produce a feeling of familiarity without reliably improving recall or understanding.
Why this feels different from how it actually is
Rereading and highlighting feel like learning because they feel easy and familiar. Each pass over the text is smoother than the last, and the brain reads that growing fluency as growing knowledge. But fluency with the material in front of you is not the same as being able to produce it later from memory, which is what tests, and life, actually require.
The better techniques feel worse for the opposite reason. Retrieval, spacing, and interleaving all introduce what researchers call desirable difficulties — they make practice harder and slower and surface your errors. That friction is precisely where the learning happens, but it feels like struggling, so people abandon it for methods that feel more reassuring.
There is also a prediction problem: people are poor judges of their own learning. Studies repeatedly find that learners rate the easy, fluent methods as more effective than the hard ones — the exact opposite of what the test results show. We trust the feeling of confidence, and the feeling is most misleading exactly when we are doing the least durable kind of study.
The feeling of fluency while rereading is not the same as having learned — and confusing the two is one of the most common study mistakes.
What the research says to do about it
Test yourself instead of rereading. After reading a section, close it and try to recall the key points from memory — with flashcards, practice questions, or a blank-page recall — before checking. The effort of retrieval, including the parts you get wrong, is what builds durable memory. Treat self-testing as a learning tool, not just a way to check what you already know.
Spread study across time rather than massing it. The same total hours, distributed over several days or weeks, produce better long-term retention than the same hours in one block. Build in deliberate review of older material so it gets revisited just as it is starting to fade.
Mix it up. Rather than drilling one topic or problem type to fluency before moving on, interleave related types within a session. It will feel harder and your in-session performance may dip, but interleaving improves your ability to choose the right approach later — which is usually the real goal. Combining retrieval, spacing, and interleaving works better than any one alone.
What the research says does not help
Rereading is the clearest example of effort that doesn't pay. Reading the same passage again and again produces a strong sense of knowing it and little durable gain. The Dunlosky review rated it low precisely because its felt benefit so badly outruns its measured one.
Highlighting and underlining fare little better on their own. They feel active but mostly mark text without engaging memory or understanding, and they can even hurt by drawing attention to isolated phrases at the expense of how ideas connect. As a standalone technique, the evidence for them is weak.
Matching study to your supposed 'learning style' — visual, auditory, and so on — has repeatedly failed to show benefits in controlled studies, despite being widely believed. And massed cramming, while it can get you through a test the next day, is one of the worst approaches for remembering anything beyond it.
The techniques that make studying feel smooth tend to teach you the least; the ones that feel effortful tend to teach you the most.
What this looks like in real life
The smooth reread that teaches nothing
Reading a chapter a third time feels great — each pass is smoother, and the brain reads that growing fluency as growing knowledge. But being able to follow the text in front of you isn't the same as producing it later from memory. The confidence rereading builds is exactly the trap: it's most reassuring precisely when you're doing the least durable kind of study.
Swapping rereading for recall
Instead of rereading notes, close them after a section and try to write down the key points from memory before checking. It feels harder and surfaces your errors — that friction, a desirable difficulty, is where the learning happens. Spreading those recall sessions across several days and mixing related topics compounds the effect beyond any one of them alone.
Real numbers in context
The gap between testing and rereading is large in the studies that find it. In classic retrieval-practice experiments, learners who practised recall substantially outperformed those who restudied on delayed tests, and the advantage widened as the delay grew. The precise sizes vary by material and design, so they are best read as a consistent, sizeable pattern rather than a fixed multiplier — but the direction is unusually robust.
The Dunlosky review's rankings are the headline number to remember: of the common techniques assessed, practice testing and distributed practice were rated high-utility and broadly generalisable, while rereading and highlighting were rated low. That those low-rated methods are also the most popular is the central, well-replicated irony of how most people study.