How Do You Study Effectively? What the Evidence Backs

Studying effectively means testing yourself instead of rereading, and spreading sessions over days instead of cramming. Dunlosky and colleagues (2013) reviewed ten common techniques and rated only practice testing and distributed practice as high utility. Rereading, highlighting and summarising rated low. Elaboration, self-explanation and interleaving sit in between.

Why Do You Feel Ready and Still Go Blank in the Exam?

It's 11pm the night before. The chapter's been read three times. The pages are yellow where the highlighter went through them, and everything on them looks familiar. You close the notes and feel, for the first time this week, a bit calm.

The next morning, question one has a term in it you know you know. You can see the page, roughly where it sat, bottom left, near the diagram. The definition won't come.

You recognise it. You can't produce it.

And then the story starts. Bad at exams. Rubbish memory. Should have started earlier, or put in more hours, or been the sort of person who doesn't panic. Except the hours were there, probably more than the person next to you put in. The hours were never the problem.

I picked Mechanical Engineering partly to dodge reading and writing, and averaged 80% across the degree. That's a number about exams. It tells you what I could do on the day, and that's all it tells you.

There's a name for the feeling in that exam hall. Call it the knowledge phantom: the sense of knowing something that isn't there when you reach for it. Most students will recognise it, and it's built by the exact things most of us do when we study.

Why does rereading feel like it's working?

Because it is working, at one thing. Each pass makes the page more familiar, and familiarity feels like knowledge. The words come easier, the argument is predictable, nothing surprises you. That ease is real. It just isn't memory.

The cleanest demonstration is Roediger and Karpicke's 2006 experiment. Students who reread a passage beat students who tested themselves on a test five minutes later. A week later it flipped, with the testers recalling 61% against 40% for the rereaders, and the rereaders were the more confident group. Active recall versus passive reading walks through that experiment in full.

The five-minute result is the one that matters the night before an exam. When everything is fresh, rereading genuinely looks better. That's where the knowledge phantom is made, and why the feeling points the wrong way.

Does everyone really study like this?

Karpicke, Butler and Roediger (2009) asked 177 Washington University undergraduates how they studied when nobody was watching. Of them, 84% listed repeated reading, and 55% ranked it as their number one strategy. Only 11% (19 of the 177) said they practised retrieval while studying.

To be fair to the students, 40% also mentioned flashcards and 43% mentioned practice problems, which can be forms of self-testing. The authors' point was that those numbers are dwarfed by rereading. This was a selective university, full of people who are good at exams. If the default there is to read the chapter again, the default isn't a sign that you're careless.

You inherited it. Rereading is what you were shown in school, it feels like diligence, and nothing in the process ever tells you it isn't working. That's the part worth sitting with. Rereading can't tell you what you don't know. A test can. So you find out in the exam hall, which is the most expensive place to find out.

What's actually going on in your head?

Robert Bjork, a memory researcher at UCLA, separates two things we usually treat as one. Performance is what you can do right now. Learning is the durable change underneath it. They come apart all the time. You can perform well on the night before and have learned very little, because the performance was borrowed from freshness.

Bjork also describes why forgetting isn't the enemy here. Memories aren't deleted, they become harder to reach. And the easier something is to reach right now, the less restudying it adds. In his phrasing, forgetting "creates the opportunity to reach additional levels of learning." By that logic, rereading something you read an hour ago buys you very little, because it's still easy to reach. A method that feels smooth is a poor sign. The smoothness is the problem.

None of this is a fault in you. It's a fault in the method you were handed, and in the feeling that method produces.

What Did Dunlosky's Review Actually Rate?

In 2013, John Dunlosky, Katherine Rawson, Elizabeth Marsh, Mitchell Nathan and Daniel Willingham published a monograph in Psychological Science in the Public Interest that went through ten learning techniques one at a time. For each, they asked how well its benefits hold up across different learners, materials, conditions and kinds of test. Then they rated each one.

Two things worth knowing about the ratings before treating them as law.

First, "low" doesn't mean useless. It means the technique works only under limited conditions or doesn't consistently improve performance. The review's authors said students should swap in better methods, such as practice testing in place of rereading.

Second, "moderate" is partly a statement about thin evidence. For elaborative interrogation and self-explanation, the reviewers said they hadn't been well tested in real educational settings. For interleaving, systematic study had only just begun. The two high-utility techniques are the ones with the widest evidence trail across ages, materials and conditions.

Why Is Practice Testing the Strongest Study Method?

Practice testing is the strongest because it forces you to rebuild an answer from memory, and the rebuilding is what makes the memory last. That's the testing effect, and the 2006 result above is one of its cleanest demonstrations.

It also beat something that looks cleverer than rereading. In 2011, Karpicke and Blunt had students learn science texts either by building concept maps, an elaborative method that feels deep, or by reading the text and then recalling it with the text put away. Straight after studying, both groups recalled about the same. A week later, the recall group retained about 50% more, and they did better on inference questions too, the ones that needed ideas connected, not just repeated. Students had predicted the opposite. (This one comes from the Purdue team's own description of the study, so treat the 50% as a headline figure rather than a precise effect size.)

How do you do it for a course, in a way that doesn't need a special app?

• The blank page. After a lecture or a chapter, close everything and write what you can remember on a blank sheet. Messy is fine. Then open your notes and mark what you missed in a different colour. The missed bits are your next session.
• Questions from your headings. Turn each heading or lecture title into a question and answer it from memory a day or two later.
• Past papers, early. Not as a final check. As a learning method. Doing a question from memory teaches you more than reading its model answer, and the 2006 students had no feedback at all, so a check against your notes afterwards is a bonus on top of what testing alone achieved.
• Flashcards, if you will actually make them. The card is only a prompt. If you flip it before you've tried to answer, you've practised recognition.
• Explaining it out loud. To an empty room if necessary. The place your explanation stalls is where your memory has a gap.

The unpleasant part is that it feels worse. You'll get things wrong. A blank page after you thought you knew the chapter is a small humiliation. That feeling is the signal, not a failure. It's the exam hall moment arriving on a Tuesday, when it's still cheap.

How Far Apart Should You Space Your Study?

Far enough that you've partly forgotten, and the right distance depends on how long you need to remember. Cramming isn't useless for a test tomorrow. It's poor for anything beyond that.

Cepeda and colleagues pooled the spacing literature in 2006: 839 assessments of distributed practice, from 317 experiments in 184 articles. The headline finding is that the gap between sessions and the length of time until the final test work together. The best gap grows as the test gets further away.

Their 2008 study put numbers on it. More than 1,350 people learned facts, got a review after a gap of up to 3.5 months, and were tested up to a year later. Making the gap longer helped at first, then gradually hurt. And measured against the delay until the test, the best gap shrank, from roughly 20 to 40% of the delay when the test was a week away to roughly 5 to 10% when it was a year away.

Translated into a calendar, the arithmetic is simple. For an exam seven days away, 20 to 40% is a gap of about 1.4 to 2.8 days, so a day or two between sessions. For something you want to hold for 365 days, 5 to 10% is about 18 to 37 days, so a few weeks between sessions. For an exam in between, a gap somewhere in the middle is a reasonable reading, though that's my interpolation, not a published finding.

The practical rule: the further the exam, the fewer sessions you need per topic, and the further apart they can sit.

There's also an uncomfortable consequence. If you only start three days before an exam, you can't space much. You can still test yourself instead of rereading, and that alone is worth more than the extra hours. But the earlier you start, the more each session buys. For the longer version of the mechanism, what spaced repetition is and why it works goes through the algorithms, and the forgetting curve shows what you're up against.

What About Elaboration, Self-Explanation and Interleaving?

They're the second tier: worth using, with thinner evidence, and best added on top of testing and spacing rather than instead of them.

Elaborative interrogation and self-explanation. Both come down to asking why. Why is this true? How does this step follow from the last one? How does this connect to something I already know? They fit naturally into the testing session. When you get a question wrong, writing a sentence on why the right answer is right is a small piece of self-explanation, and it turns a miss into something you can learn from. Dunlosky's team rated both moderate and flagged how little testing there'd been in classrooms, so these are good habits, not guarantees.

Interleaving. Instead of doing ten problems of one type and then ten of the next, you mix them. It feels worse. You keep having to work out which method applies, and that's exactly the skill an exam tests.

The best classroom evidence I found is in maths. Rohrer, Dedrick, Hartwig and Cheung ran a randomised trial across 54 grade 7 maths classes (787 students, 5 schools). The What Works Clearinghouse record puts the main contrast at 103 days of practice and a test 43 days later, so roughly three and a half months of practice and an unannounced test about six weeks on. The interleaved group scored 61%. The control group scored 38%.

That's a large gap, but it's one subject and one age group. Interleaving looks strongest where problems resemble each other and choosing the method is the hard part: maths, physics, chemistry, statistics. For a subject that's mostly facts and essays, the evidence is thinner. A useful version for those subjects is mixing topics in your weekly test session rather than finishing one topic before starting the next.

Can You Keep Rereading, Highlighting and Summarising?

You can, as a first pass. The mistake is letting them be the whole plan. Low utility doesn't mean you have to banish them. It means they can't be what you rely on, and each of the three can be turned into a test with one small change.

That table is my synthesis, not a finding from any one study. But each row follows from the research above: the same material, the same time, with retrieval swapped in for recognition.

What Would a Weekly Study Routine Look Like?

Here's one, built from those findings. Treat it as a starting shape, not a protocol that's been tested as a whole. Nobody has run a trial on this exact schedule. Each piece has evidence behind it, and the way they combine is my design.

It assumes a few courses, an exam several weeks away, and that you're already going to lectures and doing the reading.

Two details do most of the work.

The first is the blank page on the same day as the lecture. It costs ten minutes and it's one of the cheapest ways to find out what hasn't gone in, while the lecture is still fresh enough to fix.

The second is that fixed weekly slot. Spacing doesn't need willpower if it's a calendar event, and a mixed set once a week means every topic comes back at an increasing gap without a spreadsheet.

If the exam is two weeks away rather than two months, the shape compresses. Skip the same-day step for everything you've already read, run the blank-page recall on each topic now, repeat it a day or two later, then move to past papers. You'll lose some of the spacing benefit. You'll keep most of the testing benefit, and that's the larger part.

Where Does a Tool Like Alexandria Fit?

You don't need one. A notebook, a calendar and the willingness to close the book is the whole method, and it's older than any app. The 61% against 40% result was a pencil and a blank sheet.

The honest case for a tool is the friction. Testing yourself on what you read takes a step most people skip, and skipping a step is what the knowledge phantom feeds on. I built the knowledge and review layer in Alexandria because I wanted the time I spent reading and listening to stop being wasted, so that what I focused on comes back to me without digging through my own notes. The point was understanding and applying it, not collecting it.

For coursework specifically, Alexandria's web reader opens PDFs, so lecture slides and papers can sit in the same library as the rest of your reading. And one caveat I'd rather say than leave out: a review layer can feel harder than rereading a highlight. By the logic of this whole post, that's a feature.

If you want the reading side of this in more depth, active recall versus passive reading goes through the retrieval research, and how to learn faster covers the general version of the method beyond courses and exams.

The Short Version

Test yourself. Space it. Mix it. And when it feels worse than rereading, that's usually the sign it's doing something.

What's the last thing you were sure you knew, until someone asked you to say it out loud?

---

Frequently Asked Questions

What is the most effective way to study?

Test yourself on the material without looking at it, and spread those tests across several days. Dunlosky and colleagues (2013) rated practice testing and distributed practice as the only two high-utility techniques out of ten they reviewed. Everything else, including rereading and highlighting, rated moderate or low.

How do you study effectively for exams?

Read each topic once, then close the book and write down everything you can recall. Check the gaps against your notes. Repeat the recall a day or two later, then weekly, mixing topics together. Finish with past papers. Your revision becomes a series of small tests rather than a series of rereads.

Is rereading your notes a waste of time?

Mostly, yes, as your main method. Dunlosky and colleagues (2013) rated rereading low utility. In Roediger and Karpicke's 2006 experiment, students who reread a passage about 14 times remembered 40% a week later, while students who tested themselves remembered 61% after reading it about 3 times. Rereading is fine for a first pass.

How far apart should study sessions be?

A day or two apart for next week's exam, a few weeks for next year's. The best gap depends on how long you need to remember. Cepeda and colleagues (2008) found it was roughly 20 to 40% of the time until the test when that was a week, and roughly 5 to 10% when it was a year.

Does highlighting help at all?

Highlighting rated low utility in Dunlosky and colleagues' 2013 review, meaning it works only under limited conditions or doesn't reliably improve performance. It's harmless as a first pass if you then use the marked passages as the source of questions to answer from memory. The highlighting itself isn't the learning.

What is interleaving and should I use it?

Interleaving means mixing problem types in one session instead of practising one type at a time. In a 2020 trial across 54 seventh-grade maths classes, interleaved students scored 61% on an unannounced test about six weeks later, versus 38% for the control group. Most of the evidence is in maths, where picking the right method is the skill.

Why does effective studying feel harder?

Because retrieval takes effort, and the effort is part of what builds memory. Robert Bjork separates performance now from durable learning, and the two often come apart. Rereading raises how fluent the page feels. It barely changes what you can produce later. A method feeling easy is weak evidence that it's working.

Do these techniques work for every subject?

Practice testing and spacing have been tested across many ages, materials and subjects, which is why Dunlosky and colleagues rated them high. The middle-tier techniques have thinner evidence, and Dunlosky's team said so. Interleaving looks strongest where problems resemble each other, such as maths and science. For other subjects the evidence is less settled.

---

Last updated: 2026-10-22T00:00:00Z