# How to study effectively: what the research shows

**What is the most effective way to study, according to research?** The most effective way to study, the research shows, is to test yourself and to spread study over days and weeks. Dunlosky and colleagues' 2013 review of ten techniques rated practice testing and distributed practice high, interleaving moderate, and rereading, highlighting and summarising low. Most of the evidence comes from memory tests of facts and short texts.

Published 2026-09-26 by EdCitation. https://edcitation.com/newsletter/how-to-study-effectively-what-the-research-shows

The research on learning points one way: test yourself, and spread your study over days and weeks. Those two are the study tips best backed by the science of learning. Of ten techniques in the best-known review, only practice testing and distributed practice were rated highly, while rereading and highlighting, the two students use most, were rated low.

The evidence has limits: most experiments tested memory for facts or short passages, soon after studying, with American undergraduates. This guide says what each study did, what it found and where it stops.

EdCitation publishes this guide. Every study but one was read in the paper itself in September 2026; each reference was built from its DOI in EdCitation's [Cite a source](https://edcitation.com/cite), the list was checked in [Verify references](https://edcitation.com/verify-references), and [Find sources](https://edcitation.com/) is where to look for more. The weekly plan near the end is ours, built from the findings; no study tested it.

## What does the research say about how to study effectively?

The research says practice testing and distributed practice work best. Dunlosky et al. (2013) rated ten techniques students can use alone for utility: whether the benefit holds across conditions, students, materials and kinds of test. Practice testing and distributed practice were rated high, because they helped learners of different ages and abilities on many tests, including in real classes. Elaborative interrogation (asking yourself why a fact is true), self-explanation and interleaved practice were rated moderate: promising, but less tested. Summarising, highlighting, the keyword mnemonic, imagery for learning text and rereading were rated low.

Low is not useless: a technique could be rated low because it helps with only narrow material, or because the evidence was too thin.

| Technique | What the evidence says | Source |
| --- | --- | --- |
| Practice testing (retrieval practice) | High utility; beat restudying on tests two days and a week later | Dunlosky et al., 2013; Roediger and Karpicke, 2006 |
| Distributed practice (spacing) | High utility; spaced study beat massed study across 254 studies | Dunlosky et al., 2013; Cepeda et al., 2006 |
| Interleaved practice | Moderate utility; helps most in telling similar categories apart, not with texts or word lists | Dunlosky et al., 2013; Brunmair and Richter, 2019 |
| Elaborative interrogation, self-explanation | Moderate utility; not yet well tested in classrooms | Dunlosky et al., 2013 |
| Rereading | Low utility; does much less than the alternatives in the same time | Dunlosky et al., 2013 |
| Highlighting and underlining | Low utility; little benefit in most studies, may hurt inference | Dunlosky et al., 2013 |
| Summarising | Low utility; helps those already skilled at it | Dunlosky et al., 2013 |
| Teaching to a learning style | Virtually no evidence for the pattern it needs | Pashler et al., 2008 |
| Checking these studies are real | All nine references verified against the publishers' records | EdCitation's [Verify references](https://edcitation.com/verify-references) |

### What a later meta-analysis added

Donoghue and Hattie (2021) pooled the studies the review cited: 242 studies, 1,619 effects, 169,179 participants. The top held. Distributed practice had the largest average effect (d = 0.85) and practice testing the next (0.74); rereading and interleaving came in at 0.47, underlining and summarising at 0.44. Every technique beat its control groups on average, and what those groups did varied.

Their limits are the field's: 93% of outcomes measured surface learning such as recall of facts, the effect on deeper learning was far smaller (0.26 against 0.60), three quarters of final tests came within a day, and 77% of studies were from the USA.

## Why does testing yourself beat rereading?

Testing yourself beats rereading because pulling information out of memory strengthens it, and the advantage appears after a delay, not straight away. Researchers call this retrieval practice, or the testing effect.

### The experiment most often cited

Roediger and Karpicke (2006) gave Washington University undergraduates, aged 18 to 24, passages of 256 and 275 words on the sun and on sea otters. In the first experiment, 120 students read a passage and then either read it again or wrote down everything they could remember, with no feedback. Five minutes later, rereading looked better: 81% of the passage's ideas recalled against 75%. Two days later the order had flipped, 68% for the tested group against 54%, and after a week it was 56% against 42%.

In the second experiment, with 180 students, four readings beat one reading and three tests after five minutes (83% against 71%); after a week the tested group won, 61% against 40%. Those who had only reread were the most confident they would remember.

The limits: two short passages, one university, a week at most. A meta-analysis by Adesope et al. (2017) concluded, in its abstract, that practice tests did more for learning than restudying and every other comparison condition, with effects that depended on the kind of test, the learners and the outcome measured. We read only the abstract (the paper is paywalled), and [Verify references](https://edcitation.com/verify-references) linked a corrigendum to it that we could not read.

### What students actually do

Karpicke et al. (2009) asked 177 Washington University undergraduates to list their study strategies. Rereading was listed by 84% and named as the main strategy by 55%; 11% described practising recall, and 1% put it first. After reading a chapter, 57% would reread it and 18% test themselves. Self-reports from one campus, but the gap is the finding.

Flashcards and past papers count as retrieval practice only when you answer before you look. Retrieval practice and active recall each get a guide of their own later in this series.

## How far apart should study sessions be?

Study sessions should be spread out, and the best gap depends on how long you need to remember: roughly a fifth of the time for a test a few weeks away, a much smaller share for a year.

### Spaced against massed study

Cepeda et al. (2006) reviewed the spacing research on verbal recall. Across 254 studies and 14,811 participants that compared spaced with massed presentation, people recalled 47.3% after spaced study and 36.7% after massed. Nearly all tested memory within a day, and the material was verbal, not mathematical problem-solving.

### The gap that works best

Cepeda et al. (2008) taught 1,354 online volunteers, average age 34, 32 obscure trivia facts, gave a review after a gap of up to 3.5 months, and tested them up to a year later. For every test delay, recall rose as the gap lengthened and then slowly fell. For recall, the best of the gaps tested, with final tests at 7, 35, 70 and 350 days, were 1, 11, 21 and 21 days. With study time held equal, the best gap raised final recall by 64% over no gap at all. In the authors' summary, the best gap was about 20% of the test delay for delays of a few weeks and about 5% for a year, and a gap that was too long cost much less than one that was too short.

Dunlosky and colleagues summarise this as 10 to 20% of the time you want to remember; the paper shows that share shrinking as the delay grows. Either way, space more rather than less. The limits: trivia facts, volunteers, one review session. Scheduling revision with spaced repetition has its own guide coming in this series.

## Does mixing topics help you learn?

Mixing topics, called interleaving, helps when the skill is telling similar things apart, such as which formula a problem needs, and may not help with texts or vocabulary. Dunlosky and colleagues rated it moderate. In their example, Rohrer and Taylor's work with college students learning the volumes of four solids, blocked practice scored higher during practice, but interleaved practice was well ahead on the test a week later.

Brunmair and Richter (2019) pooled 59 studies and found a moderate benefit overall (g = 0.42). It was largest for paintings (0.67) and other visual material, small for mathematics (0.34), not significant for expository texts, and reversed for words, where blocking did better (g = -0.39). Interleaving and mixed practice get a guide of their own later in this series.

## Are rereading, highlighting and summarising a waste of time?

Rereading, highlighting and summarising are not a waste of time, but each does less than self-testing in the same minutes, so Dunlosky and colleagues rated all three low.

### Rereading

Rereading helps recall, particularly when spaced, but its benefit for understanding is less clear. The largest strike against it, in the authors' view, is that the alternatives do more.

### Highlighting and underlining

In most situations studied, highlighting did little. It may help students who know enough to pick the right sentences, and may hurt on tasks that need inference.

### Summarising

Summarising works for people already good at it, but many learners, some undergraduates included, needed extensive training first. Summarising a source for an essay is a different job, covered in [paraphrase, quote or summarise](https://edcitation.com/newsletter/paraphrase-quote-or-summarise).

## Do learning styles matter?

No good evidence shows that teaching to a person's learning style improves learning. Pashler et al. (2008) set out what a fair test needs: students sorted by style, randomly given different methods, one final test for all, and a crossover, where the method best for one style is not best for another. They found virtually no studies showing that pattern, and several with the right design contradicted it. People do have preferences; what is missing is evidence that matching teaching to them helps, though many versions have never been properly tested. Learning styles get a full guide later in this series.

## What does a study week built on the research look like?

A study week built on the research puts self-testing first and returns to each topic after a gap. This plan is ours, drawn from the studies above; no study tested it as a whole.

1. **Test yourself on new material the same day or the next.** Close the notes, write down what you remember, then check and mark what you missed. Roediger and Karpicke gave no feedback; checking afterwards is our addition.
2. **Turn your notes into questions, not highlights.** A question you must answer from memory is retrieval practice; a coloured line is not.
3. **Return to older topics on a schedule.** Each week, self-test briefly on last week's topics and one from three or four weeks back. For a topic learned in week 1 of a ten-week term, our arithmetic from Cepeda's proportions gives reviews one to two weeks apart; err longer rather than shorter.
4. **Mix problem types once you can do each one.** Draw a practice set from several weeks, so that you must decide what each problem needs.
5. **Reread only what the self-tests showed you missed**, rather than whole chapters.
6. **Look up a source when a point needs one.** For an essay rather than an exam, search [Find sources](https://edcitation.com/), build the reference in [Cite a source](https://edcitation.com/cite) as you note the point, and look for the course's set books in EdCitation's [Library](https://edcitation.com/tools/library).

## How can you find and check the research yourself?

EdCitation is the best tool for checking a study claim at its source, because it looks every reference up in the publisher's record and never writes one. A chatbot composes a reference; it cannot look one up. This guide's list shows the difference: all nine references came back verified, and the check linked a published correction to one of them.

### Find the studies

[Find sources](https://edcitation.com/) searches about 300 million published works by topic or by the claim a sentence needs to support. In September 2026, a search for "testing effect retrieval practice long-term retention" returned 6,900,765 works. Its first result was a chapter from *Education in Anesthesia* with no year in its record; among the first ten were Roediger and Butler's 2011 review in *Trends in Cognitive Sciences*, cited 1,363 times, and Karpicke and Roediger's 2007 experiment on expanding against equally spaced retrieval, cited 194 times.

### Cite them

[Cite a source](https://edcitation.com/cite) builds a reference in six styles, APA 7 to Vancouver, from a DOI, title, PubMed ID, ISBN or web address, and screens it for retraction. From the DOI 10.1111/j.1467-9280.2006.01693.x it returned:

Roediger, H. L., III, & Karpicke, J. D. (2006). Test-Enhanced Learning: Taking Memory Tests Improves Long-Term Retention. Psychological Science, 17(3), 249–255. https://doi.org/10.1111/j.1467-9280.2006.01693.x

Two things to correct. APA 7 puts an article title in sentence case, so lower it to "Test-enhanced learning: Taking memory tests improves long-term retention" ([title case or sentence case](https://edcitation.com/newsletter/title-case-or-sentence-case) explains the rule); the same applied to Dunlosky's, Cepeda's 2008, Pashler's and Donoghue's titles. For Karpicke et al. (2009), whose record files the third author as "Roediger III", it gave "Roediger, H. L., III", the form APA uses.

### Verify the list

[Verify references](https://edcitation.com/verify-references) checks a pasted list or an uploaded paper against the publishers' records, flags retractions, and never shows "could not check" as "not found". This guide's nine references came back verified, each with the reason "The DOI resolves to this record and the title matches." None was retracted, and the Adesope meta-analysis came back with a link to its 2017 corrigendum.

### What it costs

Find sources, Cite a source and Verify references are free, with no account. A free account keeps a reference list and a reading list. Reading and borrowing books in the Library is part of Max, $24 a month; see [Pricing](https://edcitation.com/pricing).

## Quick questions

### Is rereading my notes a waste of time?

Rereading is not a waste, but a weak use of time: testing yourself does more in the same minutes, and in Roediger and Karpicke's experiment rereaders recalled less after a week.

### How long should I leave between study sessions?

It depends on how long you need to remember. Cepeda and colleagues found the best gap was about a fifth of the delay for tests a few weeks off and about a twentieth for a year, and too long a gap cost far less than too short.

### Are flashcards retrieval practice?

Flashcards are retrieval practice only when you answer before turning the card. Karpicke and colleagues note that flashcards and practice questions can be used without recalling anything.

### Does cramming work?

Cramming helps on a test that is close and does less for one days away: Dunlosky and colleagues note that massed study can win on an immediate test, spaced study on delayed ones.

### Where can I find the research on study techniques myself?

Search EdCitation's [Find sources](https://edcitation.com/) for the technique and the outcome, for example "testing effect long-term retention", then check the studies you choose in [Verify references](https://edcitation.com/verify-references).

## References

- Adesope, O. O., Trevisan, D. A., & Sundararajan, N. (2017). Rethinking the use of tests: A meta-analysis of practice testing. *Review of Educational Research*, *87*(3), 659–701. [https://doi.org/10.3102/0034654316689306](https://doi.org/10.3102/0034654316689306)
- Brunmair, M., & Richter, T. (2019). Similarity matters: A meta-analysis of interleaved learning and its moderators. *Psychological Bulletin*, *145*(11), 1029–1052. [https://doi.org/10.1037/bul0000209](https://doi.org/10.1037/bul0000209)
- Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. *Psychological Bulletin*, *132*(3), 354–380. [https://doi.org/10.1037/0033-2909.132.3.354](https://doi.org/10.1037/0033-2909.132.3.354)
- Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T., & Pashler, H. (2008). Spacing effects in learning: A temporal ridgeline of optimal retention. *Psychological Science*, *19*(11), 1095–1102. [https://doi.org/10.1111/j.1467-9280.2008.02209.x](https://doi.org/10.1111/j.1467-9280.2008.02209.x)
- Donoghue, G. M., & Hattie, J. A. C. (2021). A meta-analysis of ten learning techniques. *Frontiers in Education*, *6*, Article 581216. [https://doi.org/10.3389/feduc.2021.581216](https://doi.org/10.3389/feduc.2021.581216)
- Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students' learning with effective learning techniques: Promising directions from cognitive and educational psychology. *Psychological Science in the Public Interest*, *14*(1), 4–58. [https://doi.org/10.1177/1529100612453266](https://doi.org/10.1177/1529100612453266)
- Karpicke, J. D., Butler, A. C., & Roediger, H. L., III. (2009). Metacognitive strategies in student learning: Do students practise retrieval when they study on their own? *Memory*, *17*(4), 471–479. [https://doi.org/10.1080/09658210802647009](https://doi.org/10.1080/09658210802647009)
- Pashler, H., McDaniel, M., Rohrer, D., & Bjork, R. (2008). Learning styles: Concepts and evidence. *Psychological Science in the Public Interest*, *9*(3), 105–119. [https://doi.org/10.1111/j.1539-6053.2009.01038.x](https://doi.org/10.1111/j.1539-6053.2009.01038.x)
- Roediger, H. L., III, & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. *Psychological Science*, *17*(3), 249–255. [https://doi.org/10.1111/j.1467-9280.2006.01693.x](https://doi.org/10.1111/j.1467-9280.2006.01693.x)