On this page
Students spend most of their study time rereading notes and highlighting textbooks. Are those good uses of time? In 2013 five researchers published a long review that rated ten common techniques on the strength and breadth of their evidence. Many of our own study-skills posts cite it. Here is a look at the review itself: how the ratings were made, what they were, and what they leave out.
What question did the review ask?
John Dunlosky, Katherine Rawson, Elizabeth Marsh, Mitchell Nathan and Daniel Willingham published Improving Students' Learning With Effective Learning Techniques in Psychological Science in the Public Interest (volume 14, issue 1, pages 4 to 58). Their aim was practical: which techniques that students could easily use on their own are worth recommending?
They chose techniques that are easy to use, plus some, like highlighting and rereading, because students rely heavily on them. This post uses the review's abstract and the publisher's summary at the Association for Psychological Science.
How were the ratings made?
For each technique, the authors asked whether its benefits generalise across four kinds of variable:
Learning conditions: for example, studying alone or in a group.
Student characteristics: age, ability and prior knowledge.
Materials: from simple concepts to maths problems to complex science texts.
Criterion tasks: what the final test measures, such as memory, problem solving or comprehension.
They also looked at whether the technique had been tested in real educational settings. A technique rated high had to help many kinds of learners, with many materials, on many types of test.
What were the ratings?
According to the abstract:
High utility
Practice testing (testing yourself, including flashcards and practice questions).
Distributed practice (spreading study over time rather than cramming).
These helped learners of different ages and abilities and improved performance on many kinds of test, including in classrooms.
Moderate utility
Elaborative interrogation (asking yourself why a fact is true).
Self-explanation (explaining each step to yourself while learning).
Interleaved practice (mixing problem types in one session).
The authors judged these promising but limited by evidence. Elaborative interrogation and self-explanation had "not been adequately evaluated in educational contexts", and interleaving had "just begun to be systematically explored".
Low utility
Summarisation.
Highlighting or underlining.
The keyword mnemonic.
Imagery use for text learning.
Rereading.
The reasons differed. Summarisation and imagery helped some students on some tasks, but under limited conditions. The keyword mnemonic was hard to use in some contexts and helped mostly with limited materials over short retention intervals. Rereading and highlighting "do not consistently boost students' performance", and the authors suggest using other techniques in their place, "e.g., practice testing instead of rereading".
What does the review not show?
It is a rating of evidence, not an experiment. The authors judged existing studies; they did not run a new trial comparing all ten.
Ratings reflect 2013 evidence. The authors themselves said interleaving had only begun to be studied, so its rating was about how much was known, not a verdict that it works less well.
Combinations were not rated. Real study mixes techniques, such as spaced practice tests with self-explanation.
It is not specific to entrance exams. Maths problems and science texts appear among the materials, but no study of JEE or NEET-UG preparation is described.
What does it mean for students in India?
The pattern behind the ratings is that techniques which make you produce something from memory or reason about it did better than techniques where you look at the material again.
Swap one rereading session for a test. Close the book and write everything you remember about a chapter, then check.
Spread it out. Revisit each chapter after a day, a week and a month. Our guide to a spaced revision schedule gives a timetable.
Explain as you study. For each step in a derivation, ask why it follows. See the self-explanation technique.
Mix problem types once you have learned two or three chapters, instead of finishing one chapter's exercise in one go.
Keep highlighting minimal. If you highlight, turn each highlight into a question you test yourself on later.
Our post on why rereading feels effective but is not explains the trap behind the low ratings.
If you want to test yourself with someone checking your reasoning, Fermi's AI tutor checks understanding as it goes and decides whether to explain, ask or show, which turns a session into retrieval rather than rereading.
How to read a study like this
A narrative review rates evidence; it does not produce a single effect size. Read the rating criteria before the ratings. Here, "utility" means generality across learners, materials and tests, so a low score can mean "not yet shown widely" as well as "does not work".
Next step: this week, replace one rereading session with a closed-book recall of the same chapter, and compare how much you can actually produce.
