Retrieval practice, and how to actually do it
One of only two study techniques rated high-utility across ages, subjects and abilities — and the one students are least likely to use.
Here is the uncomfortable finding at the centre of this: the study method that feels most productive is close to the least effective, and the one that feels like failing is close to the most.
Re-reading a chapter produces a strong sense of familiarity. The words go down easily the second time, you recognise everything, and you finish feeling that you know it. That feeling is generated by the fluency of reading, not by anything about your memory. It is why students consistently rate re-reading highly and consistently do worse than students who spent the same time testing themselves.
What retrieval practice is
Close the book and try to produce the answer from memory.
That is the whole technique. Not recognising the answer in a list, not checking whether it looks right — producing it, from nothing, before you look. The act of pulling something out of memory changes the memory, making it easier to pull out next time. Reading it again does not do this, or does far less of it.
In a large review of ten study techniques, practice testing and distributed practice were the only two rated high utility — meaning they worked across age groups, subject areas, and levels of prior knowledge. Highlighting, re-reading and summarising were rated low.
| Technique | Utility | What it means |
|---|---|---|
| Practice testing | Testing yourself, in any format | |
| Distributed practice | The same hours, spread across days | |
| Elaborative interrogation | Asking why a stated fact is true | |
| Self-explanation | Explaining a step as you take it | |
| Interleaved practice | Mixing problem types in a session | |
| Summarisation | Writing summaries of what you read | |
| Highlighting | Marking text while reading it | |
| Keyword mnemonic | Linking terms to keyword images | |
| Imagery for text | Forming mental images while reading | |
| Re-reading | Reading the material again |
Dunlosky, Rawson, Marsh, Nathan & Willingham (2013), Psychological Science in the Public Interest; an accessible summary is available from the AFT. Confirmed by a 2021 meta-analysis.
Why it feels wrong
Because it is harder, and because you fail at it visibly.
When you re-read, you never encounter evidence that you do not know something. When you test yourself, you encounter it immediately and repeatedly. Twenty minutes of retrieval practice produces a much less pleasant experience than twenty minutes of re-reading, and a considerably better outcome.
This is worth naming clearly, because the discomfort is the single biggest reason students abandon the technique.
The feeling of struggling to recall is not a sign the method is failing. It is the method.
How to actually do it
Five steps, and none of them needs an app.
Turn your material into questions before you study it
As you read a chapter the first time, write a question for anything you would be annoyed to forget. Do not write the answer. You now have a test instead of a highlighted document.
Answer from a blank page
Not multiple choice, not “does this look right”. Write the answer out. Free recall is harder than recognition and is worth more.
Check immediately, and look properly at what you missed
Feedback matters. Getting something wrong and finding out why is a large part of the benefit; getting something wrong and moving on is not.
Keep going until you get it right — then stop for now
The evidence favours retrieving to a criterion rather than a fixed number of repetitions. Once you have produced it correctly, further repetition in that same session adds little.
Come back to it days later, not hours later
This is the part almost everyone skips, and it is where most of the durability comes from.
Spacing is not optional
Retrieval practice and spacing are usually discussed separately, but combining them — retrieving to criterion, then returning after a gap, repeatedly — has a name and a track record. It is called successive relearning, and in a real university course it produced an improvement of more than a letter grade.
Rawson & Dunlosky (2022), Successive Relearning: An Underexplored but Potent Technique, Current Directions in Psychological Science.
You do not need a sophisticated algorithm for the gaps. An expanding ladder captures most of the benefit: try it again after about two days, then a week, then three weeks, then two months, resetting to the bottom of the ladder if you miss it.
- First return
- Second return
- Third return
- Fourth return
A session you can run tonight
Take one topic. Spend ten minutes writing questions from your notes without writing answers. Close the notes. Spend twenty minutes answering from a blank page, checking each one as you go and reading the explanation properly for anything you missed. Write tomorrow’s date next to anything you got wrong, and a date about a week out next to anything you got right.
That is a complete implementation. Everything else is bookkeeping.
Where Mento fits
The bookkeeping is the part that breaks down first, which is most of what we automate. Mento generates practice questions from material you upload, so step one stops being a chore. Questions come with graduated hints rather than a reveal button, because a hint you had to work with still counts as retrieval and an answer you were handed does not. Problems you solve come back on an expanding schedule.
None of that is necessary. You can run this with paper and a calendar, and people did for decades. What matters far more than the tool is not being talked out of the technique because it feels harder than the alternative.
Mento turns your own notes into practice questions, with graduated hints instead of answers and spaced review so what you learn stays learned.