Is Recalling Better Than Rereading? The Testing Effect Experiment

Students who closed the book and wrote down what they could recall remembered clearly more after 2 days and after 1 week than students who reread the passage. That is the result of an experiment by Roediger and Karpicke, published in Psychological Science in 2006.

In the Brain Mondo short video, I covered this experiment in about 40 seconds. In this post I go through the conditions the video didn’t mention, what it means that rereaders were more confident, and what to actually do with this when you study.

Video: The testing effect (Japanese)

The video is in Japanese.

What the testing effect is

We usually think of a test as a way to check whether you learned something. But taking a memory test also strengthens your memory for later. Psychologists call this the testing effect.

There has long been a question about it, though. When you take a test, you see the questions and material again. Maybe the testing effect is really just the effect of that extra look. Roediger and Karpicke used the kind of text you might read in a class to check this.

What the experiment did

There were two experiments, both with students. First, the students read and studied short prose passages.

After that, the groups did different things. Some closed the book and took a free-recall test, writing down everything they could remember. In some cases there was one test and in others three, and in both cases no feedback was given. Others, instead of taking tests, reread the passage the same number of times.

The final test came 5 minutes, 2 days or 1 week later.

The important detail is that the recall group never saw the correct answers. They had no chance to look again at what they failed to write down. The rereading group, meanwhile, saw the whole passage again. Judged only by how much material each group saw, the rereaders had the advantage.

The result: the order flipped between 5 minutes and 1 week

When the final test came after 5 minutes, the group that had reread remembered more. Right after reading, that seems natural.

On the tests 2 days and 1 week later, though, the order reversed. The group that had been tested and had to recall remembered substantially more. The paper ends by saying that testing is a powerful means of improving learning, not just assessing it.

This also answers the question from earlier. The recall group never saw the answers, so they lost to the rereaders on how much material they saw. Even so, after time had passed, the recall group came out ahead. The testing effect cannot be explained as just another look at the material.

If you only looked at the 5-minute results, rereading would have seemed like the better way to study. No advantage at first, then a clear one after time passes.

Rereaders felt more confident

As the video mentioned, the other striking result is about confidence.

Students who reread the passage became more confident that they would remember it. Yet it was the recall group whose memory actually lasted to the one-week test.

When you reread, the text starts to flow easily. It is easy to mistake that ease for having learned it. This result can be read as showing that feeling like you studied and actually remembering are two different things.

What other research says

In 2008, Karpicke and Roediger published a further experiment in Science. Students learned foreign language vocabulary. Once a student had produced a word correctly, the word was either studied again, tested again, or dropped from both.

Repeated studying after a word was learned had no effect on later recall, but repeated testing had a large positive effect. On top of that, students’ predictions of how well they would do were unrelated to how well they actually did. The authors write that even university students seem unaware of how important retrieval practice is.

There is also a 2014 meta-analysis by Rowland in Psychological Bulletin, which pulled together studies comparing testing with restudying. The results supported the idea that effortful processing contributes to the testing effect. In particular, tests where you recall the answer yourself produced larger benefits than recognition tests where you pick from options.

Limits and common misreadings

A few points are worth keeping in mind.

First, the participants were students and the material was short prose. This experiment alone can’t tell us whether the gap would be just as large for things that aren’t text, like methods for solving math problems or playing an instrument.

Second, this does not mean rereading is useless. At 5 minutes, rereading came out ahead. Reading the material once to understand it is of course necessary. What the experiment shows is how to choose what to do after you read.

Third, the form of recall matters. As Rowland’s meta-analysis shows, recalling from nothing has a bigger effect than choosing from options. There is also a view that flipping a flashcard, seeing the answer and thinking “I knew that” looks like recall but is closer to rereading.

Using it in daily life

Here are three ways to apply this study to daily life.

• Close the book once you’ve read it. After a chapter, write the main points on paper without looking. What you couldn’t recall is what to read next.
• Don’t judge by how it feels. The more you feel “I’ve got this,” the more it is worth closing the book and checking. Confidence after rereading is not reliable.
• Write, don’t pick. Don’t stop at multiple-choice drills. Add practice where you recall from nothing, like writing on a blank page or explaining it to someone.

Brain Mondo is a series of short daily videos about the brain, each based on a research paper. BrainDojo is an app for a few minutes a day of training in calculation, memory, attention and more. Many of the trainings have you answer by handwriting with your finger or by speaking, so you produce the answer yourself instead of picking from options. You also check yourself with a score rather than by how it feels.

References

• Roediger HL, Karpicke JD. Test-enhanced learning: taking memory tests improves long-term retention. Psychol Sci. 2006;17(3):249-255.
• Karpicke JD, Roediger HL. The critical importance of retrieval for learning. Science. 2008;319(5865):966-968.
• Rowland CA. The effect of testing versus restudy on retention: a meta-analytic review of the testing effect. Psychol Bull. 2014;140(6):1432-1463.