The Learning Scientists

View Original

When Restudying Trumps Retrieval

by Cindy Nebel

(Cover image from Pixabay by squarefrog.)

We have so many blogs about retrieval practice. In fact, it is the most common tag we use on our blog. If you’re new to this conversation, you can find some summary information about the benefits of retrieval practice here and some additional resources here. In short, there is a wealth of research demonstrating that retrieving information helps individuals retain information better than restudying.

But not always…

One thing that we try to emphasize is that there are no silver bullets. It’s so important to understand the mechanisms that underlie any strategy we’re recommending because when you go to apply it the why it works becomes super important. No strategy works all the time. Not even retrieval practice. (You can see other times we’ve talked about that fact here and here.)

Today I want to share a new Open Access article (so you can go read it for yourself if you’d like!) that *spoiler alert! found a situation in which restudying was actually better for retention than retrieval (1).

This study used Swahili-English word pairs for their materials. This is a pretty common strategy when we’re trying to figure out basic mechanisms. Most of our participants don’t have a background in Swahili so it allows us to see how they learn something from scratch instead of worrying about people who might have already known something when the experiment began. Still, there are some limitations here because most of us do not ask our students to memorize foreign language vocab pairs exclusively or at all.

In this study, there were a few experiments. I’m just going to tell you about the first one here. In the first experiment, people were recruited online and participated in five sessions. During the first session, they studied the 40 word pairs. Then on Days 3, 5, and 7 they were given the words in different ways. Some of the word pairs were restudied, but for others the Swahili word was presented and students either just thought about or wrote down the English pair before being shown the correct answer. They labeled these Restudy, Covert Retrieval, and Overt Retrieval. After each restudied word, participants were asked, “Would you have remembered the correct translation?” and rated it out of 100. For the Retrieval groups they were asked how close they were to the correct answer (again, out of 100).

On Day 9, everyone took a final test. And the strangest thing happened…

Image created from data reported in cited source

The Restudy group outperformed both Retrieval groups by about 10%.

What happened here? Well, to be honest, we don’t know. One sticky part is that this is within-subjects. That means that all participants answered questions in all of these conditions. It’s possible that they could have started retrieving or that overt retrieval helped them realize they needed to be more effortful in the other conditions.

It’s also possible that the spaced restudy was really beneficial and that using really excellent spacing makes up for the benefit of retrieval over time. But that doesn’t explain why Restudy is actually higher and significantly so.

Without going into too much detail, the researchers tried the experiment again, this time removing the question that they asked the restudy group (to predict how well they would have remembered this). And then they got the standard retrieval practice effect.

Image created from data reported in cited source

The authors admit that more research is needed here. Clearly, the effect has to do with the metacognition happening when participants are asked to predict how much they will remember, on each item, with an opportunity to restudy. There are other studies that ask participants how much they will remember later on (2), but these are global estimates and not quite the same thing as what we’re seeing here, where the estimates are happening during study. One suggestion made by the researchers is that this could be due to something called the “changed-goal hypothesis”. This means that the participants might have received those ratings and changed their strategy such that they were hyper-focused on retaining the restudied items and put less effort into the retrieved items (possibly because they were metacognitively aware that they were getting them wrong!).

Image from Pixabay

What does this mean for you? Well, be on the lookout for some follow up here. We don’t recommend changing your strategies based on one piece of evidence, but there is enough here to remind and encourage you to think about how you’re implementing retrieval and to make sure that you’re seeing positive benefits instead of just assuming it should work. The authors also suggest that situations in which retrieval (a quiz) is followed immediately by self-judgment of accuracy (self-grading) might be an area for caution as that situation mimics what is seen here. I would argue that it’s apples and oranges though and the jury is out on whether and when this is problematic.


References:

(1) Higham, P. A., Fastrich, G. M., Potts, R., Murayama, K., Pickering, J. S., & Hadwin, J. A. (2023). Spaced Retrieval Practice: Can Restudying Trump Retrieval?. Educational Psychology Review, 35(4), 1-41. Available at https://link.springer.com/article/10.1007/s10648-023-09809-2

(2) Roediger III, H. L., & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. Psychological science, 17(3), 249-255.