Rewards Help Memory During Learning (But Not Later)

Source: iStock image by fotoSipSak licensed to Art Markman

Your ability to remember something involves three phases. There is a learning phase (which psychologists call encoding) in which you encounter some piece of information. There is a retention interval in which processes like sleep may improve memory and other factors may lead you to forget some or all of what you encountered. Finally, there is a retrieval phase in which some piece of information (a cue) reaches into memory and leads you to recall some of what you saw initially.

There are many things that can influence the effectiveness of all three of these phases. For example, paying more attention to information during encoding can make it easier to remember what you learned better than paying less attention. When retrieving information, you might say that you saw something previously if you have a feeling that it is familiar, or you might require having a vivid memory of what you saw to say that you actually saw it.

Do potential rewards influence how much you remember about situations? If so, do you have to find out about those rewards when you first encounter the information, or can you find out about the rewards at the time that you are going to retrieve the information?

This question was explored in a paper from 2022 in the journal Cognition by Kevin da Silva Castanhiera, Azara Lalla, Katrina Ocampo, Signy Sheldon, and my former graduate student Ross Otto.

They did two parallel studies in which people viewed sixteen brief videos of naturalistic scenes like people cleaning up or talking on the phone. They viewed the scenes twice. Then, after a brief 10-minute delay in which participants answered unrelated questions, they were given information about the scenes and asked whether that information reflected something that was true of the scenes. They might be asked about the gist of the event (something about what the people did), something perceptual about the scene (like the color of what someone was wearing), or something about the context (like another individual or object in the scene that was not central to the action).

In one study, half the scenes were identified as behind high-reward scenes—meaning that participants would get 25 cents for each question they answered correctly about the scene, and half were identified as low-reward scenes—meaning that participants were informed they would get 1 cent for each question they answered correctly. To make it easy to separate the low- and high-reward scenes, they were each presented in a group, so that participants saw all of the low-reward scenes together and all of the high-reward scenes together. Participants were not reminded about any of the rewards when they were asked the retrieval questions.

In this study, rewards had a significant impact on performance. Participants were significantly more accurate at recalling information about high-reward scenes than about low-reward scenes. Some types of information were harder to retrieve than others, so participants were more accurate about the gist of the event than about specific perceptual details. But the advantage for high-reward scenes was true for all types of information. The researchers also analyzed the speed with which participants made their judgments about the information they had seen and found no systematic effect of rewards, suggesting that the rewards did not influence the strategies people used to retrieve information.

In the second study, participants studied the scenes without getting any information about rewards. Then, participants engaged in a block of high-reward and a block of low-reward retrieval trials. Unlike the previous study, rewards that people found out about at the time they were retrieving information had no impact on their performance. Again, though, rewards had no impact on the strategies people used to retrieve information.

These findings suggest that when participants find out about rewards at the time they are learning the information, they put in more effort and pay more attention to items that will get high rewards than to information that will get low rewards. But there isn’t much that people can do at the time of retrieval to improve their ability to remember.

This pattern of findings makes a lot of sense. It is valuable to be able to direct effort and attention to information that is most likely to be useful later, though some amount of other information that isn’t known to be critical or high-value also gets stored. People can take their time learning or studying information. Retrieving information from memory is fairly automatic. It is useful to be able to retrieve whatever you know about the past quickly, because the information you recall affects how you interpret new situations. The more strategic you can be at the time you retrieve information, the harder it might be to get at the information you might need immediately to decide what to do.