Rehearsal aids memory. This is a basic psychological law. We remind students to study and then study some more because rehearsal aids memory. But sometimes psychological laws get broken. Sometimes our data go flying past that psychological speed limit sign. Sometimes rehearsal, the very thing that we know aids memory, leads people to make more memory errors.
I always enjoy counter-intuitive findings. Last year, Jason Chan, Ayanna Thomas, and John Bulevich published a counter-intuitive finding concerning rehearsal. When you read the paper, they actually sound a bit surprised by their finding (and yes they replicated this counter-intuitive finding 3 times, so I'm pretty sure the effect is real). In their study, rehearsal led to worse memory performance.
Here's how their experiment worked. First they conducted a classic misinformation experiment; the type used to explore eyewitness memory errors. They started by showing people a complex video. After the video, they provided people with post-event information - some accurate, some misleading. Finally, they gave the participants a memory test. As is typically the case, Chan, Thomas, and Bulevich found that giving people misleading post-event information caused memory errors - a phenomenon called the misinformation effect.
What Chan, Thomas, and Bulevich added to the typical misinformation experiment was an attempt to stop the misinformation effect from occurring. After watching the original video, they gave some people a memory test - an opportunity to rehearse their memories. People did pretty well on this first test. Now this is straightforward: The first test gave people a chance to rehearse their memories before they were provided with any misleading information about the event. In that first test, the participants demonstrated that they knew the answers. The extra rehearsal should have improved their memory. Studying should help.
But for some of those rehearsed items, Chan, Thomas, and Bulevich then gave the research participants misleading information. They clearly expected that since they had just rehearsed the correct answers, the people should be less suggestible, less likely to accept the misleading information. That was my expectation too, when Ayanna Thomas described the experiment to me.
We were all wrong. Rehearsing made people more susceptible. After rehearsing the correct answers, people were more likely to adopt the misleading information and make errors on subsequent memory tests. The size of the effect was substantial as well: People made about 20% more errors if they rehearsed their memories than if they didn't. Rehearsal did not help memory. Instead, rehearsal caused people to make more errors!
So what's going on? Chan, Thomas, and Bulevich suggested that rehearsing the memories made the memory more vulnerable through reconsolidation. Consolidation is a well known memory phenomenon. After you experience an event, your memory goes through an original consolidation period - a time when the memory continues to be processed even if you aren't actively thinking about the event. This is why a bump on the head disrupts memory for things that happened before the bump on the head. The process of consolidation is disrupted by that head injury.
Chan, Thomas, and Bulevich argued that each time you bring a memory to mind you create another consolidation period - a time when the memory is reconsolidated, reconstructed, and vulnerable. If you learn additional information during that time, the reconsolidation may lead that new information to be included in the memory. In other words, we constantly reconstruct our memories every time we bring them to mind. Thus rehearsal is great if you rehearse accurate information. But rehearsal can also decrease memory accuracy even if you simply consider erroneous information during the reconsolidation period.
Chan, Thomas, and Bulevich concluded that psychologists may have underestimated the number of errors that real eyewitness make. Eyewitnesses are frequently asked to remember the crime and are potentially exposed to misleading information during each round of questioning. Thus eyewitnesses experience multiple periods of reconsolidation and may have multiple opportunities to construct erroneous memories.