Improving Accuracy in Memory Recall: The Particular Case of Alibi Generation

Imagine you are standing trial for a crime you didn’t commit. The prosecutor asks, “Where were you at 9pm on the night of the murder?” The night feels like ages ago, and you aren’t entirely sure. Giving your best guess, you guess wrong.

In 1985, Ronald Cotton had to provide an alibi for the night of a crime. However, being confused, he provided the place he was the week before the crime. This memory error was largely to blame for Cotton’s sentence of life plus 50 years in prison. It wasn’t until ten years later that Cotton was exonerated based on DNA evidence.

What made Cotton confused about where he was? Was it just an unlucky day, or is this a mistake that people can commonly make? And would there be a way to fix these errors, or at least help people make fewer mistakes?

Unfortunately, until recently, not many scientific studies have directly examined this important problem. This was mainly because it was hard to establish the ground truth when asking where people were. However, now, almost everybody carries a tracking device that allows researchers to know precisely where they were at any given time: a smartphone.

Relying on this modern convenience, my colleagues and I designed a study where participants used a smartphone app to record their daily whereabouts. The app automatically and continuously recorded (every 10 minutes) the participant’s location and also their environment (such as the sound and their movements). This recording happened for 4 weeks.

A week after the recording period, participants were tested on the accuracy of their memory. We gave the participants a Google Map with 4 location pins, and asked which pin showed where they were at a certain day and time. One pin was the actual location, while the other pins were randomly selected from places they visited during the 4-week period. Each participant received 72 test questions, which asked a location for a different day and time.

Overall, people were accurate 64% of the time, which was above the accuracy level if they randomly guessed (25%, four options). However, the interesting part was in the errors that the participants made, as they showed certain patterns.

How Did People Mis-Remember?

We saw three common mistakes. When the participants picked a wrong location pin, it tended to be geographically closer to the correct pin. They also wrongly picked location pins that had a similar environment (such as sound, movement) to the correct pin. Lastly, the wrongly picked locations were visited at a similar time to the correct location.

This last type of memory error regarding time is especially interesting, as it was shown in two ways. First, participants wrongly picked locations that they visited right before or after visiting the correct location. For example, if the correct location was the grocery store, when people made an error, it tended to be the location before or after visiting the grocery store (maybe a  coffee shop).

Second, participants wrongly picked locations that they visited on the same day of week, or at the same time of day as the correct locations. For example, when making an error, some participants wrongly chose a location that they visited on the same day of the week as the correct location (e.g., Monday), even though it was a week before. Others chose a location that was visited at the same time as the correct location (e.g., 5pm), even though it was not on the same day. These examples illustrate a type of error regarding the so-called ‘categorical time’ —and this error is exactly the one Cotton made. Our study supports that this is a common error everyone makes!  These results together show that people are most confused when trying to remember events that happened in a similar location, time, and/or environment.

Is There A Way To Avoid These Memory Errors?

Avoiding these memory errors is essential when the outcome has a critical consequence. Is there a way to decrease, if not perfectly avoid, cases like Ronald Cotton’s?

The results of our study provide some hints.

First, investigators such as police interrogators should be fully aware of these natural memory errors that people make. Second, it will also help if questioners check for possible memory errors by asking specific questions. For example, if the person said he was at the bar last Monday at 5pm, the investigator can ask whether it was not the week before last, or whether it was not Tuesday or Friday at 5pm, to find out whether he has a pattern of visiting that bar that could confuse his memory. We believe that these ‘right questions’ can help people with more accurate recall, and—in the case of crimes—produce more accurate alibis.

Faulty memory isn’t always a huge deal, but as the case of Ronald Cotton shows, it can be the difference between freedom and imprisonment.

For Further Reading

Dennis, S. J., Garrett, P., Yim, H., Hamm, J., Osth, A. F., Sreekumar, V., & Stone, B. (2019). Privacy versus open science. Behavior Research Methods, 51, 1839–1848. https://doi.org/10.3758/s13428-019-01259-5

Laliberte, E., Yim, H., Stone, B., & Dennis, S. J. (2021). The fallacy of an airtight alibi: Understanding human memory for “where” using experience sampling. Psychological Science, 32(6), 944–951. https://doi.org/10.1177/0956797620980752
 

Hyungwook Yim is an Assistant Professor in the Department of Cognitive Sciences at Hanyang University, Seoul, Korea. His research focuses on human learning and memory, and their development using computational models. 

Simon Dennis is a Professor in the School of Psychological Sciences, and the Director of the Complex Human Data Hub at the University of Melbourne, Australia. He is also the CEO of Unforgettable Research Services Pty Ltd. His research utilizes large scale real world data, experimental paradigms and computational modelling techniques to investigate the cognitive architecture underlying memory and language.