Misinformation, False Memory, and Breonna Taylor

Why a changed answer on a critical question matters in this police shooting case

Posted Oct 15, 2020

The witnesses all agree that the police failed to announce themselves before breaking down Breonna Taylor’s door and killing her. Except for one. One witness now reports the police did announce themselves. Originally, however, he said they didn’t. Which version of his memory is likely to be correct? In this case, his recent memory may reflect a misinformation effect.

Eyewitness memory is often critical. And for the Breonna Taylor case, one memory detail is particularly important: Did the police announce themselves before breaking down her door?

Let’s start with a short version of how Breonna Taylor was killed by police in her own apartment. You can find a more extensive version at The New York Times and many other outlets. The police had a warrant to search Ms. Taylor’s apartment. The warrant was connected to the behavior of her ex-boyfriend. The police were required to announce themselves. They broke down the door and entered. Ms. Taylor’s current boyfriend thought someone was breaking into the apartment and he fired a warning shot. This shot reportedly hit one officer. The police then fired multiple shots, resulting in the police officers shooting and killing Breonna Taylor.

A critical question is whether the police announced themselves before breaking down the door. The boyfriend said neither he nor Ms. Taylor heard the police do so, and that’s why they thought someone was attacking them. Witnesses, nearly a dozen nearby neighbors, said they did not hear the police announce themselves.

But now let’s consider the critical witness: Aarin Sarpee. According to an article in Vice, Mr. Sarpee was picking up his daughter from his brother’s apartment above Ms. Taylor’s. When he was interviewed by the police on March 21st, he reported that the police did not announce themselves. This was one week after the killing of Breonna Taylor.

Originally, Mr. Sarpee’s memory was consistent with all the other witnesses. The police didn’t announce themselves. He had actually interacted with the police since he was outside the apartment where he was picking up his daughter. He knew they were police. But when asked, he stated directly that the police did not identify or announce themselves before breaking into Ms. Taylor’s apartment (see a review of the police interview in the Louisville Courier Journal).

If we stopped at this point, things would be clear. Although the police claimed to have announced themselves, they did not use body cameras, so there is no external evidence. But all the witnesses agreed that they did not. Even Mr. Sarpee stated no announcement in his first interview. But this isn’t the end of the story.

Mr. Sarpee was interviewed several more times. How many isn’t actually clear in the records I’ve seen so far. The content of some interviews may not have been released yet. By his interview on May 15th, two months later, Mr. Sarpee’s story had changed. At this point, he stated that the police did announce themselves. But he also noted that it had been a long time, indicating a lack of confidence.

Now we have a changed claim, by a single witness, that is inconsistent with other witnesses. This was the evidence that was used by the Kentucky Attorney General to determine that the police had followed appropriate procedures.

How should we understand a changed memory? Memories can change for a variety of reasons. But a critical factor is the role of the misinformation effect. The misinformation effect is shorthand for the effect of misleading post-event information on someone’s memory. First someone witnesses an event. Afterwards, they may be exposed to a variety of other information. Some may be consistent with both the original event and their own memory. But other information may be inconsistent; it may actually be erroneous information. When exposed to misleading post-event information, people sometimes will adopt the false information. They will then include the misinformation in their memories when they tell the story later. And this is the misinformation effect. Research by Elizabeth Loftus and many others shows this quite clearly. People see an event, are exposed to misleading information, and then include the misleading information in their memories.

For these reasons, we should almost always trust the original memories that people provide. The original reports are more likely to be correct. These memories are given sooner after an event, before as much forgetting has occurred. These memories are also given before people are exposed to additional, and potentially misleading, information. Original memories are generally more reliable.

When someone has been repeatedly questioned, especially if they are asked repeatedly about an issue, they may adopt information into their memories. In this case, Mr. Sarpee was interviewed on more than one occasion and asked about the police announcing themselves multiple times. We shouldn’t be surprised if he later adopts that suggested information.

Of course, it may be that his later memory is the accurate one. But there is one other point that makes this less likely. Mr. Sarpee seemed to express a lack of confidence in the memory, noting that it had been a long time. People are able to express their confidence in various details of their memory for an event. Importantly, if you give witnesses the option to not answer a question, they will often provide more accurate memories—giving the information they originally saw and withholding the misinformation they received later. People can stop themselves from including misinformation. Ayanna Thomas and her colleagues have argued that giving people the option to not answer questions encourages them to more carefully evaluate their memories (Bulevich and Thomas, 2012; Thomas et al., 2020).

In this case, the only witness who stated that the police announced themselves changed the story after multiple interviews, he expressed a lack of confidence, and his memory is inconsistent with other witnesses. While that does not prove that a false memory is involved, it seems distinctly possible. I would recommend care in relying on a changed recollection when someone has been interviewed multiple times.


Bulevich, J. B., & Thomas, A. K. (2012). Retrieval effort improves memory and metamemory in the face of misinformation. Journal of Memory and Language, 67(1), 45-58.

Thomas, A. K., Smith, A. M., & Mazerolle, M. (2020). The unexpected relationship between retrieval demands and memory performance when older adults are faced with age-related stereotypes. The Journals of Gerontology: Series B, 75(2), 241-250.