Exploring how brains learn, remember, and decide.

When Memory Becomes Evidence: Reconstructive Memory and Eyewitness Testimony

Since about 1989, hundreds of people convicted in California have later been exonerated, according to the National Registry of Exonerations. In many of these wrongful convictions, eyewitness testimony played a role.

My question for you is: How can someone remember something so clearly and still be wrong?

When a crime is witnessed, the process of remembering what happened begins with sensation. Your eyes take in visual details such as clothing, facial features, and movements. However, your brain doesn't store everything. Only some of the information we perceive gets processed and stored in long-term memory. For example, if you see a weapon during a robbery, your attention may be drawn toward the weapon instead of the perpetrator’s face.

Even though it would be useful, our memory doesn’t work like a perfect video recording. Eyewitness information is encoded, stored, and retrieved. Your memories can become less stable over long intervals of time. This instability makes your memory more susceptible to new information and experiences over time.

You may ask yourself, what kinds of things can actually interfere with memory? Memory is reconstructive because when we remember something, our brain often rebuilds that memory using our perception of both the original situation and additional information we receive later.

Stress and Fear

The conditions of a crime can affect identification. When a witness is under high levels of stress or fear, their ability to focus on facial characteristics can be impaired, making it more difficult for them to accurately identify the person they saw.

Weapon Focus

Another factor that could impact eyewitness identification would be weapon focus. When a weapon is present during a crime, it is likely to capture a witness’s attention, drawing attention away from the perpetrator’s face. Research also suggests that the presence of a weapon can reduce perpetrator identification accuracy, although the strength of this effect can vary depending on the circumstances surrounding the crime (National Research Council 93–94).

Time (Retention Interval)

The term “retention interval” refers to the amount of time that passes between witnessing an event and trying to recall what happened later. Research indicates that longer retention intervals are usually associated with lower identification performance, although again variations exist between different situations (National Research Council 99). Longer periods of time between the event and identification allow more opportunity for other information or events to interfere. During this time, a witness may encounter news reports, discuss the event with others, or receive additional information that can become mixed with their original memory.

Misinformation and Memory

Memory may also be influenced by information a witness hears after a crime, including reports from the media and discussions about the event with people. Even being asked “suggestive” questions could alter one’s memory to include details that are not part of the original experience.

Over time, this information may become mixed with the original memory. A witness could believe they remember seeing some detail that was actually introduced to them afterwards. This scenario covers only one reason that eyewitness memories may change without the person realizing their memory was influenced. The dynamic nature of memory can become especially important in criminal investigations because a witness may be completely sincere and confident in reporting what they remember without realizing their memory has been affected by information they encountered after the crime.

Confidence vs. Accuracy

People often assume that when a witness is confident telling a story, their account must be completely accurate. However, this is often not the case.

The case of Ronald Cotton illustrates how eyewitness memory can change over time. In 1984 Jennifer Thompson identified Cotton in a photo array as the man who attacked her. However, she was not initially certain. After police gave confirming feedback and Cotton appeared in phase two of the identification procedure, her confidence increased.

Cotton was eventually convicted and spent more than 10 years in prison. In 1995, DNA testing demonstrated that Cotton was not the perpetrator and identified another man as responsible for the crime. This case reveals an important point about eyewitness testimony: a person can become increasingly confident in their memory even when that memory is incorrect.

What can be done

That being said, if eyewitness memory can be influenced so easily, then what can be done to reduce the chance of these errors?

One way is to avoid providing witnesses with confirming feedback. This can reduce the chances of a witness developing greater confidence in an identification because of information they received after making it. Another important measure is giving standardized instructions and explaining that “the perpetrator may or may not be in this lineup.” Adding this procedure would help prevent witnesses from assuming they are expected to choose someone. This leads me to what is called double-blind identification. Research has not reached consensus that there exists one guaranteed identification procedure that is always ultimately best (National Research Council 2014; Wells, Steblay, and Dysart 2015). However, double-blind identification, which means the person administering the lineup does not know who the suspect is, can help prevent the administrator from unintentionally giving the witness clues about who to choose. In practice, this can be done by ensuring the officer or administrator conducting the identification procedure does not know which person in the lineup is the suspect (Wells et al. 2019).

A common concern is that an administrator could unconsciously give clues through their behavior. Even something as small as a head nod, facial expression, or change in tone could potentially influence a witness’ response and confidence. Thus, the last strategy I would like to discuss is documentation and recording. Investigators should document how confident the witness is at the time of the initial identification rather than later in the process. Later interactions and information can influence how a witness remembers the event and how confident they feel about their identification. Additionally, some research recommends videotaping the identification process to ensure there is always a record of what happened (National Research Council 2014, 109).

Conclusion

So, how could someone remember something quite clearly and still be incorrect? My answer for you is that memory is not a perfect recording. Eyewitness memories can be affected by fear, stress, weapons, and information received after an event. The memory of a witness who is completely honest and confident can be changed over time. Understanding how memory works is important because eyewitness testimony can have serious consequences in the criminal justice system. By using procedures that reduce outside influence, such as avoiding confirming feedback, using standardized instructions, and documenting the identification process, investigators can contribute to reducing the chance that memory errors influence trial outcomes.

Memory may not be perfect, but understanding how it can change gives us a better chance of making sure that what someone remembers does not become evidence of something that never happened.