Eyewitness Memory: How Reliable Is It?
Eyewitness testimony is often unreliable, with over 70% of exonerations linked to mistaken identifications. Recent research highlights memory malleability, the influence of police procedures, and evidence‑based guidelines to improve accuracy. Understanding these factors is crucial for fair justice.
By Felo News Desk · Published
Eyewitness testimony has long been a cornerstone of criminal investigations, yet a growing body of research shows that human memory is far from infallible. Between 1989 and 2019, more than seventy percent of convictions overturned by DNA evidence were rooted in mistaken eyewitness identification, underscoring the stakes of relying on memory alone.
Memory’s Malleable Nature
When we retrieve a memory, it undergoes a process called reconsolidation. Each time a memory is recalled, it can be altered by new information, a phenomenon first demonstrated by Elizabeth Loftus in the early 1970s. By simply rephrasing a question—using words like “smashed” versus “collided”—Loftus showed that witnesses could be led to remember details that never occurred, such as broken glass in a car crash that was never present.
Flashbulb memories, the vivid recollections of major events, also reveal the fallibility of memory. Studies of September 11, 2001, show that while people feel they remember exact details, they actually get about forty percent of the facts wrong. Factors that influence accuracy include distance from the event, lighting conditions, and even the racial match between witness and suspect.
When Witnesses Get It Wrong
One of the most damaging ways memory can be distorted is through police lineups. John Wixted and colleagues at UC San Diego found that a poorly conducted lineup—such as a “showup” where a suspect is presented alone—can implant the suspect’s image into a witness’s mind. Even if a witness was uncertain immediately after the crime, confidence can rise after a second lineup, giving a false sense of certainty in court.
Confidence, however, is a poor standalone indicator of accuracy. Research by Laura Mickes at the University of Bristol and Wixted shows that when witnesses are given a fair photo lineup and asked to rate their confidence, higher confidence does correlate with correct identifications. This nuance is critical for police procedures and courtroom testimony.
Improving Lineup Practices
In 2020, experts agreed on nine evidence‑based procedures for conducting lineups. These include: selecting lineup members who match the suspect’s basic description, limiting witnesses to a single viewing, asking for confidence ratings immediately after the identification, and using a double‑blind administration where officers do not know who the suspect is. Despite federal guidelines in the U.S. recommending confidence data collection, the U.K. lacks similar mandates, leading to inconsistent practices across jurisdictions.
The case of Ervin Harris illustrates the consequences of deviating from these guidelines. Convicted in 1974 and imprisoned for forty‑two years, Harris was exonerated in 2026 after DNA proved his innocence. The police lineup that led to his conviction included a suspect with facial hair—contrary to the victim’s description—and a live lineup that reinforced the witness’s initial uncertainty. This case highlights that procedural errors, not memory itself, often underlie wrongful convictions.
What Researchers Are Doing Next
Researchers are exploring additional cues that may signal reliable testimony. Andrew Smith at Iowa State University proposes combining behavioral indicators—such as the speed of identification, use of elimination strategies, and internal debate—to assess accuracy. He also notes that when witnesses fail to identify anyone, their confidence levels can provide valuable insight into the integrity of the lineup.
Similarly, Lilian Kloft‑Heller at Maastricht University is investigating how intoxication affects eyewitness accuracy. Preliminary findings suggest that while intoxicated witnesses report fewer details, the information they do provide can still be reliable if properly questioned.
As science advances, memory experts like Travis Seale‑Carlisle from the University of Aberdeen emphasize that the blame should shift from witnesses to flawed procedures. By integrating evidence‑based guidelines into training and policy, law enforcement can reduce the risk of wrongful convictions and uphold the integrity of the justice system.
In sum, eyewitness memory is a powerful but fragile tool. Understanding its limitations, refining lineup protocols, and incorporating confidence assessments are essential steps toward a fairer legal process.
Key facts
- Eyewitness memory can be altered with each recall.
- Police lineups can implant false memories if poorly conducted.
- Confidence is a better accuracy indicator only when lineups are fair.
- Nine evidence‑based lineup procedures exist but are inconsistently applied.
- Researchers are developing behavioral cues to gauge testimony reliability.
- Procedural errors, not memory flaws, often cause wrongful convictions.
Why it matters
Wrongful convictions cost lives and erode public trust in justice. Accurate eyewitness procedures protect the innocent and strengthen the legal system.
Frequently asked questions
What is reconsolidation?
Reconsolidation is the process by which a recalled memory becomes labile and can be altered by new information before it is stored again.
Why is confidence a poor indicator of accuracy?
Confidence can rise after a second lineup or due to suggestive questioning, making it unreliable as a sole measure of correctness.
What are the nine evidence‑based lineup procedures?
They include matching suspect description, single viewing, confidence rating, double‑blind administration, and other guidelines designed to reduce bias.
Sources
- [1] scientificamerican.com — originally reported as “What science reveals about the reliability of eyewitness memory”
