In 1984, Ronald Cotton, a Black man working as a dishwasher in North Carolina, was wrongly convicted of rape and burglary largely based on eyewitness testimony from the victim, Jennifer Thompson‑Cannino. Thompson had been attacked in her apartment and later participated in a police photo lineup. After viewing the array, she selected Cotton’s photograph and expressed confidence that he was her assailant. She repeated this identification in a live lineup and later testified at trial that she was certain Cotton was the man who raped her. Based on her identification, a jury convicted Cotton, and he was sentenced to life plus 50 years in prison. It was not until 1995 — after Cotton had spent more than a decade behind bars — that DNA testing definitively proved that the biological evidence did not match him but instead matched another man, Bobby Poole. Only after this scientific breakthrough was Cotton exonerated and released (Scientific American; Wikipedia).The Cotton case became emblematic of a broader problem: faulty eyewitness identification is one of the leading contributors to wrongful convictions. According to research compiled by innocence advocacy groups and psychologists, eyewitness misidentification has played a role in a large majority of criminal cases later overturned by DNA evidence. Scientists now understand that a witness’s confidence — even when expressed firmly in court — does not reliably indicate accuracy (Ars Technica).
Contrary to the common perception that memory functions like a video recording, modern cognitive science demonstrates that human memory is reconstructive, malleable, and highly susceptible to distortion. Memories are formed through processes of encoding, storage, and retrieval, and at each stage they can be influenced by stress, suggestion, time, and post‑event information. One of the most influential bodies of research in this field comes from psychologist Elizabeth Loftus, whose experiments on the misinformation effect showed that introducing misleading information after an event (e.g., suggestive questions or incorrect details) can become integrated into people’s memories, causing them to recall events that never happened or to remember them inaccurately. For example, Loftus’s classic studies on car accidents showed that simply changing a question’s wording (e.g., using “smashed” instead of “hit”) altered participants’ recollections of how fast the vehicles were moving and whether broken glass was present — even when it was not. These findings highlight how susceptible human memory is to distortion.
Other research shows that confirmation and reinforcement can further cement false memories. In Ronald Cotton’s case, once Thompson made an initial tentative choice from a lineup, police feedback (“We thought that might be the one”) likely reinforced her belief that she had identified her attacker correctly, increasing her confidence even as her memory remained flawed. This phenomenon — where confirmatory feedback inflates confidence — has been demonstrated in multiple psychology studies. Scientists also document specific influences on memory accuracy: the cross‑race effect, where people are generally less accurate identifying faces of a different race than their own; the weapon‑focus effect, where attention fixates on a weapon at the expense of the perpetrator’s features; and the basic fact that stress and trauma can disrupt how details are encoded in the first place.
Loftus’s work on the misinformation effect is perhaps the most widely cited research challenging the reliability of eyewitness recall. In these experiments, participants exposed to false post‑event information often incorporated that misinformation into their memories as if it were real, illustrating how easily memory can be contaminated. This has critical implications for criminal justice: every step in an investigation — questioning, lineup procedures, repeated recounting of events — can unintentionally change a witness’s memory. Other research highlights how simple lineup procedures can influence outcomes. Studies show that when police fail to use double‑blind lineup administrators (where the officer presenting the lineup doesn’t know who the suspect is), witnesses may pick the person they subconsciously believe authorities want them to choose. Training and procedural reforms, such as telling witnesses that the culprit may not be present, can significantly reduce false identifications. The growing body of scientific evidence has forced the legal system to wrestle with the limitations of eyewitness testimony. While jurors often find eyewitness accounts highly convincing — sometimes more so than physical evidence — psychological research shows that confidence and clarity do not guarantee accuracy. Thus, courts and legislatures have increasingly considered reforms: allowing expert testimony on memory science, implementing improved lineup protocols, and educating jurors on the factors that can distort recall. Some researchers argue that eyewitness evidence can still be valuable when carefully collected and contextualized, especially when accompanied by corroborating evidence, and that memory science should inform but not entirely displace human testimony.
