A witness pointing to a defendant and saying “that’s the person I saw” has long carried enormous weight with juries. It feels like direct, firsthand proof, the closest thing to a video recording a courtroom could get without an actual camera.
DNA testing has told a different story
What DNA Evidence Has Revealed
Since the late 1980s, DNA testing has been used to exonerate hundreds of people who were wrongfully convicted. Many of those came after spending decades in prison for crimes they did not commit. According to data compiled by the Innocence Project, eyewitness misidentification played a role in a substantial majority of these overturned convictions, making it one of the single largest known causes of wrongful conviction in the country.
That statistic points to something specific:
This isn’t just about witnesses occasionally being wrong. Identifying someone involves several separate mental steps, and each one can fail in ways a witness never notices happening.
How Identification Breaks Down, Step by Step
Psychologists describe eyewitness memory as happening in three stages, and each stage has its own vulnerabilities that can lead to a false memory being created.
1 | Taking in information in the moment
2 | Storing that information
3 | Retrieving it later when asked
Stage 1 | Taking In Information Under Stress
When something frightening happens, especially something involving a weapon or an object being used as one, like a bottle or tool grabbed off a table, the brain doesn’t record the scene equally. Attention narrows sharply onto whatever feels most threatening. Researchers call this weapon focus.
In practical terms, a witness’s attention gets pulled toward the threat itself, and away from other details at the scene, including the face of the person holding it. This isn’t a failure to pay attention. It’s an automatic stress response the witness has no conscious control over. The result: the one detail an identification depends on most, the face, is often the detail that received the least attention to begin with.
Stage 2 | Storing That Information
Psychologists use the word encoding to describe how something a person sees gets converted into a memory that can actually be stored in the brain.
Think of it like this: Imagine taking a photo in a dark room with shaky hands. The camera still captures something, but the image comes out blurry and parts of the frame are cut off or too dark to make out. Encoding under stress works similarly. The brain isn’t failing to try, but it’s working under bad conditions. What gets stored afterward is often a partial, gap-filled version of the event.
A witness may walk away having stored very little usable detail about a face, even though they experienced the event as vivid and unforgettable.
Stage 3 | Retrieving the Memory Later
This is where most legal problems arise because memory retrieval doesn’t happen in a vacuum. By the time a witness is shown a lineup, days, weeks, or years may have passed. In that time, they may have seen a news report, talked to a family member, or been shown a photo by an investigator.
Each of those events can quietly blend into the original memory. Psychologists call this source confusion: a person accurately remembers a detail, but incorrectly recalls where that detail came from, whether they actually saw it during the crime, or picked it up somewhere afterward.
Why Questioning Procedure Matters as Much as the Memory
A witness’s memory isn’t the only variable. How police administer a lineup or photo array can introduce errors that have nothing to do with the witness at all.
If the officer running a lineup already knows which photo shows the suspect, that knowledge can unintentionally leak through, a pause, a change in tone, a follow-up question after a specific photo. This isn’t about dishonesty. It happens below the level of conscious awareness for both the officer and the witness.
Two reforms have been shown in research to reduce this risk:
1 | Double-Blind Administration
The officer conducting the lineup does not know who the suspect is, so there’s no cue to give, even by accident.
2 | Sequential Lineups
Witnesses view one photo at a time and decide on each before seeing the next, rather than viewing all photos side by side.
The second reform matters because side-by-side viewing encourages a different (and weaker) question. Instead of asking “does this person match my memory?”, witnesses tend to ask “who here looks the most like my memory, compared to everyone else in front of me?” That subtle shift, comparing photos to each other instead of comparing a photo to memory, has been linked to higher rates of misidentification.
Confidence Is Not a Measure of Accuracy
One of the most counterintuitive findings in this research is how confident a witness sounds has little bearing on how accurate their identification actually is. A witness who says “I am absolutely certain” can be wrong just as often as one who expresses hesitation.
Confidence can also grow over time, particularly after a witness has committed to an identification, discussed it with others, or seen media coverage confirming their choice, even when nothing about the accuracy of that identification has changed. In today’s environment, that also includes true crime content, online discussion, and social media commentary about a case, all of which can reinforce a witness’s confidence without doing anything to make the underlying memory more accurate. As we discussed in our previous article on recovered memory, this is part of a broader pattern: repeated exposure to information after the fact can shape what someone believes they remember, regardless of how that information was encountered.
Jurors, however, have historically treated confident testimony as far more credible. That mismatch, between how sure someone sounds and how accurate they actually are, has been identified as one of the most persistent problems in how this evidence gets used at trial.
DNA testing is what changed that, and it’s worth remembering why that matters. Most of what researchers now understand about eyewitness error didn’t come from a lab alone. It came from studying real cases where a DNA test proved that a confident, sincere witness had identified the wrong person, sometimes decades after that identification sent someone to prison. Each exoneration became a data point that gave the justice system a way to check its own work. What it found was a pattern serious enough to change how courts across the country handle this evidence today.
How Courts Have Responded
Many states have changed how eyewitness identification is conducted and presented at trial, including:
1 | Double-Blind Administration
2 | Standardized instructions given to witnesses before viewing a lineup of suspects
3 | Jury instructions explaining the known limitations of eyewitness memory.
Courts have also increasingly allowed expert testimony explaining this research to juries. A generation ago, this kind of testimony was often excluded as unnecessary or speculative. Many courts now treat it as a legitimate and often necessary tool for helping jurors evaluate identification evidence accurately.
Recognizing this isn’t an attack on victims or witnesses. It’s a recognition that fair outcomes depend on evaluating evidence based on how it actually works, not on how convincing it sounds in the moment.
This article is provided for general educational purposes and does not constitute legal advice. It does not reference, and is not intended to reference, any specific pending litigation. If you have questions about a specific legal matter, please contact our office directly.
Bannister Wyatt & Stalvey LLC.
24 Cleveland St. Suite 100, Greenville, SC 29601
(864) 298-0084
