Case File 254 · Investigation Psychology

Why Can a Confident Witness Still Be Wrong?

By Murder Book Editorial TeamPublished by Gungrounds, the studio making Murder BookPublished 3 October 2026Updated 3 October 2026

A confident witness can be wrong because confidence and accuracy are produced by the same memory system, and that system is reconstructive and open to influence. A person can hold a vivid, detailed, sincerely believed memory that is partly or wholly mistaken, and feel just as sure about the wrong parts as the right ones. Confidence can also grow after the fact, through feedback, repeated questioning, conversations with others and exposure to new information, without the underlying memory becoming any more accurate. This is not rare or exotic. Mistaken identification by confident witnesses has been the single most common factor in wrongful convictions later overturned by DNA evidence. A confident witness who turns out to be wrong is usually not lying; they are reporting, in good faith, a memory that feels certain but was built on shaky foundations.

Key takeaways

Confidence is generated by the same fallible memory system that produced the memory, so it is not an independent check on accuracy.
A false memory can feel identical to a true one, which is why sincerity and certainty are not proof of accuracy.
Confidence can be inflated after an identification by feedback and repeated questioning, while accuracy stays the same.
Mistaken eyewitness identification has been the leading contributing factor in DNA-overturned wrongful convictions.
A confident but mistaken witness is typically honest, not deceptive.

How can someone be sure and still be wrong?

The 2014 National Research Council report Identifying the Culprit explains the core reason. Memory does not work like a recording. Perceptual experiences are stored by a system that is highly malleable, and memories are forgotten, reconstructed, updated and distorted without the person being aware of it. Because the feeling of confidence comes from that same system, it can be high even when the memory has drifted from what actually happened.

Several ordinary processes can leave a witness sure but mistaken:

Poor viewing conditions at the time. Brief exposure, distance, poor light, stress or a weapon drawing attention can all weaken the original memory, yet the witness may still form a firm impression later.
The misinformation effect. Misleading details encountered after the event can be absorbed into the memory and reported with confidence.
Source-monitoring error. A real memory can be misattributed to the wrong source, so a face seen elsewhere feels like the face of the culprit.
Memory contamination. Conversations, media and suggestive questioning can reshape an account before it is ever recorded.

Does feedback make a witness more confident?

Yes, and this is one of the most direct explanations for the confident-but-wrong pattern. In a 1998 study, Wells and Bradfield had people identify someone from a lineup that did not include the real culprit, so every choice was wrong. Those told "Good, you identified the suspect" afterwards remembered being more confident, having had a better view and having paid closer attention than those given no feedback. The feedback changed how certain they felt about a mistaken choice.

A later meta-analysis of this post-identification feedback effect confirmed that it is robust. The lesson is that a witness's courtroom confidence may reflect what happened after the identification as much as the quality of the original memory. The timing and conditions of a confidence statement are central, as explained in does eyewitness confidence mean the memory is accurate?.

How often does this cause real harm?

The consequences are documented. According to the Innocence Project, mistaken eyewitness identification has been the most common contributing factor in wrongful convictions later overturned by DNA testing, playing a role in a large share of those cases. In a meaningful number of them, more than one witness identified the same innocent person, and a substantial proportion involved cross-racial identification, which research links to higher error rates.

These are cases where confident, sincere witnesses were simply mistaken. The witnesses were not the villains of the story; they were part of a process that treated certainty as if it were proof.

What reduces the risk?

Because confidence can be misleading, the safeguards focus on protecting and properly recording memory rather than on reading the witness:

Record a confidence statement in the witness's own words at the first identification, before any feedback.
Use fair lineups with one suspect and suitable fillers, as in what makes a police lineup fair?.
Run identifications double-blind, so the administrator cannot steer the witness.
Interview witnesses early and separately to limit contamination.

These steps do not make memory perfect. They make it easier to tell how much weight a confident account deserves.

Do stress and other conditions make it worse?

The quality of the original memory sets a ceiling on how accurate a later account can be, no matter how confident the witness becomes. The National Research Council report highlights that viewing conditions, duress and elevated emotion all shape what is perceived and stored in the first place. A witness who saw an event briefly, from a distance, in poor light or under threat may encode only a thin impression, yet still arrive at a firm belief weeks later once the gaps have been filled in by reconstruction.

Cross-racial identification is a further well-documented source of error: people are generally less accurate at recognising faces of a different race than their own, and confidence does not shrink to match. Time matters too. As the interval between event and recall grows, memory fades and becomes more open to the influences described above, a problem covered in how does time affect eyewitness memory?. None of these factors announces itself to the witness, which is precisely why a weak memory can still feel strong.

Is a confident witness who is wrong lying?

Almost never, and this is the point that matters most. A false memory is experienced from the inside exactly like a true one. The witness is not performing confidence to deceive; they genuinely feel it. Treating a confident mistake as a lie misreads the situation and can compound the harm, as the real error lies in the memory and the process around it, not in the witness's honesty. Confidence is a feeling, not evidence, and it is not a sign of guilt or deception on anyone's part.

How does Murder Book treat the confident account?

Murder Book: The Art of Interrogation is an upcoming narrative interrogation game from Gungrounds, an independent studio in Zagreb, Croatia. Its cases are built on unreliable stories, where a person can be completely convinced of an account that does not fit the evidence. The game invites you to weigh statements against corroboration rather than to accept certainty at face value or to assume that a confident contradiction means a lie.

That design mirrors the research: a confident account is something to test, not something to trust automatically, and a mistaken one is not proof of bad faith. Cases are designed for focused 5–10-minute sessions, and Murder Book is in development for PC (Steam) and mobile, with release currently planned for the end of Q2 2027. For the broader picture, see how reliable is eyewitness testimony?.