Case File 255 · Investigation Psychology

What Is Memory Contamination?

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

Memory contamination is what happens when a person's memory of an event is altered by information that reaches them after the event, so that their later account mixes what they actually experienced with what they learned afterwards. The outside information can come from leading questions, conversations with other witnesses, news coverage, social media, or feedback from an investigator. Once it is absorbed, the witness usually cannot tell which parts of the memory came from the event and which came from later, and they may report the whole thing with genuine confidence. Contamination is not deliberate invention and it is not a sign of a dishonest witness. It is a predictable feature of how reconstructive memory works, and it is the main reason investigators try to capture accounts early, separately and with open questions.

Key takeaways

Memory contamination is the alteration of a memory by information encountered after the event.
Common sources include suggestive questions, co-witness discussion, media coverage and post-identification feedback.
Contaminated memories often feel completely genuine, so confidence does not rule contamination out.
Co-witness discussion is a particularly strong source: in one study, most witnesses who discussed an event later reported details they only heard from someone else.
The response is better procedure, not treating the witness as a liar.

How does contamination get into a memory?

The underlying vulnerability is described in the 2014 National Research Council report Identifying the Culprit: memory is malleable and continuously evolving, and it can be compromised at every stage, from encoding to storage to retrieval, without the person noticing. That openness is what lets outside information in. Several routes are well documented:

Leading or suggestive questions. A question that contains a detail can plant it. This is the classic misinformation effect, first shown by Loftus and colleagues in the 1970s.
Co-witness discussion. When witnesses talk to each other, one person's account can become part of another's memory.
Post-identification feedback. Being told an identification was correct can reshape how a witness remembers the event, as Wells and Bradfield demonstrated in 1998.
Media and rumour. News reports, images and online discussion can supply details that a witness later recalls as their own.

How strong is the co-witness effect?

Contamination through conversation is one of the most powerful forms, and it has a clear experimental demonstration. In a 2003 study, Fiona Gabbert, Amina Memon and Kevin Allan had pairs of people watch different versions of the same event, each version containing details the other person had not seen. Pairs who were allowed to discuss the event before being tested individually often went on to report details they could only have picked up from their partner. A large share of those who discussed the event mistakenly recalled items they had never actually seen.

This became known as memory conformity or the social contagion of memory, and it is explored further in what is memory conformity? and can witnesses change each other's memories?. It is the main reason that witnesses should be kept apart, as discussed in why should witnesses be interviewed separately?.

Why can't the witness just tell what was contaminated?

Because contamination usually does not feel like anything. Once an external detail has been woven into a memory, the witness experiences it as part of the original event. They are not aware of a join between what they saw and what they learned later. This links to source-monitoring error, the difficulty of remembering where a piece of information came from. A detail might be perfectly real as a memory, while its source, and therefore its relevance, is wrong.

This is why a sincere, confident, detailed account offers no guarantee that it is uncontaminated. The confidence is real; the provenance of the details is the open question.

How do investigators limit contamination?

The safeguards are about protecting a memory before it can be reshaped and recording its condition at each stage:

Interview witnesses as early as possible, before discussion and media exposure build up.
Keep witnesses separate so they cannot blend accounts.
Use open-ended, non-leading questions and let the witness recall freely.
Avoid giving feedback that confirms or rewards particular answers.
Document what the witness said at first contact, so later changes can be traced.

These steps cannot restore a memory that has already been affected, which is why timing matters so much. The broader question of how much weight to give such accounts is covered in how reliable is eyewitness testimony?.

Does delay make contamination more likely?

Time works against a clean account. A memory that is fresh and has just been recorded is harder to reshape than one that has sat for weeks while conversations, news and questions accumulate. As the original trace fades, the mind leans more heavily on reconstruction to fill the gaps, and reconstruction readily draws on whatever information is available, including the contaminating kind. This is one reason the interval between an event and the first proper interview is treated as valuable time to protect, a theme in how does time affect eyewitness memory?.

Repeated questioning adds its own complications. Being asked about an event many times, especially with varying wording, gives more opportunities for suggestion to creep in and for a witness to settle into a version that may include borrowed details. Careful, consistent interviewing can help memory, but sloppy repetition can entrench contamination, as discussed in can repeated recall improve memory?. The common thread is that contamination is cumulative: the longer and messier the path from event to record, the more outside material a memory can absorb.

Does a contaminated account mean the witness is lying?

No. This is the most important distinction. A contaminated memory is reported in good faith; the witness believes it and has no intention to mislead. The error lives in the process that surrounded the memory, not in the witness's honesty. Treating contamination as a lie both misreads the evidence and risks punishing someone for the ordinary workings of memory. Contamination is a reason to be cautious about an account, never a reason to conclude that the person is being deceptive or that a particular detail proves guilt.

How does Murder Book use contamination?

Murder Book: The Art of Interrogation is an upcoming narrative interrogation game from Gungrounds, an independent studio in Zagreb, Croatia. Its cases revolve around unreliable stories and narrative gaps, where accounts can be shaped by what a person heard, read or was told rather than by what they witnessed. A detail that seems to come from nowhere may have been picked up elsewhere, and the game treats that as something to trace rather than as instant proof of a lie.

That approach fits the research: contamination is a reason to ask where a memory came from and to look for independent corroboration, not a verdict. 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 how fresh information can overwrite an older memory, see can new information change an old memory?.