Case File 077 · Investigation and Psychology

How Repeated Questioning Can Affect Memory

Repeated questioning can bring back genuine details or plant false ones, depending on what the interviewer adds and repeats each time.

By Murder Book Editorial TeamPublished 16 September 2026Updated 16 September 2026

Repeated questioning can affect memory in two opposite ways, and which one wins depends on how the questions are asked. Asking a witness to recall an event more than once, using open and neutral questions, often brings back genuine details that were missed the first time. Researchers call this reminiscence. But when the same misleading suggestion is repeated across interviews, people become more likely to "remember" things that never happened, and to feel confident about them. Research on the misinformation effect, led for decades by Elizabeth Loftus, shows that information met after an event can change what people later report about it. So a second or third interview is not harmful in itself. The risk comes from what the interviewer adds, repeats or implies each time.

Key takeaways

Recalling an event again with neutral questions can produce genuine new details (reminiscence) without a large loss of accuracy.
Repeating a misleading suggestion makes false memories more likely and more confidently held.
Recalling an event early can, in some conditions, make memory more vulnerable to later misinformation.
Changed answers to repeated questions are not reliable proof of lying; the questioning itself can cause the change.
Confidence expressed after several interviews says less about accuracy than confidence at the first, uncontaminated test.

What happens to memory when someone is questioned again?

Every time a person recalls an event, they rebuild it from what they stored, what they know in general and what they have heard since. That is why a second interview can go either way.

On the helpful side, a 2013 study by Odinot, Memon, La Rooy and Millen interviewed witnesses to a filmed event twice using the Cognitive Interview. Witnesses recalled new, mostly accurate details in the second interview, and the overall accuracy of their accounts held up. The number of errors rose somewhat, but the proportion of accurate details did not. The authors concluded that witnesses are unlikely to report everything they remember in a single interview, however thorough.

Earlier work on young children by La Rooy, Pipe and Murray found the same pattern of reminiscence: new information came out across repeated interviews, though less of it after a long delay. Gilbert and Fisher (2006) found that nearly all adult participants recalled something new at a second test, especially when the retrieval cues were varied, and that recalling new details was not significantly linked to lower overall accuracy.

This matters because a new detail in a later interview is sometimes treated as a warning sign. The research suggests it is often just how retrieval works. For more on this, see why truthful accounts can still be inconsistent.

How does repeated misinformation create false memories?

The problems start when the questioning carries information of its own. The misinformation effect is the impairment of memory that follows exposure to misleading information about an event. In her 2005 review of 30 years of research, Loftus describes how widely it has been observed, which people tend to be more susceptible, and the finding that warnings about misinformation help only under limited circumstances.

Repetition makes it stronger. In a study published in 1996, Zaragoza and Mitchell showed participants a video of a burglary, then asked questions containing misleading suggestions, some of them repeated. Compared with a single exposure, repeated suggestions made participants more likely to claim, with high confidence, that they had seen the suggested events in the video, and to say they consciously remembered witnessing them. The effect held over retention intervals of up to a week.

Put simply, hearing something several times makes it feel familiar, and familiarity is easily mistaken for memory. A suggestion that arrives in the first interview and returns in the second and third can end up feeling like part of the original event. This is different from a single leading question: the concern here is the cumulative effect of the same idea being reinforced across sessions.

Can recalling an event early make memory more fragile?

Normally, practising retrieval strengthens memory. That is why researchers expected an immediate recall test to protect witnesses from later misinformation. In 2009, Chan, Thomas and Bulevich found the opposite in their experiments. Immediate cued recall made the later misinformation effect worse, for both younger and older adults. They called it a "reversed testing effect".

The authors pointed to two mechanisms. The first recall seemed to help people learn the misinformation that followed, and the details they had just recalled became especially open to interference. Their point was practical: real witnesses usually do describe an event early, to a call handler or an officer at the scene, before they meet any later misinformation. The finding does not mean early accounts should not be taken. It means what happens between interviews matters.

Why do answers change when the same question is repeated?

When a question is asked again, the person may reasonably assume their first answer was wrong. That social pressure can change an answer without any change in memory.

A field study by Cederborg, Danielsson, La Rooy and Lamb (2009) analysed Swedish investigative interviews with children and young people who had intellectual disabilities. About 6% of questions were repeated at least once. When focused questions were repeated, interviewees changed their answers 40% of the time. Because this was real casework, the researchers could not tell which answers were accurate. Their conclusion was that repeating potentially contaminating questions reduced the usefulness of the testimony.

The lesson carries beyond that group: an answer that shifts under repetition tells you about the questioning as much as about the witness. It is not, on its own, a sign of deception. Research on lie detection shows people judge truth and lies correctly only about 54% of the time on average (Bond and DePaulo, 2006), so treating changed answers as proof of lying adds a second error to the first.

Can repeated suggestive interviews create memories of crimes?

Some of the strongest claims in this area come from studies that use several suggestive interviews. In 2015, Shaw and Porter reported that after three interviews using suggestive memory-retrieval techniques, 70% of participants were classified as having false memories of committing a crime in adolescence.

That figure is disputed. In a 2018 commentary, Wade, Garry and Pezdek recoded the data using approaches that separate believing a suggestion from actually remembering it. By those criteria, 26–30% of participants met the criteria for false memories. That is still a substantial number, and closer to the rest of the literature. The debate itself is useful: how you define a "memory" changes how many false memories you find.

For how memory is rebuilt rather than replayed, see why memory is not a perfect recording.

Does repeated questioning affect confidence?

Yes, and this is one of the less obvious risks. In their 2017 review, Wixted and Wells argue that an eyewitness's confidence can be a useful guide to accuracy under "pristine" conditions: an initial, uncontaminated memory test, a fair line-up, no influence from the administrator and a confidence statement taken immediately. When those conditions are not met, even a high-confidence identification can be seriously compromised.

The implication for repeated questioning is straightforward. The confidence that matters most is the confidence recorded at the first clean test. Confidence that grows over several interviews, after feedback or repeated suggestions, is a much weaker signal.

What does this mean for investigations?

The research does not say "interview once and never again". It points to a few clear principles:

Record the first account carefully. It is the least contaminated version.
Use open, neutral questions in later interviews. These are the conditions under which reminiscence is most likely to add accurate detail.
Avoid repeating specific suggestions. Repetition is what turns a suggestion into something that feels remembered.
Treat change with care. A new detail or a revised answer can come from normal retrieval, from social pressure or from contamination, as well as from deception.

When investigators compare multiple accounts, knowing the history of each interview helps explain why accounts drift.

Is there a game about stories that change under questioning?

Murder Book: The Art of Interrogation is an upcoming narrative-driven interrogation game from Gungrounds, an independent studio in Zagreb, Croatia. It is inspired by real police footage and psychological interviews, and asks players to examine unreliable stories: narrative gaps, avoidance, behavioural changes and contradictions.

Research on repeated questioning is a useful reminder when playing that kind of game. A contradiction is a reason to look closer, not a verdict. Accounts can shift for innocent reasons, and behaviour alone is not proof. Development screenshots show a case select screen that includes a "not guilty" result, a sign that pressing a suspect is not the same as proving guilt.

Murder Book is in development for PC (Steam) and mobile, with release currently planned for the end of Q2 2027. Read How Murder Book Turns Interrogation Into Gameplay to learn more.