Why Do Investigators Ask Unexpected Questions?

Asking questions that a person has not prepared for is one of the more promising interview techniques in deception research. The logic is straightforward: a truth teller can draw on genuine memory to

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

Asking questions that a person has not prepared for is one of the more promising interview techniques in deception research. The logic is straightforward: a truth teller can draw on genuine memory to answer almost any relevant question, while a liar who has rehearsed a cover story may struggle when the interviewer moves outside that script. Research by Aldert Vrij and colleagues has found that unanticipated questions can produce larger differences in verbal detail between truth tellers and liars than expected questions do. The technique does not reveal whether someone is lying. It creates conditions in which the verbal differences between truthful and deceptive accounts become easier to observe, giving the interviewer more information to work with.

Key takeaways

Unanticipated questions exploit the gap between a rehearsed cover story and genuine memory.
Truth tellers tend to maintain a similar level of detail across expected and unexpected questions, while liars often show a sharper drop in detail.
The technique is rooted in the cognitive load approach: lying is mentally demanding, and unexpected questions increase that demand.
Asking unanticipated questions does not prove deception. It highlights differences that still require independent verification.
Research suggests the approach works best when paired with other information-gathering methods rather than used as a standalone test.

Why do liars struggle with unexpected questions?

Most people who plan a deceptive account rehearse the answers they expect to be asked. They practise the chronology, the key facts and the emotional framing. What they rarely do is prepare for questions that approach the same event from an unusual angle, such as asking about spatial layout instead of sequence, or about what someone could see from a particular position rather than what they did.

Vrij and colleagues have described this as part of the cognitive load approach to lie detection. Telling the truth is largely a matter of retrieving a memory. Lying requires constructing a plausible answer, monitoring its consistency with what has already been said and suppressing the real memory at the same time. When an interviewer asks an unexpected question, a truth teller simply searches memory from a different angle. A liar, lacking that memory, has to improvise.

A 2009 study by Vrij, Leal, Granhag, Mann, Fisher, Hillman and Sperry tested this directly. Pairs of participants were sent on a mission. Truth tellers carried out the mission; liars planned a cover story. When asked anticipated questions, truth tellers and liars were hard to distinguish. When asked unanticipated questions, liars gave significantly fewer spatial details, and their answers were rated as less consistent within the pair.

What does the research show?

A key experiment published by Vrij, Leal, Mann and Fisher in 2012 had truth tellers perform five tasks in a room while liars watched through a window and prepared a story claiming to have done the same tasks. The interview schedule included pairs of questions where the first was easy to anticipate and the second involved a perspective shift, such as a temporal or spatial reframing. Compared with truth tellers, liars showed a greater decline in detail across three out of four question pairs. Liars also performed worse on a secondary task administered simultaneously, suggesting higher cognitive load.

Lancaster University and University of Portsmouth research has also examined what happens when liars are familiar with the topic they are lying about. A study by Warmelink, Vrij, Mann, Leal, Forrester and Fisher found that background questions distinguished truths from both familiar and unfamiliar lies, while detailed spatial questions distinguished truths from unfamiliar lies but not from familiar ones. This is an important limit: when a liar has genuine experience of a setting, unexpected questions about that setting may not produce the same drop in detail.

Does the technique work in digital settings?

Research has explored whether unanticipated questions can detect deception in computer-mediated communication. One study found that liars showed prolonged mouse trajectories when responding to unanticipated questions online, suggesting that the cognitive effects of unexpected questions are not limited to face-to-face interviews.

What are the limits of this approach?

The technique is not a lie detector. Several caveats are important:

Familiarity narrows the gap. If a liar has genuine knowledge of the place or event they are lying about, they can draw on real memory to answer spatial and contextual questions. The technique is strongest when the liar's account is entirely fabricated.
Accuracy is modest. Meta-analytic findings on cognitive load approaches to deception detection report overall accuracy rates around 58%, which is above chance but well below certainty. Can You Actually Tell When Someone Is Lying? reviews the broader accuracy data.
Behavioural cues remain weak. Even with unexpected questions, the differences between truth tellers and liars are statistical tendencies, not reliable individual markers. DePaulo and colleagues' 2003 meta-analysis found that most behavioural cues to deception are faint.
Innocent people can struggle too. A nervous or confused truth teller may give less detail to an unexpected question for reasons unrelated to deception. Is Nervousness a Sign That Someone Is Lying? discusses this overlap.

How does this compare with other interview techniques?

Unanticipated questions are one of several cognitive interview strategies researchers have proposed. Others include asking someone to tell their story backwards, requesting drawings or sketches, and the strategic use of evidence. Vrij, Fisher and Blank's 2017 meta-analysis of cognitive credibility assessment approaches found that they raised lie detection accuracy compared with standard interviews, but accuracy rates still varied considerably across studies.

The common thread across these techniques is that they increase the mental demands on the interviewee. The theory is that this extra demand disproportionately affects liars. In practice, each method has specific strengths and limits, and researchers generally recommend combining approaches rather than relying on any single one.

How does this relate to Murder Book?

Murder Book is an upcoming narrative interrogation game from Gungrounds, an independent studio in Zagreb, Croatia. Development screenshots show a technique menu that includes Baiting, Bluffing and Confrontation, and the game's press kit names Fabrication and Avoidance among the behaviours players examine. Asking unexpected questions fits the same investigative logic: when a suspect has prepared a story, the player's job is to find the seams.

In Murder Book, cases are designed for focused 5-10-minute sessions. Players examine unreliable stories, behavioural changes and narrative gaps. The game does not treat any single response as proof of deception, which reflects the research finding that no single technique reliably separates truth from lies. Murder Book is in development for PC (Steam) and mobile, with release currently planned for the end of Q2 2027.