Partners in crime: How liars in collusion betray themselves

The paradigmatic task for participants in studies on deception is to assess veracity on the basis of a single statement. However, in applied contexts, lie catchers are often faced with multiple statements (reported by one or several suspects). To appreciate this mismatch, we conducted a study where each member of 10 truth‐telling pairs and 10 lying pairs (reporting fabricated alibis) was interrogated twice about an alibi. As predicted, lying pair members were more consistent between themselves than were truth‐telling pair members, and single liars and truth tellers were equally consistent over time. Furthermore, truth tellers made more commissions than did liars. Although in line with our repeat vs. reconstruct hypothesis, these findings contrast sharply with beliefs held by professional lie catchers and recommendations found in literature on deception detection. The results are translated into an applied psycholegal context.

Laboratory analog study with 10 truth-telling pairs and 10 lying pairs reporting about alibis, each member interrogated twice. An additional 120 undergraduate students served as lie-catchers evaluating the statements for consistency patterns and veracity judgments.

Background & Context

This groundbreaking 2003 study by Granhag, Strömwall, and Jonsson addresses a critical gap in deception detection research. Most deception studies focus on assessing veracity from a single statement, but in applied contexts, lie catchers are often faced with multiple statements from one or several suspects. This mismatch between laboratory research and real-world investigative practice had profound implications for how police evaluate suspects in collusion.

The research challenges conventional wisdom about consistency as a hallmark of truthfulness. Traditionally, law enforcement has operated under the assumption that liars contradict themselves while truth tellers remain consistent—a belief that underpins many interrogation manuals and training programs. This study tests whether that assumption holds when suspects work together to fabricate alibis.

The work introduces the "repeat versus reconstruct" hypothesis, a theoretical framework proposing that liars rehearse and repeat prepared stories, while truth tellers reconstruct events from genuine memory—a cognitive process that naturally produces some inconsistencies across retellings.

Research Design & Methodology

The study involved 10 truth-telling pairs and 10 lying pairs (reporting fabricated alibis), with each member interrogated twice about an alibi. This yielded 20 pairs total (40 individual participants) in the primary experimental manipulation. Additionally, 120 undergraduate students were enrolled as lie-catchers to evaluate the statements and make veracity judgments.

The design examined two critical dimensions of consistency:

• Between-pair consistency: How well statements matched between the two members of each pair
• Within-person consistency: How consistent each individual remained across their two interrogation sessions
• Commissions: New details added in the second interrogation that were not present in the first

This repeated-measures design with paired suspects allowed researchers to test whether consistency patterns differed systematically between truthful and deceptive dyads—a scenario common in criminal investigations where accomplices coordinate alibis.

Results & Key Findings

As predicted, lying pair members were more consistent between themselves than were truth-telling pair members, and single liars and truth tellers were equally consistent over time. This primary finding directly contradicts the traditional consistency heuristic used by interrogators.

Key statistical outcomes included:

• Higher between-pair consistency for liars: Pairs fabricating alibis showed greater agreement with each other than pairs telling the truth
• Equal within-person consistency: Individual liars and truth tellers were equally consistent across their own repeated statements
• Truth tellers made more commissions than did liars: Honest witnesses added new details in their second interview, while liars stuck to their prepared scripts
• Modest detection accuracy: Overall deception detection accuracy was modest, and lie-catchers given access to a large number of statements did not outperform those given access to fewer statements

Although in line with the repeat vs. reconstruct hypothesis, these findings contrast sharply with beliefs held by professional lie catchers and recommendations found in literature on deception detection. The study revealed that consistency—long considered a sign of truthfulness—may actually signal collusion and rehearsal when comparing statements between suspects.

Discussion & Significance

This research fundamentally challenges how investigators should interpret consistency in statements from multiple suspects. The "repeat versus reconstruct" mechanism explains why liars appear more consistent: they memorize a fabricated story and repeat it verbatim, while truth tellers rely on genuine memory reconstruction, which naturally introduces variations and additional details upon repeated recall.

The finding that truth tellers produce more commissions aligns with memory research showing that genuine retrieval is a dynamic, reconstructive process. Each time we recall an event, we may access different aspects of the memory trace, leading to new details emerging. Liars, constrained by their need to maintain a consistent cover story, avoid adding information that might contradict their prepared narrative.

The findings contrast sharply with beliefs held by professional lie catchers and literature recommendations on deception detection. This disconnect between scientific evidence and practitioner beliefs represents a serious problem in applied forensic contexts, where misapplied consistency judgments could lead to wrongful convictions of truth tellers or failure to detect coordinated deception.

The study's contribution extends beyond polygraph examination to all forms of investigative interviewing where multiple suspects are questioned separately about shared events. It suggests that high consistency between suspects should raise suspicion rather than confirm credibility.

Limitations & Considerations

The study used a laboratory analog design with undergraduate students, raising questions about ecological validity. Real criminal suspects face higher stakes than research participants, which could affect both lying strategies and memory processes. The sample size of 10 pairs per condition, while adequate for detecting the predicted effects, limits generalizability.

The fabricated alibis involved relatively simple scenarios. In actual investigations, coordinated lies may involve more complex narratives, longer time intervals between interrogations, and opportunities for suspects to communicate between sessions. These factors could alter consistency patterns.

The research does not address how skilled interrogators might detect collusion through questioning techniques beyond simple consistency assessment. Follow-up questions probing unexpected details, as suggested in later research on unanticipated questions, might reveal differences not captured in standard repeated interrogations.

Practical Applications

This research demands a fundamental shift in how investigators evaluate statements from multiple suspects. Rather than viewing between-suspect consistency as evidence of truthfulness, investigators should treat unusually high agreement—especially on peripheral details—as a potential red flag for collusion. The presence of commissions (new details in later interviews) should be viewed more favorably for truth tellers than previously assumed.

For polygraph examiners and law enforcement, the implications are clear: consistency-based credibility assessment must be reconsidered when evaluating coordinated statements. Training programs should incorporate the repeat-versus-reconstruct framework, teaching investigators to distinguish between the rigid consistency of rehearsed lies and the natural variability of reconstructed memories. Questioning strategies should deliberately probe areas where rehearsed stories might break down, focusing on unexpected or peripheral details that suspects may not have prepared together.