Can we tell when someone is lying from their body language? Popular media (TV detectives, polygraph fans) say yes. The psychological evidence says: only modestly — and most "tells" people believe in are myths.
Common myths about deception
- "Liars avoid eye contact." Actually, careful liars often increase eye contact deliberately to seem honest. Truthful people may glance away while thinking.
- "Fidgeting means lying." Fidgeting reflects anxiety or stimulation, not lying specifically. Innocent suspects in police interviews often fidget more than guilty ones (the innocent fear being disbelieved).
- "Touching the face means lying." A weak correlation at best. Many truthful speakers do this.
Vrij (2008) reviewed hundreds of studies on deception detection. The headline finding: untrained observers detect lies at about 54% — barely above chance. Police officers, judges and even polygraphers do not perform much better.
What does correlate with lying (modestly)
- Increased cognitive load — lying requires constructing a story, monitoring consistency, and watching the listener; this can leak as longer pauses, slower speech and reduced gesture (Vrij's cognitive approach).
- Reduced detail — fabricated stories often contain fewer specific sensory and contextual details than true accounts (Statement Validity Analysis, criteria-based content analysis).
- Inconsistency with prior statements (catchable in repeated interviews).
Vrij's cognitive interview approach
Rather than looking for body-language tells, deliberately increase cognitive load on the suspect:
- Ask them to recall the event in reverse order.
- Ask unanticipated questions they couldn't have rehearsed.
- Maintain eye contact and ask probing follow-ups.
Liars under high cognitive load show more verbal slips and gaps than truth-tellers under the same load. Vrij and colleagues report detection rates climbing to ~70% with cognitive interviewing — substantially better than baseline.
Police and forensic applications
- The PEACE model (UK) — Planning & preparation, Engage & explain, Account, Closure, Evaluation. Trained, non-coercive interviewing produces more reliable evidence than aggressive interrogation.
- Cognitive interview (Geiselman & Fisher) for witnesses (not suspects) — instructions to mentally reinstate the context, recall everything, change order, change perspective. Produces more accurate detail than standard police interviews.
- Body-language reading is taught some places but evidence-light. Vrij has urged forces to drop reliance on classic "tells."
Why we are bad at detecting lies
- Overconfidence — most people rate themselves above average at lie detection.
- Reliance on stereotypes — gaze aversion = liar (false).
- Truth bias — we tend to assume people are telling the truth (especially in low-stakes contexts).
⚠Common mistakes— Common errors
- Citing classic body-language "tells" (gaze aversion, fidgeting) as reliable — they aren't.
- Confusing the cognitive interview for witnesses with the cognitive approach for suspects — they are different applications of the same general "increase cognitive demand" principle.
- Forgetting that polygraph evidence is also unreliable and inadmissible in many courts.
AI-generated · claude-opus-4-7 · v3-deep-psychology