A meta-analysis of 206 studies on lie detection found that humans average 54% accuracy at spotting lies. Barely above a coin flip (Bond & DePaulo, 2006).
Police officers, judges, customs agents, experienced therapists: none do meaningfully better than college students. The body-language industry has built a multi-billion-dollar self-help market on top of a research literature that says: you are not good at this.
The lineup
The intuitive cues that pop psychology sells as “tells” are mostly noise. Calm people lie. Nervous people tell the truth. Confident people lie. Sweating people tell the truth (they might just be warm). Practiced deceivers often look more trustworthy than honest people having a hard day.
Why single cues fail
There is no reliable single cue for deception. The largest meta-analysis of nonverbal cues to deception, DePaulo et al. 2003, found that 158 studies’ worth of candidate tells (fidgeting, eye contact, pause length, posture shifts, hand gestures) produced effect sizes close to zero. No cue reliably distinguishes liars from truth-tellers.
The reason is simple. Everything that pop psychology attributes to lying (nervousness, shifting, avoiding eye contact) is also what happens when someone is:
- Anxious for non-deception reasons
- In a cold or hot room
- Tired
- Talking to someone they find intimidating
- Thinking hard about something technical
- Neurodivergent in a completely ordinary way
The so-called tells don’t index lying. They index arousal. And arousal has a hundred causes.
What the industry sells vs what’s true
| Pop-psych claim | What research actually shows |
|---|---|
| “93% of communication is nonverbal” | Applies only to narrow conflict cases between words and face. Mehrabian himself has been correcting this for 55 years. |
| “Avoiding eye contact = lying” | If anything, backwards. Practiced liars often maintain more deliberate eye contact. |
| “Arms crossed = defensive” | Not empirically supported. People cross their arms for many boring reasons. |
| “Microexpressions reveal hidden emotions” | Much weaker evidence than Lie to Me or training programs suggest. |
| “You can spot a liar if you know what to look for” | Meta-analyses disagree. You can’t. Almost nobody can. |
Why this matters
- Your gut feels certain, and the data says it’s 54% right. That’s very close to useless, and worse than the Dunning-Kruger confidence you probably bring to it.
- Trained professionals don’t do better than amateurs. Police, TSA, judges, therapists. Training in nonverbal cues barely moves the needle.
- “Something feels off” is often reliable. “I know exactly what they’re hiding” is almost always wrong.
- You cannot read strangers. Baseline reading works on people you know. On strangers, you are guessing.
- The industry that sells you otherwise is selling a fantasy. A useful fantasy in many cases, but a fantasy.
Humility is the most advanced skill in body-language reading.