Core topic 05 of 13
Body language: what it shows, and what people invented
Nonverbal behaviour carries real information about rapport, status and attention. It carries almost no reliable information about whether someone is lying, and the famous claim that 93 percent of communication is nonverbal comes from a misreading of two small studies from 1967.
Body language is the area where popular psychology and the research have drifted furthest apart. Whole training industries rest on findings that either never said what they are quoted as saying or failed when someone repeated them. Start with the number everybody knows.
The 7-38-55 rule, and where it came from
Albert Mehrabian ran two experiments at UCLA in 1967. In one, people heard a single word spoken in different tones and judged the speaker's attitude. In the other, they matched a spoken word against a photograph of a face. When the word and the tone disagreed, listeners went with the tone. From those results came the split of 7 percent words, 38 percent tone, 55 percent facial expression.
The conditions matter. One word, no context, deliberately conflicting signals, a handful of female participants. Mehrabian has spent decades telling people the numbers do not generalise. They generalised anyway, because they fit neatly on a slide.
Signals with evidence behind them
Mirroring and rapport Replicated
People unconsciously copy the posture and mannerisms of people they are getting along with. Chartrand and Bargh named this the chameleon effect in 1999, and found that being mirrored made people rate the interaction as smoother and the other person as more likeable. Copying deliberately and obviously does not produce the same result.
Interpersonal distance carries meaning Replicated
Edward Hall's proxemics categories from the 1960s have held up in the broad sense: people keep predictable distances that shift with relationship and culture. Sorokowska's 2017 study across 42 countries found large national differences, with preferred stranger distance ranging from under 80 centimetres to well over 120. The rule is real, the specific numbers are local.
Vocal cues beat visual ones Replicated
When researchers compare channels, voice usually wins. Michael Kraus found in 2017 that people judged others' emotions more accurately from voice alone than from voice plus video. Adding the face made judgements worse, which is a problem for any training that tells you to read the face first.
Claims that did not survive
Reading deception from posture and gaze Debunked
Bond and DePaulo's 2006 meta-analysis of 206 studies found people detect lies at 54 percent accuracy. Police officers, customs agents and judges performed the same as undergraduates, though they were more confident. Gaze aversion in particular is a cultural signal of respect in much of the world and shows no consistent relationship with lying.
Microexpressions as reliable lie detectors Debunked
Brief involuntary facial movements do occur. Their value as deception evidence has not been demonstrated outside the labs that sell the training. Independent tests, including work by Burgoon in 2018, found microexpression training produced little or no improvement in detection accuracy.
Power posing changes your hormones Debunked
The 2010 study reported that expansive postures raised testosterone, lowered cortisol and increased risk taking, from 42 participants. Ranehill and colleagues repeated it in 2015 with 200 participants and found no hormonal or behavioural effect. Dana Carney, one of the original authors, later stated publicly that she no longer believes the effect is real. Feeling more powerful after standing tall is the part that survived.
How to use nonverbal information without fooling yourself
Read clusters over time rather than single gestures, and treat any change from a person's normal baseline as a question rather than an answer. Crossed arms mean the room is cold at least as often as they mean defensiveness. If you want to know what someone is thinking, the highest accuracy method in the literature remains asking them and listening to how they answer.
Watch this on the channelBody language videos that separate the trained-observer claims from what the meta-analyses found.
Open Psychology Cue on YouTubeQuestions people ask about this topic
Is communication really 93 percent nonverbal?
No. Mehrabian's 1967 studies produced that split under a narrow condition: judging feeling behind a single spoken word when tone and expression conflicted with the word. He has said repeatedly the figures were never meant to describe communication in general.
Can you tell if someone is lying from their body language?
Barely. Bond and DePaulo's meta-analysis of 24,000 judgements found 54 percent accuracy, and trained professionals scored no better than students while feeling more certain.
Does power posing work?
The hormone and risk-taking claims failed replication with larger samples. A small effect on self-reported feelings of power survives.
What nonverbal cues do carry real information?
Vocal pitch and speech rate, interpersonal distance, spontaneous mirroring, and eye contact timing during conversational turns. These predict rapport and status better than any single gesture.