Is 93% of communication really nonverbal?
Two studies from 1967, single spoken words, photographs of faces. Mehrabian has spent decades objecting to how his own finding gets used.
No. The figure comes from two small studies published in 1967. In one, women listened to a single spoken word — "maybe", "honey", "brute" — and judged the speaker's feeling from the tone. In the other, they judged it from a photograph of a face.
The numbers describe how people guess an attitude when tone and wording contradict each other. They were never a measurement of how meaning travels in a presentation, and the author has spent decades saying so.
Somewhere in almost every presentation course, a slide appears with three numbers on it: 7% words, 38% tone, 55% body language. The conclusion drawn from it is always the same — what you say barely matters, so work on your delivery.
It is the most quoted statistic in communication training, and it is a misreading of two experiments that were about something else entirely.
What the experiments actually involved
Albert Mehrabian, then at UCLA, published two studies in 1967. The first, with Morton Wiener, tested what happens when the same single word is spoken in a positive, neutral or negative tone. Nine words in total. Participants judged whether the speaker felt positively or negatively toward them.
The second, with Susan Ferris, added photographs of facial expressions and asked the same kind of question. Combining the relative weights from the two studies is where 7 / 38 / 55 comes from. It was arithmetic across two experiments, not a single result.
Mehrabian's own objection
Mehrabian has stated repeatedly that the equations apply only in a narrow case: when someone is communicating about feelings or attitudes, and when the verbal and nonverbal signals contradict each other. Neither condition holds in a quarterly update.
That restriction is not a technicality — it is the entire finding. The studies were designed around conflict between channels. If a colleague says "fine" in a clipped voice, you believe the voice. That is genuinely what was measured, it is true, and it is useful to know. It says nothing about the ratio in which content and delivery contribute when they agree with one another.
The test that settles it
If 93% of communication really were nonverbal, you could sit through a technical presentation in a language you do not speak and come away with 93% of it. Everyone knows that is false. It takes one second of thought, which suggests the number does not survive on evidence — it survives because it is repeatable and flattering to a certain kind of training.
What this does not mean: delivery is not irrelevant. A monotone kills attention, and attention is the precondition for everything else. The claim being dismantled here is a specific one — that content is 7% of the job — not the broader idea that how you speak matters. It does. It is simply not thirteen times more important than what you say.
Why the myth is expensive
Because it sends people to the wrong work. Someone who believes content is 7% will rehearse gestures, practise eye contact, and record themselves for tone — and walk into the room still unable to say in one sentence what they want to happen.
I have watched that repeatedly. The delivery is polished, the room is attentive, and fifteen minutes later nobody can reconstruct what was being asked for. Attention was never the constraint. Having something precise to be attentive to was.
What to take instead
A more useful pair of statements, both defensible:
- When your words and your manner disagree, people believe your manner. This is what Mehrabian found, and it is worth acting on. If you do not believe the number you are presenting, do not expect to sell it by saying it firmly.
- When they agree, the words carry the meaning and the manner carries the credibility. Delivery decides whether you are listened to. Content decides what is left when the listening stops.
Which is the more practical framing anyway, because it tells you what to fix. If people are not paying attention, work on delivery. If they pay attention and still leave with the wrong idea — the far more common failure, and the more expensive one — no amount of gesture practice will touch it.
Sources
- Mehrabian, A., & Wiener, M. (1967). Decoding of inconsistent communications. Journal of Personality and Social Psychology, 6(1), 109–114.
- Mehrabian, A., & Ferris, S. R. (1967). Inference of attitudes from nonverbal communication in two channels. Journal of Consulting Psychology, 31(3), 248–252.
- Mehrabian, A. (1981). Silent Messages: Implicit Communication of Emotions and Attitudes (2nd ed.). Wadsworth. Contains the author's own restriction on when the figures apply.