93%. Any corporate communications trainer worth their salt knows (and tells their clients at the beginning of a session) that 93% of what people convey to others are non-verbal. Participants usually audibly gasp when they learn that only 7% of what they convey is through content, 38% tone of voice, and 55% gestures and body language. This statistic even appears in a recent article on About.com. After returning to the original study, Inference of Attitudes from Nonverbal Communication in Two Channels (Mehrabian and Ferris, 1967), I found that the reporting of these statistics is inaccurate and broadly over generalized.
In the article, the author, Susan Heathfield, wrote: “One study at UCLA indicated that up to 93 percent of communication effectiveness is determined by nonverbal cues. Another study indicated that the impact of a performance was determined 7 percent by the words used, 38 percent by voice quality, and 55 percent by the nonverbal communication.” Ms. Heathfield erroneously iterates that one study was actually two and gives the reader no other information as to how these conclusions were formed.
The study had an experimental design. The variables in the study were positive, negative and neutral vocal and facial attitude and the degree of positive attitude inferred from two speakers and two models. Subjects were asked to listen to a recording of a female saying the single word ‘maybe’ in three tones of voice to convey liking, neutrality and disliking. The subjects were then shown photos of female faces with the same three emotions and were asked to guess the emotions in the recorded voices, the photos and both in combination. The photos got more accurate responses than the voice, by a ratio of 3:2.
What the article failed to convey is that the randomly assigned sample size was relatively small (62). 25 subjects were used in the selection of a neutral word, 17 assessed the independent effects of facial and vocal communication and 20 obtained the combined effects of facial-vocal communications. Research participants were all female UCLA undergraduates in an introductory psychology course.
It is difficult to extrapolate these findings to the general public when tall the subjects were female, all in the same limited age range, and attended the same university. Since the study was actively seeking the effects of facial and vocal communication it is not surprising that this is the effect it found.
Frequently the ‘facial expression’ variable is taken to mean all physical non-verbal communication (i.e. gestures, posture, etc.). In addition, this experiment was done asking the participants what the attitude of the speaker was to a third party. Effects would most likely be different if the participant was the direct recipient of the communication. The experiment was also designed to test the feelings of others as opposed to the impact or persuasiveness of the speaker, which is how it was presented in the About.com article.
In comparing the article to the research report, I found that the results were described too broadly and included inferences not made in the original report. This highlights the need for caution by consumers and especially communications trainers who offer these statistics. Although it is powerful to quote scientific studies to support our work, it is too easy to inaccurately present findings that support our own point of view.
Mehrabian, Albert, and Ferris, Susan R. “Inference of Attitudes from Nonverbal Communication in Two Channels,” Journal of Consulting Psychology, Vol. 31, No. 3, June 1967, pp. 248-258
Heathfield, Susan M. “Listen With Your Eyes: Tips for Understanding Nonverbal Communication,” About.com.