Albert Mehrabian’s studies in nonverbal communication
Albert Mehrabian’s famous formula expressing the dominance of nonverbal communication is derived from two studies he carried out with colleagues in 1967. They were pioneering studies in the study of nonverbal communication. But they were contrived, methodologically unsound by today’s standards, and limited in their real-world application.
This page is support for my blog post Why the stickiest idea in presenting is just plain wrong.
Here I describe Mehrabian’s experiments in detail, and quote extensively from critiques of the methodology used in the experiments.
Mehrabian’s first experiment
In the first study with Morton Wiener, he set out to investigate how people judged a speaker’s feelings where what the speaker said was inconsistent with the tone of voice used. Mehrabian wanted to find out which was more important: the content (the words the speaker used) or the tone.
Two female speakers were employed to read nine different words (three positive “dear””thanks” “honey”, three neutral “maybe” “oh” “really”and three negative “brute” “don’t” “terrible”), each spoken with three different tones (positive, neutral and negative towards an imaginary addressee). These were recorded.
Three groups of 10 participants were then asked to listen to the recordings and rate the degree of positive attitude of the speaker. They were given different instructions in making their judgement:
Group 1: Pay attention to only the content
Group 2: Pay attention to only the tone of voice
Group 3: Pay attention to all the information available.
Mehrabian and Wiener found that:
The results indicate that judgements of attitude from “inconsistent” messages involving single words spoken with intonation are primarily based on the attitude carried in the tonal component.
Mehrabian’s second experiment
This experiment was carried out with Susan Ferris. It was also about how people judged the feelings of a speaker. But this time instead of looking at the relative importance of tone versus content, Mehrabian and Ferris looked at the relative importance of tone versus facial expression. The impact of content was minimalized by choosing the most neutral word that they could find: “maybe”.
This time, three tones of voice (recordings of three female speakers saying the word “maybe” in three different ways) were combined with three facial expressions (photos of three female models).
They found that facial expression was approximately 1.5 times more important than tone of voice in judging the attitude of the speaker.
Mehrabian’s famous formula
At the end of this second research paper, Mehrabian and Ferris attempted to integrate the findings from both experiments. They say:
It is suggested that the combined effect of simultaneous verbal, vocal and facial attitude communications is a weighted sum of their independent effects – with the coefficients of .07, .38 and .55, respectively.
They don’t show how they combined the results of the two experiments and arrived at this formula. But it’s worth noting that neither experiment looked at all three communication channels together.
For other useful descriptions of Mehrabian’s studies see:
Mehrabian has always been clear that the results of these studies are limited in application. In the first research paper he says:
These findings regarding the relative contribution of the tonal component of a verbal message can be safely extended only to communication situations in which no additional information about the communicator-addressee relationship is available.
On his website Mehrabian says:
Please note that this and other equations regarding relative importance of verbal and nonverbal messages were derived from experiments dealing with communications of feelings and attitudes (i.e., like-dislike). Unless a communicator is talking about their feelings or attitudes, these equations are not applicable.
What other researchers have said
Mehrabian’s findings in these two papers (together with other papers of that era using similar methodologies that also trumpeted the dominance of nonverbal messages) have been discussed by subsequent researchers in the field. The have identified a number of methodological shortcomings in the experimental methodologies employed by Mehrabian and others in that era.
It is difficult to reach conclusions from these experiments because of a number of methodological shortcomings. Most studies presented only a few stimulus persons (usually one or two) in an artificial interpersonal situation (an actor demonstrating an emotion). The question of whether the importance of a channel might depend on or interact with the type of information being judged, the characteristics of the observer making the judgement, or the characteristics of the stimulus person being rated has not been considered in most experiments. (Ekman, Friesen, o”Sullivan and Scherer 1980)
In 1977 Archer and Akert said:
…this conclusion rests upon a highly specific experimental base, and the question has never been addressed more generally, using natural sequences of behavior. The external generalizability of most studies in this area has been undermined by three non-naturalistic design features:
(a) the use of context-free nonverbal channels, often in artificial isolation from other communication channels;
(b) the use of posed channel contradictions (eg: positive content intentionally said in a negative voice) which are uncommon outside the laboratory; and
(c) a narrow focus on judgments about emotions, ignoring other aspects of person perception and interpretation.
Krauss, Apple, Morency, Wenzel and Winton (1981) point out that the Mehrabian studies are based on the assumption that the stimulus person is capable of communicating “facial expression and vocal contours that are identified by others as representing specific affects”:
…to the extent that affective information cannot be transmitted via a particular channel, it is unlikely that the channel will contribute importantly to the perception of affect.
Walbott and Scherer (1986) consider the problem of finding natural “stimulus material’ to use in these experiments:
As far as naturalness is concerned, one of the major problems is the persistent use of monologues for the emotional portrayal. In most cases, senders are given a set of standard verbal material and are asked to deliver this text with various emotional meanings to the microphone,as it were. This is, of course, a rather untypical situation, particularly for nonprofessional actors, and it is quite possible that individual conceptions of declamatory style rather than attempts at realistic emotional portrayal are obtained.
Walbott and Scherer designed an experiment to look at the differences in the abilities of ‘senders’ to express particular emotions. They used professional actors and found that there were large differences in their abilities to express particular emotions.
Jones and LeBaron (2002) say in relation to studies of verbal and nonverbal messages in the 1960s:
Many of these early investigations were based on a “channel summation” model…Subsequent research showed that the channel summation model was too simple, especially when it was used to create formulas for the relative contributions of different channels to the overall impressions of observers.
Jeremy Dean, a psychology doctoral student and psychology blogger says:
Perhaps an even stronger criticism of these studies relates to their ‘demand characteristics’. Demand characteristics is a term psychologists use when they are referring to participants in an experiment acting in ways they think the experimenter wants them to act. People generally want to please, they want to go with the flow. So if they can work out what the experimenter is after, they’ll often try and give it to them.
So, when watching videos in these experiments it will be obvious to participants the speeches are acted, not spontaneous. Participants pick up on what the experimenter wants from the social cues provided. Indeed, one study has found that when the purpose of the experiment is actually well-camouflaged from the participants, the dominance of nonverbal communication disappears [emphasis added] (Trimboli & Walker, 1987).
All of these critiques apply to Mehrabian’s 1967 studies in nonverbal communication.
Archer D., & Akert R. M. (1977) Words and everything else: verbal and nonverbal cues in social interpretation. Journal of Personality and Social Psychology, 35, 443-449.Mehrabian A., & Wiener M. (1967) Decoding of inconsistent communications. Journal of Personality and Social Psychology, 6, 109-114.
Ekman, P., Friesen, W. V., O”Sullivan, M., &Scherer, K. (1980). Relative importance of face, body, and speech in judgements of personality and affect. Journal of Personality and Social Psychology, 38, 270-277.Mehrabian A., & Ferris S. R. (1967) Inference of attitudes from nonverbal communication in two channels. Journal of Consulting Psychology, 31, 248-452.
Jones E. J., & LeBaron C. D. (2002) Research on the relationship between verbal and nonverbal communication: emerging integrations. Journal of Communication. Special Issue, 52, 499-521.
Krauss, R. M., Apple, W., Morency, N. Wenzel, C., &Winton, W. (1981) Verbal, vocal and visible factors in judgements of another’s affect. Journal of Personality and Social Psychology, 40, 312-320.
Trimboli, A. & Walker, Michael B. (1987) Nonverbal dominance in the communication of affect: A myth? Journal of Nonverbal Behavior. 11, 180-190.
Wallbott H. G., & Scherer K. R. (1986) Cues and channels in emotion recognition. Journal of Personality and Social Psychology, 51, 690-699.