The model can best be represented in this form:

From the model, it immediately becomes apparent why messages directed to other cultures often have little effect. In such cases, the difference between S and Q on the scale AB approaches or may even reach infinity, and regardless of the magnitude of the message M, the communication cannot be effective, since approaches O as Y approaches infinity. What happens to the effectiveness of the same message on two different receivers, Q, and Q2? If Q, is located close to the sender on the hypothetical scale AB, and Q2 is located far away from the sender, the same message will have much greater effect on Q1, even when both have equal predispositions. Being dependent on the magnitude of the predispositions, a message will often have a total negative rather than a positive effect. Since

2

X will take a negative sign and will become negative when the magnitude of P2 is greater than M + P1, resulting in a "boomerang" effect. With an inverse relationship between social-psychological distance and the negative or "boomerang" effect, the size of the angle becomes the number of degrees in the complement angle (90° - 0) and takes a negative

If the message has an effect opposite to that intended by the sender, then the greater the social and psychological distance, the greater will be the Angle of effect. A message is incorporated as a force Rn from the sender's beginning assertions; therefore, in addition to its attitude changing strength, the message has a strong influence on the social-psychological distance. With the receiver in constant readjustment of his position and re-evaluation of the s-p distance from the sender, the early portions of a message may well increase this distance and directly reduce its overall effectiveness, or indeed the opposite may occur.

The introduction of this model raises many interesting questions: What units of measurement must be used if the proposed formulae are to be algebraically manipulated? Do overt and covert effect ranges overlap? At what angle does the effect become behavioral? The significance of the model lies in its apparent usefulness for predicting communication effectiveness. Theorists have developed many valid models, but the concept of an Angle of Effect is new to communication theory. Obviously the concept is useful in reaching a total Theory of Communication Effectiveness, which can only come after additional discussion.

A model that attempts to relate so many variables in a rigid, mathematical way cannot be completely free of weaknesses and errors; therefore, criticisms are justified. However, with an "anchor point" established, a more refined mathematical model can now evolve, where previously the prospects were limited at best.

NOTES

1. Shannon, E. C. and W. Weaver. The Mathematical Theory of Communication. Urbana: University of Illinois Press, 1949.

2. Harrah, David. "A Model of Communication." Philosophy of Science 23:333–42, 1956. 3. Cherry, Colin. On Human Communication. New York: Wiley, 1957.

Osgood, C. E., C. J. Sauci, and P. H. Tannenbaum. The Measurement of Meaning. Urbana: University of Illinois Press, 1957.

5. Stephenson, William. The Play Theory of Mass Communication. Chicago: University of Chicago Press, 1967.

6. Abelson, Robert P. "Lectures on Computer Simulation." Mathematics and Social Sciences. The Hague: Mounton, 1965, p. 443–79.

7. Heider, Fritz. "Attitudes and Cognitive Organization." The Journal of Psychology 21:107-12, 1946.

8. Newcomb, Theodore. “An Approach to the Study of Communicative Arts.” Psychological Review 60:393-404, 1953.

9. Festinger, Leon. A Theory of Cognitive Dissonance. Stanford, Calif.: Stanford University Press, 1957.

THE MEASUREMENT OF SPEAKER CREDIBILITY*

BY R. BARRY FULTON

The author contends that the credibility dimensions of agreeableness, conscientiousness, and culture are positively and significantly related to an independent measure of the attractiveness of a public speaker who is judged only by those overt cues which the listener perceives during the speech act.

Trustworthiness and expertness, the two factors of credibility suggested in 1953 by Hovland, Janis, and Kelley,1 have been identified more recently in factor analytic studies reported by McCroskey2 and by Bowers and Phillips. Nevertheless, there are those who insist that this representation of credibility doesn't capture the full complexity of the phenomenon. For example, in a study conducted by Schweitzer and Ginsburg, 28 different factors emerged in the rotated factor matrix for the low-credibility condition, accounting for 74 percent of the variance; under the high-credibility condition, 27 factors accounted for only 60 percent of the variance. Although Schweitzer and Ginsburg reasonably conclude from this interpretation of their data that the factors of trustworthiness and expertness do not adequately represent the complexity of the concept "credibility," their analysis does little to explicate the underlying relationship.

Lying between the reported extremes are a number of factor analytic studies which appear to systematically represent a greater portion of the complexity involved in a judgment of credibility. Berlo and Lemert reported a factor analysis study in which three dimensions were found: trustworthiness, competence, and dynamism3. In addition to these three factors, Whitehead has found a fourth major factor: objectivity. Norman and his associates have identified five factors in a series of studies which have served as the basis for the research reported here. These five factors along with the scale items used in their measurement are compared in Table 1 to factors found in two of the other studies mentioned above.

The genesis of the dimensions which emerge from any factor analytic study is not an unimportant consideration in their acceptance, for no factor analysis can extract factors which were not represented in the original scale items. The Norman scale items had their origin in Allport and Odbert's search for personality traits in a standard dictionary; in 1936 they reported finding some 18,000 terms. From the 4,504 terms which

*Excerpts from "The Measurement of Speaker Credibility," The Journal of Communication, XX (September 1970), pp. 270-279. Reprinted with the permission of The Journal of Communication, copyright holder.

Allport described as the "real" traits of personality, Cattell selected 171 terms to represent synonym groups. By means of cluster analysis, Cattell further reduced this number of 36 bipolar pairs from which he has reported finding 12 stable personality factors. Analyses by other researchers have revealed as few as five recurrent factors.7

Norman used four scales from each of the five dimensions in a number of studies in which subjects nominated one-third of the members of some peer group on Pole "A" and one-third on Pole "B" of each scale. A varimax rotation of a factor analysis of the data revealed, as hypothesized, five orthogonal personality factors. In a later study by Passini and Norman, the same five factors emerged with subjects whose contact was limited to being together for less than 15 minutes without opportunity for verbal communications.".

Table 1

8

A follow-up analysis by Norman and Goldberg revealed that, even with this minimal contact among subjects, there was some degree of ratee relevance in the choice of scale items. Their explanation for this relevance is based on what Cronbach has termed an “implicit personality theory": If, for example, it were generally held within the implicit personality theories of these raters that persons who are irresponsible and undependable are also careless, unscrupulous, fickle, and slovenly and if the shared stereotype of the person who is careless and slovenly included aspects of dress and grooming, then a ratee who gave such an appearance in this setting was, in the absence of more specifically relevant information, apt to be rated as possessing all traits in the set.io

If the five factors of personality reported by Norman and associates are accepted, it follows that these same factors might operate as underlying dimensions of credibility during the speech act; this proposition is explored in this article.

EXPERIMENT I

Through an analysis of the literature on interpersonal attraction an attempt was made to identify those credibility cues which might serve as indicators to a respondent of some set of underlying credibility dimensions. 11 The suggested pairing follows:

Two speakers, undergraduates chosen for their speaking and acting abilities, were trained to represent opposite poles of the five dimensions by speaking and behaving in such a way as to provide for their listeners the cues suggested above. Each of 246 undergraduate subjects heard one of the two speakers deliver (in person) an eight-minute persuasive speech prepared by the experimenter.

The abbreviated scale labels for the 20 items used in the Norman studies were used as labels for opposite ends of 20 semantic differentialtype scales. By assigning values of one through seven for each of the responses by subjects and summing for each of the five dimensions, responses to the two speakers could be compared. Subjects also responded to the following two scales: "I feel that I would probably like this person very much" vs. "I feel that I would probably dislike this person very much," and "I believe that I would very much enjoy working with this person in an experiment" vs. "I believe that I would very much dislike working with this person in an experiment."12 These poles were placed at opposite ends of semantic differential-type scales; responses were scored by assigning values of one through seven and summing as a measure of attraction.