Page 119 - The Social Animal

Mass Communication, Propaganda, and Persuasion 101

theirs. But the more discrepant such a communicator’s position is
from those of the audience, the more the audience might begin to
question his or her wisdom, intelligence, and sanity. The more they
question his or her wisdom, intelligence, and sanity, the less likely
they are to be influenced.
Let’s return to our example involving physical exercise: Imagine
a 73-year-old man, with the body of a man half his age, who had just
won the Boston Marathon. If he told me that a good way to stay in
condition and live a long, healthy life was to exercise vigorously for at
least 2 hours every day, I would believe him. Boy, would I believe him!
He would get much more exercise out of me than if he suggested I
should exercise for only 10 minutes a day. But suppose a person some-
what less credible, such as a high-school track coach, were delivering
the communication. If he suggested I exercise 10 minutes a day, his
suggestion would be within my own latitude of acceptance, and he
might influence my opinion and behavior. But if he advised me to em-
bark on a program of vigorous exercise requiring 2 hours every day, I
would be inclined to write him off as a quack, a health freak, a mono-
maniac—and I could comfortably continue being indolent. Thus, I
would agree with Hovland, Harvey, and Sherif: People will consider
an extremely discrepant communication to be outside their latitude of
acceptance—but only if the communicator is not highly credible.
Armed with these speculations, my students and I scrutinized
the existing experiments on this issue, paying special attention to
the ways in which the communicator was described. Lo and behold,
we discovered that each of the experiments showing a direct linear
relation between discrepancy and opinion change happened to de-
scribe the source of the communication as more credible than did
those whose results showed a curvilinear relation. This confirmed
our speculations about the role of credibility. But we didn’t stop
there: We constructed an experiment in which we systematically in-
vestigated the size of the discrepancy and the credibility of the com-
municator in one research design. 57 In this experiment, college
women were asked to read several stanzas from obscure modern po-
etry and to rank them in terms of how good they were. Then each
woman was given an essay to read purporting to be a criticism of
modern poetry that specifically mentioned a stanza she had rated as
poor. For some participants, the essayist described this particular
stanza in glowing terms; this created a large discrepancy between
the opinion of the communicator and the opinion voiced by the