Page 117 - The Social Animal

Mass Communication, Propaganda, and Persuasion 99

should remain dry, others felt strongly that it should go wet, and the
rest took a moderate position. The participants were divided into
groups of people reflecting all three positions. The members of each
group were presented with communications supporting one of the
three opinions, so that there were some participants in each group
who found the communication close to their own position, some
who found it moderately discrepant from their own position, and
some who found it extremely discrepant from their own position.
Specifically, some groups were presented with a “wet” message,
which argued for the unlimited and unrestricted sale of liquor; some
groups were presented with a “dry” message, which argued for com-
plete prohibition; and some groups were presented with a moderately
“wet” message, which argued to allow some drinking but with cer-
tain controls and restrictions. The greatest opinion changes occurred
when there was a moderate discrepancy between the actual message
and the opinions of individual members of the groups.
For a scientist, this is an exciting state of affairs. When a sub-
stantial number of research findings point in one direction and a sim-
ilarly substantial number of research findings point in a different
direction, it doesn’t necessarily mean someone has to be wrong;
rather, it suggests there is a significant factor that hasn’t been ac-
counted for—and this is indeed exciting, for it gives the scientist an
opportunity to play detective. I beg the reader’s indulgence here, for
I would like to dwell on this issue—not only for its substantive value,
but also because it provides us with an opportunity to analyze one of
the more adventurous aspects of social psychology as a science. Ba-
sically, there are two ways of proceeding with this game of detective.
We can begin by assembling all the experiments that show one re-
sult and all those that show the other result and (imaginary magni-
fying glass in hand) painstakingly scrutinize them, looking for the
one factor common to the experiments in group A and lacking in
group B; then we can try to determine, conceptually, why this factor
should make a difference. Or, conversely, we can begin by speculat-
ing conceptually about what factor or factors might make a differ-
ence; then we can glance through the existing literature, with this
conceptual lantern in hand, to see if those in group A differ from
those in group B on this dimension.
As a scientist, my personal preference is for the second mode.
Accordingly, with two of my students—Judith Turner and Merrill