Social Coordination 399The potential for groupthink exists in any group context,
informal as well as formal, but the most intriguing examples
concern decisions with far-reaching consequences by people
normally considered the best and the brightest. Janis (1982)
analyzed several such situations, including the Bay of Pigs
invasion during the Kennedy administration, the bombing of
Pearl Harbor, and the Vietnam War. Janis identified several
common factors in these instances. In each case, crucial deci-
sions were made in small groups whose members had con-
siderable respect and liking for one another. Positive regard
is certainly preferable to disinterest or disrespect, of course,
but it can also serve to inhibit criticism and close examina-
tion of one another’s suggestions. The group members also
tended to exhibit collective rationalization, systematically
discrediting or ignoring all information contrary to the
prevailing group sentiment. They also tended to develop
strong feelings that their mission (e.g., invading Cuba, imple-
menting a massive troop build-up in South Vietnam) was
moral and that the opposite side was not only immoral but
also stupid. To further cocoon the group, self-appointed
“mind-guards” precluded members from accessing informa-
tion that was inconsistent with the party line. The upshot is
something akin to tunnel vision, in which a single perspective
is seen as the only viable perspective—not because of a
rational assessment of the facts but because of the group’s
irrational espirit de corps.
Group Polarization
The groupthink phenomenon has rather straightforward im-
plications for another phenomenon—group polarization—
that was nonetheless considered surprising when first noted
by researchers (e.g., Stoner, 1961; Wallach, Kogan, & Bem,
1962). The conventional wisdom was that individuals in
groups avoid going out on the proverbial limb, and thus tend
to produce more common or popular opinions and recom-
mendations (cf. F. H. Allport, 1924). It followed from this
that a group decision is usually more conservative than the
average of the decisions generated by group members indi-
vidually. This assumption regarding group decision making
is reflected in critics’ laments about the bland and often timid
recommendations generated by committees in bureaucratic
environments. When faced with making a decision, groups
were assumed to inhibit boldness, subjugating the creative
mind to the lowest common denominator of the group. What
the research began to reveal, however, was quite the opposite
tendency—greater endorsement of risky decisions as a result
of group discussion.
This so-called risky shift is not surprising in light of theory
and research on groupthink. If anything, the sense of
superiority and certainty fostered by an emphasis on cohe-
siveness as opposed to rationality would seem to be a breed-
ing ground for bold decisions that go beyond what an
individual alone would contemplate. The shift toward risky
decisions, however, was observed in contexts that didn’t in-
volve the intellectual and emotional incest displayed by
highly cohesive groups of self-important people. Even
groups of strangers brought together for a one-shot encounter
in a laboratory setting were found to advocate courses of
action with less guarantee of success than the recommen-
dations volunteered by the group members prior to their
discussion. Because this observation flew in the face of con-
ventional wisdom, it cried out for both replication and expla-
nation. During the 1960s, neither proved to be in short supply.
This burgeoning literature demonstrated greater risk-taking
with respect to a wide variety of domains, including bargain-
ing and negotiations (Lamm & Sauer, 1974), gambling be-
havior (Blascovich, Ginsberg, & Howe, 1975; Lamm &
Ochssmann, 1972), and jury decisions (Myers, 1982).
The risky shift was observed, moreover, when the conse-
quences of a group’s decision involved real as well as hypo-
thetical consequences (Wallach et al., 1962). The research
also demonstrated that the risky shift was not limited to rec-
ommendations regarding possible courses of action. Indeed,
group discussion—again, even among strangers—seemed to
intensify all sorts of attitudes, beliefs, values, judgments, and
perceptions (Myers, 1982). Such shifts were observed for
both sexes, in different populations and cultures (e.g., United
States, Canada, England, France, Germany, New Zealand),
and with many kinds of group participants (Pruitt, 1971).
Several explanations for the risky shift achieved currency
(Forsyth, 1990). The diffusion of responsibilityperspective
suggested that people are less averse to risk in groups
because they feel less responsibility for—and hence less anx-
ious about—the potential negative outcomes of risky deci-
sions. The leadershipaccount held that risk takers tend to
emerge as leaders because of their greater confidence,
assertiveness, and involvement in the task, and that their
leadership status makes them more influential in group dis-
cussions.Familiarizationtheory maintained that group dis-
cussion increases members’ familiarity with the issue, which
reduces their uncertainty and increases their willingness to
advocate more risky alternatives. The valueperspective pro-
posed that taking risks is positively valued (in our culture, at
least) and that group members like to be perceived as willing
to take a chance; when group members discover that others in
the group favor riskier alternatives, they change their original
position to agree with the riskiest member.
During this same period, however, some research hinted at
the opposite effect of group discussion—a cautious shift.To