390 Social Influence and Group Dynamics
control in that context, for example, or perhaps to their sense
of personal competence in performing the task. For that mat-
ter, subjects might also be considering their feelings about the
experimenter, pondering the value of the experiment, or
rethinking the value of psychological research in general. In
view of the plethora of likely cognitive elements and the po-
tential for these elements to come in and out of focus in the
stream of thought, the achievement of coherence is anything
but a trivial task. What processes are at work to impart
coherence to this complex and dynamic array of information?
And what psychological dimensions capture the resultant
coherence?
There is hardly a shortage of relevant theories. Several
early models, for example, emphasized processes of causal
attribution (cf. Bem, 1972; Jones & Davis, 1965; Kelley,
1967) that were said to promote personal interpretations
favoring either internal causation (e.g., personal beliefs and
desires) or external causation (most notably, the monetary
incentive). In this view, a large incentive provides a reason-
able and sufficient cause for engaging in the activity, short-
circuiting the need to make inferences about the causal role
of one’s beliefs or desires. A small incentive, on the other
hand, is not perceived as a credible cause for taking the time
and expending the effort to engage in the activity, so one in-
stead invokes relevant beliefs and desires as causal forces
for the behavior. In effect, the counterintuitive influence
of rewards is a testament to their perceived efficacy in caus-
ing people to do things they might not otherwise do. Causal
attribution, of course, is not the only plausible endpoint of
coherence concerns. Other well-documented dimensions
relevant to higher-order integrative understanding include
evaluative consistency (cf. Abelson et al., 1968), explanatory
coherence (cf. Thagard & Kunda, 1998), narrative structure
(cf. Hastie, Penrod, & Pennington, 1983), and level of action
identification (cf. Vallacher & Wegner, 1987). It is hardly sur-
prising, then, that a number of other models have been fash-
ioned and tested in an attempt to explain why rewards
sometimes fail to influence people’s beliefs and desires in
the intended direction (e.g., Csikszentmihalyi, 1990; Deci &
Ryan, 1985; Kruglanski, 1975; Harackiewicz, Abrahams, &
Wageman, 1987; Trope, 1986; Vallacher, 1993).
Taken together, the various models emphasizing inference
and interpretation have a noteworthy advantage over the
standard dissonance reduction model in that they predict re-
verse incentive effects for any action, not just those that are
likely to be viewed in a context-free manner as aversive by
some criterion (e.g., repetitive, boring, pointless, time-
consuming, etc.). Indeed, some of the most interesting re-
search has established conditions under which otherwise
enjoyable or interesting activities can seemingly lose their
intrinsic interest by virtue of their association with material
rewards (cf. Lepper & Greene, 1978). Rewards do not always
have this effect, however, a point that has been incorporated
with varying degrees of success into many of these models.
Still, the theoretical preoccupation with the effects of rewards
has generated an unequivocal lesson: The success or failure
of attempted influence depends on how the attempt engages
the mental machinery of the target. Rewards can be perceived
as bribery and aversive consequences can mobilize resis-
tance, for example, and both can activate concerns about
one’s freedom of action and self-determination. Social influ-
ence does not operate on blank minds, but rather encounters
an active set of interpretative processes that operate accord-
ing to their own dynamics to make sense of incoming infor-
mation (Vallacher, Nowak, Markus, & Strauss, 1998).MANIPULATIONChange in people’s behavior can be imposed from the outside
by the exercise of power, but this approach to influence may
prove effective only as long as the relevant contingencies
(reward, punishment, expertise, information) are in place. To
influence people in a more fundamental sense, it is necessary
to include them as accomplices in the process. A self-sustain-
ing change in behavior requires a resetting of the person’s in-
ternal state—her or her beliefs, preferences, goals, and so
on—in a way that preserves the person’s sense of freedom
and control. Assuming the influence agent has an agenda that
does not coincide with the target’s initial preferences and
concerns, the agent may then find it necessary to employ sub-
tle strategies designed to manipulate the relevant internal
states of the target. Couched in these terms, social influence
boils down to various means by which an agent can obtain
voluntary compliance from targets in response to his or her
requests, offers, or other forms of overture. Research has
identified several compliance-inducing strategies, some of
which rely on basic interpersonal dynamics, others of which
reflect the operation of basic social norms. We discuss spe-
cific manifestations of these general approaches in the fol-
lowing sections.Manipulation Through AffinityCould you pass the broccoli? Will you marry me?Whether
the agenda at issue is mundane or life-altering, requests pro-
vide the primary medium by which people seek compliance
from one another. Requests are a fairly routine feature of
everyday social interaction and have been examined for their
effectiveness under experimental arrangements designed to