Handbook of Psychology, Volume 5, Personality and Social Psychology

Relatively High-Effort Processes of Attitude Change 369

Results that might be inconsistent with the original disso-

nance formulation include the finding that proattitudinal be-

haviors can elicit dissonance when aversive consequences

result (e.g., Scher & Cooper, 1989) and that inconsistency

might fail to arouse dissonance if the individual has low

consistency needs (e.g., Cialdini, Trost, & Newsom, 1995;

Snyder & Tanke, 1976). The original dissonance formulation

might account for such results by considering the importance

of the cognitions creating the inconsistency (Harmon-Jones

et al., 1996). Important cognitions should have greater weight

in determining dissonance magnitude than should unimpor-

tant cognitions. Therefore, if aversive (and inconsistent) con-

sequences are highly important, they could override the

effect of the proattitudinal act itself (Harmon-Jones et al.,

1996). Similarly, individual differences in consistency needs

might be accounted for by differential importance weights

across individuals.

As should be apparent, the original dissonance formula-

tion and its alternatives appear to be quite flexible in ac-

counting for the wide variety of effects one’s behavior can

have on one’s attitude. The flexibility associated with these

different ways of interpreting dissonance findings affords

greater explanatory breadth, but it comes with a cost. Specif-

ically, this flexibility makes it difficult to accurately predict

when any given individual will experience dissonance—a

criticism that has often been leveled at dissonance theory

(Aronson, 1992, 1999).

Nondissonance Alternatives

In addition to the dissonance modifications described previ-

ously, two nondissonance alternatives have been proposed to

account for the findings of dissonance researchers. One such

alternative is self-perception theory. As described earlier,

self-perception theory (Bem, 1965) holds that individuals

often infer their attitudes from their own behavior. Self-

perception theory was a formidable opponent to the disso-

nance view because it was able to account for many of the

results attributed to dissonance mechanisms (Greenwald,

1975). It later became apparent that self-perception was a dif-

ferent phenomenon that functioned in different settings and

was not simply an alternative explanation for cognitive dis-

sonance (e.g., Beauvois, Bungert, & Mariette, 1995). For in-

stance, in contrast to dissonance processes, self-perception

processes appear to operate when one’s behavior falls in

one’s latitude of acceptance and thus elicits little aversive

arousal (Fazio et al., 1977). In addition, whereas dissonance

reduction has been proposed to require considerable cogni-

tive effort (Festinger, 1957), self-perception processes appear

to involve simpler attributional decisions that operate under

relatively low-effort circumstances (see attribution section in

this chapter).

A second alternative mechanism is impression manage-

ment. Proponents of this view believe that the attitude change

observed in dissonance experiments results not from aversive

arousal associated with cognitive inconsistency, but instead

from the desire to appearconsistent to others (e.g., Tedeschi

et al., 1971). Although impression management is a motiva-

tional variable that can affect attitude reports, it cannot

account for all dissonance phenomena. For example, disso-

nance-based attitude change can also occur in situations in

which attitude reports are private and anonymous and should

therefore arouse no impression management concerns (e.g.,

Baumeister & Tice, 1984; Hoyt et al., 1972).

Combinatory Approaches

Combinatory approaches emphasize the different ways in

which individuals assign value to pieces of information and

integrate them into a structure of beliefs and attitudes. These

models differ in their emphasis on the types of information

individuals consider, as well as the means by which the in-

formation is integrated. Three different types of models have

received the most research attention.

Probabilogical Model

The probabilogical model (McGuire, 1960, 1981; Wyer,

1970, 1974) suggests that beliefs are represented in memory

in a network of syllogistic structures that have both horizon-

tal and vertical dimensions. Each syllogism contains two

premises that logically imply a conclusion.

Consider the following syllogism:

Premise 1: Drinking Brand X beer makes one popular.

Premise 2:Being popular is desirable.

Conclusion:Drinking Brand X beer is desirable.

The conclusion of this syllogism relies on Premises 1 and 2.

The vertical structure of the network is formed by related

syllogisms. For example, Premises 1 and 2 could each be

the conclusion of other syllogisms, and the premises that

lead to these conclusions could each be the conclusions of

yet other syllogisms. The vertical structure of the network

has important implications for attitude change because

changing beliefs at one point in the vertical structure can

lead to logical change in other elements within the vertical

structure.

The network of syllogisms also has a horizontal structure.

The horizontal dimension incorporates other syllogisms that

mill_ch15.qxd 10/1/02 10:04 AM Page 369