Handbook of Psychology, Volume 5, Personality and Social Psychology

Relatively High-Effort Processes of Attitude Change 365

A number of factors have been proposed to account for the

relatively low correlation between attitude change and infor-

mation recall. One argument, for example, has been that

simple learning theories do not account for the fact that dif-

ferent people form different evaluations of information con-

tained in persuasive messages—that is, although one person

may be convinced by an argument, someone else might find it

to be ludicrous (see Petty, Ostrom, et al., 1981). Yet both

might be able to recall the argument. Attitude change has been

found to correspond more closely with information recall

when individuals’ unique assessments of the information

recalled is accounted for (Chattopadhyay & Alba, 1988).

Furthermore, attitudes have been found to correlate more

strongly with learning and recall when people are not evaluat-

ing information on-line at the time of exposure. For example,

when processing is made difficult (e.g., Bargh & Thein, 1985;

Bodenhausen & Lichtenstein, 1987), when people are given

nonevaluative processing goals (e.g., Chartrand & Bargh,

1996; Hastie & Park, 1986; Lichtenstein & Srull, 1987;

Mackie & Asuncion, 1990), or when they are the type of

people who do not spontaneously engage in evaluation (low

in theirneed to evaluate;Tormala & Petty, 2001), the attitude-

recall correlation is higher. Under these conditions, when peo-

ple are asked to report their attitudes, they are forced to first

retrieve what they can from memory, and then base their atti-

tudes on the evaluative implications of this information.

Cognitive Response Approach

Following a series of early findings that attitude change and

information recall were not consistently related, researchers

developed the cognitive response approach to persuasion

(e.g., Brock, 1967; Greenwald, 1968; Petty, Ostrom, et al.,

1981). According to this approach, attitudes and message

argument recall are not always related because persuasion

typically depends largely on an individual’s idiosyncratic

thoughts in response to a persuasive message (i.e., thoughts

about message arguments or other factors such as the tone,

source, or context of the message). According to the cognitive

response view, when exposed to a persuasive message, peo-

ple reflect on the message with respect to their preexisting

knowledge and prior attitude (if they have one), considering

information not contained in the message itself. Three aspects

of people’s cognitive responses have proven important.

Extent of Thought

First, investigators have explored determinants of the extent

of issue-relevant thinking. As noted earlier in our discussion

of the elaboration likelihood model, a number of variables

have been found to affect how much people are motivated

(e.g., personal relevance) or able (e.g., distraction) to think

about a persuasive communication.

Content of Thought

Another aspect of thinking that has garnered considerable re-

search attention is the content of thought. Perhaps the most

important dimension in this regard is the overall valence of

the thinking that occurs. Researchers typically categorize

thoughts as favorable, unfavorable, or neutral, and then com-

pute an overall valence index (e.g., positive thoughts minus

negative thoughts; see Mackie, 1987). According to the cog-

nitive response approach, persuasion can be increased to the

extent that the message elicits mostly favorable thoughts

(e.g.,If we raise taxes, the roads will improve and reduce my

commute time) and few unfavorable thoughts (e.g., If we

raise taxes, I’ll have less money to go out to dinner). On the

other hand, people can resist messages to the extent that they

generate mostly unfavorable thoughts and few favorable

thoughts.

As noted earlier, people can be motivated to generate par-

ticular thoughts by external variables such as their mood

(Petty et al., 1993) or the message source (Chaiken &

Maheswaran, 1994). In a series of studies on resistance to

change, McGuire (1964) demonstrated that counterarguing

of persuasive messages could be increased by giving people

weak attacking messages prior to a stronger attack. The un-

derlying logic of this inoculation approach to resistance is

that a small dose of an attacking virus (i.e., a weak challenge

to the person’s attitude that is refuted) motivates the person to

build up antibodies (i.e., counterarguments) that can be used

against subsequent attacks. Even if an exact defense is not

anticipated, people are presumably motivated by the inocula-

tion treatment (i.e., the initial attack and refutation) to defend

their attitudes by counterarguing opposition messages in the

future. Recent research has shown that having people experi-

ence a successful defense of their attitudes can produce

greater confidence in the initial attitude. This enhanced confi-

dence renders the attitude not only more resistant to change

in the future, but also more predictive of future behavior

(Tormala & Petty, in press).

Confidence in Thoughts

In addition to extent and content of thinking, recent research

has uncovered a third aspect of thought that influences

persuasion—the confidence people have in their own cognitive

mill_ch15.qxd 10/1/02 10:04 AM Page 365