Page 175 - Media Effects Advances in Theory and Research

164 PETTY, PRIESTER, BRIÑOL

enduring attitude change (e.g., Janis & King, 1956). When engaged in role
playing (e.g., “generate a message to convince your friend to stop smok-
ing”), people engage in a “biased scanning” of evidence on the issue and
end up persuading themselves because the arguments they generate are
seen as compelling (Greenwald & Albert, 1968). More recently, Tesser and
his colleagues conducted a series of investigations of the effects of merely
thinking about an attitude object without any external information pre-
sented. These studies have shown clearly that with mere thought, peo-
ple’s reactions and impressions to other people, objects, and issues can
become more extreme, in either a positive or negative direction, depend-
ing on the valence of the initial thoughts generated (see Tesser, Martin, &
Mendolia, 1995, for a review).
The cognitive response approach holds that even when external infor-
mation is presented, people’s thoughts or cognitive responses to this
information, rather than learning the information per se, determine the
extent of influence. Most studies of cognitive responses to messages
focus on the valence and the extent of thinking. Valence refers to the
favorableness or unfavorableness of the thoughts with respect to the
message, and extent of thinking refers to the number of thoughts gener-
ated. In general, the more favorable thoughts people have to the mes-
sage, the more persuasion that results, and the more unfavorable
thoughts people have to a message, the less influence (or even change in
a direction opposite to the advocacy) that occurs (Greenwald, 1968; Petty
et al., 1981; Wright, 1973).
In addition to coding thoughts for valence and number, other catego-
rization schemes have be used (e.g., coding for the origin of the thought,
target, self-relevance, and so forth; see Cacioppo & Petty, 1981; Shavitt &
Brock, 1986). One feature of thoughts that has proven to be useful is the
confidence with which people hold their thoughts. That is, two people
can have the same thought about the message (e.g., “the proposed tax
increase should help our schools”), but one person might have consider-
ably more confidence in the thought than another. According to self-
validation theory (Petty & Briñol, 2000; Petty, Briñol, & Tormala, in press),
the relationship between thoughts and attitudes should be greater when
people have confidence rather than doubt in their thoughts, and many of
the traditionally studied source, message, recipient, and channel vari-
ables can influence persuasion by influencing the extent to which people
have confidence in the thoughts they have in response to a persuasive
message. In a series of initial studies conducted to test the basic self-
validation hypothesis, Petty, Briñol, and Tormala (in press) found that
when the thoughts in response to a message were primarily favorable,
increasing confidence in their validity increased persuasion, but increas-
ing doubt about their validity decreased persuasion. When the thoughts