Characteristics of Attitudes 309
direct effect occurs when people perceive the value itself as
relevant to their attitude (Maio & Olson, 1994, 1995). The
latter, direct process may be more likely when the value and
the reasons for its importance have been consciously articu-
lated (Maio & Olson, 1998).
CHARACTERISTICS OF ATTITUDES
Attitudes vary along numerous dimensions, orcharacteris-
tics,that have significant implications for information pro-
cessing, persistence, and behavior. A continuing issue in the
literature on attitude has been the relations among these
dimensions; some researchers have argued that the various
characteristics are distinct and should be treated as indepen-
dent, but other researchers have argued that the characteristics
are interdependent and should be treated as manifestations of
a smaller set of constructs. In this section, we briefly describe
these dimensions and address the controversy surrounding
the interrelations among them.
Extremity
Attitude extremity is the oldest and most basic dimension of
attitudes.Extremityrefers to the extent to which the attitude
deviates from a neutral midpoint—that is, the extent to which
the individual’s evaluation is strongly favorable or strongly
unfavorable. Extreme attitudes (compared to moderate atti-
tudes) are more resistant to influence (e.g., Osgood &
Tannenbaum, 1955), more likely to be projected onto others
(e.g., Allison & Messick, 1988), and more likely to predict
behavior (e.g., Fazio & Zanna, 1978). Attitude theorists have
generally assumed that extreme attitudes develop over time,
often resulting from actions that publicly commit the individ-
ual to his or her position.
Direct-Indirect Experience
Attitudes can be based on direct, personal experience with
the attitude object, or they can be based on indirect infor-
mation from others about the object. For example, students’
attitudes toward chemistry courses can be based on their own
experiences with previous chemistry courses or on things
they have heard from others who have taken chemistry
courses. Researchers have found that attitudes based on di-
rect experience (compared to those based on indirect experi-
ence) are more confidently held (e.g., Fazio & Zanna, 1978),
more stable over time (e.g., Doll & Ajzen, 1992), more resis-
tant to influence (e.g., Wu & Shaffer, 1987), and more likely
to predict behavior (e.g., Fazio & Zanna, 1981). Presum-
ably, these effects of direct experience reflect that we trust
our own senses more than we do others’ reports, which in-
creases confidence in attitudes based on direct experience.
Accessibility
Accessibility refers to the ease of activation (activation
potential) of a construct (Higgins, 1996). Highly accessible
attitudes are evaluations that come to mind quickly and spon-
taneously when the attitude object is encountered. Accessi-
bility depends at least in part on the frequency with which the
attitude has been activated in the recent past. Researchers
have found that highly accessible attitudes (compared to
less accessible attitudes) are more resistant to change (e.g.,
Bassili, 1996), more likely to influence perceptions of atti-
tude-relevant events (e.g., Houston & Fazio, 1989), and more
likely to predict behavior (e.g., Fazio & Williams, 1986).
These effects of accessibility presumably reflect that highly
accessible attitudes are always activated by the attitude ob-
ject, so they exert an impact (compared to low accessibility
attitudes, which are more likely to remain dormant).
Embeddedness
Attitudeembeddedness (also called working knowledge)
refers to the amount of attitude-relevant information, such as
beliefs and experiences, that is linked to the attitude (Scott,
1968; Wood, 1982). The more information that comes to
mind when one encounters the attitude object, the more em-
bedded is the attitude. Highly embedded attitudes are more
resistant to change (e.g., Wood, Rhodes, & Biek, 1995), more
likely to influence perceptions of attitude-relevant stimuli
(e.g., Vallone, Ross, & Lepper, 1985), and more predictive of
behavior (e.g., Kallgren & Wood, 1986) than are attitudes
with low embeddedness. These effects of embeddedness pre-
sumably reflect that attitudes based on a lot of information
are held more confidently and provide the individual with
many bits of knowledge to counteract the potential influence
of new information. Also, embedded attitudes can be more
accessible than are attitudes low in embeddedness (see Wood
et al., 1995).
Evaluative Consistency
Evaluative consistencyrefers to the degree of consistency be-
tween the overall attitude (the evaluation) and one of its compo-
nents (cognitive, affective, or behavioral information).
Evaluative consistency occurs when the favorability of the
overall evaluation of the object is similar to (a) the favorability
implied by the individual’s beliefs about the object (evaluative-
cognitive consistency), (b) the favorability implied by the indi-
vidual’s feelings toward the object (evaluative-affective