300 Attitudes in Social Behavior
groups of people (e.g., ethnic groups), controversial issues
(e.g., legalized abortion), and concrete objects (e.g., pizza).
In fact, the potentially unlimited range of attitude objects
sometimes causes confusion about the relations between atti-
tudes and other social psychological constructs. For example,
there is conceptual overlap between attitudes and values,
which are abstract ideals that people consider to be important
guiding principles in their lives (e.g., freedom; Rokeach,
1973; Schwartz, 1992). The importancecomponent of values
makes them distinct from attitudes (Feather, 1995; Maio &
Olson, 1998), because positive attitudes do not imply that the
targets are important guiding principles in life.
One fundamental attribute of attitudes is that they are
subjective—that is, they reflect how a person sees an object
and not necessarily how the object actually exists. Conse-
quently, attitudes should be considered a part of the subjective
self, which is the stream of thoughts, feelings, and actions that
govern how someone lives (James, 1890).
STRUCTURE OF ATTITUDES
The relevance of attitudes to the subjective self suggests that
attitudes may be connected to thoughts, feelings, and actions.
This hypothesis raises the question of how attitudes are struc-
tured in the human mind. Understanding the mental structure
of attitudes is potentially as important to attitude research as
identifying the structure of DNA was to biological research.
Uncovering the internal structure of attitudes can facilitate
our understanding of how attitudes form, strengthen, and
change.
In this section, we describe four well-established perspec-
tives on attitude structure and their implications for attitude
measurement. Two perspectives focus on the content of atti-
tudes. These perspectives examine how attitudes may express
more elemental psychological constructs, such as beliefs and
emotions. The other two perspectives examine the dimension-
ality of attitudes—that is, these theories consider precisely
how attitudes summarize positivity and negativity toward the
attitude object. After reviewing attitude content and dimen-
sionality, we describe some alternative attitude measures and
the concept of implicit attitudes.
Attitude Content
Two perspectives have dominated research on the content
of attitudes: the three-component model and the expectancy-
value model. For both models, we describe their chief
characteristics, implications for attitude measurement, and
supporting evidence.
Three-Component ModelGuiding Assumptions The three-component model hypoth-
esizes that attitudes express people’s beliefs, feelings, and past
behaviors regarding the attitude object (Zanna & Rempel,
1988). For example, people might form a positive attitude
toward eating spaghetti because spaghetti tastes good (affec-
tive component) and they believe that spaghetti is nutritious
(cognitive component). Moreover, through the process of
self-perception (Bem, 1972; Olson, 1992), people may decide
that they like spaghetti because they can recall eating it often
(behavioral component). Thus, this model suggests that peo-
ple have positive attitudes toward an object when their beliefs,
feelings, and behaviors express favorability toward an object,
whereas people have negative attitudes toward an object when
their beliefs, feelings, and behaviors express unfavorability to-
ward the object.
We think it important to note, however, that the three-com-
ponent view also regards attitudes as being distinct from the
beliefs, feelings, and behaviors that influence them—
following the adage that the whole is not simply the sum of its
parts (Eagly & Chaiken, 1998; Zanna & Rempel, 1988). The
attitude per se is a net evaluation of an attitude object; people
can experience this evaluation when they encounter the atti-
tude object, and they can store their attitude as a statement in
memory (e.g.,ice cream is good). Similarly, the attitude object
can evoke the component beliefs, feelings, and behaviors, and
the components can be subjectively represented in memory.
Nonetheless, these components are more circumscribed in
their focus. Beliefs are perceived associations between an ob-
ject and its attributes, which may be evaluative in nature (e.g.,
ice cream is fattening); feelings are experiences of pleasant or
unpleasant mood, which may be evoked by particular objects
(e.g.,ice cream makes me relaxed); and behaviors are overt
acts that involve approaching or avoiding the object in some
way (e.g.,I buy ice cream often).Measurement The three-component model indicates
that it is possible to obtain measures of overall attitudes with-
out attempting to assess attitude-relevant beliefs, feelings,
and behaviors. For example, attitudes are frequently mea-
sured using an attitude thermometer, which asks participants
to use a thermometer-like scale to indicate the extent
to which they feel favorable versus unfavorable toward the
attitude object (Campbell, 1971; Haddock, Zanna, & Esses,
1993; Maio, Bell, & Esses, 1996; Wolsko, Park, Judd, &
Wittenbrink, 2000). Using this scale, people can indicate a
general evaluation, which may be derived from attitude-
relevant beliefs, feelings, behaviors, or some combination of
all three.