Attitudes and Higher-Order Constructs 307homosexual persons, persons with HIV, African Americans;
Herek, 2000; Reeder & Pryor, 2000; Sears, 1988), consumer
objects (Ennis & Zanna, 2000; Prentice, 1987; Shavitt, 1990),
altruistic behaviors (Maio & Olson, 1995; Snyder, Clary, &
Stukas, 2000), and political issues (Kinder & Sears, 1985;
Lavine & Snyder, 2000).
At least three lines of research support this distinction.
First, some attitude objectselicit attitudes that are associated
primarily with one or the other of these functions. For exam-
ple, Shavitt (1990) found that people’s thoughts about air
conditioners and coffee focus on the utility of the objects,
whereas thoughts about greeting cards and flags tend to focus
on the objects’ capacity to symbolize the self and social
values.
Second, evidence indicates that people are more persuaded
by messages containing arguments that match the instrumental
or symbolic functions of their attitudes than by messages con-
taining arguments that do not match the functions of their atti-
tudes. For example, Shavitt (1990) found that instrumental ads
for instrumental products (e.g., an air conditioner) were more
persuasive than were symbolic ads for instrumental products.
Similarly, Snyder and DeBono (1985) found that low self-
monitors (who typically possess instrumental attitudes) were
more persuaded by instrumental ads for various products (e.g.,
whiskey, cigarettes) than were high self-monitors (whose
attitudes typically fulfill social-adjustive functions). Also,
Prentice (1987) found that participants who attached high
importance to symbolic values (e.g., mature love, self-respect)
and symbolic possessions (e.g., family heirlooms) were less
persuaded by messages that contained instrumental arguments
than by messages that contained symbolic arguments. Presum-
ably, these match effects occurred because people scrutinize
arguments that match the function of their attitude more
carefully than they scrutinize arguments that do not match
the function of their attitude (Petty & Wegener, 1998). As a re-
sult, match effects occur only when the persuasive arguments
are strong, but not when the persuasive arguments are weak
(Petty & Wegener, 1998).
Finally, the distinction between instrumental and sym-
bolic attitudes improves the measurement of attitudes and the
prediction of behavior. Regarding attitude measurement,
many studies have shown that attitudes toward ethnic groups
are related to beliefs about the group members’ values, over
and above beliefs about the group members’ implications for
personal well-being (e.g., Esses et al., 1993; I. Katz & Hass,
1988; see also Schwartz & Struch, 1989). Also, when an atti-
tude serves a symbolic function, personal values enhance the
prediction of attitude-relevant behavior over and above be-
liefs about the positive or negative instrumental attributes of
the behavior and perceptions of group norms (Beck & Ajzen,
1991; Maio & Olson, 1995). Values exhibit weaker relations
to attitudes and behaviors that serve utilitarian functions
(Kristiansen & Zanna, 1988; Maio & Olson, 1994, 1995; cf.
Maio & Olson, 2000).ATTITUDES AND HIGHER-ORDER CONSTRUCTSAttitudes do not, of course, exist in isolation from each other
or from other constructs. For example, people who favor so-
cial assistance payments to the poor may on average possess
positive attitudes toward other social welfare programs such
as national health care and subsidized housing. The positive
attitudes toward all of these programs may in turn arise be-
cause the person attaches high importance to the social value
of helpfulness. Such relations among attitudes and values
may have implications for stability and change in attitudes. In
this section, we consider how attitudes are structurally and
functionally related to each other and how sets of attitudes
may be related to higher-order constructs such as values and
ideologies.Interattitude StructureHeider’s (1958) balance theory is one of the earliest models
of relations between attitudes. This theory examined a situa-
tion in which a person (P) holds a positive or negative atti-
tude toward another person (O), and both people (P and O)
hold a positive or negative attitude toward a particular object
(X). According to Heider, such P-O-X triads are balanced
when P likes O and they hold the same (positive or negative)
attitude toward X, or when P dislikes O and they hold differ-
ent attitudes toward X. A state of imbalanceoccurs when P
likes O and they hold different attitudes toward X, or when P
dislikes O and they hold the same attitude toward X. In other
words, balance exists when a person agrees with someone
whom he or she likes, or when a person disagrees with some-
one whom he or she dislikes.
Heider (1958) predicted that unbalanced states create an
unpleasant tension, which causes people to prefer balanced
states. Subsequent research documented that participants
report more discomfort with hypothetical unbalanced triads
than with hypothetical balanced triads (e.g., Jordan, 1953).
Individuals can convert unbalanced states to balanced states
by using three strategies: Change the attitude toward O or X
(attitude change), change the beliefs about O’s attitude
(belief change), or focus on some aspect of O or X that bal-
ances the triad (differentiation). In cases in which attitude