The Psychology of Bigotry 523toward Ethiopian immigrants only. By exploring women’s
attitudes toward men, Stephan, Stephan, Demitrakis, Yamada,
and Clason (2000) showed that ITT is useful for target groups
other than immigrants and for attitudes of members of subor-
dinate or oppressed groups as well as dominant ones. Stephan
et al. found that for women, symbolic threat, intergroup anxi-
ety, and negative contact were predictors of negative attitudes
toward men; however, contrary to prediction, realistic threats
failed to emerge as a reliable predictor.
Because the preceding research on ITT is correlational in
nature, it does not and cannot conclusively document that the
causal sequence goes only from perceived threat to prejudice
and not the other way or in both directions. However, Maio,
Esses, and Bell (1994) experimentally manipulated perceived
realistic and symbolic threats and found increased prejudice
toward immigrants, thus validating at least the proposed
causal sequence of threats heightening prejudice that lies at
the core of ITT, at least for that target group.
Assessment of different types of threat has potential utility
for those interested in improving intergroup relations. In stud-
ies in which attitudes toward more than one target group are
assessed from an ITT perspective, one may explore which tar-
get group may deserve more attention in ameliorative efforts
(for an example, see Schwarzwald & Tur-Kaspa, 1997). Simi-
larly, in the aforementioned studies of ITT, some but usually
not all types of threat emerged as significant predictors, sug-
gesting where change attempts might profitably focus. For ex-
ample, in attitudes of U.S. university students toward Mexican
immigrants, intergroup anxiety has emerged as the most reli-
able predictor (Stephan & Stephan, 1996; Ybarra & Stephan,
1994). ITT is, therefore, especially useful for those interested
in reducing as well as understanding prejudice (see the chap-
ter on reducing prejudice by Dovidio in this volume).
The Multicomponent Approach to Intergroup Attitudes
The multicomponent approach to intergroup attitudes
(MAIA), proposed by Esses, Haddock, and Zanna (1993; see
also Haddock et al., 1993; Zanna, 1994), is the final example
of an integrative theoretical approach to be considered. Al-
though MAIA was derived independently from ITT, the two
perspectives clearly resemble one another in their mutual em-
phases on symbolic beliefs and emotional reactions to out-
groups as important predictors of prejudice and also in a
shared interest in determining if and when stereotypes of
out-groups will relate to prejudice toward them.
MAIA presumes that an intergroup attitude, like the atti-
tude concept in general, has several components (viz., evalu-
ations, cognitions, and affect). An attitude toward a social
group is an overall evaluation, either positive or negative.
Esses and her colleagues use the feeling thermometer as
their preferred measure of an intergroup attitude as a global
evaluation. The goal of MAIA is to predict prejudice and
intergroup attitudes, relying mainly on cognitive and affec-
tive factors as the key predictors. Stereotypes and symbolic
beliefs constitute MAIA’s cognitive factors. Stereotypes are
beliefs about the characteristics of groups, both those shared
with other perceivers (i.e., a consensual stereotype) and those
unique to a given perceiver (a personal stereotype), with per-
sonal stereotypes assumed by MAIA researchers to be more
useful to predict prejudice than consensual stereotypes. Sym-
bolic beliefs are a person’s ideas as to how a social group hin-
ders or facilitates her or his core values and norms. In the
MAIA the affective component consists of the specific feel-
ings and emotions evoked by a social group (see also Esses,
Haddock, & Zanna, 1994). To assess personal stereotypes,
symbolic beliefs, and emotions toward one or more groups,
MAIA researchers typically employ open-ended measures in
which respondents first list their thoughts and feelings toward
a specified group and then go over their lists in order to rate
the valence of each entry and the percentage of the social
group believed to be characterized by it.
In their initial studies MAIA researchers explored attitudes
among English-Canadian university respondents in Ontario
toward several social groups: English-Canadians, French-
Canadians, Native Indians, Pakistanis, and homosexuals (see
Esses et al., 1993). The MAIA model successfully predicted
attitudes toward the out-groups. Attitudes toward Pakistanis
and homosexuals were best predicted by symbolic beliefs, a
component of intergroup attitudes believed to be impor-
tant for assessing prejudice toward disliked or unfavorable
groups. By contrast, out-groups more favorably regarded by
the English-Canadian respondents (viz., French-Canadians
and Native Indians) were best predicted by emotions.
Esses et al. (1993) also showed that RWA is an important
moderator of out-group attitudes and their subcomponents.
English-Canadian respondents scoring high on RWA had
consistently more negative attitudes toward all four out-
groups, especially the disfavored groups, and symbolic be-
liefs were their single best predictor of attitudes toward
different groups, including French-Canadians. By contrast,
emotions best predicted the more favorable out-group atti-
tudes of those scoring low on RWA.
These conclusions, particularly regarding homosexuals as
a target group, were further reinforced in two studies by
Haddock et al. (1993). Their first study confirmed the more
negative attitude of high RWA scorers toward homosexuals
and the importance of symbolic beliefs in predicting preju-
dice toward homosexuals. Their second study replicated and
extended these findings by showing that for those scoring