The Psychology of the Victim of Prejudice and Discrimination 529ethnicities. Lifetime SSE scores enhanced prediction (over
and above generic stress measures) of total HSC symptoms
for older women but not for younger ones.
Buffers for Discrimination-Related Stress
Not all members of oppressed groups will suffer the stress of
discrimination in the same way or to the same extent. The
personality construct of hardiness—a composite of self-
esteem and sense of control—may be one factor that buffers
the stress of experiencing or perceiving discrimination to-
ward oneself and one’s group. Dion et al. (1992) explored the
role of personality-based hardiness in a study of Toronto’s
Chinese community. As they predicted, the relationship of
discrimination to psychological symptoms was markedly
higher among Chinese community respondents who were
low in hardiness than among those high in hardiness. Indeed,
for those scoring high in hardiness, discrimination and re-
ported psychological symptoms were effectively unrelated,
whereas they related reasonably strongly for those low in har-
diness. In addition, alternative interpretations in terms of
differential life stresses or differential exposure to discrimi-
nation in the two hardiness groups were ruled out as rival ex-
planations (see Dion et al., 1992).
Foster and Dion (2001) explored whether the beneficial
relationship of personality-based hardiness to discrimination-
related stress is due to buffering or denial in an experiment in
which women confronted gender discrimination on an exam-
ination. The findings favored a buffering interpretation and
suggested that the buffering was due to the types of attribu-
tions that hardy women made relative to their less hardy
counterparts. Specifically, hardy women made specific, un-
stable attributions rather than global, stable ones; that is, they
tended to see the gender discrimination as a unique and un-
usual occurrence, even though there were no differences
between the hardy and nonhardy women in perceived un-
fairness of the discrimination.
Whereas hardiness may provide a personality-based
buffer and coping dimension, in-group identification has
been hypothesized to be important in predicting reliance on
group-based responses to coping with discrimination and
buffering self-esteem. Branscombe and Ellemers (1998) pro-
posed a rejection-identification modelsuggesting that greater
willingness to make attributions to prejudice among Black
Americans heightens their minority-group identification as
well as hostility toward the dominant White group but has a
negative effect on personal and collective sense of well-
being. Minority-group identification, however, has a buffer-
ing effect in sustaining well-being. Branscombe and her
colleagues tested and supported this model with SEM
procedures. Some alternative theoretical models failed to re-
ceive support.Stereotype ThreatNot only do women and minority members confront preju-
dice and discrimination, but they also must deal with broadly
shared, negative stereotypes about their groups by majority
group members, which can have pernicious and deleterious
effects upon their academic and athletic performance. Black
Americans, for example, confront low expectations in the
realm of academic ability, whereas women in the United
States, Canada, and some other societies are presumed by
consensually shared stereotypes to be inferior in mathematics
compared to men.
Steele (1997; Steele & Aronson, 1995) and his colleagues
contended that negative stereotypes impugning the abilities
of stigmatized group members constitute a powerful situa-
tional threat with two notable consequences. First, in a testing
situation involving an ability where one’s group is negatively
stereotyped, the performance of those members who care
about the ability and doing well on the test can be adversely
affected. Second, chronic experiences of stereotype threat
can lead members of stigmatized groups to disidentify by
denying the importance of the ability for themselves. At the
college level, this disidentification can lead to academic
dropouts among Black Americans and proportionally fewer
women enrolling in math, science, and engineering programs
where mathematical ability is prerequisite.Initial StudiesSteele and Aronson (1995) reported the first set of four ex-
periments documenting the impact of stereotype threat on the
performance of Black American university students, relative
to their White American counterparts, at Stanford University,
an elite U.S. university. These investigators told participants
that difficult and challenging items from the Graduate Record
Examination (GRE) either were diagnostic of their intellec-
tual ability (the diagnostic or stereotype threat condition) or
were a test of problem-solving with no implications for diag-
nosing their intellectual ability (the nondiagnostic or no-
stereotype-threat condition). In all four studies, participants’
previous Scholastic Achievement Test (SAT) scores in high
school were statistically controlled in the analyses by means
of analysis of covariance (ANCOVA) procedures.
The first two studies by Steele and Aronson (1995) demon-
strated that Black American participants in the diagnostic
or stereotype threat condition completed fewer items and