UNDERSTANDING THE ASSOCIATION BETWEEN PERCEIVED DISCRIMINATION,
RACIAL IDENTITY AND WELL-BEING
By
Stevie C.Y. Yap

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF ARTS
Psychology
2011

ABSTRACT
UNDERSTANDING THE ASSOCIATION BETWEEN PERCEIVED DISCRIMINATION,
RACIAL IDENTITY AND WELL-BEING
By
Stevie C. Y. Yap

An extensive body of literature suggests that perceived racial discrimination is related to
negative well-being outcomes. However, the role that racial identity plays in this relationship is
unclear. Past research has suggested three roles for racial identity in the relationship between
perceived racial discrimination and well-being: racial identification mediates the relationship
between perceived discrimination and well-being (Rejection Identification Model), racial
identification moderates the relationship between perceived discrimination and well-being
(Buffering Hypothesis), and racial identification leads to perceptions of racial discrimination,
which in turn relate to higher well-being (Discounting Hypothesis). Given the mixed evidence
for these theorized relationships, the aim of this study was to test and compare these three
theoretical models using a sample of 3,613 children of immigrants in the Children of Immigrants
Longitudinal Study (Portes & Rumbaut, 2006). Overall, the results of this study lent the greatest
support for the Rejection Identification Model and some support for the Discounting Hypothesis,
but these results must be interpreted with caution given the relatively small effect sizes garnered
across these analyses.

ACKNOWLEDGEMENTS

I would like to thank my committee members, Isis Settles, Richard Lucas and NiCole Buchanan,
for their assistance in the completion of this thesis.

iii

TABLE OF CONTENTS

LIST OF TABLES

vii

LIST OF FIGURES

viii

INTRODUCTION
Racial Discrimination
Identity
Racial Identity
Racial Identity Development
Theoretical Models
Rejection Identification Model
The Buffering Hypothesis
The “Discounting” Perspective
The Proposed Study
Hypotheses

2
5
7
8
12
13
14
17
20
22

METHOD
Participants and Procedures
Measures
Perceived Discrimination
Self Esteem
Depression
Racial Identity
Job Satisfaction
Monthly Income
Educational Attainment

24
24
25
25
26
26
26
27
27
27

RESULTS
Overview of Analysis
Cross-Sectional Analyses
Cross-Sectional Analyses: Rejection Identification Model
Self-Esteem (Time 2)
Depression (Time 2)
Monthly Income (Time 3)
Job Satisfaction (Time 3)
Years of Education (Time 3)
Cross-Sectional Analyses: Buffering Hypothesis
Self-Esteem (Time 2)
Depression (Time 2)
Monthly Income (Time 3)
Job Satisfaction (Time 3)
Years of Education (Time 3)

27
27
28
28
28
29
29
30
30
30
31
31
31
32
32

iv

Cross-Sectional Analyses: Rejection Discounting Hypothesis
Self-Esteem (Time 2)
Depression (Time 2)
Monthly Income (Time 3)
Job Satisfaction (Time 3)
Years of Education (Time 3)
Summary
Cross-Wave Analyses
Cross-Wave Analyses: Rejection Identification Model
Cross-Wave Analyses: Buffering Hypothesis
Self Esteem
Depression
Educational and Career Outcomes
Cross-Wave Analyses: Discounting Hypothesis

32
33
33
33
34
34
35
35
36
39
39
40
40
40

DISCUSSION
41
Cross-Sectional Results
Longitudinal Results
Limitations and Future Directions
Conclusion

41
45
49
52

APPENDIX

53

REFERENCES

70

v

LIST OF TABLES

Table 1. Demographic Characteristics of the Full (Wave 1) Sample

58

Table 2. Variables Measured by Wave of Data Collection

59

Table 3. Descriptive Statistics and Bivariate Correlations

60

Table 4. Cross Sectional Results: Rejection Identification Model

61

Table 5. Cross Sectional Results: Buffering Hypothesis

62

Table 6. Cross Sectional Results: Discounting Hypothesis

63

Table 7. Cross Wave Results: Rejection Identification Model

64

Table 8. Cross Wave Results: Buffering Hypothesis

65

Table 9. Cross Wave Results: Discounting Hypothesis

66

Table 10. Results Summary

67

Table 11. Descriptive Statistics and Bivariate Correlations (Hispanic Only)

68

Table 12. Descriptive Statistics and Bivariate Correlations (Asians Only)

69

Table 13. Descriptive Statistics and Bivariate Correlations (Blacks Only)

70

vi

LIST OF FIGURES

Figure 1. The Rejection Identification Model

55

Figure 2. The Buffering Hypothesis

56

Figure 3. The Discounting Hypothesis

57

vii

Introduction
Discrimination of people on the basis of group membership is unfortunately a large part
of U.S. history and has been common throughout modern society, with recent research
suggesting that major lifetime discrimination is experienced by as much as a third of the total
U.S. population (Kessler, Michelson, & Williams, 1999). Indeed, racial discrimination, which is
defined as the unfavorable treatment of a person or a group of people because of their race and
characteristics associated with their race, has been seen as a particularly problematic form of
discrimination and a large body of past research has shown that the experience of discrimination
is related to many negative physical and psychological outcomes (Branscombe, Schimitt, &
Harvey, 1999; Sellers, Copeland-Linder, Martin, & Lewis, 2006). Although it is clear that there is
a relationship between racial discrimination and negative well-being outcomes, the role that
racial identity plays in this relationship is unclear. Recent research has posited several ways of
modeling the role of racial identity on the relationship between perceived discrimination and
well-being (e.g., Branscombe et al., 1999; Yip, Gee, & Takeuchi, 2008) and there is empirical
evidence for each of these models and hypotheses. Given the equivocal nature of these research
findings, we aim to examine and critically compare the fit of three theoretical models that have
been used to describe the observed relationship between perceived racial discrimination and
adjustment, and the particular role that identity plays in this relationship. Each of these models
offers different explanations and frameworks for understanding the relationship between
perceived discrimination, identity, and well-being, and our goal is to identify the best fitting
model in order to garner a better understanding of the interrelationship between these important
psychological constructs. In particular, this proposed study will examine the nature of the
associations between perceived racial discrimination, racial identity, psychological adjustment

1

(i.e., self esteem, depression), career, and educational attainment. This question will be explored
cross-sectionally and longitudinally among a sample of 3,613 children of immigrants in the
Children of Immigrants Longitudinal Study (CILS; Portes & Rumbaut, 2006).
Racial Discrimination
Discrimination is defined as unjustified negative behavior toward a particular social
group and the members of that group. Indeed, past research has suggested that discrimination
typically stems from prejudiced attitudes and stereotypes towards other groups (Dovidio,
Kawakami, & Gaertner, 2002). Racial discrimination is experienced by many members of racial
and ethnic minority groups and is associated with numerous negative outcomes. There are
several underlying reasons why some individuals racially discriminate against members of other
groups. Some individuals are more likely to engage in racial discrimination because they simply
hold more bias and prejudice against minority racial groups. For example, an experiment by
Ziegert and Hanges (2005) indicated that when given justification for discrimination by an
authority in an organizational setting, individuals with higher levels of unconscious racial bias
rated Black job applicants more negatively compared to those low in unconscious racial bias.
Further, a study by Norton, Sommers, Vandello, and Darley (2006) showed evidence that
individuals will provide non-racial justifications for race-related selection decisions, and that this
is especially true for individuals high on racial prejudice.
Group based prejudice and discrimination have also been explained by Social Dominance
Theory (Sidanius & Pratto, 1999). This theory suggests that some individuals may have more of
a social dominance orientation, which is an individual’s tendency to support group-based
hierarchies, including the support of higher status groups and their domination of lower status
groups (Sidanius & Pratto, 1999). Thus, individuals high on social dominance orientation are

2

thought to have a set of beliefs or values that justify inequality between groups, which in turn are
associated with a greater likelihood to discriminate against subordinate groups. Indeed, research
by Parkins, Fishbein, and Ritchey (2006) demonstrated this by finding that undergraduate
students who were higher on self-rated social dominance orientation reported a greater likelihood
to discriminate.
Past research has shown that there are clear relationships between perceived racial
discrimination and psychological well-being and adjustment. For example, discrimination has
been associated with lower self esteem (Jasinskaja-Lahti & Liebkind, 2001; Wong, Eccles, &
Sameroff, 2003), decreased life satisfaction (Liebkind & Jasinskaja, 2000), lower perceived selfcompetence and psychological resiliency (Wong et al., 2003). Discriminatory experiences have
been associated with increased distress, traumatic stress (e.g., Branscombe, Schmitt, & Harvey,
1999; Broman, Mavaddat, & Hsu, 2000; Forman, 2003; Kessler et al., 1999), anxiety, depression
and obsessive-compulsive disorder (Kessler et al., 1999; Klonoff, Landrine, & Ullman, 1999;
Sellers et al., 2006). Perceived racial discrimination has also been associated with negative
physical health outcomes, such as increased mortality rates and elevated incidences of several
chronic health conditions, including high blood pressure, heart disease, cancer, and diabetes
(Barnes, Mendes de Leon, Lewis, Bienias, Wilson, & Evans, 2008; Guyll, Matthews, &
Bromberger, 2001; Harrell, 2000; Williams, Neighbors, & Jackson, 2003). Further, experience of
discrimination is negatively associated with several achievement outcomes. For instance, racial
discrimination is associated with decreased motivation to achieve goals and increased emotional
discomfort, anger, and behavioral problems among adolescents and college students (Sellers et
al., 2003; Swim, Hyers, Cohen, Fitzgerald, & Bylsma, 2003; Wong et al., 2003).
One of the primary ways we can understand the association between perceived racial

3

discrimination and many of the negative mental and physical health outcomes discussed here is
that perceived discrimination has been shown to be a stressful experience for many individuals.
In particular, the type of stress that results from the experience of discrimination is largely
considered unpredictable and uncontrollable, characteristics that make discrimination related
stress especially detrimental to health (Pascoe & Smart Richman, 2009; Williams & Mohammed,
2009). Some lines of research have conceptualized the negative physical health outcomes
associated with perceived discrimination as a form of stress response (Clark, Anderson, Clark, &
Williams, 1999; Dion, Dion & Pak, 1992). According to these models, discrimination creates
stress in the individual, which results in physiological and psychological reactivity that are
associated with a stress response (e.g., increased heart rate, increased blood pressure; Anderson,
1989; Armstead, Lawler, Gordon, Cross, Gibbons, 1989). In turn, these exaggerated
physiological responses over time have a negative impact on one's physical and psychological
health.
Yet, despite the negative effect that perceived discrimination has on well-being, relatively
little research has examined its impact on adolescent well-being and development. Indeed, the
little research that does exist suggests that negative well-being outcomes associated with
discrimination related distress are experienced by adolescents from many ethnic and racial
backgrounds (Fisher, Wallace, & Fenton, 2000; Rosenbloom & Way, 2004; Wong et al., 2003).
For example, a study by Fisher and colleagues (2000) examined the extent to which adolescents
of several racial and ethnic groups experienced distress in response to perceived instances of
racial discrimination. The results of this study indicated that members of all racial and ethnic
groups experienced distress as a result of perceived racial discrimination occurring within
educational contexts. This study also found that distress was negatively associated with self

4

esteem for these adolescents. Given the strong association between racial discrimination and
negative outcomes, researchers have sought to learn more about individual-difference factors that
may be involved in the relation between discrimination and outcomes. One such factor that has
received a great deal of attention is identity. Particularly, racial identity may be especially
important for racial minority adolescents, and past research supports this idea. For example,
Wong et al. (2003) showed that perceived racial discrimination from teachers and peers were
associated with several negative outcomes for African American adolescents, but that a positive
connection to one’s racial group protected adolescents from some of these negative outcomes. As
such, the construct of racial identity may serve as an important psychological resource for racial
and ethnic minority adolescents.
Identity
Recent psychological research on identity has largely conceptualized this construct within
the framework of Social Identity Theory (Tajfel & Turner, 1986). An individual's social identity
is defined as part of the self concept that is derived from one's self-categorizations, that is, one's
perceived group membership and one's positive or negative evaluation of the group. In other
words, a social identity reflects aspects of one's self that are associated with one's internalized
group memberships rather than aspects of the self that stem from an individual's unique attributes
and characteristics (i.e., personal identity). Since identification with a social group depends on
how individuals categorize themselves, the process of self-categorization is at the core of Social
Identity Theory. This framework suggests that social perception is guided by how one
categorizes individuals into social categories (e.g., race, sex, nationality) and the stereotypes and
prototypes associated with these groups (Turner, 1999). For example, when one encounters an
Asian-American woman, one’s perceptions and beliefs about this person will be influenced by

5

the beliefs and stereotypes that the individual holds about the social groups used to categorize
her (i.e., Women and Asian-Americans). Similarly, individuals also use this process to categorize
the self into various social groups. One then may apply the stereotypes and beliefs about these
groups to the self. Once individuals categorize the self as part of a particular social group, in
order for that group membership to become a social identity, they must engage in the process of
identifying with that group (Hogg, 2005). The process of identifying with a group involves
acceptance of the group membership as a core part of the self-concept and defining oneself based
(at least in part) on characteristics of the group.
Social Identity Theory (Tajfel & Turner, 1986 ) suggests that individuals derive a sense of
belongingness from membership in groups, and their self esteem and self image are contingent
on positive perceptions of the groups they belong to and identify with. Individuals also
experience more attachment and connection to the group as a result of increased identification
with a group. Social identity theory was originally conceived as a framework to understand
intergroup relations and discrimination between groups. Indeed, a central aspect of this theory is
the notion that group membership causes individuals to have a tendency to maximize the
perceived differences between ingroup members (i.e., members of the group to which one
belongs to or identifies with) and outgroup members (i.e., members of opposing or outside
groups). In turn, research has shown that this tendency causes one to show favoritism and
positive bias towards ingroup members and perceive a greater number of differences between
ingroup and outgroup members than actually exist (Tajfel & Turner, 1986). This notion of
ingroup favoritism is also theorized to be one of the primary mechanisms individuals use to
bolster individual self esteem and positive self image. According to this view, our positive sense
of self is established or maintained at least in part by the groups that we associate with.

6

It is important to note that according to Social Identity Theory, an individual's group
memberships and social identifications are not necessarily based on formal group memberships
(e.g., clubs, political parties). Rather, the perceived membership and self-categorization in a
particular group is the psychologically important factor. For example, studies using minimal
groups (e.g., Tajfel & Turner, 1986) suggest that simply categorizing oneself in an arbitrary
group is sufficient to elicit ingroup favoritism and outgroup derogation in participants. Tajfel and
Turner (1986) suggest that this finding results from individuals seeking positive self esteem by
differentiating their ingroup from an outgroup on some important dimension and making positive
comparisons to the group.
Racial Identity
Although individuals typically identify with several concurrent social identities, race is a
salient and important identity for many people and serves as a core element for many people's
sense of self (Frable, Blackstone, & Scherbaum, 1990). This is particularly true for racial
minorities. In terms of self-categorization (e.g., Turner, 1995), individuals tend to categorize
themselves based on salient characteristics and groups. As such, racial groups are indeed one of
the most salient group memberships for many racial minorities, and past research indicates that
racial group membership has important implications for the physical and psychological wellbeing of individuals from minority racial groups (e.g., Rowley, Sellers, Chavous, & Smith, 1998)
Racial identity is defined as an individual’s sense of being a member of a particular racial
group. Racial identity has been conceptualized as a stable, multidimensional construct, which
includes feelings and beliefs associated with one’s race, such as evaluations of one’s racial group
and the importance of race to the self (Ashmore, Deaux, & McLaughlin-Volpe, 2004; Sellers,
Smith, Shelton, Rowley, & Chavous, 1998). In other words, racial identity concerns the degree

7

to which an individual relates to and identifies with his or her own racial subgroup as well as an
individual’s involvement with, and attitudes towards her or his racial group (Phinney, 1990).
Although researchers differ in the dimensions they theorize comprise racial identity, several
dimensions emerge across these conceptualizations. One such common dimension is the
importance of racial group membership to the self-concept, a construct that Sellers et al. (1998)
termed racial centrality.
Several studies suggest that identification with a minority racial group is associated with
several positive outcomes. In particular, this research has found that minority group
identification is positively associated with several indicators of psychological adjustment and
well-being such as self esteem (Bat-Chava, 1994; Grossman, Wirt, & Davids, 1985; Phinney,
1990; Rowley, Sellers, Chavous, & Smith, 1998), depression (Munford, 1994), and life
satisfaction (Yap, Settles, & Pratt-Hyatt, 2011). In addition, experimental research also suggests
that increased identification with one’s minority group membership can have similarly positive
psychological outcomes. For example, a study by Major, Spencer, Schmader, Wolfe, and Crocker
(1998) showed that African Americans who were reminded of their race had increased self
esteem following a failure compared to those who were not primed with race. In a similar vein,
other experimental work by McKenna and Bargh (1998) showed that the being in the mere
presence of other individuals from a shared marginalized group increased the degree to which
one identified with that particular identity group. This in turn raised esteem and self acceptance
in these stigmatized individuals.
Racial Identity Development
Examining the foundations of the identity development literature is a useful starting point
to understanding the development of racial identity. Erikson (1968) posited that the formation of

8

one’s sense of identity is the main challenge of psychosocial development during adolescence.
During this time, Erikson believed that individuals struggled with forming a consistent selfconcept and a sense of self. This notion of identity development was further expanded upon by
Marcia (1966, 1980) who theorized that the presence and absence of exploration and
commitment could be crossed to form four distinct identity statuses. Foreclosure is characterized
by the presence of commitment to a set of values or beliefs associated with a particular identity
without engaging in exploration, typically stemming from values and roles prescribed by one’s
parents or society. Moratorium is characterized by the presence of identity exploration without
commitment and involves the active struggle for a sense of identity. Identity diffusion
characterizes an individual who neither engages in exploration or commitment. Diffuse
individuals simply are not engaging in the processes of identity exploration development.
Finally, identity achievement involves commitment to an identity following a period of
exploration and consideration of alternative possibilities. Although this model of identity
development does not posit that these identity statuses necessarily form progressive stages,
identity achievement is seen as the most sophisticated identity status.
The development of racial identity in adolescents and young adults has been at the core
the identity literature since its inception. For example, Erikson (1968), in his classic monograph
on identity, discussed how oppressed racial minority groups are at risk of internalizing the
negative views of other dominant racial groups and developing a racial identity that incorporates
these negative views as a result. Several models of racial identity development (e.g., Cross,
1978; Phinney 1989) overlap with Marcia’s model of identity status in that achieving a sense of
identity results from a period of exploration and experimentation that culminates in a firm
commitment to particular identity. However, one way that that these racial identity models

9

diverge from Marcia’s model is that they suggest that racial identity development occurs in a
series of progressive stages.
For instance, Cross’s (1978) model of Black identity development proposed five
progressive stages through which the development of Black racial identity occurs. Each
progressive stage is characterized by a distinct pattern of attitudes towards one’s own racial
identity. Cross’s (1978) model was originally conceptualized to describe the process through
which African American individuals attain self-actualization when faced with explicit oppression
1

from other, more privileged racial groups . Cross’ (1978) model suggests that individuals in the
first stage, the pre-encounter stage, views the world from the perspective of a Euro-American.
This includes the idealization of whiteness and the devaluation of Blackness. In the second stage,
encounter, individual begin to shed their old views on race as a result of some important
experience that prompted them to question their racial views. In immersion-emersion, the third
stage of Cross’s model, the individual begins to value Blackness and has a positive attitude
towards Blackness. An individual in this stage may begin to engage in Black activities and
experiences and join Black organizations. Additionally, individuals in this stage may also show
an increased tendency to denigrate Whites. However, this stage is characterized by minimal
commitment and internalization of one’s own Blackness and racial identity. The fourth stage of
Cross's (1978) model is characterized by the internalization of one’s experiences with racial
identity stemming from the immersion-emersion stage. It is in this stage where individuals
incorporate aspects of her or his immersion-emersion experience into the self-concept. This stage
is also marked by a decline in anti-White attitudes. Finally, the fifth stage of Cross's model,
1 This model speaks specifically to the development of racial identity in Black Americans. It is
also important to note that this model was developed during the time of the civil rights
movement in the United States. Thus, the applicability of Cross’s model of racial identity
development to other racial groups or to contemporary African Americans is questionable
10

internalization-commitment, is characterized by a continued identification with one's Blackness
and a commitment to continued political involvement with groups and organizations associated
with one’s race.
Cross’s (1978) model shows clear parallels with Marcia’s (1980) model of identity
statuses. For instance, Cross’s pre-encounter stage is comparable to Marcia’s identity foreclosure
status given that both describe individuals who adopt racial attitudes that they have been exposed
to without question. However, these two models differ in their view as to whether individuals in
this stage or status necessarily endorse negative attitudes towards one’s own group. Similarly,
both models suggest that that individuals go through a period of exploration (e.g., exploring the
meaning of one’s race) and that this type of exploration is central to the development of one’s
identity.
These apparent parallels between the racial identity development literature and Marcia’s
(1980) model led Phinney (1989) to develop a stage model of racial identity development drawn
directly from Marcia’s identity status model that could be applied to multiple racial and ethnic
groups. Phinney’s (1989) model posits that racial identity development is characterized by a
progression through Marcia’s (1980) four identity statuses in the following order: diffusion,
foreclosure, moratorium, and achievement. Overall, research examining this developmental
model (e.g., Phinney, 1989; Phinney & Alipuria, 1990; Phinney & Tarver, 1988) has indicated
that the majority of 15 year old racial minority adolescents have engaged in some form of
exploration of their racial identity and that exploring issues related to their racial or ethnic group
is of central concern to identity for these minority adolescents. Indeed, some research suggests
that the exploration of one’s own racial identity not only is a primary concern for racial minority
adolescents’ identity development, but racial identity is seen by these individuals as the most

11

important aspect of one’s self-definition (Aries & Morehead, 1989) Other research by Phinney
and Chavira (1992) examining the developmental trajectories of racial identity in racial minority
adolescents suggests that the period from age sixteen to nineteen is associated with marked
change in the racial identity status of these individuals. Most individuals experienced shifts to
higher racial identity statuses during this time. This research has also shown that adolescents of
multiple racial groups show evidence of racial identity development during early adolescence
including African Americans, Hispanics, East Asian Americans and South Asian Americans (e.g.,
Fisher et al., 2000; Phinney, 1989; Phinney & Alipuria, 1990; Phinney & Tarver, 1988)
Theoretical Models
As discussed above, past research has shown that racial identity and perceived
discrimination are both important variables that have important implications for several
psychological adjustment outcomes for adults and adolescents. Because of these associations, it
is important to understand the nature of the relationship between identity and discrimination and
how the interplay of these variables affects individuals’ psychological well-being. Past research
has posited three major models describing the relationship between racial discrimination, racial
identity and psychological well-being. Each model substantially differs in the hypothesized role
that racial identity plays in the relationship between discrimination and well-being. The first
model, known as the Rejection Identification Model (Branscombe et al., 1999) suggests that
when individuals perceive themselves to experience racial discrimination, they react by
increasing identification with their racial group, which in turn leads to increased well-being. The
second model, referred to in the current paper as the Buffering Hypothesis (Sellers, CopelandLinder, Martin, & Lewis, 2006), hypothesizes that racial identity buffers the negative effects of
perceived discrimination on well-being. Specifically, the negative effects of perceived

12

discrimination on well-being are attenuated in highly identified individuals, but exacerbated for
individuals low in racial identity. The final model, referred to as the Discounting Hypothesis
(Crocker & Major, 1989), posits that racial identification is associated with the tendency to
attribute ambiguous negative events to racial discrimination against oneself; this perceived racial
discrimination is, in turn, associated with negative psychological outcomes. Below, each of
these models is discussed in greater detail.
Rejection Identification Model
The Rejection Identification Model (Branscombe et al., 1999; Schmitt & Branscombe,
2002; see Figure 1) posits that an individual's willingness to make attributions to prejudice will
lead to direct and negative effects on his or her psychological well-being. This model also
predicts that an individual's willingness to make attributions to prejudice will increase minority
group identification, which in turn has a positive effect on psychological well-being. That is,
discrimination is conceptualized as having a direct negative effect on well-being, but an indirect
positive effect on well-being, mediated by minority group identification.
The Rejection Identification Model draws upon Social Identity Theory (Tajfel & Turner,
1986) with the notion that the experience of prejudice and discrimination from racial outgroup
members will increase an oppressed group member’s level of identification with their racial
ingroup as a way to increase one’s self esteem. Thus, Social Identity Theory predicts that
increases in one’s racial identity have a positive effect on the individual’s psychological wellbeing. In this model, minority group identification is predicted to be a positive mediator of the
negative relationship between discrimination and psychological well-being. In other words, this
model predicts that when a member of a disadvantaged group experiences unfair treatment,
increased psychological commitment and investment into ones group membership results in

13

benefits for one’s well-being (Branscombe et al., 1999).
In a study by Branscombe, Schmitt, and Harvey (1999), this model was tested in a sample
of African Americans. The results of this study indicated that for African Americans, perceptions
of anti-African American prejudice was associated with increased racial identification, which in
turn partially attenuated the direct negative effects of perceived discrimination on well-being.
Indeed, a further study by Schmitt, Branscombe, Kobrynowicz, and Owen (2002) found similar
results in a sample of women; perceptions of gender discrimination were associated with
increased gender identification, which in turn attenuated the direct negative effects of gender
discrimination on well-being.
Other past research supports the notion that perceptions of discrimination result in
increased group identification. For example, work by Turner, Hogg, Oakes and Smith (1984)
suggests that threats to the status of an ingroup can increase identification with one’s ingroup.
Other research by Ellemers (1993) using a minimal group paradigm found that when members of
low-status minimal groups were under the impression that they would not be able to move up in
group status, members' identification with their group was higher compared to members of lowstatus minimal groups that were not led to believe that their group membership would be a
barrier to their success. Further research suggests that perceptions of prejudice and minority
group identification are positively related among many other disadvantaged minority groups
including religious groups (e.g., Jewish people; Rollins, 1973), “counter-culture” groups (e.g.,
college punks and hippies; Cozzarelli & Karafa, 1998) and sexual minorities (e.g., gay women;
Crosby et al., 1989)
The Buffering Hypothesis
The Buffering Hypothesis (e.g., Sellers, Copeland-Linder, Martin, & Lewis, 2006; Yip,

14

Gee, & Takeuchi, 2008; see Figure 2), like the Rejection Identification Model, predicts that
perceived discrimination will have a negative direct effect on well-being. Where this framework
differs from the Rejection Identification Model is in the predicted role of racial identity. In this
buffering framework, racial identity is viewed as a protective factor that buffers an individual
from the negative effects of perceived discrimination. In other words, racial identity is
conceptualized as having a moderating role on the association between racial discrimination and
well-being. Rather than viewing discrimination as causing increased racial identification (i.e., a
mediating role), which in turn results in positive effects on well-being, the Buffering Hypothesis
suggests that ethnic identification and perceptions of discrimination have an interactive effect on
well-being outcomes. More specifically, this model predicts that perceived discrimination will
have negative direct effects on psychological well-being for individuals that are low in racial
identity. However, this model predicts that the negative relationship between perceived
discrimination and well-being is attenuated in individuals that are highly identified with their
racial group.
Like the Rejection Identification Model, Social Identity Theory can also offer an
explanation for why identification with one’s racial group would buffer the effects of racial
discrimination. Social Identity Theory suggests that individuals have a tendency to place an
emphasis on the positive aspects of one’s ingroup. In turn, this leads to increased self esteem and
self worth (Tajfel & Turner, 1986). Individuals for whom membership in their racial group is an
important part of the self will be especially motivated to feel good about their group membership
in the face of discrimination from others and these individuals may simply be more committed to
focus on positive attributes of their in-group.
There is ample evidence that racial identity protects individuals from the negative impact

15

of discrimination. For example, a study by Sellers, Caldwell, Schmeelk-Cone, and Zimmerman
(2003) found that there was a positive relationship between perceived discrimination and distress
for African American adolescents with low or moderate racial identification. In contrast, there
was no substantial relationship between perceived discrimination and distress for individuals
who were highly identified with their racial group. A later study by Neblett, Shelton, and Sellers
(2004) found similar results indicating that perceived discrimination was linked to depression,
stress and anxiety for weakly identified African Americans but the relationship was not found for
highly identified individuals. A further study by McCoy and Major (2003) found evidence that
group identification moderates the effect of experimentally induced perceived prejudice on self
esteem. Similarly, the results of a study by Yip et al. (2008), using a nationally representative
sample of Asian people living in the United States, indicated that racial identity buffered the
association between perceived discrimination and psychological distress in a sample American
born Asian that were 41-50 years of age. It should be noted, however, that opposite results were
found for other age groups. Specifically, for American born Asians between 31- 40 years of age
or 51-75 years of age, racial identity exacerbated the effects of discrimination on distress. The
reasons for this difference are not clear. However, Yip et al. (2008) suggested that these results
reflect developmental trends in racial identity across the life span. For individuals 31-40 years of
age, Yip et al. (2008) suggested that several significant life changing events are particularly
likely to happen during this period in life (e.g., marriage, having children). In turn, these types of
events prompt a period of identity search and development during this time. Thus, according to
Yip et al. (2008), experiencing discrimination during this period of identity search may be
especially detrimental. Similarly, Yip et al. (2008) suggest that individuals 51-75 years of age
begin to suffer from the realization that they are aging and may engage in the reorganization of

16

the life goals. As a result of these changes, this age group is posited to be especially susceptible
to the negative effects of discrimination.
Despite some contradictory evidence, taken as a whole, there is a consistent body of
evidence that racial identity buffers the negative effects of discrimination. However, there is
some evidence that identity exacerbates this effect as well.
The “Discounting” Perspective
The Discounting Hypothesis is a substantial departure from the two previous models
discussed above. Although the previous models suggest that perceptions of discrimination are
negatively related to well-being, this perspective maintains that being able to perceive
discrimination towards one’s group may have positive outcomes for psychological well-being.
Thus, the Discounting Hypothesis predicts that an individual's perception of discrimination will
lead to a direct positive effect on her or his well-being. The predicted role of racial identity in
this model also differs from the two previous models. The Discounting Hypothesis views racial
identity as an antecedent to perceived discrimination. That is, rather than being a consequence of
perceived discrimination, the degree to which one identifies with her or his racial group predicts
the degree to which one construes negative events as racial discrimination. Thus, racial
identification is hypothesized to increase one's tendency to attribute negative events to
discrimination. This attribution, in turn, increases one's perception of well-being (see Figure 3).
According to the Discounting Hypothesis, if one is able to attribute negative events one
experiences to prejudice and discrimination, this may serve to protect self esteem and other
important well-being outcomes. That is, it may be psychologically beneficial for individuals to
be able to “discount” one's own behavior and personality as causes for negative treatment, and
instead attribute negative treatment to the prejudice of another individual. One can then

17

effectively avoid blaming the self for the negative outcome (Crocker & Major, 1989). In other
words, being able to attribute negative events to discrimination based one's ethnic group may
have a positive effect on one's self esteem because it allows people to discount internal, stable
aspects of the self as causes of the underlying negative event. Rather, individuals are able to
externalize the cause for their negative experiences.
In line with this view, the results of several studies have demonstrated that, consistent
with the Discounting Hypothesis, perceived discrimination of one's group may actually enhance
one's well-being and adjustment by reminding people that others share this perceived
discrimination against them (Bourguignon, Seron, Yzerbyt & Herman, 2006; Major, Kaiser, &
McCoy, 2003; Major, Quinton & Schmader, 2003). However, other studies find no relationship
between perceived discrimination and self esteem among both women and African Americans
(Crocker, Luhtanen, Blaine, & Broadnax, 1994). These mixed results highlight the importance of
identifying factors that affect the complex relationship between discrimination and self esteem.
One potential explanation for this inconsistency is that studies that show a positive perceived
discrimination to well-being relationship are typically experimental studies that examine a single
attribution to prejudice in response to a contrived, laboratory stimuli or a hypothetical written
scenario. This creates differences in how the notion of perceived discrimination is
operationalized across these studies compared to studies that show a negative perceived
discrimination to well-being association (single attribution vs. general beliefs about the extent of
discrimination over time).
The proponents of this model typically discuss the importance of emotional and
psychological responses to negative life events as a function of how one cognitively appraises
the event (i.e., appraising an event as discriminatory or not). As such, past research (e.g.,

18

Eccelston & Major, 2006) has identified group identification and the importance of group
identity (i.e., centrality) as an important predictor of such appraisals. Indeed, experimental
research has indicated that group identification can increase the likelihood that individuals will
attribute ambiguous negative events to discrimination (Major, Quinton & Schmader, 2003;
Operario & Fiske, 2001).
Although Crocker and Major's (1989) Discounting Hypothesis was originally conceived
as a model that accounts for a single attribution to prejudice for a single negative event, this
framework has also been applied to more general attributions of discrimination (e.g., Eccelston
& Major, 2006 ). However, given the substantial body of existing evidence suggesting that
general perceived discrimination is a major stressor that is negatively associated with various
physical and psychological well-being outcomes (e.g., Barnes et al., 2008; Branscombe et al.,
1999; Broman et al., 2000; Forman, 2003; Guyll et al., 2001; Harrell, 2000; Jasinskaja-Lahti &
Liebkind, 2001; Kessler et al., 1999; Klonoff et al., 1999; Liebkind & Jasinskaja, 2000; Sellers et
al. 2006; Williams et al., 2003; Wong et al., 2003), the plausibility of the Discounting Hypothesis
in representing the relationship between these general attributions and well-being is questionable.
It may be the case that attributing a negative experience to discrimination is associated with
certain protective processes in single, stand alone instances (as originally outlined by Crocker &
Major, 1989), but perceptions of chronic discrimination and mistreatment from others still may
be associated with negative outcomes. Even if one where able to discount the self, the chronic
nature of this type of perceived discrimination and the fact that this general perceived
discrimination likely involves important life domains (as opposed to artificial aspects of the lab
environment), suggests that general perceived discrimination may have negative associations to
well-being. As such, the present study will explore whether this discounting perspective can be

19

applied to general perceptions of discrimination in the sample to be used in this proposed study.
The Proposed Study
Past research has suggested three roles for racial identity in the relationship between
perceived racial discrimination and well-being: racial identification mediates the relationship
between perceived discrimination and well-being (Rejection Identification Model), racial
identification moderates the relationship between perceived discrimination and well-being
(Buffering Hypothesis), and racial identification leads to perceptions of racial discrimination,
which in turn relate to higher well-being (Discounting Hypothesis). Given the mixed evidence
for these theorized relationships, the aim of this study was to test and compare these three
theoretical models. Each of the models discussed here make markedly different predictions about
the nature of the relationships between these variables and this study examined which of these
models was most consistent with data derived from a large longitudinal study of children of
immigrants to the United States.
This study makes several important theoretical contributions. To the best of our
knowledge, this was the first study to examine these models simultaneously and examine the
predictions of each of these theoretical models in a large longitudinal community sample of
individuals from multiple ethnic minority groups living in the United States. In one exception, a
study by Eccleston and Major (2006) examined these models using a relatively small cross
sectional study of Latino American college students. These researchers asked participants to read
five ambiguous scenarios and report on the degree to which they felt that each scenario could be
attributed to racial prejudice if it happened to them. Each participant’s responses were averaged
across the scenarios and this composite measure was used as a measure of discrimination
attributions. Participants were then asked to report on their global self esteem and the degree to

20

which they identified with their ethnic group. These researchers were primarily interested in
testing the Discounting Hypothesis, the Buffering Hypothesis, and the Rejection Identification
Model as models of the relationships between attributions to discrimination, racial identification,
and self esteem. The results of this study showed that the models representing the Discounting
Hypothesis and the Buffering Hypothesis fit their observed data. However, this study failed to
find evidence for the Rejection Identification Model. Although these researchers were not able to
examine whether the Discounting Hypothesis or the Buffering Hypothesis was a better fit to their
data, one thing that we can draw from this study is that multiple theoretical models may hold true
in explaining the relationships between racial identity, perceived discrimination and self esteem.
Ecceleston and Major’s (2006) study makes an important contribution to our understanding of
the relationship between racial identity, perceived discrimination and well-being. However, a
notable limitation of this study was that it used a relatively small cross-sectional sample of
undergraduate students. The present study extends the study by Eccelston and Major in several
regards. Not only did our data set offer a substantially larger sample size, but it also allowed the
test of these models over time across the different waves of data collection. Although a true test
of causality is not possible in either Ecceleston and Major’s or the present study, the longitudinal
nature of the present data allowed for a stronger test of some of the causal assumptions of these
three models.
This study also examined these questions in a sample of second generation immigrants to
the United States, a population in which this line of research has not been addressed. This
population has many unique characteristics (e.g., increased feelings of biculturality; BenetMartínez, Leu, Lee, & Morris, 2002) that make the dynamics and interplay of racial identity,
perceived discrimination and psychological well-being theoretically interesting. For this

21

population, identification and involvement in an ethnic culture along with the mainstream U.S.
culture is common (Phinney, 1996) and relatively little research has examined identity processes
and perceived discrimination among this group. Given the recent rising trends in cultural and
ethnic diversity in North America (Citizenship and Immigration Canada, 2006; Fix & Passel,
2001), understanding how the associations between these variables differ in immigrant
populations remains an important avenue of research. Finally, in addition to psychological wellbeing outcomes, this study allowed us to consider how career and educational outcomes are
related to racial identity and perceived discrimination.
Hypotheses
Although this study was largely exploratory, to determine which model was most
consistent with the present data we evaluated whether the direction and strength of the
relationships between variables were consistent with the predictions of each model. As with any
correlational study, we were not able to test the causal direction of the relationship between racial
identity, perceived discrimination and well-being. However, using this dataset, we were able to
examine whether the relationships observed in this data were consistent with one or more of the
models examined in this paper. Further, we were also able to use three waves of longitudinal data
to examine these models in the current study. This provides a stronger test of these causal models
than compared to just using cross-sectional data.
Based on the Rejection Identification Model, one would have expected a negative direct
effect between perceived discrimination and well-being both within wave and across waves.
Moreover, this model also predicted that there was an indirect positive effect between perceived
discrimination and well-being that was mediated by racial identity. The positive relationship
between perceived discrimination and racial identity was predicted to occur both within waves

22

and between the perceived discrimination scores of one wave and the racial identity scores of
subsequent waves.
Based on the Buffering Hypothesis model, one would have predicted negative direct
effects between perceived discrimination and the outcomes of well-being, career attainment and
educational attainment. These effects would be predicted to occur both between waves and
within waves. Regarding the role of racial identity, this model predicted that racial identity would
serve to buffer the negative relationship between perceived discrimination and well-being. That
is, this model predicted that these relationships were moderated by the level of racial identity.
Higher racial identity was predicted to buffer the negative effect but lower racial identity was
predicted to exacerbate the negative impact of perceived discrimination on well-being, career
attainment and educational attainment. Although we were able to examine whether the Buffering
Hypothesis accounts for the relationship between racial identity, perceived discrimination and
well-being, this model did not speak directly to any longitudinal predictions.
Based on the Discounting Hypothesis model, one would have predicted that racial
identity was positively associated with the tendency to attribute negative events to
discrimination, thus these individuals should report higher levels of perceived discrimination.
Further, this model predicted a direct positive association between racial identity and the
outcomes of well-being, career attainment and educational attainment. These effects were
predicted to occur both between waves and within waves.
The Discounting Hypothesis and the Rejection Identification Model are distinguished
mainly by their predictions regarding the association between racial identity and perceived
discrimination. The Discounting Hypothesis predicts that racial identity is an antecedent to
perceived discrimination and that perceived discrimination is positively related to well-being. In

23

contrast, the Rejection Identification Model predicts that perceived discrimination precedes
racial identity and that perceived discrimination is negatively related to well-being. Although
examination of whether these variables were truly causally related and causal ordering of
perceived discrimination and racial identity was not possible in this study, we were able to
examine whether the relationship between perceived discrimination and well-being was positive
or negative. Thus, examining the nature of the relationship between perceived discrimination and
well-being was key evidence for distinguishing between these models. It was also possible that
the data would be consistent with more than one of these models. For example, if racial identity
mediated and moderated the association between perceived discrimination and well-being, this
would have suggested that the Rejection Identification Model and Buffering Hypothesis were
operating simultaneously. Thus, in addition to examining the specific predictions that
differentiate the Rejection Identification Model and the Buffering Hypothesis, we also examined
whether there was simultaneous evidence for multiple models.
Overall, we predicted that the Discounting Hypothesis was the least plausible of the three
models given that this study examined generally perceived discrimination, and we did not predict
that we would find a positive relationship between perceived discrimination and well-being,
which is a key component of this model. However, it was plausible that individuals that were
higher in racial identity were more likely to perceive negative events as discrimination (which is
predicted by this model).
Method
Participants and Procedures
The data in this study was obtained from the three waves of the Children of Immigrants
Longitudinal Study (CILS; Portes & Rumbaut, 2006), a panel study of United States born

24

children with at least one foreign born parent or foreign born children that immigrated to the
United States at an early age (M = 14 years). This study involved three waves of data collection.
The first wave of data was collected in 1992 and sampled 8th and 9th grade children in schools
around the urban areas of Miami, Florida and San Diego, California. This survey was conducted
in classrooms using paper and pencil questionnaires. The total sample size for this wave of data
collection was 5262. The second wave of data was collected three years later in 1995 when the
sample was about to graduate high school. This second wave was able to capture 4288
participants (81.5%) from the original sample. As with the first wave, data were collected at the
schools attended by the participants. The third and final wave of data was collected between
2001 to 2003, approximately 10 years following the initial survey. At this time, the sample had a
mean age of 24 years. As the majority of the respondents had finished schooling, they had to be
contacted individually using a mailed questionnaire. A total of 3613 individuals responded to the
final survey, accounting for 68.9% of the original wave 1 sample. In order to maximize our
power to detect potential effects, each analysis in the present study utilized all participant data
available to examine each unique question, regardless of whether or not the participant provided
data at other waves. Thus, cross sectional and longitudinal analyses contain overlapping samples,
with some differences due to attrition over the course of data collection. Demographic
characteristics for the entire sample are displayed in table 1.
Measures
Table 2 displays an overview of the measures assessed at each wave of the study.
Perceived discrimination. This measure was not uniform across the three waves of data
collection. In the first two waves of the study, participants were asked a single item indicating
whether or not they have ever felt discriminated against. Following this, participants were then

25

asked to indicate, using an open ended response, the main reason they felt that this
discrimination occurred. These open ended responses were then coded by the primary
investigators of this dataset and codes were provided in the publically available dataset. We
computed a binary perceived racial discrimination variable where “1” represented individuals
who indicated that they have been discriminated against and whose main reason for the
discrimination was coded as due to “race,” or “ethnic or national origin.” Individuals who either
reported that they had not felt discriminated against or did feel discriminated against but did not
report the main reason to be race or ethnicity related were coded as “0”. In the third wave of the
study, participants were asked a single item, “have you ever felt discriminated against because of
your race or ethnicity?” Participants responded with a yes or no, which were coded “1” and “0”,
respectively.
Self esteem. Self esteem was measured using the 10-item Rosenberg Self Esteem
Inventory (Rosenberg, 1965). Participants rated the degree to which they agreed or disagreed
with 10 statements (e.g., “I am a person of worth, at least on an equal basis with others”) on a
scale ranging from 0 (not at all) to 6 (very much). This scale was included in both the first and
second waves of data collection.
Depression. Depression was measured using four items from the CES-D Depression
Scale (Radloff, 1977). Participants rated how they felt during the past week in response to four
items (e.g., I felt sad; I could not get going; I did not feel like eating/my appetite was poor; I felt
depressed) on a scale ranging from 1 (rarely) to 4 (most of the time). This scale was included in
both the first and second waves of data collection.
Racial Identity. Racial identity was measured using a single item. Participants were
asked to indicate the ethnic self categorization, followed by the question, “And how important is

26

this identity to you, that is, what you call yourself?” Participants rated the degree of importance
on a scale ranging from 0 (not important) to 3 (very important). This scale was included in both
the second and third wave of data collection.
Job satisfaction. Job satisfaction was measured using a single item, “How satisfied are
you with this job, as a whole?” Participants rated their degree satisfaction on a scale ranging
from 1 (very dissatisfied) to 5 (very satisfied). This item was included in the third wave of data
collection only.
Monthly income. Participants were asked to report a dollar amount indicating their
approximate gross income, per month, from all sources. This item was included in the third wave
of data collection only.
Educational attainment. Educational attainment was measured using a single item
asking participants to indicate what “the highest grade or level of school” they had completed.
Responses to this item were recoded into the number of years of education completed by the
primary investigators of the dataset. Participant responses ranged from “some high school” to
“professional/doctoral degree.” This item was included in the third wave of data collection only.
Results
Overview of Analysis
The goal of the present study was to test the Rejection Identification Model, Buffering
Hypothesis, and Discounting Hypothesis as theoretical explanations of the relationship between
perceived discrimination, racial identity and well-being. In order to test these theories, we
examined the predictions of each theory regarding the specific relationships between the
variables available in this dataset using linear regression. All the following analyses were
conducted using Mplus5 (Muthén & Muthén, 2007).

27

Bivariate correlations, descriptive statistics and Cronbach's alphas for variables at each
wave are presented in Table 3. For a summary of the overall findings from these analyses, please
refer to Table 10.
Cross-Sectional Analyses
Each theoretical model was examined cross-sectionally at time 2 and time 3. At time 2,
the outcome variables of interest were self esteem and depression. At time 3, the outcome
variables of interest were monthly income, job satisfaction, and years of educational attainment.
Each outcome variable was examined in separate analyses for each model at each wave.
Cross-sectional analyses: Rejection Identification Model. In the following analyses,
we assessed the indirect effects of perceived discrimination on each respective dependent
variable via racial identity in a single mediator model. In each of these analyses, perceived
discrimination was the predictor variable, racial identity was the mediator variable, and the
outcome variable was either self esteem, depression, monthly income, job satisfaction, or years
of education. Each of these analyses was cross sectional, and the perceived discrimination and
racial identity variable within each individual analysis came from the same wave that each
outcome variable came from.. A bootstrapping procedure recommended by Shrout and Bolger
(2002) was utilized in all analyses, in which a 95% confidence interval is created for the size of
the specific indirect effects of each mediating path. We used 5000 bootstrap resamples to create
the 95% CI for each estimate.
Self-Esteem (Time 2). Results, shown in Table 4, indicated that perceived discrimination
was negatively associated with self esteem at time 2. Further, perceived discrimination was
positively associated with racial identity, which in turn, was positively associated with self
esteem. Finally, these results showed that racial identity was a significant mediator of the

28

association between perceived discrimination and self esteem. In sum, these results suggest that
higher perceived discrimination is associated with higher racial identity and lower self esteem.
Further, higher racial identity is also associated with higher self esteem. These results indicate
that racial identity suppresses the association between perceived discrimination and self esteem
such that after racial identity is taken into account, the direct effect of perceived discrimination
on self esteem increases in magnitude (see Table 4; unstandardized total effect = -0.04;
unstandardized direct effect = -0.05).
Depression (Time 2). Results, shown in Table 4, indicated that perceived discrimination
was positively associated with depression at time 2. Perceived discrimination was also positively
associated with racial identity. However, racial identity was not significantly associated with
depression. Finally, we failed to find evidence that racial identity mediated the association
between perceived discrimination and depression. When racial identity is taken into account, the
direct effects of perceived discrimination and depression do not change.
Monthly Income (Time 3). Results, shown in Table 4, indicated that perceived
discrimination was negatively associated with monthly income at time 3. Further, perceived
discrimination was positively associated with racial identity, which in turn, was negatively
associated with monthly income. Finally, these results showed that racial identity was a
significant mediator of the association between perceived discrimination and monthly income. In
sum, these results suggest that higher perceived discrimination was associated with higher racial
identity and lower monthly income. Further, higher racial identity was also associated with lower
monthly income. These results also indicate that racial identity was a significant mediator of the
association between perceived discrimination and monthly income, such that after racial identity
is taken into account, the direct effects of perceived discrimination on monthly income decrease

29

in magnitude. However, these results do not support the Rejection Identification Model since
racial identity and income are negatively related.
Job Satisfaction (Time 3). Results, shown in Table 4, indicated that perceived
discrimination was negatively associated with job satisfaction at time 3. Further, perceived
discrimination was positively associated with racial identity, which in turn, was positively
associated with job satisfaction. However, we failed to find a significant indirect effect of
perceived discrimination on job satisfaction via racial identity. This suggests that after racial
identity is taken into account, the direct effects of perceived discrimination on job satisfaction do
not change.
Years of Education (Time 3). Results, shown in Table 4, indicated that perceived
discrimination was positively associated with years of education at time 3. Further, perceived
discrimination was positively associated with racial identity, which in turn, was negatively
associated with years of education. Finally, these results showed that racial identity was a
significant mediator of the association between perceived discrimination and years of education.
These results indicate that racial identity is a significant suppressor of the association between
perceived discrimination and years of education, such that after racial identity is taken into
account, the direct effects of perceived discrimination and years of education increase in
magnitude (see Table 4; unstandardized total effect = 0.32; unstandardized direct effect = 0.33).
Cross-sectional analyses: Buffering Hypothesis. In the following analyses, we assessed
whether the association between perceived discrimination and well-being is qualified by
individuals' level of racial identity. Support for the Buffering Hypothesis would be given by a
significant interaction between perceived discrimination and racial identity, such that when racial
identity is low, the association between perceived discrimination and well-being is negative. In

30

contrast, when racial identity is high, the negative association between perceived discrimination
and well-being is buffered and becomes less negative.
Self Esteem (Time 2). Results, shown in Table 5, indicated that perceived discrimination
is negatively associated with self esteem at time 2. That is, higher perceived discrimination was
associated with lower self esteem. Results also indicated that racial identity positively predicts
self esteem, suggesting that higher racial identity was associated with higher self esteem.
However, the interaction term between racial identity and perceived discrimination failed to
reach significance, indicating that there was no evidence that perceived discrimination and racial
identity significantly interact to predict self esteem.
Depression (Time 2). Results, shown in Table 5, indicated that perceived discrimination
was positively associated with depression at time 2. That is, higher perceived discrimination was
associated with higher depression. However, the coefficients for racial identity and the
interaction between perceived discrimination and racial identity failed to reach significance.
Thus, we failed to find evidence for a relationship between racial identity and depression, and
there is no evidence that perceived discrimination and racial identity significantly interacted to
predict depression.
Monthly Income (Time 3). Results, shown in Table 5, indicated that perceived
discrimination negatively predicts monthly income at time 3. That is, higher perceived
discrimination was associated with lower income per month. Results also indicated that racial
identity negatively predicts monthly income, suggesting that higher racial identity was associated
with lower income per month. However, the interaction term between racial identity and
perceived discrimination failed to reach significance, indicating that there was no evidence that
perceived discrimination and racial identity significantly interact to predict monthly income.

31

Job Satisfaction (Time 3). Results, shown in Table 5, indicated that perceived
discrimination negatively predicts job satisfaction at time 3. That is, higher perceived
discrimination was associated with lower job satisfaction. Results also indicated that racial
identity positively predicts job satisfaction, suggesting that higher racial identity was associated
with higher job satisfaction. However, the interaction term between racial identity and perceived
discrimination failed to reach significance, indicating that there was no evidence that perceived
discrimination and racial identity significantly interacted to predict job satisfaction.
Years of Education (Time 3). Results, shown in Table 5, indicated that perceived
discrimination positively predicts years of education attained at time 3. That is, higher perceived
discrimination at time 3 was associated with more years of education completed at time 3.
Results also indicated that racial identity negatively predicts years of education, suggesting that
higher racial identity was associated with attaining a lower number of years of education.
However, the interaction term between racial identity and perceived discrimination failed to
reach significance, indicating that there was no evidence that perceived discrimination and racial
identity significantly interacted to predict years of education.
Cross-sectional analyses: Discounting Hypothesis. In the following analyses, we
assessed the indirect effects of racial identity on each respective dependent variable via perceived
discrimination in a single mediator model. In each of these analyses, racial identity was the
predictor variable, perceived discrimination was the mediator variable, and the outcome variable
was either self esteem, depression, monthly income, job satisfaction, or years of education. Each
of these analyses was cross sectional, and the perceived discrimination and racial identity
variable within each individual analysis came from the same wave that each outcome variable
came from. A bootstrapping procedure recommended by Shrout and Bolger (2002) was utilized

32

in all analyses, in which a 95% confidence interval is created for the size of the specific indirect
effects of each mediating path. We used 5000 bootstrap resamples to create the 95% CI for each
estimate.
Self Esteem (Time 2). Results, shown in Table 6, indicated that racial identity was
positively associated with self esteem at time 2. Further, racial identity was positively associated
with perceived discrimination, which in turn, was negatively associated with self esteem. In sum,
these results suggest that higher racial identity was associated with higher perceived
discrimination and higher self esteem. Finally, these results showed that perceived discrimination
was a significant mediator of the association between racial identity and self esteem.
Depression (Time 2). Results, shown in Table 6, indicated that racial identity was
positively associated with perceived discrimination, which in turn, was positively associated with
depression. The direct effect from racial identity to depression was non-significant when
perceived discrimination was included in the model but there was evidence for a significant
indirect effect from racial identity to depression via perceived discrimination.
Monthly Income (Time 3). Results, shown in Table 6, indicated that racial identity was
negatively associated with monthly income at time 3. Further, racial identity was positively
associated with perceived discrimination, which in turn, was negatively associated with monthly
income. Finally, these results showed that perceived discrimination was a significant mediator of
the association between racial identity and monthly income. In sum, these results suggest that
higher racial identity was associated with higher perceived discrimination and lower monthly
income. Further, higher perceived discrimination was also associated with lower monthly
income. These results also indicate that perceived discrimination was a significant mediator of
the association between racial identity and monthly income, such that after perceived

33

discrimination is taken into account, the direct effects of racial identity on monthly income
decrease in magnitude.
Job Satisfaction (Time 3). Results, shown in Table 6, indicated that racial identity was
positively associated with job satisfaction at time 3. Further, racial identity was positively
associated with perceived discrimination, which in turn, was negatively associated with job
satisfaction. Finally, these results showed that perceived discrimination was a significant
mediator of the association between racial identity and job satisfaction. In sum, these results
suggest that higher racial identity was associated with higher perceived discrimination and higher
job satisfaction. Further, higher perceived discrimination was also associated with lower job
satisfaction. These results also indicate that perceived discrimination was a significant suppressor
of the association between racial identity and job satisfaction, such that after perceived
discrimination is taken into account, the direct effects of racial identity on job satisfaction
increase in magnitude (see Table 6; unstandardized total effect = 0.06.; unstandardized direct
effect = 0.07).
Years of Education (Time 3). Results, shown in Table 6, indicated that racial identity was
negatively associated with years of education at time 3. Further, racial identity was positively
associated with perceived discrimination, which in turn, was positively associated with years of
education. Finally, these results showed that perceived discrimination was a significant mediator
of the association between racial identity and years of education. In sum, these results suggest
that higher racial identity was associated with higher perceived discrimination and less years of
education. Further, higher perceived discrimination was also associated with lower years of
education. These results also indicate that perceived discrimination was a significant mediator of
the association between racial identity and years of education, such that after perceived

34

discrimination is taken into account, the direct effects of racial identity on years of education
increase in magnitude.
Summary. The results showed that racial identity was positively associated with
perceived discrimination and self esteem and suppressed the negative association between
perceived discrimination and self esteem. This finding provides the clearest support for the
Rejection Identification Model. Although there were significant indirect effects from perceived
discrimination to monthly income and educational attainment via racial identity, these findings
are inconsistent with the Rejection Identification Model given that racial identity was negatively
related to the two outcome variables. Regarding the Discounting Hypothesis, these results also
indicated that there was a significant indirect effect from racial identity on depression, self
esteem, income, job satisfaction, and education via perceived discrimination. However, these
models diverge from the predictions of the Discounting Hypothesis given that racial identity was
negatively associated with income and education, and perceived discrimination was negatively
associated with self esteem and job satisfaction, and positively associated with depression.
Overall, these cross sectional analyses provide the clearest evidence for the Rejection
Identification Model and some limited support for the Discounting Hypothesis. In contrast, we
did not find any support for the Buffering Hypothesis in regard to any of the outcomes.
Cross-Wave Analyses
Next, we examined the specific predictions of each theoretical model longitudinally
across the three waves of the data. The nature and direction of the relationship between each
study variable was examined across waves and compared with predictions of each theoretical
model. These analyses offer an opportunity to examine the relations of these variables across
time, which allows one to test the meditational effects that are predicted by these models in a

35

more rigorous manner. Statistical mediation makes predictions about a causal chain in which, at
minimum, it is assumed that one variable causes another, which in turn has a causal effect on
some outcome. An essential component of this notion of causality is that the variable that is said
to cause an outcome must precede that outcome in time (Sobel, 1990). Given the importance of
temporal precedence in causal statements, the relative strength of longitudinal data over cross
sectional data in testing meditational relationships becomes clear. The following analyses
examined the degree to which the associations between the variables in this dataset across time
were consistent with specific predicted causal relationships of the three theoretical models.
In order to be thorough, we examined the key predictions of the Buffering Hypothesis
longitudinally, although it is important to note that the Buffering Hypothesis does not make clear
predictions regarding the temporal causal precedence of the study variables. The critical test of
the Buffering Hypothesis involves a simple interaction (i.e., racial identity interacting with
perceived discrimination to predict an outcome), which does not suggest that either the predictor
or the moderator precedes one another.
Cross-wave analyses: Rejection Identification Model.
As mentioned earlier, a longitudinal design allows for a stronger test of mediational
models due to the causal predictions inherent in mediation models (Cole & Maxwell, 2003). In
the case of the Rejection Identification Model, perceived discrimination is predicted to cause
racial identity, which in turn is predicted to cause increased psychological well-being. The
finding that racial identity can statistically explain the relation between perceived discrimination
and well-being (e.g., significant indirect effect) is not sufficient to make racial identity a
mediator. Racial identity must also be a dependent variable relative to perceived discrimination
and an independent variable relative to the well-being outcome. This implies that the perceived

36

discrimination precedes racial identity in time, which in turn precedes the outcome variable in
time (Cole & Maxwell, 2003; Sobel, 1990). Ideally, a longitudinal test of this mediational model
would involve measurement of perceived discrimination at one time point, followed by
measurement of racial identity at a second time point, which in turn is followed by measurement
of the well-being outcome variable at a third subsequent time point. As Cole and Maxwell (2003)
pointed out, research designs where a predictor variable and supposed mediator variable are
measured concurrently do not truly test a meditational model. When measured concurrently, one
is not able to disentangle the causal ordering of the two variables, which is central to the
2

predictions of the model .
The design of the present study deviates from this ideal design in several important ways.
Racial identity was only measured during the last two waves of data collection. Although
perceived discrimination was measured at all three waves of data collection, the main outcome
variables of interest (self esteem and depression) were only measured at the first two waves of
data collection and were not measured at the third wave. Thus, a true test of a meditational model
that includes either self esteem or depression as an outcome is not possible given that these
variables were not measured at time 3.
Despite the limitations with the design of the current study, examining the predictions of
the Rejection Identification Model with the available longitudinal data is informative since it still
allows one to test some of the predicted relationships between variables of this model
longitudinally. When one has an incomplete design where only two waves of data are collected,
2 It is important to note that simply measuring two constructs at two different time points does
not ensure that the construct measured first actually precedes or causes the construct measured
at the second time point, especially when the timing of the measure does not reflect the timing
of the construct and changes in the construct (Cole & Maxwell, 2003). It is entirely possible
that the measure of a variable at a later time point actually assesses a construct or condition
that occurs prior to the variable at the first time point.
37

Cole and Maxwell (2003) suggest that the most appropriate way to test a meditational model is to
conduct two analyses; one where the wave 2 mediator is regressed on to the wave 1 predictor
while controlling for the wave 1 mediator, and one where the wave 2 outcome is regressed onto
the wave 1 mediator, while controlling for wave 1 outcome. The product of these two estimates
would be the estimate of the meditational (indirect) effect. This procedure is advantageous
because it allows one to statistically control for prior levels of the mediator and outcome
variables in a regression but requires that both the mediator and the outcome variables to be
measured at both waves being examined. Thus we are unable to perform this procedure given
that racial identity and the various outcome variables were never measured concurrently for two
consecutive waves in this study. However, educational attainment, monthly income and job
satisfaction were measured at time 3, this design allows for an actual test of the Rejection
Identification Model which includes these latter variables as outcomes in the model.
In order to test the predictions of the Rejection Identification Model across waves, we
examined whether racial identity at time 2 meditated any associations between time 1 perceived
discrimination and time 2 self esteem, time 2 depression, time 3 education, time 3 income, and
time 3 occupational satisfaction. We also examined whether racial identity at time 2 meditated
any associations between time 2 perceived discrimination and time 3 education, time 3 income,
and time 3 occupational satisfaction. In each of these models, perceived discrimination was
entered as the predictor, time 2 racial identity was entered as the mediator, and one of the
outcome variables listed above was entered as the criterion variable. Overall, the results of these
analyses indicated that the indirect effect via racial identity was not significant across any of the
relationships between the variables listed here (see Table 7). Thus, we failed to find evidence that
racial identity mediated any of the aforementioned cross-wave predictor-criterion associations, as

38

predicted by the Rejection Identification Model.
The failure to find significant indirect effects in these analyses may be attributable to the
fact that the variables themselves were relatively weakly correlated across waves (see Table 1).
For instance, we failed to find evidence of a significant association between perceived
discrimination at time 1 and racial identity at time 2, or between perceived discrimination at time
2 and racial identity at time 3, which were key predictions of these meditational models. The
associations between perceived discrimination at all time points and all outcome variables were
either very small or non-significant.
Cross-wave analyses: Buffering Hypothesis.
In order to test the predictions of the Buffering Hypothesis across waves, we examined
whether racial identity at time 2 interacted with time 1 perceived discrimination to predict time 2
self esteem, time 2 depression, time 3 education, time 3 income, and time 3 occupational
satisfaction. We also examined whether racial identity at time 2 interacted with time 2 perceived
discrimination to predict time 3 education, time 3 income, and time 3 occupational satisfaction.
Finally, we also examined whether time 3 racial identity interacted with either time 1 or time 2
perceived discrimination to predict time 3 education, time 3 income, and time 3 occupational
satisfaction.
Self Esteem. Results (shown in Table 8) indicated that, in addition to a significant main
effect of time 1 perceived discrimination and time 2 racial identity, the interaction between time
1 perceived discrimination and time 2 racial identity was significant in predicting time 2 self
esteem. In order to break down the interaction, we examined the simple slopes for individuals
who reported low and high racial identity, respectively. These analyses indicated that for
individuals who reported low racial identity at time 2, perceived racial discrimination at time 1

39

was associated with increased self esteem at time 2. In contrast, for individuals who reported
high racial identity at time 2, perceived racial discrimination at time 1 was also associated with
increased self esteem at time 2, but this association was weaker.
Depression. Results (shown in Table 8) indicated that there was a significant main effect
of time 1 perceived discrimination in predicting time 2 depression. However, the main effect of
time 2 racial identity did not reach significance. Further, the interaction between time 1
perceived discrimination and time 2 racial identity was significant in predicting time 2
depression. . In order to break down the interaction, we examined the simple slopes for
individuals who reported low and high racial identity, respectively. These analyses indicated that
for individuals who reported high racial identity at time 2, perceived racial discrimination at time
1 was associated with increased depression at time 2. In contrast, for individuals who reported
low racial identity at time 1, perceived racial discrimination at time 1 was not significantly
associated with depression at time 2.
Educational and career outcomes. Overall, results indicated that there were no
significant interactions between time 1 perceived discrimination and time 2 racial identity, time 2
perceived discrimination and time 2 racial identity, time 1 perceived discrimination and time 3
racial identity, and time 2 perceived discrimination and time 3 racial identity in predicting
income, job satisfaction, and education at time 3 (see Table 8).
Cross-wave analysis: Discounting hypothesis
The data available in this dataset did not allow for a cross wave analysis of the
Discounting Hypothesis. The predictor variable in this model is racial identity, and this variable
was not measured at time 1. However, we did examine the extent to which perceived
discrimination at time 2 meditated any associations between time 2 racial identity and time 3

40

education, time 3 income, and time 3 occupational satisfaction. In each of these models, time 2
racial identity was entered as the predictor, time 2 perceived discrimination was entered as the
mediator, and one of the outcome variables listed above was entered as the criterion variable.
Overall, the results of these analyses indicated that the indirect effect via perceived
discrimination was not significant across any of the relationships between the variables listed
here (see Table 9). Thus, we failed to find evidence that perceived discrimination mediated any
of the aforementioned cross-wave predictor-criterion associations, as predicted by the
Discounting Hypothesis.
Discussion
The present study examined the Rejection Identification Model, Buffering Hypothesis,
and Discounting Hypothesis as theoretical explanations of the relationship between perceived
discrimination, racial identity and well-being. In order to test these theories, we examined the
predictions of each theory regarding the specific relationships between perceived discrimination,
racial identity and various well-being outcomes both cross-sectionally and across time using a
large, longitudinal sample of children of immigrants living in the United States. Overall, the
results of this study lent the greatest support for the Rejection Identification Model and some
support for the Discounting Hypothesis, but these results must be interpreted with caution given
that all the effect sizes garnered in these analyses were so small. We failed to find any support for
the Buffering Hypothesis across all analyses. In contrast, our longitudinal results indicated that
racial identity actually exacerbated the negative effects of perceived discrimination on wellbeing. The theoretical implications of these latter findings are discussed below.
Cross-Sectional Results
The clearest support for the Rejection Identification Model comes from the time 2 finding

41

that racial identity was positively associated with perceived discrimination and self esteem and
suppressed the negative association between perceived discrimination and self esteem. Although
this pattern of results is consistent with the predictions of the Rejection Identification Model, one
must be cautious in the interpretation of these results given that these analysis garnered such
small effect sizes. For example, the standardized parameter estimates for the Rejection
Identification Model in Branscombe et al.’s (1999) original study were approximately ten times
larger than the coefficients found in the current study.
The results of this study also indicated that there was a significant indirect effect from
racial identity on depression via perceived discrimination. Although at first glance this finding is
consistent with the Discounting Hypothesis, the size of this indirect effect is so small (<.001) that
these results are not likely not meaningful or informative. Thus, this effect will not be discussed
any further.
One factor contributing to the weakness of these effects could be weaknesses in the
measures of the study variables. For example, the perceived discrimination measure used in the
first two waves is an imprecise measure of perceived discrimination. It is a dichotomous measure
that measures the presence or absence of perceived discrimination, which is problematic because
it fails to account for the magnitude of perceived discrimination. Based on this classification
scheme, individuals who perceive themselves to experience a large degree of discrimination and
those who perceive themselves to experience little discrimination are categorized the same and
are analytically indistinguishable, which discards meaningful variance in the data. Another
problem with this measure is that it does not directly reflect perceived racial discrimination.
After indicating whether or not they have felt discriminated against, participants were asked to
indicate the main reason they felt that this discrimination occurred. Assessing racial

42

discrimination in this manner may fail to capture individuals that have experienced a substantial
amount of racial discrimination, but do not consider racial discrimination to be the main form of
discrimination they experience. Thus, instead of reflecting whether or not individuals perceive
themselves to experience racial discrimination, this measure reflects whether or not individuals
have experienced general discrimination and whether they feel that racial or ethnic background is
the main reason, among others, for this discrimination.
Although this study, and virtually every other past study that was reviewed in the current
paper, has examined the relationship between racial identity, well-being, and perceived
discrimination, there is a conceptual distinction between the notions of perceived discrimination
and actual experiences of discrimination. Perceived discrimination is characterized by a
subjective process where negative events and experiences are attributed to discrimination by the
target. In contrast, assessments of actual experiences of discrimination and discriminatory
behaviors (which may be measured using various forms of behavioral assessment) are seldom
used to operationalize racial discrimination. One possible reason for this may be the difficulty in
measuring actual experiences of discrimination. Reports of actual discriminatory behavior are
subjective and may be difficult to separate from perceived discrimination. Further, perceived
discrimination measures are likely influenced by both actual discrimination experiences and an
individuals’ perceptions of discrimination from others, even though they are considered two
theoretically distinct constructs. The measures of perceived discrimination used in this study
capture both perceived discrimination and actual discrimination. This is because in order to
perceive discrimination and attribute negative experiences to discrimination, an individual must
have actually encountered negative, potentially discriminatory experiences to make attributions
about. Thus, it is the case that anyone who reports that they perceive themselves to experience

43

discrimination also has likely experienced some form of negative treatment or actual
discriminatory behaviors. Both the Rejection Identification Model and the Discounting
Hypothesis are conceptualized using the notion of perceived discrimination. However, the
Buffering Hypothesis does not explicitly speak to whether it conceptualizes racial discrimination
as perceived discrimination or actual discrimination, so one may argue that it could use both
constructs. Further, all three models implicitly assume that perceived discrimination and actual
discrimination are separable constructs; hence, our inability to empirically separate these
constructs in the current dataset limits the interpretability and applicability of the current results
to the three theoretical models being examined. Thus, it is possible that the weak effects garnered
in the present analyses (discussed above), and the lack of significant effects in other analyses,
could be due to our inability to separate out actual experiences of discrimination from
attributions to racial discrimination in our perceived discrimination measure.
The cross sectional analyses examining the Rejection Identification Model also indicated
that there were significant indirect effects of perceived discrimination on monthly income and
years of education via racial identity at time 3. Analyses examining the predictions of the
Discounting Hypothesis at time 3 also revealed significant indirect effects from racial identity to
income, job satisfaction, and education via perceived discrimination. However, these results are
not consistent with either the Rejection Identification Model or the Discounting Hypothesis
given that racial identity was negatively related to monthly income and educational attainment,
and perceived discrimination was negatively associated with job satisfaction (which is
inconsistent with the Discounting Hypothesis).
It is important to note that none of the theories examined in the present study made
specific predictions regarding career or educational outcomes and the results regarding these

44

outcomes were not consistent with the models examined in the present study. However, these
results were consistent with other theory drawn from past research, so considering the
implications of these results still may have theoretical and informational value. For example,
research has shown that there are associations between racial self-categorization and income
(Darity, Mason, & Stewart, 2000; Mason, 2004), but that the nature of this relationship depends
on the prevailing social norms surrounding racial self-categorization in that context. Particularly,
Mason (2004) posits that when acculturation norms are present (i.e., when the norm is to
acculturate to larger society), acculturation to the racial majority and self categorizing as White
will increase the income of persons of Mexican decent whereas identifying with a unique
Hispanic identity will decrease income. However, when contextual norms dictate identification
with one's ethnic group, acculturation and self-categorization as white has a negative impact on
the income of persons of Mexican decent because this type of self-categorization contradicts
prevailing social conventions. Although there is no measure of what the normative attitudes
towards acculturation surrounding the individuals in this sample, if an acculturation norm was
present for this sample, then these findings would be consistent with the findings of Mason's
(2004) study regarding the negative relationship between racial identity and income. Thus,
identifying with one's racial or ethnic group in this case may contradict the acculturation norms
surrounding these individuals, which in turn leads to decreases in income.
The results regarding the negative relationship between years of education, income and
racial identity may also be explained by an alternative causal model where increases in income
and years of education cause decreases in racial identity. It is possible that as individuals gain
more education and more income, the resulting increases in social status may lead to less contact
with other members of their racial minority groups, which in turn makes expression of one's

45

racial identity less normative. Individuals that occupy the social situation surrounding increased
education and income may not have developed a strong racial identity if they have had less
opportunity to explore their racial identity or it was not necessary to explore or commit to a
particular racial identity because they have had little need to consider the implications of their
racial or ethnic group. Further, low identification with one's racial or ethnic group may also
benefit individuals in that it may allow increased assimilation into the their social surroundings
and greater ease of following any acculturative norms that exist in these social groups (Mason,
2004). Other research also suggests that individuals high in racial identity may be the target of
increased prejudice and discrimination (Kaiser & Pratt-Hyatt, 2009). In turn, this increased
discrimination from others may lead to fewer opportunities in education and higher earning jobs,
which would be consistent with the negative relationship between education, income and racial
identity observed in the present data.
Longitudinal Results
Although there was evidence for the Rejection Identification Model in the cross-sectional
analyses, we failed to find any longitudinal evidence for this model. Further, the lack of self
esteem or depression measures in the third wave of the study (two key outcomes in the current
study) prevents testing of the true longitudinal mediation model with these particular variables as
outcomes. The lack of a racial identity measure at time 1 also prevented a true longitudinal test
of the mediational model predicted by the Discounting Hypothesis as racial identity is the
predictor variable in this model. Other longitudinal analyses regarding the Discounting
Hypothesis also failed to find a significant indirect effect from time 2 racial identity to income,
education and job satisfaction via perceived discrimination at time 2. As discussed previously,
failure to find significant indirect effects in these analyses may be attributable to limitations with

46

the measures used in this study, particularly the perceived discrimination measure. Further, the
fact that the variables themselves were also weakly correlated across waves (see Table 1) may be
another factor explaining why there were no significant effects. For instance, we failed to find
evidence of a significant association between perceived discrimination at time 1 and racial
identity at time 2, or between perceived discrimination at time 2 and racial identity at time 3,
which were key predictions of these models. The associations between perceived discrimination
at all time points and all outcome variables were either very small or non-significant.
The longitudinal nature of these data does not strengthen the test of the Buffering
Hypothesis, given that it predicts that racial identification and perceptions of discrimination have
an interactive effect on well-being outcomes and does not make explicit longitudinal predictions
regarding the relationship between these variables. Despite this, we examined the predictions of
the Buffering Hypothesis across waves simply as a further test of this model. This set of analyses
garnered two significant interactions; one being the interaction between time 1 perceived
discrimination and time 2 racial identity in predicting time 2 self esteem, the other being the
interaction between time 1 perceived discrimination and time 2 racial identity in predicting time
2 depression. Neither of these patterns of results is consistent with the Buffering Hypothesis. The
results involving self esteem as an outcome suggest that perceived discrimination at time 1 is
positively related to self esteem at time 2, and that racial identity attenuates this positive
relationship. In other words, individuals that report that they experience time 1 perceived
discrimination also report higher self esteem at time 2 in general, but the association between
perceived discrimination and self esteem is lower, but still positive, for individuals with strong
racial identity. This finding is not predicted by the Buffering Hypothesis, and stands in contrast
with the vast majority of research which indicates that perceived discrimination is associated

47

with decreased well-being (e.g., Jasinskaja-Lahti & Liebkind, 2001; Wong et al., 2003). This
finding is consistent with one of the primary predictions of the discounting perspective
suggesting that perceptions of discrimination are negatively related to well-being (Crocker &
Major, 1989). Proponents of the Discounting Hypothesis suggest that attributing negative
treatment from others to the prejudice of others against your racial group allows one to discount
the self as the reason for this negative treatment. In other words, if one is able to attribute
negative events one experiences to prejudice and discrimination, this protects self esteem and
other important well-being outcomes. This may, at least in part, explain the positive association
between perceived discrimination and well-being observed in the current data. The fact that
racial identity attenuates the positive relationship between perceived discrimination and self
esteem may also reflect an inability of the individuals with high racial identity to discount the
self when attributing negative experiences to racial discrimination given that for these
individuals, race and ethnicity is a central part of the self-concept.
The results involving depression as an outcome suggest that racial identity at time 2
exacerbates the association between perceived discrimination at time 1 and higher depression at
time 2, which stands in contrast to the prior results (described above) showing that perceived
discrimination is associated with benefits to one’s self esteem. These findings regarding the role
of racial identity are also not consistent with the predictions of the Buffering Hypothesis. Despite
this, these results are consistent with past research that suggests that negative experiences related
to important aspects of the self (characterized by Thoits [1991] as “identity-related stressors”) are
especially distressing for individuals. Further, some past research has also found that racial
identity exacerbates the negative effect of perceived discrimination (e.g., Yip et al., 2008). For
example, Yip and colleagues (2008) found that racial identity exacerbated the effects of

48

perceived discrimination on distress for Asian Americans between 31- 40 years of age or 51-75
years of age. Although the reasons for why racial identity serves to buffer the negative effects of
perceived discrimination in some instances and not in others is not clear, Yip et al. (2008)
surmised that these may be sensitive periods of development and several significant life
changing events are particularly likely to happen during these periods in life. This reasoning
could also explain why racial identity exacerbates the negative effects of perceived
discrimination on depression in this study. It could be argued that adolescence is a particularly
sensitive period given that it is theorized to be the time where individuals engage in identity
exploration and development. Individuals that report increased racial identity during this period
may be especially susceptible to the negative outcomes associated with perceived discrimination
such as depression if they have not fully explored this racial identity or developed other
protective aspects of their identities. Thus, when adolescents are high in racial identity,
experiencing racial discrimination may be particularly damaging to psychological well-being
since they may be ill prepared to face this type of identity related stressor.
It is important to note that the explanations for these effects are merely speculative and
further work would need to be done in order to understand these relations. Additionally, it is
unclear why perceived discrimination is positively related to both self esteem and depression,
and it is unclear if any conclusions can be made regarding the role of racial identity. Thus, given
the small sizes and the contrasting nature of the effects across these analyses, we are hesitant to
draw firm conclusions as to the status of the Buffering Hypothesis (or any of the three models)
based on the data presented here.
Limitations and Future Directions
Despite the strengths of using a large, longitudinal dataset to examine the research

49

questions presented in the current paper, this study is not without its limitations. Use of a
preexisting dataset limits the control a researcher has as to the variables that are included in the
study and the way in which constructs are assessed and coded. As discussed above, there are
several reasons that the perceived discrimination measure used in this study is problematic.
These reasons, coupled with the fact that the measure was not worded the same way across
waves, leads one to question the utility of this measure. Further, many other relevant variables,
such as measures of attributions of prejudice, would have helped evaluate these models but these
constructs were not included in the current dataset. Measures of key variables, including racial
identity, self esteem, and depression were not included at each wave of data collection, which
limited the analyses we were able to perform to examine this study’s main research questions.
Further, because this was a longitudinal study, selective attrition may be a concern as it is
possible that there were important differences between individuals that remained in the study
versus those who dropped out. These issues are typically seen as trade-offs inherent in any study
that involves secondary data analysis where one is trading off lack of control over the design of
the study for the benefits associated with use of relatively large, longitudinal dataset without the
burden of collecting such a sample. Future research may benefit from making use of other large
longitudinal datasets to explore the relationship between racial identity, perceived discrimination
and well-being. For instance, the German Socio-Economic Panel Study (GSOEP) is a large,
nationally representative sample of German households which includes data from over 40,000
individuals spanning multiple decades. The GSOEP includes a diverse body of questions,
including questions related to experiences of prejudice and discrimination, racial identity and
subjective well-being. It may be fruitful to explore the relationship between variables using
samples of immigrants and racial minorities in the GSOEP.

50

Another potential limitation is of this study involves the use of single item measures of
racial identity. Typically, single item measures are criticized because they are not as internally
consistent as expected in psychological measurement, especially when compared to other
measures of greater length. Use of brief measures in large scale panel studies are often necessary
to minimize participant burden given the usual length of these type of studies. Some critics
would also argue that a single item measure of racial identity lacks the bandwidth necessary to
adequately capture the construct of racial identity. These issues are important to keep in mind
when interpreting the results as there has been no research examining the reliability of this single
item measure of racial identity, nor has there been research examining the validity of such a
measure of racial identity. When comparing the effect sizes garnered using multi-item measures
of racial identity in past research to the results in the present study using a single item measure,
one can see that the effect sizes are substantially smaller in the present study. This may be
indicative of problems with the measurement of the racial identity in this study. Although use of
this single item measure may not necessarily be problematic, it is important for future research to
examine the validity and reliability of the brief racial identity measures like the one used in the
current study. Research showing that these measures have acceptable psychometric properties
would not only lend weight to the results of research using these type of measures, but may
promote the use of these measures in future research. As brief measures like this are typically the
ones included in panel studies such the CILS or the GSOEP (discussed above), validating the use
of such measures may also encourage scholars interested in the construct of racial identity to use
available panel data to examine the research questions they are interested in.
Another potential limitation of the present study is that it includes participants from
multiple racial and ethnic groups, thus it is possible that there are differences in how racial

51

identity, perceived discrimination and well-being relate to one another among individuals from
different racial and ethnic groups that were not explored in this study. As a preliminary
examination of this, we computed the correlations across study variables for Hispanic, Asian, and
Black participants independently and found a few minor differences in the correlations between
the outcome variables across racial groups, but in general the correlations between the study
variables were the same direction and size across groups. Although at first glance this may
suggest that there are no differences in how racial identity, perceived discrimination, and wellbeing are related across groups, more detailed analyses should be conducted in order to examine
this question. Future research may examine how racial or ethnic group membership moderates
the associations between these variables and the processes through which these variables are
related. It is also possible that the processes through which racial identity relates to perceived
discrimination and well-being differs by generation status of an individual (e.g., whether one is
foreign born, born in the United States to immigrant parents, or born within the United States to
US born parents). Thus, it would likely be fruitful for future research to also examine whether
there is a moderating effect of generational status on the relationship between racial identity,
perceived discrimination and well-being.
Conclusion
The results of this study were mixed and the findings did not show strong support for any
particular model. The data did not support all the key predictions for any of the theoretical
models examined here. Cross-sectionally, there was the most support for the Rejection
Identification Model, some limited support for the Discounting Hypothesis, and no support for
the Buffering Hypothesis. However, it is important to note that the effect sizes garnered in all the
analyses presented here were very small and prevent firm conclusions from being drawn from

52

this study. Longitudinal analyses did not provide evidence for any of the models. Future research
would benefit from examine the relationships between perceived discrimination, racial identity,
and well-being outcomes in greater detail as the way in which these variables are related remains
an open empirical question.

53

APPENDIX

54

Figure 1. The Rejection Identification Model.

55

Figure 2. The Buffering Hypothesis

56

Figure 3. The Discounting Hypothesis

57

Table 1
Demographic Characteristics of the Full (Wave 1)
Sample
Variable
Gender
Female
Male

n (%)
2687 (51.1)
2575 (48.9)

Citizenship Status
Citizen
Non-Citizen
Unknown

3335 (63.4)
1410 (26.8)
517 (9.8)

Born in the United States
Yes
No
Unknown

2630 (50.0)
2628 (49.9)
4 (0.1)

US Stay Length
Whole Life
Over 10 years
5 - 9 years
Less than 5 years
Unknown

2420 (46.0)
1387 (26.4)
1127 (21.4)
326 (6.2)
2 (0.0)

58

Table 2
Variables Measured by Wave of Data Collection
Measure

Wave 1

Wave 2

Wave 3

Racial Identity

-

X

X

Perceived
Discrimination

X

X

X

Well-being
Self esteem
Depression

X
X

X
X

-

Job Satisfaction
Monthly Income
Years of Education

-

-

X
X
X

59

60

61

62

63

64

65

66

67

68

69

70

REFERENCES

71

References
Anderson, N. B. (1989). Racial differences in stress-induced cardiovascular reactivity and
hypertension: Current status and substantive issues. Psychological Bulletin, 105, 89-105.
Aries, E., & Moorehead, K. (1989). The importance of ethnicity in the development of identity of
Black adolescents. Psychological Reports, 65, 75-82.
Armstead, C., Lawler, K., Gordon, G., Cross, J., & Gibbons, J. (1989). The relationship of racial
stressors to blood pressure responses and anger expressions in ebon college students. Health
Psychology, 8, 542-556.
Ashmore, Deaux & McLaughlin-Volpe (2004). An organizing framework for collective identity:
Articulation and significance of multidimensionality. Psychological Bulletin, 130, 80-114.
Bat-Chava, Y. (1994). Group identification and self-esteem of deaf adults. Personality and Social
Psychology Bulletin, 20, 473-483.
Barnes, L. L., Mendes de Leon, C. F., Lewis, T. T., Bienias, J. L., Wilson, R. S., & Evans, D. A. (2008).
Perceived discrimination and mortality in a population-based study of older adults. American
Journal of Public Health, 98, 1241-1247.
Benet-Martínez, V., Leu, J., Lee, F., & Morris, M. (2002). Negotiating biculturalism: Cultural
frame-switching in biculturals with 'oppositional' vs. 'compatible' cultural identities. Journal
of Cross-Cultural Psychology, 33, 492-516.
Bourguignon, D., Seron, E., Yzerbyt, V., & Herman, G. (2006). Perceived group and personal
discrimination: Differential effects on self-esteem. European Journal of Social Psychology,
36, 1-24.
Branscombe, N., Schmitt, M., & Harvey, R. (1999). Perceiving pervasive discrimination among
African Americans: Implications for group identification and well-being. Journal of
Personality and Social Psychology, 77, 135-149.
Broman, C. L., Mavaddat, R., & Hsu, S. (2000). The experience and consequences of perceived racial
discrimination: A study of African Americans. Journal of Black Psychology, 26, 165-180.
Clark, R., Anderson, N. B., Clark, V. R., & Williams, D. R. (1999). Racism as a stressor for African
Americans: A biopsychosocial model. American Psychologist, 54(10), 805– 816.
Cole, D. A., & Maxwell, S. E. (2003). Testing meditational models with longitudinal data:
Questions and tips in the use of structural equation modeling. Journal of Abnormal
Psychology, 112, 558-577.
Cozzarelli, C., & Karafa, J. A. (1998). Cultural estrangement and terror management theory.
Personality and Social Psychology Bulletin, 24, 253-267.
Crocker, J., Luhtanen, R. Blaine, B., & Broadnax, S. (1994). Collective self-esteem and
psychological well-being among White, Black, and Asian college students. Personality and
72

Social Psychology Bulletin, 20, 503-513.
Crocker, J., & Major, B. (1989). Social stigma and self-esteem: The self-protective properties of
stigma. Psychological Review, 96, 608-630.
Cross, W. (1978). The Thomas and Cross models of psychological Nigrescence: A review. The
Journal of Black Psychology, 5, 13-31.
Darity, Jr., W., Mason., P. L., & Stewart, J. (2006). The economics of identity: The Origin and
persistence of racial identity norms. Journal of Economic Behavior and Organization, 60, 285305.
Dion, K. L., Dion, K. K., & Pak, A. W. (1992). Personality-based hardiness as a buffer for
discrimination-related stress in members of Toronto's Chinese community. Canadian Journal of
Behavioural Science, 24, 517-536.
Dovidio. J. F., Kawakami, K., & Gaertner, S. L. (2002). Implicit and explicit prejudice and interracial
interaction. Journal of Personality and Social Psychology, 82, 62-68.
Eccleston, C.P. & Major, B. (2006). Attributions to discrimination and self-esteem: The role of group
identification and appraisals. Group Processes and Intergoup Relations. 9,147-162.
Ellemers, N. (1993). Influence of socio-structural variables on identity enhancement strategies.
European Review of Social Psychology, 4, 27–57.
Erikson, E.H. (1968). Identity: Youth and Crisis. New York: Norton.
Fisher, C. B. Wallace, S. A. & Fenton, R. E. (2000). Discrimination distress during adolescence.
Journal of Youth and Adolescence, 29, 679-695.
Frable, D. E. S, Blackstone, T., & Scherbaum, C. (1990). Marginal and mindful: Deviants in social
interactions. Journal of Personality and Social Psychology, 59, 140-149.
Forman, T. A. (2003). The social psychological costs of racial segmentation in the workplace: A study
of African Americans' well-being. Journal of Health and Social Behavior, 44, 332-352.
Grossman, B., Wirt, R. & Davids, A. (1985). Self-esteem, ethnic identity, and behavioral adjustment
among Anglo and Chicano adolescents in West Texas. Journal of Adolescence, 8, 57-68.
Guyll, M., Matthews, K.A., & Bromberger, J.T. (2001). Discrimination and unfair treatment:
Relationship to cardiovascular reactivity among African American and European American
women. Health Psychology, 20, 315-325.
Harrell, S. P. (2000). A multidimensional conceptualization of racism-related stress: Implications for
the well-being of people of color. American Journal of Orthopsychiatry, 70, 42-57.
Hogg, M. A. (2005). The social identity perspective. In S. Wheelan (Ed.), The handbook of group
research and practice (pp.133-157). Thousand Oaks, CA: Sage.
73

Hu, L., & Bentler, P. M. (1999). Cuttoff criteria for fit indexes in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1-55.
Jasinskaja-Lahti, I., & Liebkind, K. (2001). Perceived discrimination and psychological adjustment
among Russian-speaking immigrant adolescents in Finland. International Journal of
Psychology, 36, 174–185.
Kessler, R. C., Mickelson, K. D., & Williams, D. R. (1999). The prevalence, distribution, and mental
health correlates of perceived discrimination in the United States. Journal of Health and Social
Behavior, 40, 208–230.
Klonoff, E. A., Landrine, H., & Ullman, J. B. (1999). Racial discrimination and psychiatric symptoms
among Blacks. Cultural Diversity and Ethnic Minority Psychology, 5, 329-339.
Liebkind, K., & Jasinskaja, I. (2000). The influence of experiences of discrimination on psychological
stress: A comparison of seven immigrant groups. Journal of Community and Applied Social
Psychology, 10, 1–16.
Marcia, J. E. (1966). Development and validation of ego identity status. Journal of Personality and
Social Psychology, 3, 511-558.
Marcia, J. E. (1980). Identity in adolescence. In J. Andelson (Ed.), Handbook of Adolescent
Psychology (pp. 109-137). New York: Wiley.
Major, B., Kaiser, C. R., & McCoy, S. K. (2003). It’s not my fault: When and why attributions to
prejudice protect self-esteem. Personality and Social Psychology Bulletin, 29, 772-781.
Major B., Quinton W.J., Schmader T. (2003). Attributions to discrimination and self-esteem: Impact of
social identification and group ambiguity. Journal of Experimental Social Psychology, 39, 220231.
Major, B., Spencer, S. J., Schmader, T., Wolfe, C. T., & Crocker, J. (1998). Coping with negative
stereotypes about intellectual performance: The role of psychological disengagement.
Personality and Social Psychology Bulletin, 24, 34-50.
Mason, P. L. (2004). Annual income, hourly wages, and identity formation among Mexican Americans
and other Latinos. Industrial Relations, 43, 817-834.
McCoy, S. K. & Major, B. (2003). Group identification moderates emotional responses to perceived
prejudice. Personality and Social Psychology Bulletin, 29, 1005-1017.
McKenna, K. Y. A., & Bargh, J. A. (1998). Coming out in the age of the Internet: Identity demarginalization' from virtual group participation. Journal of Personality and Social
Psychology, 74, 681-694.
Munford, M. (1994). Relationship of gender, self-esteem, social class, and racial identity to depression
in Blacks. Journal of Black Psychology, 20, 157-174.
74

Mokgatlhe, B. P., & Schoeman, J. B. (1998). Predictors of satisfaction with life: The role of racial
identity, collective self-esteem and gender-role attitudes. South African Journal of Psychology,
28, 28-35.
Muthén, L. K., & Muthén, B. O. (1998 –2007). Mplus user’s guide (5th ed.). Los Angeles: Author.
Neblett, E.W., Shelton, J.N., & Sellers, R.M. (2004). The role of racial identity in managing daily
hassles. In G. Philogene (Ed.). Race and Identity: The Legacy of Kenneth Clark (pp. 77-90).
Washington, DC: American Psychological Association Press.
Norton, M. I., Sommers, S. R., Vandello, J. A., & Darley, J. M. (2006). Mixed motives and racial
bias: The impact of legitimate and illegitimate criteria on decision making. Psychology,
Public Policy, and Law, 12, 36-55.
Oakes, P. J., Haslam, S. A., & Turner, J. C. (1994). Stereotyping and social reality. Oxford, UK &
Cambridge, MA: Blackwell.
Operario, D., & Fiske, S. T. (2001). Ethnic identity moderates perceptions of prejudice: Judgments
of personal versus group discrimination and evaluations of subtle versus blatant bias.
Personality and Social Psychology Bulletin, 27, 550-561.
Parkins, I. S., Fishbein, H. D., & Ritchey, P. N. (2006). The influence of personality on workplace
bullying and discrimination. Journal of Applied Social Psychology, 36, 2554-2577.
Phinney, J. S. (1989). Stages of ethnic identity development in minority group adolescents. Journal of
Early Adolescence, 9, 34-49.
Phinney, J. S. (1992). The multigroup ethnic identity measure: A new scale for use with diverse
groups. Journal of Adolescent Research, 7, 156–176.
Phinney, J. S. (1996). When we talk about American ethnic groups, what do we mean? American
Psychologist, 51, 918–927.
Phinney, J. S. & Alipuria, L. L. (1990). Ethnic identity in college students from four ethnic groups.
Journal of Adolescence, 13, 171–183.
Phinney, J. S. & Chavira, V. (1992). Ethnic identity and self-esteem: An exploratory longitudinal
study. Journal of Adolescence, 15, 271-281.
Phinney, J. S., & Tarver, S. (1988). Ethnic identity search and commitment in Black and White
eighth graders. The Journal of Early Adolescence, 8, 265-277.
Pascoe, E. A., & Smart Richman, L. (2009). Perceived discrimination and health: A meta-analytic
review. Psychological Bulletin, 135, 531-554.
Portes, A.,& Rumbaut, R.G. (2006). Children of Immigrants Longitudinal Study (CILS), 1991-2006
[Data file and code book].Retrived from Inter-university Consortium for Political and Social
75

Research [distributor], Ann Arbor, MI. doi:10.3886/ICPSR20520.
Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general
population. Applied Psychological Measurement, 1, 385-401.
Rosenberg, M. (1965). Society and the Adolescent Self-Image. Princeton, New Jersey: Princeton
University Press.
Rosenbloom, S. R., & Way, N. (2004). Discrimination among African American, Asian American, and
Latino adolescents in an urban high school. Youth & Society, 35, 420-451.
Rowley, S. A. J., Sellers, R. M., Chavous, T. M., & Smith, M. (1998). The relationship between racial
identity and self-esteem in African American college and high school students. Journal of
Personality and Social Psychology, 74, 715-724.
Schmitt, M.T., & Branscombe, N.R. (2002). The causal loci of attributions to prejudice. Personality
and Social Psychology Bulletin, 28, 620-628
Schmitt, M.T., Branscombe, N.R., Kobrynowicz, D., & Owen, S. (2002). Perceiving discrimination
against one’s gender group has different implications for well-being in women and men.
Personality and Social Psychology Bulletin, 28, 197-210.
Schmitt, M.T., Spears, R., & Branscombe, N.R. (2003). Constructing a minority group identity out
of shared rejection: The case of international students. European Journal of Social
Psychology, 33, 1-12.
Sidanius, J., & Pratto, F. (1999). Social dominance: An intergroup theory of social hierarchy and
oppression. New York, NY: Cambridge University Press.
Sellers, R.M., Caldwell, C.H., Schmeelk-Cone, K.H., & Zimmerman, M.A. (2003). The role of
racial identity and racial discrimination in the mental health of African American young
adults. Journal of Health and Social Behavior, 44, 302-317.
Sellers, R. M., Copeland-Linder, N., Martin, P. P., & Lewis, R. L. (2006). Racial identity matters:
The relationship between racial discrimination and psychological functioning in African
American adolescents. Journal of Research on Adolescence, 16 (2), 187-216.
Sellers, R. M., Smith, M. A., Shelton, J. N., Rowley, S. A. J., & Chavous, T. M. (1998).
Multidimensional model of racial identity: A reconceptualization of African American racial
identity. Personality and Social Psychology Review, 2, 18-39.
Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies: New
procedures and recommendations. Psychological Methods, 7, 422–445.
Sobel, M. E. (1990). Effect analysis and causation in linear structural equation models.
Psychometrika, 55, 495–515.
Swim, J. K., Hyers, L. L., Cohen, L. L., Fitzgerald, D. C., & Bylsma, W. H. (2003). African American
college student’s experiences with everyday racism: Characteristics of and responses to these
incidents. Journal of Black Psychology, 29, 38-67.
76

Tajfel, H. & Turner, J. C. (1986). The social identity theory of intergroup behavior. In S. Worchel &
W. G. Austin (Eds.), Psychology of intergroup relations (2nd ed., pp.7-24). Chicago: NelsonHall.
Thoits, P. A. (1991). On merging identity theory and stress research. Social Psychology Quarterly,
54, 101-112.
Turner, J. C. (1999). Some current issues in research on social identity and self-categorization
theories. In N. Ellemers, R. Spears, & B. Doosje (Eds.), Social identity (pp.6-34). Oxford,
UK: Blackwell.
Williams, D. R., & Mohammed, S. A. (2009). Discrimination and racial disparities in health:
Evidence and needed research. Journal of Behavioral Medicine, 32, 20 – 47.
Williams, D. R., Neighbors, H. W., & Jackson, J. S. (2003). Racial/Ethnic Discrimination and Health:
Findings from Community Studies. American Journal of Public Health, 93, 200-208.
Wong, C. A., Eccles, J. S., & Sameroff, A. (2003). The influence of ethnic discrimination and ethnic
identification on African American adolescents’ school and socioemotional adjustment. Journal
of Personality, 71, 1197–1232
Yap, S. C. Y., Settles, I. H., & Pratt-Hyatt, J. S. (2011). Mediators of the relationship between racial
identity and life satisfaction among a community sample of African American women and
men. Cultural Diversity and Ethnic Minority Psychology, 17, 89-97.
Yip, T., Gee, G. C., Takeuchi, D. T. (2008). Racial discrimination and psychological distress: The
impact of ethnic identity and age among immigrant and United States-born Asian adults.
Developmental Psychology, 44, 787-800.
Ziegert, J. C. & Hanges, P. J. (2005). Employment discrimination: The role of implicit attitudes,
motivation, and a climate for racial bias. Journal of Applied Psychology, 90, 553-562.

77