-
The Effects of Changes in Racial Identity and Self-Esteem on
Changesin African American Adolescents Mental Health
Jelani MandaraNorthwestern University
Noni K. Gaylord-HardenLoyola University Chicago
Maryse H. RichardsLoyola University Chicago
Brian L. RagsdaleWalden University
This study assessed the unique effects of racial identity and
self-esteem on 259 African American adolescentsdepressive and
anxiety symptoms as they transitioned from the 7th to 8th grades
(ages 1214). Racial identityand self-esteem were strongly
correlated with each other for males but not for females. For both
males andfemales, an increase in racial identity over the 1 year
was associated with a decrease in the prevalence ofdepressive
symptoms over the same period, even with self-esteem controlled. It
was concluded that racialidentity may be as important as
self-esteem to the mental health of African American adolescents,
and itexplains variance in their mental health not associated with
feelings of oneself as an individual.
A critical developmental task of adolescence is theformation of
a cohesive and positive sense of self.Adolescents higher in
self-esteem tend to have bet-ter mental health and are more
resilient in the faceof adversity compared to those low in
self-esteem(Compas, Hinden, & Gerhardt, 1995; DuBois et
al.,2002). In the early research on African Americanidentity
development, self-esteem was commingledwith racial preferences and
racial identity to such adegree that it was difficult to determine
the impor-tance of ones beliefs and feelings about their per-sonal
attributes relative to their beliefs and feelingsabout their racial
group (Whaley, 1993). Counter tothis perspective, several theorists
have long arguedthat ones self-concept consists of an
individualidentity domain and a group identity domain,which are
related but also conceptually distinct(Cross, 1991; McAdoo, 1985;
Spencer, 1982). Thesetwo-factor models of identity suggest that
Afri-can American youth have the ability to separate
orcompartmentalize feelings toward their race from
feelings about themselves as individuals. Thus,many theorists
questioned the assumption thatones perceptions and feelings of
their racial group,or racial identity, is a proxy for ones
self-esteem(Cross, 1991; McAdoo, 1985; Spencer, 1982).
Once self-report measures became more stan-dard, two important
strands of research emerged.One strand directly assessed the
effects of self-reported self-esteem on African American
adoles-cent and adult mental health, without regard toracial
identity. Another strand attempted to furtherdisentangle racial
identity from self-esteem andexamine their interrelation, as well
as the impor-tance of racial identity to African American
adoles-cent and adult mental health in general (see Cross,1991, for
a review). Many important findings aboutthe identity development of
African Americansemerged from this research (Murray &
Mandara,2001). However, many of the most theoretically
andpractically important questions have yet to beanswered. For
instance, few studies have examinedthe joint influences of racial
identity and self-esteemon African American adolescents mental
health(McMahon & Watts, 2002). Their affects could becumulative
or one component could explain theeffects of the other.
Furthermore, virtually no stud-ies have examined the effect of
changes in racial
This research was funded by Grant R01-MH57938 from theNational
Institute of Mental Health awarded to Maryse H. Rich-ards. The
authors acknowledge the contributions of the projectdirectors,
Cathy Flynn, and Phil Hammack, and the students,school
administrators, teachers, and parents who volunteeredtheir time to
participate in this research.
Correspondence concerning this article should be addressed
toJelani Mandara, Program in Human Development and Social Pol-icy,
Northwestern University, 2120 Campus Drive, Evanston, IL60208.
Electronic mail may be sent to [email protected].
Child Development, November/December 2009, Volume 80, Number 6,
Pages 16601675
! 2009, Copyright the Author(s)Journal Compilation! 2009,
Society for Research inChildDevelopment, Inc.All rights reserved.
0009-3920/2009/8006-0007
-
identity and self-esteem on changes in mentalhealth.
Understanding how racial identity and self-esteem each affects
adolescent mental health overtime may have important implications
for develop-mental theory and prevention interventions. Thepurpose
of this study was to help fill these gapsand assess the unique
effects of racial identity andself-esteem on African American
adolescents men-tal health as they transitioned between the
seventhand eighth grades.
Self-Esteem Development During Adolescence
Most modern theorists define self-esteem as theaffective and
evaluative dimension of ones self-concept. It essentially refers to
ones generalizedappraisal of their positive or negative
attributesacross different domains (Harter, 1999). Researchsuggests
that self-esteem becomes most salient dur-ing the transition to
adolescence (Greene & Way,2005). A number of studies have
examined thedevelopmental trajectory of self-esteem during
ado-lescence, but little consensus exists. For instance,some
studies examined mean-level changes in self-esteem at multiple time
points and found increasesin self-esteem during adolescence (Hirsch
&Rapkin, 1987; OMalley & Bachman, 1983), whereasothers show
decreases (Brown et al., 1998). Onestudy of self-esteem using
growth curve analysisfound that self-esteem increased in
preadolescence,began decreasing in early adolescence, andincreased
again during late adolescence (Baldwin &Hoffman, 2002). Another
growth curve analysis ofself-esteem in a multiracial sample
provided evi-dence of a general increase in self-esteem for
ethnicminority adolescents during adolescence (Greene &Way,
2005). In a study of over 800 students fromvarious racial
backgrounds, researchers used theexperiential sampling method to
assess feelings ofself-worth and found that it dramatically
reducedbetween the 6th and 10th grades, and then began toslowly
increase by the end of high school (Moneta,Schneider, &
Csikszentmihalyi, 2001). Thus, the evi-dence suggests that
self-esteem greatly changesfrom early adolescence to later
adolescence, but thetrajectory of those changes are still
unclear.
Self-Esteem and African American AdolescentMental Health
Self-esteem is generally thought to be one of themost important
psychological constructs for adoles-cent mental health (Mann,
Hosman, Schaalma, & deVries, 2004). One prevalent theory
suggests that self-
esteem benefits adolescent mental health by actingas a
psychological buffer from negative environmen-tal stressors (Compas
et al., 1995). This is thought tooccur because high self-esteem
makes adolescentsmore emotionally resilient and feel as if they
havethe capability to overcome obstacles (Mann et al.,2004).
Therefore, having a positive image of oneselfas an individual may
be a particularly importantresiliency factor for African American
adolescentsgiven their exposure to environments that oftendevalue
their worth (Spencer, 1995).
Beginning in the early 1960s with the develop-ment of
self-report measures, a slew of studies onAfrican American
self-esteem and well-being testedthese assumptions. The accumulated
research isstrikingly clear. Self-esteem has been
negativelycorrelated with cigarette smoking, drug use, andother
risky behaviors among African Americanyouth (Botvin et al., 1993;
Unger, Kipke, Simon,Montgomery, & Johnson, 1997). Most
importantlyfor the current study, self-esteem is also
negativelyrelated to suppressed anger (Johnson & Greene,1991),
anxiety symptoms (Matthews & Odom, 1989;Youngstrom, Weist,
& Albus, 2003), and depressivesymptoms (Caldwell, Antonucci,
Jackson, Wolford,& Osofsky, 1997) in African American children
andadolescents. One recent study assessed over 1,500African
American adolescents and found that self-esteem was negatively
related to depressive symp-toms, even with family socioeconomic
status (SES),grades, and exposure to violence accounted
for(Fitzpatrick, Piko, Wright, & LaGory, 2005). More-over,
self-esteem was a stronger correlate of depres-sive symptoms than
any of the other variables.Another study of 681 low-achieving
AfricanAmerican adolescents found that self-esteem anddepressive
symptoms were related each of 4 years(Repetto, Caldwell, &
Zimmerman, 2004). Thus, aswith other adolescents, self-esteem has a
strong andconsistent relation with African American adoles-cents
mental health.
Racial Identity Development During Adolescence
Racial identity is a multidimensional constructthat has been
defined in a variety of ways. Themajor assessment devices include
subscales thatmeasure: (a) an understanding and acceptance ofones
socially constructed racial label, (b) knowl-edge of ones racial
groups history and culturalnorms, (c) a sense of shared activities,
(d) beliefsabout the social position of ones racial group, and(e)
feelings of pride regarding ones racial group(Phinney, 1992;
Sellers, Rowley, Chavous, Shelton,
Racial Identity and Self-Esteem 1661
-
& Smith, 1997; Vandiver, Cross, Worrell, & Fhagen-Smith,
2002). The nature of African Americansracial identity development
during adolescence iseven less clear than self-esteem
developmentbecause very few studies have assessed changes inracial
identity during this period. However, thereare two related
perspectives on African Americanracial identity development that
guide mostresearch in the area.
The cognitive development perspective suggeststhat racial
identity is developed as one becomescognitively able to make sense
of the multitude ofsocial messages about race they encounter
(Murray& Mandara, 2001; Wardle, 1992). Children in theearly
stages of cognitive development can identifyracial symbols such as
skin color and even haverudimentary understandings that such
symbols areassociated with social currency, but they do nothave
complex ideas about their own racial identity.The cognitive
perspective further suggests that aschildren mature into early
adolescence (i.e., ages1014), they begin to process societal
messagesabout preferences for certain phenotypic
features,race-based hierarchies in wealth and academictracks, and
the plethora of racial stereotypes theywill encounter (Murray &
Mandara, 2001; Spencer,1982, 1984). These messages are either
challengedor reinforced by parents racial socialization
strate-gies, peers, teachers, and media. Thus, as Murrayand Mandara
(2001) argue, ones cognitive readi-ness determines when they are
capable of develop-ing a more sophisticated notion of racial
identity,but the social context determines the form of
racialidentity they will develop.
A second set of theories propose stage or statusmodels of racial
identity development similar toEriksons (1968) stages of identity
development,although they have a few different assumptions(e.g.,
Cross, 1991; Helms, 2007; Phinney, 1989).These models suggest that
one is in the first stagewhen parental and societal beliefs about
ones raceand racial identity are accepted without
criticalreflection. African American parents can raise
theirchildren to have positive racial pride, but this per-iod is
usually described as a time of uncriticallyaccepting negative
stereotypes and beliefs aboutAfrican Americans (Phinney, 1989).
This initialstage is usually followed by some period of
explo-ration, where individuals question their earlier,usually
negative views and assumptions aboutAfrican Americans. The
exploration, or immersionstage in Crosss (1991) model, is
characterized by agreat desire to learn about ones heritage and
con-nect with members of their group. Individuals tend
to focus on the positive stereotypes of their groupwhile
downplaying and often reacting to negativestereotypes. According to
the stage models, the bestoutcome is for one to reach an achieved
stage inwhich they have attained a more secure sense ofoneself as a
member of their racial group. This finalstage is characterized by a
lack of defensivenessregarding racial matters and a genuine
comfortable-ness with the racial aspects of their identity
(Cross,1991; Phinney, 1989).
Although little evidence exists regarding the ageat which one
typically enters each stage, a fewstudies have shed light on this
question. Forinstance, Phinney & Tarver (1988) found thatalmost
30% of 8th-grade students could be classi-fied into the exploration
stage. Another study of10th-grade non-White students found that
55%could be classified as being in pre-explorationstages, 23% in
exploration, and 22% in an achievedstage of racial identity
(Phinney, 1989). Also consis-tent with most stage models, Pahl and
Way (2006)followed a small group of African American andLatino
adolescents throughout high school andfound that exploration began
to decelerate after the10th grade. Another recent study used
cluster ana-lytic methods and found they could classify
AfricanAmerican adolescents into four groups similar tothose
suggested by the stage models (Seaton,Scottham, & Sellers,
2006). They also found thatover a 1-year period, adolescents
changed acrossstages at a fairly high rate. Other longitudinal
stud-ies did not focus on the stage of identity develop-ment but on
growth in racialethnic pride duringthe adolescent years. One large
longitudinal studyfound that Canadian ethnic minority adolescents
inthe 7th to 10th grades exhibited a linear increase intheir
perceptions of ethnic pride over a 15-monthperiod (Perron,
Vondracek, Skorikov, Tremblay, &Corbie`re, 1998). Similarly, a
recent study usingcohorts of middle- and high-school-aged
adoles-cents found a significant linear increase in racialethnic
group esteem among African Americans andLatinos over a 3-year
period in both cohorts(French, Seidman, Allen, & Aber, 2006).
Therefore,although the evidence is limited, the few studies inthis
area suggest that African Americans racialidentity significantly
changes throughout theadolescent years.
Racial Identity and African American AdolescentMental Health
Similar to self-esteem and other aspects of onesidentity, racial
identity is considered by most
1662 Mandara, Gaylord-Harden, Richards, and Ragsdale
-
theorists to be of particular importance during theexplorative
phases of adolescence (Phinney, 1989).Many modern theorists suggest
that a positiveracial identity helps adolescents cope with
thestresses of discrimination (Caldwell, Zimmerman,Bernat, Sellers,
& Notaro, 2002; Rowley, Sellers,Chavous, & Smith, 1998;
Sellers, Copeland-Linder,Martin, & Lewis, 2006). For instance,
Spencers(1995) phenomenological variant of ecological sys-tems
theory (PVEST) suggests that normativedevelopment for African
American adolescentsinvolves exposure to environments that often
deva-lue their racial group. Chronic exposure to thesehigh-risk
environments can be damaging to AfricanAmerican adolescents
self-esteem and generalmental health. According to PVEST, a
positive viewof ones racial group is as an adaptive copingresponse
to such environments and buffers AfricanAmerican adolescents from
the negative psychoso-cial consequences of discrimination
(Spencer,Dupree, & Hartmann, 1997).
A related notion is that racial identity helps indi-viduals
avoid internalizing negative stereotypes ofones racial group
(McLoyd, Cauce, Takeuchi, &Wilson, 2000; Spencer, 1995). Media
exposureamong American adolescents is quite high (Roberts,Foehr,
Rideout, & Brodie, 1999a), and AfricanAmerican youth tend to be
exposed to media highin negatively stereotypical portrayals of
AfricanAmericans (Ward, 2004). According to the prevail-ing social
identity and reflected appraisal theories,if society views a group
negatively, then membersof the group tend to view themselves
negatively(Tajfel & Turner, 2004). This negative media
expo-sure should then result in more negative stereotypi-cal group
perceptions, which is correlated withmental health problems (Arroyo
& Zigler, 1995).However, a positive racial identity may allow
indi-viduals to not be as concerned about the percep-tions of
others and thus avoid defining their groupin terms of the medias
portrayals (McLoyd et al.,2000). Therefore, modern notions of
racial identitysuggest that it facilitates mental health among
Afri-can American adolescents because it helps protectthem from the
difficult social circumstances theyhave to navigate.
Nonetheless, unlike the studies of self-esteem,the relation
between African American child andadolescent racial identity and
mental health hasbeen much less consistent in the empirical
research(Sellers et al., 2006). For instance, in a sample ofover
1,200 African American sixth- to eighth- grad-ers, D. Roberts et
al. (1999) and R. E. Roberts et al.(1999) found that a revised
version of Phinneys
(1992) Multigroup Ethnic Identity Measure (MEIM)was positively
correlated with their coping but onlyhad small correlations with
their self-esteem anddepressive symptoms (i.e., .14 and ).07,
respec-tively). Another recent study of 10- to 12-year-oldAfrican
Americans also found that the MEIMassessed at the individual level
did not correlatewith depressive symptoms (Simons et al.,
2002).Arroyo and Zigler (1995) found that the endorse-ment of
stereotypical beliefs about African Ameri-cans was positively
correlated with lower collectiveesteem among African American high
school stu-dents, but it had no relation with their depressiveor
anxiety symptoms. A study that used the racialcentrality and
private regard subscales of the Multi-dimensional Inventory of
Black Identity (MIBI; Sell-ers et al., 1997) also found that they
did notcorrelate with the depressive or anxiety symptomsof African
American adolescents (Caldwell et al.,2002). Similarly, Sellers et
al. (2006) found that thepublic regard subscale of the MIBI did not
correlatewith African American adolescents depressivesymptoms,
stress, or general well-being.
Counter to these findings, other studies havefound significant
relations between adolescentracial identity and mental health. A
study of 10- to15-year-old African American adolescents foundthat
the MEIM was negatively related to depressivesymptoms and
aggressive behaviors, as well as pos-itively related to self-esteem
and coping (McMahon& Watts, 2002). Similarly, the affirmation
andbelonging subscale of the MEIM was negativelyrelated to
depressive symptoms and general psy-chological distress among
African American adoles-cents around 12 years old in another study
(Yasui,Dorham, & Dishion, 2004). In the same study thatdid not
find significant effects of public regard,Sellers et al. (2006) did
find that the private regardsubscale of the MIBI was related to
depressivesymptoms and general well-being. Studies thatused similar
measures also found that they werepositively and strongly related
to African Americanadolescents self-esteem (Rowley et al.,
1998;Townsend & Belgrave, 2000) and psychologicalresiliency
(Wong, Eccles, & Sameroff, 2003).
One of the main differences between the two setsof findings is
that the studies that found relationswith racial identity tended to
use measures of thedegree to which adolescents feel prideful of
AfricanAmericans, whereas those that did not find anysignificant
effects of racial identity tended to usemeasures of the degree to
which one has culturalknowledge, ethnic identity exploration,
ethnicspecific behaviors, or adolescent thoughts about
Racial Identity and Self-Esteem 1663
-
what others think of their racial group. Rowleyet al. (1998)
also argued that measures of racialpride or racial esteem were the
subscales that bestrelate to African Americans mental health.
Theytheorized that the affective and evaluative aspectsof the
racial pride subscales may be tapping intothe affective components
of the self-esteem anddepressive symptoms measures. Thus, much of
theinconsistency in findings regarding the relationbetween racial
identity and African Americanadolescent mental health may be
explained by thedifferent measures and subscales used to
assessracial identity (Helms, 2007).
There are also other limitations with prior stud-ies that
prevent more definitive statements aboutthe effects of racial
identity on African Americanadolescents mental health. One glaring
omission isstudies that examine the joint or simultaneous
rela-tions of racial identity and self-esteem with mentalhealth.
This is critical to understanding whatunique features of
self-esteem and racial identityaccount for variance in mental
health. If, forinstance, self-esteem accounts for the effects
ofracial identity on mental health, then it can be con-cluded that
racial identity does not add anythingsocially relevant over and
above self-esteem. Ifracial identity does account for variance in
mentalhealth beyond that accounted for by self-esteem,then this
will have important implications for ourtheoretical understanding
of the identity develop-ment of African American youth (McMahon
&Watts, 2002).
Only a few studies have examined this question,and they have had
ambiguous results. In theMcMahon and Watts (2002) study, racial
identityand self-esteem were simultaneously entered into aset of
regression models to assess their effects onadolescent coping,
depressive, and anxiety symp-toms. They found that racial identity
was correlatedwith coping, but self-esteem was not. Conversely,the
relation between racial identity and depressivesymptoms was
nonsignificant once self-esteem wasaccounted for. With both in the
model, self-esteem,but not racial identity, significantly
correlated withanxiety symptoms. Thus, it is still unclear if
racialidentity has a unique effect on African Americanadolescents
mental health.
Probably the major limitation with the researchin this area is
that all but a few studies have beencross-sectional. An important
advantage of usingmultiple waves is that one is able to assess
theeffects of changes in self-esteem and racial identityon changes
in mental health (Duncan & Duncan,2004). One study examined the
1-year effects of Afri-
can American adolescents racial salience and theirperceptions of
what others think about AfricanAmericans on a composite of
depressive and anxietysymptoms (Sellers, Caldwell, Schmeelk-Cone,
&Zimmerman, 2003). No correlations above .13 werefound between
the variables at the same time orbetween changes in racial identity
and changes inmental health. However, as discussed earlier,
theseparticular subscales usually do not correlate withmental
health; therefore, no conclusions about theeffects of changes in
racial pride or racial esteem canbe derived from that study.
Another issue that has almost never beenaddressed in prior
studies is the potential moderat-ing effects of gender (Chavous,
Rivas-Drake,Smalls, Griffin, & Cogburn, 2008). This is a
criticalconcern given that male and female African Ameri-can
adolescents tend to experience the constructsdifferently as
evidenced by their different means onmany of the variables. Some
studies show thatfemale African American adolescents have
higherlevels of racial pride compared to male adolescents(Martinez
& Dukes, 1997). Conversely, female Afri-can American
adolescents are slightly more at riskfor experiencing depressive
(Fitzpatrick et al., 2005;Repetto et al., 2004) and anxiety
(Kingery, Gins-burg, & Alfano, 2007) symptoms than male
AfricanAmerican adolescents. Therefore, it is possiblethat the
relations between these variables are alsodifferent for male and
female adolescents. Asalmost none of the prior studies analyzed
theresults separately by gender, gender differencesmay explain much
of the reason why some studiesfind significant effects of racial
identity and othersdo not.
The Current Study
The purpose of this study was to assess theunique effects of
self-esteem and racial identity onAfrican American early
adolescents depressive andanxiety symptoms while addressing some of
thelimitations in prior studies. The participants wereassessed
during early adolescence because it iswidely considered to be one
of the most criticalperiods in the development of ones
identity(Brinthaupt & Lipka, 2002). Although the exactform of
identity development during early adoles-cence is still unclear,
significant changes in racialidentity and self-esteem are expected
during thisperiod (French et al., 2006; Moneta et al., 2001).Thus,
this period when greater cognitive aware-ness, physical maturation,
and novel race-basedsocial pressures converge is an opportune time
to
1664 Mandara, Gaylord-Harden, Richards, and Ragsdale
-
assess the effects of changes in racial identity andself-esteem
on mental health.
Based on Cross (1991) and other theorists(McAdoo, 1985; Spencer,
1982) two-factor modelsof identity described earlier, it was
expected thatself-esteem and racial identity would each accountfor
unique variance in depressive and anxietysymptoms. These models
imply that AfricanAmerican adolescents have the ability to
compart-mentalize positive or negative feelings about theirracial
group from feelings about themselves asindividuals. As a result,
most of the variance inAfrican American adolescents mental
healthaccounted for by self-esteem should be separatefrom the
variance accounted for by racial identity,and vice versa.
A further purpose was to determine if 1-yearchanges in
self-esteem and racial identity betweenthe seventh and eighth
grades would correlate withchanges in depressive and anxiety
symptoms overthe same period. Based on the models of racialidentity
development and the consistent self-esteemliterature described
earlier, it was expected thatracial identity and self-esteem would
correlate witheach other and depressive and anxiety symptomsat both
time points. We also expected increases inracial identity and or
self-esteem to correlate withreductions in symptoms. Similar to
Rowley et al.(1998), we argued that the inconsistent
findingsregarding the relation between racial identity andmental
health are due in large part to the differentdimensions of racial
identity assessed in priorstudies. The racial pride or racial
esteem compo-nent of racial identity was the most
consistentcorrelate of African American adolescent mentalhealth in
prior studies. Thus, this study includedreliable measures of racial
pride and self-esteem.The potential moderating effects of gender
werealso assessed.
Method
Participants
Participants included 100 male (39%) and 159female (61%)
sixth-grade African American stu-dents from six public schools in
Chicago. Of the519 asked to participate, 301 (58%) agreed to
partic-ipate in the first wave. Of the 301 participants inthe sixth
grade, 259 of the youth were able to belocated in either the
seventh or eighth grades. Thenonretained youth were similar to the
retainedyouth in terms of annual household income(p = .92),
depressive symptoms (p = .43), anxiety
symptoms (p = .46), and self-esteem (p = .96) dur-ing the sixth
grade. Because racial identity was notassessed in the sixth grade,
this study onlyincluded data from the seventh- and
eighth-gradewaves. The students had a mean age of 12.55(SD = 0.69)
in the seventh grade. The youth wereprimarily from low-income
households with a med-ian family income between $10,000 and
$20,000.Most parents (83%) had at least a high schooldegree, with
10% reporting a college or graduate orprofessional degree.
Procedure
Letters and a short demographics survey weresent home asking
parents for their consent to allowtheir early adolescent to
participate in a multiyearstudy. Parents and students were informed
thatstudents would receive prizes such as gift certifi-cates,
sports equipment, and games at the end ofeach data collection. Each
student returned to theirschool the completed demographics survey
andsigned consent form in a sealed envelope. Research-ers went to
each school everyday for 1 week toadminister surveys in small
groups. All the partici-pating students in each school were
assessed by theend of the week. One year later, when
participantswere in the eighth grade, they were invited to
againparticipate in the study and the same procedurewas
repeated.
Measures
Depressive symptoms. Participants completed the27-item Childrens
Depression Inventory (CDI;Kovacs, 1985), a self-report instrument
on whichyouth rated their level of depressive symptoms forthe 2
weeks prior to administration on a scale from1 (I am sad once in a
while) to 3 (I am sad all the time).The widely used CDI was chosen
because of itsease of use and excellent psychometric propertiesin
multiracial samples of early adolescents (Kovacs,1985). The
seventh- and eighth-grade Cronbachsalphas for the current sample
were .84 and .80,respectively.
Anxiety symptoms. The trait subscale of the State-Trait Anxiety
Inventory for Children (STAIC;Spielberger, Edwards, Montuori, &
Lushene, 1973)was used to assess anxiety symptoms. The STAICis a
20-item measure on which youth rate the fre-quency of anxiety
symptoms on a scale from 1(hardly ever) to 3 (often). Example items
include, Iworry about school and I get a funny feelingin my
stomach. The instrument was used in the
Racial Identity and Self-Esteem 1665
-
current study because it was specifically designedfor late
elementary and middle school and has beenshown to have very good
psychometric propertieswith African American children (Papay &
Hedl,1978). The scale yielded a Cronbachs alpha of .90each year of
the current study.
Racial identity. A modified version of the five-item affirmation
and belonging subscale of theMEIM (Phinney, 1992) was used to
assess racialethnic identity. This subscale was chosen because
itrepresents the affective and prideful component ofracial identity
and because it consistently relates toAfrican American adolescents
self-esteem andmental health in most prior studies. To more
specif-ically assess racial pride or esteem, the MEIM wasmodified
by replacing phrases such as, my ethnicgroup with phrases such as
Black people. Theitems were: (a) I have a lot of pride in Black
peo-ple; (b) I have a strong sense of belonging toBlack people; (c)
I feel a strong attachmenttowards Black people; (d) I am happy to
be amember of the Black group; and (e) I feel goodabout my cultural
or ethnic background. Itemswere rated using a 4-point Likert-type
scale rangingfrom 1 (strongly disagree) to 4 (strongly
disagree).Cronbachs alphas in the seventh and eighth gradeswere .80
and .73, respectively.
Self-esteem. To assess self-esteem, adolescentscarried watches
for 1 week that were programmedto signal at random times. After
each signal, theyouth completed a brief self-report form in a
book-let they carried each day. The youth in this study
responded to a median number of 42 signals,yielding an 82%
compliance rate. Respondents wereprompted with the question, How
were you feel-ing when you were signaled? and rated feelingsabout
how important, tough, respected, and pride-ful they felt on a
4-point scale ranging from 1 (notat all) to 4 (very much). This
experiential samplingmethod was used to assess self-esteem
becauseprior research suggests that it is more reliable andmore
sensitive to changes between the 6th and 12thgrades compared to
dispositional measures (Mone-ta et al., 2001). The scale had a very
strong reliabil-ity in the 7th (a = .94) and 8th grades (a =
.93).
Analysis Plan
To test the main hypotheses of the study, latentvariable growth
models were assessed with struc-tural equation modeling (SEM) using
Amos 5.0with maximum likelihood estimation (see Figure 1).These
models assessed the effects of the racialidentity and self-esteem
intercepts and slopes (i.e.,linear changes between the seventh and
eighthgrades) on linear changes in mental health over thesame
period. Although the purpose was not to testa theoretical model,
but get better estimates ofeffects than could be provided by
ordinary leastsquares (OLS) regression, model fit was assessedwith
the goodness-of-fit v2, the comparative fitindex (CFI), and the
root-mean-square error ofapproximation (RMSEA). Established
criteria sug-gest that a nonsignificant v2, an RMSEA < .06,
and
Figure 1. Latent variable growth model illustrating the paths to
test the effects of baseline and changes in racial identity and
self-esteem on changes in mental health.
1666 Mandara, Gaylord-Harden, Richards, and Ragsdale
-
a CFI of 0.95 or better indicates a very good fit ofthe model to
the data (Hu & Bentler, 1999).
The measurement models for racial identity, self-esteem, and
depressive symptoms at each wavewere tested with exploratory and
then confirmatoryfactor analyses. Each was found to be
unidimen-sional and had high internal consistencies (see theMethod
section). The composites of each factorwere used as single
indicators of their latent vari-able in the SEM analyses at each
wave. To accountfor measurement error, the residuals were fixed
totheir error variances times their variances (Loehlin,1998).
To assess the potential moderating effects of gen-der, the
multiple-group analysis facility in Amos5.0 was used to fit the
data simultaneously to bothgenders. This analysis helped answer two
ques-tions. First, the multiple-group analysis assessed ifthe path
diagram fit the data equally well for bothmales and females, and if
so, if the regressionweights were different between the genders.
Themultiple-group analysis in Amos 5.0 proceeds byautomatically
generating several nested models.Each lower level model adds
additional constraintsand can be tested for differences from the
previousmodels (Arbuckle, 2003). For our purpose, only
thecompletely unconstrained model and the modelthat constrained the
regression weights to be equal(i.e., the structural weights model)
were important.A test of the unconstrained or first level
modelassesses the fit of the data to both groups. If
theunconstrained model fits the data well, the test ofthe
structural weights model gives an estimate ofthe degree to which
the regression weights are sim-
ilar in both groups. The facility also computes criti-cal ratio
tests for each parameter in both samples(Arbuckle, 2003).
Before any formal analyses, missing data wereanalyzed. To be
included in the study, participantshad to have data at one of the
two time points foreach variable. This resulted in a maximum
sampleof 259 adolescents. The percentages missing foreach variable
are presented in Table 1. Overall,16.9% and 16.7% of data points
were missing formales and females, respectively. The pattern
ofmissing was very similar for males and females aswell. Littles
MCAR test showed that the data weremissing completely at random,
v2(102) = 126.45,p = .05. We therefore imputed missing data withthe
expectation maximization (EM) algorithm. Thismethod replaces
missing values with iterative max-imum likelihood estimations based
on the availabledata. Schafer (1997) discusses more details of
theEM algorithm.
Results
Descriptive Statistics
The means, standard deviations, and zero-ordercorrelations by
gender are presented in Table 1.Contrary to our predictions,
females racial identitywas not related to their self-esteem or
anxiety symp-toms. By modern effect size standards (Hemphill,2003),
racial identity had a small relation with girlsdepressive symptoms
in the seventh grade and amoderate effect on depressive symptoms in
the
Table 1
Means, Standard Deviations, and Zero-Order Correlations Among
Study Variables in the Seventh and Eighth Grades
Variables (grade) 1 2 3 4 5 6 7 8
1. Racial identity (7) .04 ).16* ).09 .34** ).03 .03 .042.
Self-esteem (7) .28** ).24** ).18* .04 .81** ).28** ).23*3.
Depression (7) ).35** ).17 .50** ).19* ).21** .65** .44**4. Anxiety
(7) .06 .04 .54** ).08 ).19* .31** .59**5. Racial identity (8)
.56** .39** ).44** ).21* .03 ).32** ).156. Self-esteem (8) .27**
.84** ).23* .03 .45** ).17* ).18*7. Depression (8) ).32** ).29**
.63** .40** ).55** ).28** .52**8. Anxiety (8) .13 ).18 .38* .47**
).20* ).22* .42***M 3.30 3.30 1.18 10.99 3.44 3.29 1.08 10.33
SD 0.67 0.65 1.24 7.79 0.52 0.68 1.02 7.59
Percent missing 21 20 17 23 6 17 10 18
Note. Females correlations are above the diagonal (n = 159) and
males correlations are below (n = 100). Values in
parenthesesrepresent seventh- and eighth-grade assessments. Means,
standard deviations, and percent missing are for the total
sample.*p < .05. **p < .01.
Racial Identity and Self-Esteem 1667
-
eighth grade. Self-esteem in the seventh and eighthgrades was
negatively related to depressive andanxiety symptoms in both years
for girls. The resultswere somewhat different for boys. The
correlationbetween racial identity in the seventh and eighthgrades
was slightly larger for boys than it was forgirls. However, racial
identity was a very strong cor-relate of boys self-esteem and
depressive symptomsin both years. Racial identity in the eighth
grade hada small effect on anxiety symptoms in the
expecteddirection, but they did not correlate in the seventhgrade.
Self-esteem also had small negative relationswith boys depressive
and anxiety symptoms ineighth grade, but not in the seventh
grade.
Independent samples t tests were conducted toassess gender
differences in mean levels on each ofthe variables in the study. No
significant differ-ences were found between males and
females.Paired samples t tests were also computed to assessmean
differences on the identity and mental healthvariables between the
seventh and eighth grades.Those results showed that racial identity
signifi-cantly increased between the seventh and eighthgrades,
t(258) = 3.43 p = .001. No other changeswere found.
Effects of Changes in Identity on Changes in
DepressiveSymptoms
The first set of analyses assessed the effectsof racial identity
and self-esteem on depressivesymptoms as participants transitioned
from theseventh to the eighth grade. The results of the mul-
tiple-group analysis showed that the unconstrainedmodel fit the
data adequately, v2(2) = 3.60, p = .17,CFI = 1.0, RMSEA = .06. This
indicates that theoverall path model fit well for both genders.
How-ever, when constraining the regression weights tobe equal,
there was a significant decrease in modelfit, Dv2(13) = 27.35, p =
.01, which indicates that theoverall pattern of regression weights
were signifi-cantly different between girls and boys (seeFigures 2
and 3). An examination of the criticalratios for the differences
between the regressionweights showed that the relation between
racialidentity and self-esteem was significantly larger forboys
than for girls (z = 1.99, p < .05). Furthermore,the path from
the self-esteem intercept to the racialidentity slope was
significantly larger for boys(z = 2.24, p < .05). The depressive
symptoms inter-cept also had a significantly larger negative
effecton the depressive symptoms slope for boys com-pared to girls
(z = 3.04, p < .01). There were noother significant differences
between males andfemales for the same path.
The overall pattern of results generally supportedthe main
hypotheses. The higher the seventh-graderacial identity (e.g., the
intercept), the lower the sev-enth-grade depressive symptoms for
both malesand females, even after seventh-grade self-esteemwas
accounted for. The reverse was only true forfemales. Initial
self-esteem did not correlate withdepressive symptoms for boys.
Seventh-grade racialidentity also had a strong negative effect on
thedepressive symptoms slope for boys. The highertheir racial
identity in the seventh grade, the less
Figure 2. Latent variable growth model illustrating the effects
of baseline and changes in racial identity and self-esteem on
changes indepressive symptoms for males, v2(1) = 3.25, p = .07,
comparative fit index = 1.0, root mean square error of
approximation = .15.**p < .01.
1668 Mandara, Gaylord-Harden, Richards, and Ragsdale
-
likely they were to increase in depression betweenthe seventh
and eighth grades. Self-esteem in theseventh grade did not predict
changes in depressivesymptoms for boys or girls.
The results also showed that an increase in racialidentity from
the seventh to the eighth grade wasassociated with a large decrease
in depressivesymptoms over the same period once baseline fac-tors
and increases in self-esteem were accountedfor. This was consistent
for both males and females.Conversely, an increase in self-esteem
was associ-ated with a significant increase in depressive symp-toms
only for girls.
Effects of Changes in Identity on Changes in AnxietySymptoms
The same analyses were conducted for anxietysymptoms. As with
the depressive symptoms, theresults of the multigroup analyses
showed that themodel fit equally well for males and females,v2(2) =
3.41, p = .18, CFI = 1.0, RMSEA = .05. Alsolike the depressive
symptoms data, when constrain-ing the regression weights to be
equal, there was asignificant decrease in model fit, Dv2(13) =
25.78,p = .02 (see Figures 4 and 5). The critical ratios forthe
differences between the regression weightswere also similar. The
relations between racial iden-tity and self-esteem (z = 1.99, p
< .05), between theintercept of self-esteem and the racial
identity slope(z = 2.25, p < .05), between the anxiety
symptomsintercept and slope (z = 2.12, p < .05), and betweenthe
anxiety symptoms intercept and racial identity
slope (z = 2.16, p < .05) were significantly strongerfor
boys.
However, the overall pattern of results wassomewhat different
for anxiety symptoms com-pared to depressive symptoms. The
seventh-graderacial identity and self-esteem did not predict
theanxiety intercept or slope. This held up for malesand females.
Partially in support of our prediction,the results showed that an
increase in racial iden-tity between the seventh and eighth grades
wasassociated with a significant decrease in anxietysymptoms over
the same period. This effect wassignificant for females only, but
the coefficient wasthe same size for the males. Counter to our
expecta-tions, changes in self-esteem did not relate tochanges in
anxiety symptoms.
Discussion
Prior research shows that higher levels of self-esteem and the
racial esteem or pride component ofracial identity are related to
lower levels of mentalhealth problems among African American
adoles-cents. This study was designed to build on thisprior
research and address some voids in the litera-ture by assessing the
unique effects of racialidentity and self-esteem on African
Americanadolescents depressive and anxiety symptoms asthey
transitioned between the seventh and eighthgrades. The results
tended to support the generaltheme of prior research, but several
distinctive find-ings also emerged.
Figure 3. Latent variable growth model illustrating the effects
of baseline and changes in racial identity and self-esteem on
changes indepressive symptoms for females, v2(1) = 0.35, p = .56,
comparative fit index = 1.0, root mean square error of
approximation = .00.*p < .05. **p < .01.
Racial Identity and Self-Esteem 1669
-
We found some support for the main hypothesisthat both racial
identity and self-esteem wouldhave unique effects on mental health,
but thisdepended on the gender of the adolescent andwhether the
outcome was depressive or anxietysymptoms. The higher males racial
identity in theseventh grade, the lower their depressive symp-toms
in the seventh or eighth grades tended tobe. Moreover, the higher
males racial identity inthe seventh grade, the less likely it was
that theyexperienced an increase in depressive symptomsover the 1
year. These findings persisted after
accounting for self-esteem. The results for femaleswere in the
same general direction, but they werenot as strong. Self-esteem did
not have the samestrong unique effect on depressive symptoms asdid
racial identity. Seventh grade self-esteem didcorrelate with
depressive symptoms in the seventhgrade for girls, even with racial
identity accountedfor, but this was not found for boys. For boys,
racialidentity accounted for all of the self-esteem effecton
depressive symptoms.
A related finding was that 1-year increases inadolescents racial
identity were associated with
Figure 5. Latent variable growth model illustrating the effects
of baseline and changes in racial identity and self-esteem on
changes inanxiety symptoms for females, v2(2) = 3.60, p = .17,
comparative fit index = 1.0, root mean square error of
approximation = .06.*p < .05. **p < .01.
Figure 4. Latent variable growth model illustrating the effects
of baseline and changes in racial identity and self-esteem on
changes inanxiety symptoms for males, v2(1) = 3.04, p = .17,
comparative fit index = .99, root mean square error of
approximation = .14.**p < .01.
1670 Mandara, Gaylord-Harden, Richards, and Ragsdale
-
decreases in the prevalence of depressive symp-toms over the
same period. Those adolescents whoincreased their prideful feelings
of African Ameri-cans as they transitioned from the seventh to
theeighth grades also experienced a relatively largedecrease in
depressive symptoms over the sameperiod. This finding held up even
with baselineself-esteem and changes in self-esteem accountedfor.
This was the same pattern of results for themales and females.
Given the cross-sectional andthe change model findings, it seems
clear that thepride young African American adolescents have intheir
racial group is an important correlate of theircurrent depressive
symptoms and changes indepressive symptoms beyond their general
feelingsof personal esteem. The better they feel about theirracial
group, the less depressive symptoms theyseem to experience.
Unlike the results with depressive symptoms,racial identity did
not have the same clear effectson anxiety symptoms. Racial identity
in the seventhor eighth grades did not correlate with
anxietysymptoms in the same year. This is consistent withmost
previous studies that have found that racialidentity does not
correlate with anxiety symptoms(e.g., Arroyo & Zigler, 1995;
Caldwell et al., 2002).McMahon and Watts (2002) also found that
racialidentity did not correlate with anxiety symptoms,but it did
with depressive symptoms. However, wedid find that increases in
racial identity were corre-lated with reductions in anxiety
symptoms for girls.Although the relation was not as strong as it
wasfor depressive symptoms, the effect was significantand
important. The effect size was the same forboys, but it was not
significant for them. Therefore,changes in racial identity may have
a unique effecton changes in anxiety symptoms, but the
overallrelation between racial identity and anxiety symp-toms is
not very strong.
One plausible reason for the differences betweenthe depressive
and anxiety symptoms findings isthat racial identity, self-esteem,
and depressivesymptoms share an affective component. Forinstance,
each measure assesses the degree to whichone feels positive about
themselves, their racialgroup, or their emotional experiences. This
affectivecomponent is less prevalent among anxiety symp-toms, which
tend to be more focused on physicalsymptoms and fears (Spielberger
et al., 1973). Thismay explain why racial identity did not account
foras much variance in anxiety symptoms as it didwith depressive
symptoms. Future studies mayfind that other components of racial
identityaccount for variance in anxiety symptoms not
accounted for by affective components of racialidentity.
Overall, the findings from this study support thenotion that
developing positive feelings about onesracial group is related to
reductions in mentalhealth problems. As some African American
ado-lescents develop more prideful feelings about theirracial
group, they simultaneously experience fewersymptoms of some mental
health problems. Asdescribed earlier, theorists suggest that racial
iden-tity is associated with mental health because it mayhelp
African American adolescents avoid internal-izing negative
stereotypes (McLoyd et al., 2000;Spencer, 1995), and it may buffer
them from theeffects of racial discrimination (Caldwell et
al.,2002; Cross, Parham, & Helms, 1998). In support ofthese
theories, a few recent studies have found thatvarious aspects of
racial identity moderate theeffects of perceived racial
discrimination on AfricanAmerican youths mental health (Sellers et
al., 2006;Wong et al., 2003). Thus, a positive racial identitymay
allow individuals to not feel as if discrimina-tory practices are a
result of shortcomings of theirgroup. There may be other mediators
of racial iden-tity as well. This will be an important topic
forfuture studies.
Important gender differences were also found inthis study.
Although males and females did nothave mean differences on any of
the variables, therelation between racial identity and
self-esteemwere different. Racial identity and self-esteem
hadmoderate to large positive correlations with eachother for the
boys in both the seventh and eighthgrades. They did not correlate
at all for the girls ineither year. One possible explanation is
that raceduring this age may be more salient for AfricanAmerican
males than it is for females, and this dif-ference explains the
stronger racial identity andself-esteem relation among boys. For
instance,some studies suggest that self-esteem and racialidentity
are only correlated when race is a salientaspect of ones identity
(Phinney, 1991; Rowleyet al., 1998). Moreover, other studies imply
thatrace may be more salient for African American ado-lescent boys
compared to girls. For instance, a fewstudies found that African
American boys are morelikely than girls to receive racial
socialization mes-sages concerned with preparing them for racial
dis-crimination (Bowman & Howard, 1985; Thomas &Speight,
1999). Other studies found that AfricanAmerican adolescent males
perceive more discrimi-nation from peers and teachers than
AfricanAmerican girls (Chavous et al., 2008). Thus, if
thesedifferent racial socialization and discrimination
Racial Identity and Self-Esteem 1671
-
experiences made race more salient for AfricanAmerican males, it
could explain why racial iden-tity and self-esteem only correlated
for boys duringthis age.
Quite surprisingly, increases in self-esteem werecorrelated with
increases in depressive symptomsfor girls once racial identity was
accounted for. Thisdoes not imply that self-esteem is a negative
con-struct for this population, because self-esteem wasnegatively
correlated with depressive symptoms ateach time point for males and
females. However,there is some evidence that at least a few
AfricanAmerican girls experience both higher than
averageself-esteem and depressive symptoms in the gen-eral
population. For instance, in the general popula-tion, female
African American adolescents tend tohave high levels of self-esteem
compared to otherfemale adolescents (Twenge & Crocker, 2002),
butthey are equal to or even slightly more likely toexperience
depressive symptoms than otherfemales during middle to later
adolescence (Frankoet al., 2005). Thus, there is some evidence that
thetwo are positively related for many African Ameri-can girls.
This could be due to the amount ofresponsibilities expected of
African American ado-lescent females. They tend to have a great
burdenof household chores, caring for younger siblings,and other
adult responsibilities compared toAfrican American boys and other
adolescents(Burton, 2007; Dodson & Dickert, 2004).
Theseresponsibilities may result in the girls feeling moreconfident
in their abilities but also more pressuredand stressed from taking
on adult responsibilitiesat a young age (Burton, 2007; Dodson &
Dickert,2004). If this is the case, then it could result in a
sit-uation in which at least some girls simultaneouslyincrease in
self-esteem and depressive symptoms.Future studies should examine
this possibility.
Limitations of this study should be noted. Aswith the majority
of studies on this topic, the adoles-cents in this study were
primarily from lower SESinner-city backgrounds. Adolescents in this
situa-tion may have different views of racial identity
andself-esteem compared to African American adoles-cents from
middle-class backgrounds and thoseliving in less threatening
environments. Forinstance, issues of race at school may be less
salientfor adolescents in this study as almost all of
theirclassmates were African American, whereas thosereared in the
suburbs are likely exposed to moreracially heterogeneous
environments. Futureresearch should begin to compare these results
todifferent subpopulations of African Americanyouth.
The two waves of data in this study have manyadvantages over
cross-sectional studies, but thisdesign still does not allow one to
make the claimthat the changes in identity caused changes
insymptomology. Moreover, only having two timepoints limits our
ability to describe nonlineargrowth curve models (Duncan &
Duncan, 2004). Inlieu of randomized experiments, finding that
differ-ent types of identity growth curves predict similartypes of
growth curves in mental health would bethe strongest evidence for
causation. Thus, studiesthat measure the variables from an earlier
periodand follow the participants for a longer period oftime have
the potential to significantly increase ourunderstanding of these
important developmentalfactors.
In spite of these limitations, there are someimportant
implications of this study. It has beenestablished that self-esteem
is a critical feature ofoptimal adolescent development and mental
health.As suggested by the two-factor models of identity(Cross,
1991), this study implies that racial identitycorrelates with
mental health as strongly as self-esteem and accounts for variance
in African Ameri-can adolescents mental health not associated
withself-esteem. Although they are correlated in adoles-cent males,
it does not seem as if racial identity andself-esteem are proxies
for each other. Thus, thesefindings underscore the importance of
includingboth personal and group-level components of iden-tity in
research on adolescent development.
We believe that the results of the change modelsin this study
are some of the strongest evidence todate of the correlation
between racial identity andmental health for African American
adolescents.Although the causality of the relations are
stillquestionable, and thus it is not clear exactly howprotective
racial identity actually is, it is evidentfrom this study and the
extant literature that racialidentity is not a risk factor for
mental health prob-lems. Therefore, randomized prevention
interven-tions that incorporate sound models of racialidentity
development should be implemented(McMahon & Watts, 2002; Wills
et al., 2007). Suchexperiments will help answer questions about
thecausal effects of racial identity and may help reducethe
prevalence of symptomology in this population.Based on the research
findings in this area, the pre-vention interventions should
primarily focus on thedevelopment of pride and esteem in ones
heritage.Those who focus on teaching African Americanadolescents
about racism, public regard for othersbeliefs, and or just
performing ethnic specificbehaviors are likely not going to be
efficacious.
1672 Mandara, Gaylord-Harden, Richards, and Ragsdale
-
References
Arbuckle, J. (2003). Amos 5.0 update to the Amos usersguide.
Chicago: SmallWaters.
Arroyo, C. G., & Zigler, E. (1995). Racial identity,
aca-demic achievement, and the psychological well-beingof
economically disadvantaged adolescents. Journal ofPersonality and
Social Psychology, 69, 903914.
Baldwin, S. A., & Hoffman, J. P. (2002). The dynamics
ofself-esteem: A growth-curve analysis. Journal of Youthand
Adolescence, 31, 101113.
Botvin, G. J., Baker, E., Botvin, E. M., Dusenbury, L.,Cardwell,
J., & Diaz, T. (1993). Factors promoting ciga-rette smoking
among Black youth: A causal modelingapproach. Addictive Behaviors,
18, 397405.
Bowman, P. J., & Howard, C. (1985). Race-related
sociali-zation, motivation, and academic achievement: Astudy of
Black youths in three-generation families.Journal of the American
Academy of Child Psychiatry, 24,134141.
Brinthaupt, T. M., & Lipka, R. P. (2002). Understandingearly
adolescent self and identity: Applications and interven-tions.
Albany: State University of New York Press.
Brown, K. M., McMahon, R. P., Biro, F. M., Crawford,
P.,Schreiber, G. B., Similo, S. L., et al. (1998). Changes
inself-esteem in Black and White girls between the agesof 9 and 14.
Journal of Adolescent Health, 23, 719.
Burton, L. (2007). Childhood adultification in
economicallydisadvantaged families: A conceptual model.
FamilyRelations, 56, 329345.
Caldwell, C. H., Antonucci, T. C., Jackson, J. S., Wolford,M.
L., & Osofsky, J. D. (1997). Perceptions of parentalsupport and
depressive symptomatology among Blackand White adolescent mothers.
Journal of Emotional andBehavioral Disorders, 5, 173183.
Caldwell, C. H., Zimmerman, M. A., Bernat, D. H.,Sellers, R. M.,
& Notaro, P. C. (2002). Racial identity,maternal support, and
psychological distress amongAfrican American adolescents. Child
Development, 73,13221336.
Chavous, T. M., Rivas-Drake, D., Smalls, C., Griffin, T.,
&Cogburn, C. (2008). Gender matters, too: The influencesof
school racial discrimination and racial identity onacademic
engagement outcomes among African Ameri-can adolescents.
Developmental Psychology, 44, 637654.
Compas, B. E., Hinden, B. R., & Gerhardt, C. A.
(1995).Adolescent development: Pathways and processes ofrisk and
resilience. Annual Review of Psychology, 46,265293.
Cross, W. E. (1991). Shades of Black: Diversity in
AfricanAmerican identity. Philadelphia: Temple University
Press.
Cross, W. E., Parham, T. A., & Helms, J. E. (1998).
Nigres-cence revisited: Theory and research. In R. L. Jones(Ed.),
African American identity development: Theory,research, and
intervention (pp. 371). Hampton, VA:Cobb and Henry.
Dodson, L., & Dickert, J. (2004). Girls family labor in
lowincome households: A decade of qualitative research.Journal of
Marriage and Family, 66, 318332.
DuBois, D. L., Burk-Braxton, C., Swenson, L. P.,Tevendale, H.
D., Lockerd, E. M., & Moran, B. L. (2002).Getting by with a
little help from self and others: Self-esteem and social support as
resources during earlyadolescence. Developmental Psychology, 38,
822839.
Duncan, T. E., & Duncan, S. C. (2004). An introduction
tolatent growth curve modeling. Behavior Therapy, 35,333363.
Erikson, E. H. (1968). Identity: Youth and crisis. New
York:Norton.
Fitzpatrick, K. M., Piko, B. F., Wright, D. R., & LaGory,M.
(2005). Depressive symptomatology, exposure toviolence, and the
role of social capital among AfricanAmerican adolescents. American
Journal of Orthopsychia-try, 75, 262274.
Franko, D. L., Striegel-Moore, R. H., Bean, J., Barton, B.A.,
Biro, F., Kraemer, H. C., et al. (2005). Self-reportedsymptoms of
depression in late adolescence to earlyadulthood: A comparison of
African-American andCaucasian females. Journal of Adolescent
Health, 37, 526529.
French, S. E., Seidman, E., Allen, L., & Aber, J. L.
(2006).The development of ethnic identity during
adolescence.Developmental Psychology, 42, 110.
Greene, M. L., & Way, N. (2005). Self-esteem
trajectoriesamong ethnic minority adolescents: A growth
curveanalysis of the patterns and predictors of change. Jour-nal of
Research on Adolescence, 15, 151178.
Harter, S. (1999). The construction of the self: A
developmen-tal perspective. New York: Guilford.
Helms, J. E. (2007). Some better practices for measuringracial
and ethnic identity constructs. Journal of Counsel-ing Psychology,
54, 235246.
Hemphill, J. F. (2003). Interpreting the magnitudes
ofcorrelation coefficients. American Psychologist, 58, 7880.
Hirsch, B. J., & Rapkin, B. D. (1987). The transition
tojunior high school: A longitudinal study of
self-esteem,psychological symptomatology, school life, and
socialsupport. Child Development, 58, 12351243.
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for
fitindexes in covariance structure analysis: Conventionalcriteria
versus new alternatives. Structural EquationModeling, 6, 155.
Johnson, E. H., & Greene, A. F. (1991). The
relationshipbetween suppressed anger and psychosocial distress
inAfrican American male adolescents. Journal of BlackPsychology,
18, 4765.
Kingery, J. N., Ginsburg, G. S., & Alfano, C. A.
(2007).Somatic symptoms and anxiety among African Ameri-can
adolescents. Journal of Black Psychology, 33, 363378.
Kovacs, M. (1985). The Childrens Depression Inventory(CDI).
Psychopharmacology Bulletin, 21, 995998.
Loehlin, J. C. (1998). Latent variable models: An introductionto
factor, path, and structural analysis (3rd ed.). Mahwah,NJ:
Erlbaum.
Mann, M., Hosman, C. M. H., Schaalma, H. P., &de Vries, N.
K. (2004). Self-esteem in a broad spectrum
Racial Identity and Self-Esteem 1673
-
approach for mental health promotion. Health EducationResearch,
19, 357372.
Martinez, R. O., & Dukes, R. L. (1997). The effects of
eth-nic identity, ethnicity, and gender on adolescent well-being.
Journal of Youth and Adolescence, 26, 503516.
Matthews, D. B., & Odom, B. L. (1989). Anxiety: A com-ponent
of self-esteem. Elementary School Guidance andCounseling, 24,
153159.
McAdoo, H. P. (1985). Racial attitude and self-concept ofyoung
Black children over time. In H. P. McAdoo & J.L. McAdoo (Eds.),
Black children: Social, educational, andparental environments (pp.
213242). Beverly Hills, CA:Sage.
McLoyd, V. C., Cauce, A. M., Takeuchi, D., & Wilson,
L.(2000). Martial processes and parental socialization infamilies
of color: A decade review of research. Journalof Marriage and the
Family, 62, 10701093.
McMahon, S. D., & Watts, R. J. (2002). Ethnic identity
inurban African American youth: Exploring links withself-worth,
aggression, and other psychosocial vari-ables. Journal of Community
Psychology, 30, 411431.
Moneta, G. B., Schneider, B., & Csikszentmihalyi, M.(2001).
A longitudinal study of the self concept andexperiential components
of self-worth and affectacross adolescence. Applied Developmental
Science, 5,125142.
Murray, C. B., & Mandara, J. (2001). Racial identity
devel-opment: Cognitive and experiential antecedents. In H.P.
McAdoo (Ed.), Black child development (3rd ed.,pp. 7396). Newbury
Park, CA: Sage.
OMalley, P. M., & Bachman, J. G. (1983). Self-esteem:Change
and stability between ages 13 and 23. Develop-mental Psychology,
19, 257268.
Pahl, K., & Way, N. (2006). Longitudinal trajectories
ofethnic identity among urban Black and Latino adoles-cents. Child
Development, 77, 14031415.
Papay, J. P., & Hedl, J. J. (1978). Psychometric
characteris-tics and norms for disadvantaged third and fourthgrade
children on the State-Trait Anxiety Inventory forChildren. Journal
of Abnormal Child Psychology, 6, 115120.
Perron, J., Vondracek, F. W., Skorikov, V. B., Tremblay,C.,
& Corbie`re, M. (1998). A longitudinal study of voca-tional
maturity and ethnic identity development. Jour-nal of Vocational
Behavior, 52, 409424.
Phinney, J. S. (1989). Stages of ethnic identity develop-ment in
minority group adolescents. Journal of EarlyAdolescence, 9,
3449.
Phinney, J. S. (1991). Ethnic identity and self-esteem: Areview
and integration. Hispanic Journal of BehavioralSciences, 13,
193208.
Phinney, J. S. (1992). The Multigroup Ethnic IdentityMeasure: A
new scale for use with adolescents andyoung adults from diverse
groups. Journal of AdolescentResearch, 7, 156176.
Phinney, J. S., & Tarver, S. (1988). Ethnic identity
searchand commitment in Black and White eighth graders.Journal of
Adolescent Research, 8, 265277.
Repetto, P. B., Caldwell, C. H., & Zimmerman, M. A.(2004).
Trajectories of depressive symptoms amonghigh risk African American
adolescents. Journal of Ado-lescent Health, 35, 468477.
Roberts, D., Foehr, U., Rideout, V., & Brodie, M.
(1999).Kids and media at the new millennium. Palo Alto, CA:Henry J.
Kaiser Family Foundation.
Roberts, R. E., Phinney, J. S., Masse, L. C., Chen, Y.
R.,Roberts, C. R., & Romero, A. (1999). The structureof ethnic
identity of young adolescents from diverseethnocultural groups.
Journal of Early Adolescence, 19,301322.
Rowley, S. J., Sellers, R. M., Chavous, T. M., & Smith, M.A.
(1998). The relationship between racial identity andself-esteem in
African American college and highschool students. Journal of
Personality and Social Psychol-ogy, 3, 715724.
Schafer, J. L. (1997). Analysis of incomplete multivariate
data.London: Chapman & Hall.
Seaton, E. K., Scottham, K. M., & Sellers, R. M. (2006).The
status model of racial identity development inAfrican American
adolescents: Evidence of structure,trajectories, and well-being.
Child Development, 77,14161426.
Sellers, R. M., Caldwell, C. H., Schmeelk-Cone, K. H.,
&Zimmerman, M. A. (2003). Racial identity, racial
dis-crimination, perceived stress, and psychological dis-tress
among African American young adults. Journal ofHealth and Behavior,
43, 302317.
Sellers, R. M., Copeland-Linder, N., Martin, P. P., &Lewis,
R. L. (2006). Racial identity matters: The rela-tionship between
racial discrimination and psychologi-cal functioning in African
American adolescents.Journal of Research on Adolescence, 16,
187216.
Sellers, R. M., Rowley, S. A. J., Chavous, T. M., Shelton, J.N.,
& Smith, M. A. (1997). Multidimensional Inventoryof Black
Identity: A preliminary investigation of reli-ability and construct
validity. Journal of Personality andSocial Psychology, 73,
805815.
Simons, R. L., Murray, V., McLoyd, V., Lin, K., Cutrona,C.,
& Conger, R. D. (2002). Discrimination, crime, eth-nic
identity, and parenting as correlates of depressivesymptoms among
African American children: A multi-level analysis. Development and
Psychopathology, 14, 371393.
Spencer, M. B. (1982). Preschool childrens social cogni-tion and
cultural cognition: A cognitive developmen-tal interpretation of
race dissonance findings. Journalof Psychology: Interdisciplinary
and Applied, 112, 275286.
Spencer, M. B. (1984). Black childrens race awareness,racial
attitudes and self-concept: A reinterpretation.Journal of Child
Psychology and Psychiatry, 25, 433441.
Spencer, M. B. (1995). Old issues and new theorizingabout
African American youth: A phenomenologicalvariant of ecological
systems theory. In R. L. Taylor(Ed.), Black youth: Perspectives on
their status in the UnitedStates (pp. 3769). Westport, CT:
Praeger.
1674 Mandara, Gaylord-Harden, Richards, and Ragsdale
-
Spencer, M. B., Dupree, D., & Hartmann, T. (1997).
Aphenomenological variant of ecological systems theory(PVEST): A
self-organization perspective in context.Development and
Psychopathology, 9, 817833.
Spielberger, C. D., Edwards, C. D., Montuori, J., &
Lush-ene, D. (1973). How I feel questionnaire (State-Trait Anxi-ety
Inventory for Children). Palo Alto, CA: CounselingPsychologists
Press.
Tajfel, H., & Turner, J. C. (2004). The social identity
the-ory of intergroup behavior. In J. T. Jost & J.
Sidanius(Eds.), Political psychology: Key readings (pp. 276293),New
York: Psychology Press.
Thomas, A. J., & Speight, S. L. (1999). Racial identity
andracial socialization attitudes of African American par-ents.
Journal of Black Psychology, 25, 152170.
Townsend, T. G., & Belgrave, F. Z. (2000). The impact
ofpersonal identity and racial identity on drug attitudesand use
among African American children. Journal ofBlack Psychology, 26,
421436.
Twenge, J. M., & Crocker, J. (2002). Race and
self-esteem:Meta-analyses comparing Whites, Blacks,
Hispanics,Asians, and American Indians and comment on Gray-Little
and Hafdahl (2000). Psychological Bulletin, 128,371408.
Unger, J. B., Kipke, M. D., Simon, T. R., Montgomery, S.B.,
& Johnson, C. J. (1997). Homeless youths and youngadults in Los
Angeles: Prevalence of mental healthproblems and the relationship
between mental healthand substance abuse disorders. American
Journal ofCommunity Psychology, 25, 371394.
Vandiver, B. J., Cross, W. E., Jr., Worrell, F. C.,
&Fhagen-Smith, P. E. (2002). Validating the Cross
RacialIdentity Scale. Journal of Counseling Psychology, 49,
7185.
Ward, L. M. (2004). Wading through the stereotypes:Positive and
negative associations between media useand Black adolescents
conceptions of self. Developmen-tal Psychology, 40, 284294.
Wardle, F. (1992). Supporting biracial children in the
schoolsetting. Education and Treatment of Children, 15, 163172.
Whaley, A. L. (1993). Self-esteem, cultural identity,
andpsychosocial adjustment in African American children.Journal of
Black Psychology, 19, 406422.
Wills, T. A., McBride-Murry, V., Brody, G. H., Gibbons, F.X.,
Gerrard, M., Walker, C., et al. (2007). Ethnic prideand
self-control related to protective and risk factors:Test of the
theoretical model for the strong AfricanAmerican families
program.Health Psychology, 26, 5059.
Wong, C. A., Eccles, J. S., & Sameroff, A. (2003).
Theinfluence of ethnic discrimination and ethnic identifi-cation on
African American adolescents school and so-cioemotional adjustment.
Journal of Personality, 71,11971232.
Yasui, M., Dorham, C. L., & Dishion, T. J. (2004).
Ethnicidentity and psychological adjustment: A validity anal-ysis
for European American and African Americanadolescents. Journal of
Adolescent Research, 19, 807825.
Youngstrom, E., Weist, M. D., & Albus, K. E.
(2003).Exploring violence exposure, stress, protective factorsand
behavioral problems among inner-city youth.American Journal of
Community Psychology, 32, 115129.
Racial Identity and Self-Esteem 1675