Вы находитесь на странице: 1из 12

J Youth Adolescence (2007) 36:77–88

DOI 10.1007/s10964-006-9115-9

ORIGINAL PAPER

Perceived Support and Internalizing Symptoms in African


American Adolescents: Self-Esteem and Ethnic Identity
as Mediators
Noni K. Gaylord-Harden · Brian L. Ragsdale ·
Jelani Mandara · Maryse H. Richards ·
Anne C. Petersen

Received: 7 April 2006 / Revised: 26 May 2006 / Accepted: 6 June 2006 / Published online: 25 July 2006

C Springer Science+Business Media, Inc. 2006

Abstract Existing research leaves a gap in explaining why accounted for more of the social support effect for males,
African American adolescents do not exhibit more anxiety whereas self-esteem had more impact for females. The me-
and depression than other youth, at the same time that they diation model for anxiety was supported in females, with
experience more contextual risk factors. The current study self-esteem more important than ethnic identity. The re-
examined the roles of social support as well as possible me- sults suggest that ethnic identity and self-esteem function
diators self-esteem and ethnic identity (sense of belonging as important links in how social support reduces internaliz-
to one’s ethnic group) in reducing internalizing symptoms ing symptoms in African American youth.
in 227 African American adolescents (mean age = 12.55).
Structural equation models indicated that self-esteem and Keywords Self-esteem . Ethnic identity . Social support .
ethnic identity partially mediated the relation between so- African American adolescents . Depression . Anxiety
cial support and depression. For depression, ethnic identity

N. K. Gaylord-Harden () M. H. Richards


Assistant Professor, Clinical Psychology, Loyola University Professor, Clinical and Developmental Psychology, Loyola
Chicago. Received Ph.D. in Psychology from The University of University Chicago. Received Ph.D. in Human Development from
Memphis. Current interests include coping and resilience in the University of Chicago. Current research interests include the
African American youth and the role of family characteristics in developmental stage of adolescence with a focus on the daily
children and adolescents’ stress and coping processes. experience of urban African American young adolescents and
Department of Psychology. how this relates to their psycho- social well being. Dr. Richards
Loyola University Chicago, served as a Predoctoral Adolescent Fellow (1979–1981) and
6525 N. Sheridan Rd., Chicago, Illinois, 60626 USA Postdoctoral Adolescent Fellow (1984–1985) at the Clinical
e-mail: ngaylor@luc.edu Research Training Program in Adolescence in Chicago, IL, which
was co-directed by Dr. Daniel Offer., Loyola University Chicago,
B. L. Ragsdale 6525 N. Sheridan Rd., Chicago, IL, 60626 USA
Teaching Associate, School of Education and Social Policy,
Northwestern University. Received Ph.D. in Psychology from A. C. Petersen
University of Rhode Island. Research interests include ethnic Visiting Professor, Humphrey Institute of Public Affairs;
identity in African American youth and the effects of exposure to President, University of Minnesota and Global Philanthropy
violence on well-being. Northwestern University, Alliance. Received Ph.D. in Measurement, Evaluation, and
633 Clark Street, Evanston, IL 60208 USA Statistical Analysis from the University of Chicago. Research
interest is in adolescent development. Dr. Petersen served as
J. Mandara Coordinator of the Clinical Research Training Program in
Assistant Professor, Human Development and Social Policy, Adolescence (1978–1982) and Associate Director (1976–80) and
Northwestern University. Received Ph.D. in Psychology from Director (1980–82) of the Laboratory for the Study of
University of California, Riverside. Primary research examines Adolescence at Michael Reese Hospital and Medical Center
the nature and effects of socialization, father’s involvement, and (Chicago, IL) where Dr. Daniel Offer served as Director of the
how they interact with gender, race, and SES to impact youths’ Department of Psychiatry. Dr. Petersen and Dr. Offer collaborated
academic and social development. on numerous research papers while working together at Michael
Northwestern University, Reese Hospital., University of Minnesota and Global
633 Clark Street, Evanston, IL 60208 USA Philanthropy Alliance USA

Springer
78 J Youth Adolescence (2007) 36:77–88

Adolescence is a developmental transition period marked A number of studies demonstrate that perceptions of higher
by rapid physical, psychological, social, and environmental support received from family and peers are associated with
change, and these life changes have been found to contribute fewer internalizing symptoms, including less depression
to the onset or exacerbation of internalizing difficulties, such and anxiety (e.g., Caldwell et al., 1998).
as anxiety and depression, during adolescence (Burton et al., In samples that are predominantly or exclusively African
2004; Petersen et al., 1991). African American adolescents American, researchers have examined the association be-
report experiencing more stressful events during this devel- tween perceived support and internalizing symptoms us-
opmental period than their White counterparts (Gonzales ing a variety of measures for internalizing symptoms. Re-
and Kim, 1997; Weist et al., 1995), potentially heightening searchers have utilized a composite measure of depressive
their risk of internalizing problems. Interestingly, a number and anxiety symptoms as an internalizing variable, and find-
of studies demonstrate that African American youth do not ings have demonstrated that higher levels of family and peer
exhibit significantly higher rates of internalizing problems support predict lower internalizing scores (Kliewer et al.,
than White youth (Maag and Irvin, 2005; Safren et al., 2000). 1998; Youngstrom et al., 2003). Although the creation of a
Thus, consistent with a resilience framework, it becomes im- composite internalizing variable is a common practice with
portant to understand the factors that protect African Ameri- measures of depression and anxiety due to overlap in the
can adolescents, despite elevated contextual risks. The com- constructs, there is literature suggesting distinctive, as well
parable rates of internalizing symptoms have been attributed as overlapping, features of the two (e.g., Kendall and Watson,
to strong social ties in the African American community 1989). Other studies have examined the effects of perceived
(Collins, 1995), and previous research confirms that social support on depression and anxiety separately. Higher levels
support received from family and peers is a factor associ- of support from family and peers have been shown to be re-
ated with lower rates of depression and anxiety in African lated to lower levels of depressive symptoms (Caldwell et al.,
American adolescents (Davis et al., 1997; Youngstrom et al., 1998). In a sample of African American youth, higher lev-
2003). What is largely missing is a clear understanding of els of parental support predicted lower levels of depressive
the mechanisms that help to link social support to healthy symptoms (Johnson and Kliewer, 1999). Also, in a sample
psychological functioning for African American youth. Pos- of pregnant and parenting adolescent females, family sup-
itive views of the self have been proposed as a mechanism port was inversely related to symptoms of depression (Davis
through which socioemotional experiences may influence et al., 1997). Low levels of parental support have also been
developmental outcomes (Harter, 1999; Offer et al., 1984). found to predict the onset of major depressive disorder in
However, positive views of one’s ethnic group are considered African American adolescent females (Stice et al., 2004).
to be equally important for functioning in African American In a sample of adolescent males, less parent and peer sup-
youth (Cross, 1991). Thus, the purpose of the current study port predicted depression longitudinally (Zimmerman et al.,
was to examine self-esteem and ethnic identity as mediators 2000).
of the relation between perceived family and peer support and Consistent findings have been reported for studies exam-
internalizing difficulties in African American adolescents. ining anxiety, with peer and family support associated with
Specifically, it is proposed that perceived support exerts its lower levels of anxiety symptoms. A longitudinal investiga-
protective influence on depression and anxiety through its tion demonstrated that family and peer support was related to
relation to higher self-esteem and ethnic identity. less anxiety for African American females (Hammack et al.,
2004). Another longitudinal investigation of African Amer-
Social support and internalizing symptoms ican youth showed that lower levels of family support were
associated with an increase in anxiety symptoms six months
Social support has been studied extensively as a psychoso- later (White et al., 1998). In a sample of African American
cial determinant of well-being in children, adolescents, and male adolescents, Zimmerman et al. (2000) demonstrated
adults (Cohen and Wills, 1985). Early research on social that perceived support measured at baseline predicted less
support utilized broad and vague conceptualizations of sup- anxiety six months later. Although it has been suggested
port, but researchers now distinguish between more specific that the relation between support and internalizing difficul-
support constructs, including the construct of perceived ties may be due to symptoms of anxiety and depression
support (Barrera, 1986; Sandler et al., 1989). Perceived (e.g., maladaptive cognitions) eroding support, findings with
support is conceptualized as the evaluation or appraisal of African American youth demonstrated that increases in anx-
the quality of support, and, compared to objective measures iety and depression symptoms do not predict lower levels of
of support, is more strongly associated with psychological perceived support over time (Zimmerman et al., 2000).
functioning (Wetherington and Hessler, 1986). Research The relation between perceived support and internalizing
on the effects of perceived support in adolescence largely symptoms may differ as a function of adolescent gender
focuses on perceptions of family support and peer support. (Hops, 1995; Kavanagh and Hops, 1994). In the studies

Springer
J Youth Adolescence (2007) 36:77–88 79

above, findings have been less consistent for adolescent self-esteem in African American adolescents (Bean et al.,
males than for females. Using the same data as examined 2003). Both parental and peer support have been found to
in the current study, Hammack et al. (2004) found that contribute to higher self-esteem in African American pread-
perceived support predicted less anxiety for females, but olescents (Franco and Levitt, 1998). A cluster analytic profile
not males. Another study found no relation between family of family and peer typologies indicated that adolescents in
support and depressive symptoms in African American groups characterized by high family and peer support re-
adolescent males (Paxton et al., 2004). Females may be ported higher levels of self-esteem (Roberts et al., 2000).
more relationship-oriented than males (Gilligan, 1982), Positive views of the self have been shown to be important
thus strengthening the association between support and in helping African American adolescents evade internalizing
outcomes. Unfortunately, little research has examined this difficulties (Harter, 1999). Depressed adolescents report sig-
association separately for males and females in the same nificantly lower self-esteem than non-depressed adolescents
study, making it difficult to draw firm conclusions regarding (Repetto et al., 2004) and this inverse association has oc-
gender differences. curred in African American adolescents (Fitzpatrick et al.,
Nevertheless, the evidence of a protective function of per- 2005). Further, in African American males, higher levels of
ceived support from family and peers in the reduction of de- self-esteem in adolescence predict lower levels of depression
pression and anxiety symptoms highlights its importance in in adulthood (Mizell, 1999). Although considerably less re-
the daily lives of African American youth. A model explain- search has been conducted on the association between self-
ing the protective effects of support suggests that support esteem and anxiety, findings show that higher self-esteem
may strengthen intervening variables that lead to psycholog- is related to fewer anxiety symptoms for African American
ical outcomes (Sandler et al., 1989). A greater understanding adolescents (Matthews and Odom, 1989; Youngstrom, et al.,
of the mechanisms or intervening variables responsible for 2003; Zimmerman et al., 1995).
the effects of perceived support is needed for African Amer- Evidence that perceived support is related to higher self-
ican youth. esteem, and evidence that higher self-esteem is related to
symptom reduction, suggests that self-esteem may serve as a
Self-esteem as a mediator of perceived support mediator of the relation of perceived support to internalizing
symptoms. Despite the evidence supporting the conditions
Perceived support has been proposed to relate to positive psy- for mediation and the calls to examine self-esteem as an
chological functioning through a number of mechanisms, intervening variable, little research has examined a media-
including provision of a sense of predictability and stabil- tional model of self-esteem. In two exceptions, adolescent
ity, recognition of self-worth, opportunity to process ex- self-reports of self-esteem were found to mediate the relation
periences with significant others, and appraisals of stress of perceived support to psychological well-being (Yarcheski
(Cohen and Wills, 1985; Hammack et al., 2004; Schmeelk- et al., 2001), and adolescent self-report of support was related
Cone and Zimmerman, 2003). Among the many mechanisms to lower internalizing symptoms through self-esteem scores
proposed to explain the relation between perceived support over time (DuBois et al., 2002). Although these investiga-
and functioning, the role of self-esteem has been examined tions were conducted on diverse ethnic samples, additional
empirically (DuBois et al., 2002; Yarcheski et al., 2001). research is warranted to examine the mediational effects of
Self-esteem is conceptualized as the overall evaluation of self-esteem in African American adolescent samples.
one’s worth or value (Rosenberg, 1965). Due to the role of
self-esteem in positive youth development, most research Ethnic identity as a mediator of perceived support
has examined it as an outcome variable or as a correlate of
psychosocial outcomes. Very little research has incorporated Social identity theory asserts that there is an underlying need
self-esteem as a mediator, despite calls to examine its in- to maintain self-esteem and that this need is linked to eth-
tervening role (DuBois and Tevendale, 1999; Sandler et al., nic identity in non-White ethnic groups (Tajfel and Turner,
1989). Consistent with conceptualizations of mediation ef- 1986). Thus, when examining the influence of self-esteem on
fects (Baron and Kenny, 1986), a mediational model of self- African American adolescents, it is critical to also examine
esteem states that various psychosocial and environmental the role of ethnic identity. Ethnic identity refers to one’s sense
factors influence self-esteem, which in turn, influences psy- of belonging to an ethnic group and the part of one’s think-
chological and behavioral outcomes (DuBois and Tevendale, ing, perceptions, feelings, and behavior that is due to ethnic
1999). group membership (Rotheram and Phinney, 1987). Sociolo-
Consistent with the conditions necessary for a mediation gists distinguish between race and ethnicity, with the former
model, perceived support has been found to be associated encompassing groups who share similar physical attributes
with self-esteem in adolescence. For instance, higher lev- (e.g., Black) and the latter describing groups who share sim-
els of parental support have been shown to predict higher ilar values, customs, language, etc. (e.g., African American,

Springer
80 J Youth Adolescence (2007) 36:77–88

Senegalese, Haitian). Although the terms ethnic identity and tions for a mediational role of ethnic identity (Baron and
racial identity are used interchangeably in the literature, the Kenny, 1986). Similar to self-esteem, a paucity of research
current study utilized the term ethnic identity because only has examined the intervening function of ethnic identity on
African American adolescents were included. It is generally the relation of perceived support to internalizing outcomes.
agreed that ethnic identity is a multidimensional construct One study with African American adolescents examined
assessing a person’s inner feelings, daily experiences, and these relations and found that ethnic identity did not me-
social and interpersonal interactions with others (Steinberg diate the association between maternal support and symp-
and Morris, 2001). Ethnic identity and self-esteem may show toms of depression and anxiety, primarily due to the lack of
minimal associations to one another in younger children, but direct effects from ethnic identity to the internalizing symp-
research suggests a convergence during adolescence as indi- toms (Caldwell et al., 2002). Instead, ethnic identity was
viduals become more aware of the social significance of their indirectly associated with depression and anxiety through its
ethnicity (Smith et al., 1999; Spencer and Dupree, 1996) relation with perceived stress. Additional studies are needed
and begin to encounter ethnic group barriers (Oyserman to examine the mediating role of ethnic identity.
et al., 2001). Empirical evidence consistently demonstrates
a positive relation between ethnic identity and self-esteem in The current study
African American adolescents (Phinney et al., 1997; Row-
ley et al., 1998; Smith et al., 1999; Spencer et al., 2001). The purpose of the current study was to examine the extent
This relation supports the perspective that self-perceptions to which self-esteem and ethnic identity mediate the relation
are linked to the perceptions of one’s group and positive lev- of peer and family support to anxiety and depression. Prior
els of both are necessary for adaptive functioning (Akbar, research demonstrates that support from family and peers is
1985). Thus, ethnic identity may also function as a mediator important for protecting African American youth from el-
of the relation between perceived support and internalizing evated levels of depression and anxiety, despite higher risk
difficulties. for these difficulties. Existing research falls short in provid-
Unlike self-esteem, less research has examined the role ing a clear understanding of the mechanisms that help link
of “non-racial socialization” determinants, such as perceived perceived support to healthy psychological functioning for
support, on ethnic identity. Consistent with self-esteem, the African American youth. The current study builds on prior
findings that exist suggest that support is associated with research demonstrating that perceived support from fam-
positive ethnic identity. In a sample of African American ily and peers is predictive of fewer depression and anxiety
adolescents, perceived support from family was found to be symptoms (Caldwell et al., 1998; Zimmerman et al., 2000)
positively related to adolescents’ beliefs that racial socializa- by examining mechanisms of this relation. Existing research
tion (i.e., teaching of cultural pride) was important (Steven- shows evidence for self-esteem as a mediator of perceived
son et al., 1996). In another study with African American support and internalizing symptoms, but almost no research
adolescents, perceived maternal support was significantly has examined this model in an African American sample.
related to adolescents’ positive feelings about their ethnic Further, the positive association between self-esteem and
group (Caldwell et al., 2002). ethnic identity in ethnic minority youth suggests that eth-
Research on the relation between ethnic identity and inter- nic identity may also be an important intervening variable.
nalizing symptoms in adolescence shows that ethnic identity The hypotheses centered on the general notion that factors
is consistently linked to internalizing symptoms. For ex- related to self-esteem and ethnic identity would represent
ample, a study of adolescents from diverse ethnic groups an important mechanism by which support reduces anxiety
showed a significant negative relation between positive eth- and depression. Consistent with literature suggesting both
nic identity and depression (Roberts et al., 1999), consistent distinctive and overlapping features of anxiety and depres-
with other findings with diverse adolescent samples (Yasui sion (e.g., Kendall and Watson, 1989), the two constructs
et al., 2004). Other research demonstrated that negative eth- (e.g., depression and anxiety) were assessed by separate
nic identity was associated with higher levels of depression instruments; therefore the current study focuses on the dis-
and anxiety in African American adolescents (Arroro and tinctiveness of the outcomes with regard to the hypothesized
Zigler, 1995), consistent with findings in African American protective factors. Consistent with past research, it was pre-
college students (Munford, 1994; Plant and Yanico, 1991). dicted that higher levels of perceived support from family
Still, another investigation of African American youth found and peers would be predictive of lower levels of depression
that ethnic identity was negatively related to depression, but and anxiety. Building on past research, it was predicted that
showed no relation to anxiety (McMahon and Watts, 2002). perceptions of higher support would be related to lower de-
Taken together, findings suggest that perceived support is pression and anxiety scores via its relation to higher levels of
predictive of ethnic identity, and ethnic identity is predictive self-esteem and ethnic identity. Because no studies have ex-
of internalizing difficulties, providing the necessary condi- amined these relations separately for adolescent males and

Springer
J Youth Adolescence (2007) 36:77–88 81

females in an African American sample, the models were gift certificates, sports equipment, and games. According to
tested for invariance across gender; however, exiting research the Experience Sampling Method (ESM), participants car-
provided little support for making specific predictions about ried watches for one week that were programmed to signal
gender differences. at random times approximately every 90 minutes when the
youth were not in school. Designed to minimize disruption
in the school context, participants received two signals dur-
Method ing the school day. This signaling schedule was designed to
optimize sampling of time when the youth may have been
Participants at risk for exposure to violence. After each signal, the youth
completed a brief self-report form in a booklet they carried
Participants in the current study were part of a larger, longi- each day. Trained research assistants carefully instructed par-
tudinal project focused on exposure to community violence. ticipants in how to complete these forms. To ensure accurate
Since the purpose of the study was to examine the predictors compliance with the ESM, research staff visited participants
and effects of exposure to community violence, six public each day at school to answer participant questions and to
schools in Chicago were selected based on high neighbor- monitor the quality and completeness of the ESM forms.
hood crime statistics obtained from the Chicago Police De- Over the course of the week, youth responded to a maxi-
partment for the year preceding data collection. Of youth mum of 51 signals. To be included in the sample, youth had
recruited from the schools, 58% agreed to participate. This to complete 15 forms as a minimum. Consistent with ac-
participation rate is consistent with other studies conducted ceptable standards of compliance with the ESM (see Larson,
with inner-city youth (e.g., Allison et al., 1999; Cooley- 1989; Larson et al., 2001), the youth in this study responded
Quille and Lorion, 1999). During the first year of this three- to a median number of 42 signals, yielding an 82% signal
year study, a sample of 301 African-American sixth grade response rate. During the daily school visits, research person-
students completed questionnaires and the time sampling in- nel also administered surveys. To ensure confidentiality and
strument. Wave 2 data collection occurred one year later and the quality of the data, students were administered question-
wave 3 data collection occurred one year after wave 2. Be- naires in small groups under close supervision by research
cause the ethnic identity measure was not assessed in the first personnel. One and two years later, when participants were
year of the study, data from the 6th grade was not utilized in in the seventh and then the eighth grade, they were invited
the analyses. to again participate in the study. The same procedures were
The final sample in the current study included 227 ado- repeated in the second and third year of data collection.
lescents for which data was available for the second (7th
grade) and third (8th grade) years of the study. Data on these Measures
227 youth were combined across 7th and 8th grades (i.e.,
the average of each variable across 7th and 8th grade was Perceived support
derived) to retain the largest possible sample. The sample
was composed of 85 males (37%) and 142 females (63%), To obtain a report of adolescents’ perceived support, youth
and had a mean age of 12.55 (SD = .69). The non-retained completed an adaptation of the Survey of Children’s Social
youth were similar to the retained youth in terms of parental Support (SCSS; Dubow and Ullman, 1989) which assesses
marital status, annual household income, and parental educa- the perceived availability of social support and the subjective
tion. There were no significant differences between retained appraisal of the quality of support received. Two subscales
and non-retained youth in depressive symptoms or anxiety at were used from this measure: perceived family support and
Time 1. The youth came from low-income households, with perceived peer support. The two subscales together contained
a median family income between $10,000 and $20,000 as 10 items rated on a 5-point scale (“never” to “always”) and
reported by a parent or guardian. Most parents (83%) had at included questions such as “I can count on family for advice
least a high school degree, with 10% reporting a college or and help.” Consistent with prior research, the 10 items were
graduate/professional degree. Nearly half the sample (48%) averaged to yield a summary score of perceived support for
lived in single-parent households. The median number of each participant (Hammack et al., 2004). Reliability for the
people living in the household was five. scale was .70 in the current sample.

Procedure Ethnic identity

Participation was not allowed without written child assent Ethnic identity was measured with the positive ethnic atti-
and parent consent. As an incentive to participate, students tudes and sense of belonging scale of the Multigroup Eth-
received prizes at the end of each data collection such as nic Identity Measure (MEIM; Phinney, 1992). Due to time

Springer
82 J Youth Adolescence (2007) 36:77–88

constraints of data collection, only one subscale was utilized higher levels of depressive symptoms. In the current sam-
during data collection. The positive ethnic attitudes and sense ple, internal consistency, as indicated by Cronbach’s alpha,
of belonging subscale is represented by five items that as- was .83.
sess feelings of belonging and attachment to one’s ethnic
group, ethnic pride, and good feelings about one’s ethnic-
ity. The subscale consists of 5 items rated on a four-point Results
scale ranging from “strongly disagree” to “strongly agree.”
Consistent with Phinney (1992), the scale score was derived Structural equation modeling (SEM) was utilized to test the
by obtaining the mean score across items, with higher scores theoretical model for depressive symptoms and anxiety. Be-
indicating higher ethnic identity. In the current sample, inter- cause of missing data at time 1 or time 2 for various ado-
nal consistency, as indicated by Cronbach’s alpha coefficient, lescents, the current study utilized the average of 7th grade
was .81. and 8th grade data for the indicator variables of each factor
in the model, and the average score was used as a single
Self-esteem indicator of its latent variable in the SEM analyses. This
approach is equivalent to using multiple indicators in the
Self esteem was examined with five items from the ESM model as long as the variables are standardized first, the
booklet. Respondents were prompted with the following paths from the composites to the latent variables are fixed to
question: “How were you feeling when you were signaled?” the square root of the internal consistency reliability of the
and rated feelings about self, including importance, tough- composites, and their residuals are fixed to the error vari-
ness, calmness, pride, and respect. Four of the feelings about ance (Loehlin, 1998). This approach has a few important
self were rated on a 4- point scale ranging from “not at all” advantages for our purpose. First, utilizing the averages of
to “very much” and one feeling about self was rated on a time 1 and time 2 allows for the increase of the sample size
7-point scale. These items were chosen after conducting a without imputing missing data. Given the relatively small
principal components factor analysis with a varimax rotation sample of males (n = 85), using as much available data
on the ESM items related to individual identity. The factor as possible was important. Second, the averaging approach
analysis showed one strong factor for the 7th grade students reduces the complexity of the model and interpretation of
and when the analysis was repeated for the 8th grade stu- the results because only the structural model is assessed
dents, the exact same factor structure emerged. The average (Loehlin, 1998).
of the 7th and 8th grade scores were derived to create the
total score for self esteem. Reliability was established at .94 Descriptive statistics, t-tests, and correlations
in the current sample.
It was essential to first test whether or not the data from boys
Anxiety symptoms and girls differed significantly, to determine whether to fit
one model or two. The means, standard deviations, and tests
The trait subscale of the State-Trait Anxiety Inventory for of gender differences for the study variables are presented in
Children (STAI-C; Spielberger et al., 1973) was used to Table 1. The results of the independent-samples t-tests
assess anxiety symptoms. The STAI-C is a 20-item measure showed that females perceived significantly more support
on which youth rate the frequency of anxiety symptoms on from family and peers (p < .01), as well as significantly
a 3-point scale (hardly ever to often). Questionnaire items more anxiety than did males (p < .01). No gender difference
include “I worry about school” and “I get a funny feeling in emerged on depression or self-esteem scores.
my stomach.” Higher scores indicate higher levels of anxiety
symptoms. The scale yielded an alpha of .90 in the current
Table 1 Means and Standard Deviations of all Study Variables by
sample.
Adolescent Gender

Depressive symptoms Adolescent Gender


Female (n = 142) Male (n = 85) t (225)

Participants completed the Children’s Depression Inventory Social Support 12.54 (2.37) 11.73 (2.38) 2.49∗
(CDI; Kovacs, 1985), a 27-item, self-report instrument on Ethnic Identity 3.46 (.045) 3.33 (0.61) 1.68
which youth rate their level of symptoms for the two weeks Self-Esteem 3.95 (0.96) 3.92 (0.96) 0.24
prior to administration on a 3-point scale (e.g., “I am sad Depression 6.25 (4.96) 5.29 (5.04) 1.40
Anxiety 11.10 (6.84) 8.50 (6.20) 2.86∗
once in a while” to “I am sad all the time”). Responses to

the items are added to obtain a summary score of depres- p < .01.
sive symptoms for each participant. Higher scores indicate

Springer
J Youth Adolescence (2007) 36:77–88 83

Table 2 Zero-Order Correlations among Study Variables for Fe- gender differences, we also used the multiple-group analysis
males and Males facility in Amos 5.0 to fit the data simultaneously to both
1 2 3 4 5 genders. This analysis helped answer two questions. First,
the multiple-group analysis assessed if the path diagram fit
1. Social Support 1.00 .18∗ .31∗∗ −.48∗∗ −.22∗∗
2. Ethnic Identity .35∗∗ 1.00 −.02 −.15 −.13 the data equally well for both females and males; and if
3. Self Esteem .42∗∗ .28∗∗ 1.00 −.29∗∗ −.26∗∗ so, it assessed if the regression weights were different be-
4. Depression −.59∗∗ −.48∗∗ −.27∗∗ 1.00 .42∗∗ tween the genders. The multiple-group analysis in Amos
5. Anxiety −.30∗∗ .06 −.03 .34∗∗ 1.00 5.0 proceeds by automatically generating several nested
models. Each lower level model adds additional constraints
Note. Females’ correlations are above the diagonal and males’ corre-
lations are below. and can be tested for differences from the previous models

p < .05. ∗∗ p < .01. (Arbuckle, 2003). For the purposes of the current study, only
the completely unconstrained model and the model that con-
strained the regression weights to be equal (i.e., the structural
Table 2 presents the zero-order correlations of the com- weights model) were important. A test of the unconstrained
posite variables for males and females separately. Consistent or first level model assesses the fit of the data to both groups.
with the predictions, perceived support was positively corre- If the unconstrained model fits the data well, the test of the
lated with self-esteem and ethnic identity for both genders, structural weights model gives an estimate of the degree to
indicating that the more support the adolescents perceived, which the regression weights are similar in both groups. The
the more positive they tended to feel about themselves and facility also computes critical ratio tests for each parameter
other African Americans. Further, the correlation coefficients in both samples (Arbuckle, 2003).
between perceived support and internalizing difficulties were The results of the multiple-group analysis with de-
the strongest among the study variables. As expected, per- pressive symptoms serving as the outcome showed that
ceived support was also negatively correlated with depressive the unconstrained model fit the data well, χ 2 (2) = 2.77,
symptoms and anxiety for both groups. Unexpectedly, eth- p = .25, CFI = .99, RMSEA = .04, indicating that the path
nic identity did not correlate significantly with self-esteem model fit equally well for both genders. In fact, for males,
for females; whereas ethnic identity showed a moderately χ 2 (1) = 1.61, p = .20, CFI = .99, RMSEA = .08, and for
sized positive relation with self-esteem for males, which im- females, χ 2 (1) = 1.15, p = .28, CFI = 1, RMSEA = .03,
plies that those who had more positive feelings about African the model fit very well (see Figs. 1 and 2). However, when
Americans also tended to have more positive feelings about constraining the regression weights to be equal, there was a
themselves. Ethnic identity also had a negative relation with significant decrease in model fit, χ 2 (11) = 31.1, p = .001,
depression for males, but no relation to their anxiety scores. which indicates that the overall pattern of regression weights
Self-esteem was moderately and negatively correlated with were significantly different between females and males.
anxiety and depression for females, but only depression for
males.
MEIM
Thus, the correlation matrices implied that the path model
might be different for females and males. The strong and .96
consistent effect of perceived support for both genders also
Ethnic
implied that self-esteem and ethnic identity only partially ity
Identity
(R 2 -.32
mediate its effect on internalizing problems. Furthermore, R2 =.15)
= 15
anxiety was not related to male’s ethnic identity or self- .39
esteem at all, so formal tests with anxiety as the outcome Social -.55
Depressive
for males were unnecessary. Despite some contrary findings, Symptoms
Support
(R2 =.51)
formal tests of a modified version of the model were con-
.47
ducted with SEM. .94 .96
Self- .06
SCSS ity CDI
Esteem
Structural equation analyses R22 = 22
(R

.97
Because of the obvious direct effect of perceived support,
the mediation model was modified to reflect a partial medi- ESM
ation model. The modified theoretical model was assessed Fig. 1 Maximum likelihood estimation of the model predicting de-
with latent variable SEM using Amos 5.0 with maximum pressive symptoms for males (N = 85). Solid lines indicate statistically
likelihood estimation. Because of the predicted and possible significant paths. χ 2 (1) = 1.61, p = .20, CFI = .99, RMSEA = .08

Springer
84 J Youth Adolescence (2007) 36:77–88

MEIM MEIM
.92 .92

Ethnic Ethnic
ity ity
Identity
Identity
(R2 = .04 -.08 R2 = .04 -.12
.20 .20
Depressive
Social -.46 Social -.14 Anxiety
Symptoms
Support Support (R2 =.12)
(R2 =.30)
.33 .33
.94 .96 .94 .97
Self- -.15 Self- -.23
SCSS Identity CDI SCSS STAI-C
Esteem Esteem
R22 ==.11)
(R .11 R2 = .11

.97 .97
ESM ESM

Fig. 2 Maximum likelihood estimation of the model predicting depres- Fig. 3 Maximum likelihood estimation of the model predicting anxi-
sive symptoms for females (N = 142). Solid lines indicate statistically ety for females (N = 142). Solid lines indicate statistically significant
significant paths. χ 2 (1) = 1.15, p = .28, CFI = 1, RMSEA = .03 paths. χ 2 (1) = 1.15, p = .28, CFI = 1, RMSEA = .03

Discussion
An examination of the critical ratios for the differences
between the regression weights showed that the path from The protective effects of perceived support on the manifes-
perceived support to ethnic identity was larger for males tation of internalizing symptoms in African American youth
than for females with a one-tailed test (z = 1.8, p = .03). are well-documented. The current study proposed that per-
The path from self-esteem to depressive symptoms was ceived support from family and peers would exert its pro-
significant for females, but not males (z = 1.65, p = .05). tective effect on depression and anxiety through its positive
Also, the size of the path from ethnic identity to depressive influence on adolescents’ perceptions of self and percep-
symptoms was slightly larger for males than for females tions of their ethnic group. Thus, the purpose of the current
(z = 1.4, p = .08). For males, ethnic identity had a sig- study was to examine the ability of self-esteem and ethnic
nificantly larger impact on depressive symptoms than did identity to mediate the relation of perceived support to inter-
self-esteem (z = 2.6, p = .005); but for females, there was nalizing difficulties of depression and anxiety in a sample of
no difference between the two (z = .45, p = .33). Further, African American adolescents. In addressing this possibility,
perceived support, self-esteem, and ethnic identity explained the current study built on previous research by examining a
over 30% of the variance in females’ depressive symptoms, more complex model of the protective role of perceived sup-
but they explained over 50% of the variance for males. port for African American youth. Consistent with the first
This same multiple-group analysis was then repeated hypothesis, perceived support from family and peers was
while using anxiety as the main outcome. The results of predictive of fewer depressive and anxiety symptoms. The
this analysis showed that the unconstrained model was a larger model implied that a combination of self-esteem and
poor fit to the data, χ 2 (7) = 16.74, p = .02, indicating that ethnic identity can account for some, but not all of the effects
the theoretical model does not apply equally well to females of perceived support on African American adolescents’ de-
and males. This is not surprising given that for males, eth- pressive symptoms, showing partial support for the second
nic identity and self-esteem are not significantly related to hypothesis. Even when self-esteem and ethnic identity were
anxiety. Thus, the analyses were conducted for females only. accounted for, perceived support still had a very large direct
These results showed that the general theoretical model fit relationship with depressive symptoms. Furthermore, ethnic
the females data very well, χ 2 (1) = 1.15, p = .28, CFI = 1, identity seems to be more relevant and accounted for more of
RMSEA = .03, accounting for around 12% of the variance the perceived support effect for males, whereas self-esteem
in their anxiety scores (see Fig. 3). showed more of an impact on females’ depressive symptoms.
The mediational model for anxiety showed partial support
for the hypothesis in female participants. In contrast to the
hypotheses, the mediation model of anxiety symptoms was
not supported for males. Self-esteem had similar impact on

Springer
J Youth Adolescence (2007) 36:77–88 85

females’ anxiety and depression scores. For males, neither males are monitored (Beck et al., 2003; Svensson, 2003),
self-esteem nor ethnic identity related to their anxiety scores. which may result in females being protected from some of
the more negative experiences that, according to ethnic iden-
The protective effects of perceived support tity process models, would initiate the awareness of ethnic
identity (Cross, 1991; Phinney, 1990). Therefore, it may be
The current study proposed a mediational model that in- expected that adolescent females in early adolescence have
cluded adolescents’ perceived support from both family and less awareness of the importance of ethnic identity than males
peers. The findings support previous research demonstrating and measures of ethnic identity will show weaker relations
that African American adolescents perceptions of support to other determinants and outcomes for females at this age.
from family and peers is related to fewer symptoms of However, additional research is warranted before definitive
depression and anxiety (Caldwell et al., 1998; Johnson and conclusions can be made.
Kliewer, 1999; Kliewer et al., 1998), providing additional Another explanation may be the increased salience of
backing for the importance of social ties in the African group membership for African American males versus fe-
American community. The findings were also consistent males. Because African American males spend more time
with prior research demonstrating that higher levels of with peers than family (Coates, 1987), they may gain more
perceived support are related to more positive perceptions of exposure to experiences that require group members to show
self (Bean et al., 2003; Franco and Levitt, 1998) and percep- more similarity to one another and less similarity to other
tions of one’s ethnic group (Caldwell et al., 2002; Stevenson, ethnic groups (Spencer et al., 1995). These group dynamics
1996). A more supportive network composed of family and may increase the ethnic identity of African American ado-
peers may buffer the effects of events that directly threaten lescent males, providing the link from perceived support to
self-esteem or reduce the distress associated with stressors outcomes. Finally, a sense of belonging to one’s ethnic group
in adolescence and ethnicity-linked stressors (e.g., racism) may be more salient for other outcomes in African Ameri-
by promoting appraisals of stressors that maintain esteem can adolescent females. For example, Belgrave et al., (2000)
(Sandler et al., 1989). The examination of the association found that high ethnic identity, operationalized as sense of
between perceived support and ethnic identity in the current belonging, was associated with less risky sexual attitudes for
study represents a growing area of research investigating African American female adolescents.
the impact of “non-ethnic socialization” processes on ethnic Inconsistent with our predictions, self-esteem and ethnic
identity (Caldwell et al., 2002). Specifically, research has identity did not show a mediating role on anxiety symp-
consistently shown that direct ethnic socialization behaviors toms for males in the current study, primarily due to the
by individuals in one’s social network, especially parents, lack of associations of self-esteem and ethnic identity to
are predictive of one’s ethnic identity. What this study has anxiety symptoms. The lack of associations is incongru-
shown is that more general processes, such as support from ent with most research on African American adolescents
others, are associated with adolescents’ feelings about of (Arroro and Zigler, 1995; Zimmerman et al., 1995); how-
their ethnic group. Consistent with the attachment theory ever, one study demonstrated that ethnic identity was related
(Bowlby, 1969), supportive relationships with family and to depression, but not anxiety (McMahon and Watts, 2002).
friends may become part of adolescents’ internal working Perhaps, other components of ethnic identity, such as engag-
models of relationships, and these working models may then ing in ethnic behaviors, may be more important for males’
be generalized to other members of their ethnic group, thus anxiety than a sense of belonging to one’s ethnic group
contributing to positive feelings towards their ethnic group. or self-perceptions. For example, African American ado-
lescents who engage in behaviors perceived to be reflective
Self-esteem and ethnic identity as a mediators of non-Black ethnic behaviors attitudes experience elevated
levels of anxiety (Arroro and Zigler, 1995). Because African
Tests of the hypothesized model showed that both self- American males spend more time outside of the home than
esteem and ethnic identity partially mediated the effects of females, it may be important to exhibit ethnic behaviors to
perceived support on depression and anxiety. The most strik- gain acceptance and conform to the expectations of other
ing findings of the current study are the gender differences in group members. Still, other research has demonstrated that
this mediational model. The gender differences demonstrated the relation between ethnic identity and anxiety was indi-
that ethnic identity was a more salient mediator for the ef- rect through perceived stress (Caldwell et al., 2002). African
fects of perceived support on males’ depression; whereas, American males experience more stressful events than fe-
self-esteem was a more salient mediator for perceived sup- males (Gibbs, 1988), and the inclusion of perceived stress
ports’ effects on depression in females. One plausible ex- in future research may provide clarification on the role of
planation may stem from research demonstrating that ado- self-esteem and ethnic identity on anxiety symptoms.
lescent females are monitored more closely than adolescent

Springer
86 J Youth Adolescence (2007) 36:77–88

Drawbacks of the current study should be noted. First, the hensive models can lead to a greater understanding of the
findings of the path analyses in the current study prevented contextual determinants on the development of perceptions
causal inferences. The data indicated the existence of these of self and perceptions of one’s group in African American
relations, and further analysis is necessary to definitively adolescents.
state causation. Also, the participants included in the current
study were from low-income, inner-city neighborhoods with Acknowledgments This research was funded by grant R01-MH57938
from the National Institute of Mental Health awarded to Maryse H.
high rates of community violence, a contextual issue that Richards. The authors acknowledge the contributions of the Project
cannot be overlooked in regards to the generalizability of Directors, Cathy Flynn, Brian Ragsdale, and Phil Hammack, and the
the findings. The current study utilized a dimension of eth- dedicated team of research assistants at Loyola University Chicago
nic identity that, despite its theoretical importance, has been and the University of Illinois at Urbana-Champaign without whom
collection of this valuable data would not have been possible. We are
relatively neglected in past research (Phinney, 1990). Future extremely grateful for the students, school administrators, teachers, and
research will benefit from inclusion of additional dimen- parents who volunteered their time to participate in this research.
sions of ethnic identity. The assessment of other dimensions
of ethnic identity, which may reflect other stages in the pro-
cess of ethnic identity development, may be important for References
understanding the nature of this process in adolescence.
Although numerous studies have demonstrated that Akbar N (1985) Our destiny: Authors of a scientific revolution. In
perceived support is related to more positive outcomes in McAdoo HP, McAdoo JL (Eds) Black children: Social, educa-
African American adolescents, the current study is one of tional, and parental environments. Sage Publications, California,
pp 17–31
only a few studies to extend these findings by examining Allison KW, Burton L, Marshall S, Perez-Febles A, Yarrington J, Kirsh
possible mediators of the relation between support and LB, Merriwther-DeVries C (1999) Life experiences among urban
outcomes. Further, the current study may be the only study adolescents: Examining the role of context. Child Deve 70:1017–
to date that assesses the influence of both perceptions of 1029
Arbuckle J (2003) Amos 5.0 update to the Amos user’s guide. Small
self and perceptions of one’s ethnic group as explanatory Waters Corporation, Chicago.
mechanisms for African American adolescents in the same Arroyo C. G, Zigler E (1995) Racial identity, academic achievement,
study. Theoretical writings on African American children and the psychological well-being of economically disadvantaged
and adolescents have consistently discussed the importance adolescents. J Pers and Soc Psychol 69:903–914
Baron R. M, Kenny DA (1986) The moderator-mediator variable dis-
of both constructs to well-being (e.g., Cross, 1991). Previous tinction in social psychological research: Conceptual, strategic,
research examining the mediating role of ethnic identity and statistical considerations. J Pers and Soc Psychol 51:1173–
did not include adolescents’ perceptions of self-esteem 1182
(Caldwell et al., 2002), and the results of the current Barrera M (1986) Distinctions between social support concepts, mea-
sures, and models. Am J Community Psychol 14:413–445
study show differences in the mediating roles of these two Bean R. A, Bush KR, McKenry PC, Wilson SM (2003) The impact of
constructs. The finding that self-esteem and ethnic identity parental support, behavioral control, and psychological control on
account for some of the relation between perceived support the academic achievement and self-esteem of African American
and psychological outcomes is encouraging and suggests and European American adolescents. J Adolesc Res 18:523–541
Beck KH, Malele JR, Boekeloo BO (2003) Parental monitoring and
that further research should explore the mediating effects of adolescent alcohol risk in a clinic population. Am J Health Behav
these variables for other dimensions of ethnic identity. 27:108–113
In summary, the current study highlights the importance Belgrave FZ, Van Oss Marin B, Chambers DB (2000) Cultural, contex-
of considering mechanisms that may explain the relation tual, and intrapersonal predictors of risky sexual attitudes among
urban African American females in early adolescence. Cultur
of perceived support from family and peers to internaliz- Divers Ethni Minor Psychol 6:309–322
ing difficulties in African American adolescents. By provid- Bowlby J (1969) Attachment and loss. Vol. 1. Attachment. Basic Books,
ing data on the role of both self-esteem and ethnic identity New York
in the mediation process among an under-researched pop- Burton E, Stice E, Seeley JR (2004) A prospective test of the stress-
buffering model of depression in adolescent females: No support
ulation, the current research generates valuable questions once again. J Consult Clin Psychol 72:689–697
for future research. The results imply that a combination of Caldwell CH, Antonucci TC, Jackson JS (1998) Supportive/conflictual
self-esteem and ethnic identity can account for some of the family relations and depressive symptomatology: Teenage mother
effects of perceived support on African American adoles- and grandmother perspectives. Family Relat 47:395–402
Caldwell CH, Zimmerman MA, Bernat DH, Sellers RM, Notaro PC
cents’ internalizing difficulties. Of critical importance to fu- (2002) Racial identity, maternal support, and psychological dis-
ture studies on ethnic identity are the implications for gender tress among African American adolescents. Child Dev 73:1322–
differences. Future research may benefit from further exam- 1336
ining gender differences in the influence of ethnic identity Coates DL (1987) Gender differences in the structure and support
characteristics of black adolescents’ social networks. Sex Roles
and the salience of different dimensions of ethnic identity 17:667–687
for adolescent males and females. Perhaps more compre-

Springer
J Youth Adolescence (2007) 36:77–88 87

Cohen S, Wills TA (1985) Stress, social support, and the buffering Larson R (1989) Beeping children and adolescents: A method for study-
hypothesis. Psychol Bull 98:310–357 ing time use and daily experience. J Youth and Adolesc 18:511–
Collins RL (1995) Issues of ethnicity in research on the prevention of 530
substance abuse. In Botvin G, Schinke S, Orlandi M (Eds) Drug Larson R, Richards M, Sims B, Dworkin J (2001) How urban African
Abuse Prevention with Multi-Ethnic Youth. Sage Publications, American young adolescents spend their time: Time budgets for
Thousand Oaks, CA, pp 28–45 locations, activities, and companionship. Am J Community Psy-
Cooley-Quille M, Lorion R (1999) Adolescents’ exposure to com- chol 29:565–597
munity violence: Sleep and psychophysiological functioning. J Loehlin JC (1998) Latent variable models: An introduction to factor,
Community Psychol 27:367–375 path, and structural analysis (3rd). Lawrence Erlbaum Associates,
Cross WE (1991) Shades of black: Diversity in African-American iden- Mahwah, NJ
tity. Temple University Press, Philadelphia Maag JW, Irvin DM (2005) Alcohol use and depression among African
Davis AA, Rhodes JE, Hamilton-Leaks J (1997) When both parents American and Caucasian adolescents. Adolesc 40:87–101
may be a source of support and problems: An analysis of pregnant Matthews DB, Odom BL (1989) Anxiety: A component of self-esteem.
and parenting female African American adolescents’ relationships Elementary School Guid Couns 24:153–159
with their mothers and fathers. J Res on Adolesc 7:331–348 McMahon SD, Watts RJ (2002) Ethnic identity in urban African Amer-
DuBois DL, Burk-Braxton C, Swenson LP, Tevendale HD, Lockerd ican youth: Exploring links with self-worth, aggression, and other
EM, Moran BL (2002) Getting by with a little help from self and psychosocial variables. J Community Psychol 30:411–431
others: Self-esteem and social support as resources during early Mizell CA (1999) Life courses of African American men’s depression:
adolescence. Dev Psychol 38:822–839 Adolescent parental composition, self-concept, and adult earnings.
DuBois DL, Tevendale HD (1999) Self-esteem in childhood and ado- J Black Stud 29:467–490
lescence: Vaccine or epiphenomenon? Appl Prev Psychol 8:103– Munford MB (1994) Relationship of gender, self-esteem, social class,
117 and racial identity to depression in blacks. Journal of Black Psy-
Dubow EC, Ullman DG (1989) Assessing social support in elementary chology 20:157–174
school children: The Survey of Children’s Social Support. J Clin Offer D, Ostrov E, Howard KI (1984) Patterns of Adolesc, self-image.
Child Psychol 181:52–64 Jossey-Bass, San Francisco, CA
Fitzpatrick KM, Piko B, Wright D, LaGory M (2005) Depressive symp- Oyserman D, Harrison K, Bybee D (2001) Can racial identity be pro-
tomatology, exposure to violence, and the role of social capi- motive of academic efficacy? Int J Behav Dev 25:379–385
tal among African-American adolescents. Am J Orthopsychiatry Paxton KC, Robinson WL, Shah S, Schoeny ME (2004) Psychologi-
75:125–137 cal distress for African American adolescent males: Exposure to
Franco DL, Striegel-Moore RH (2002) The role of body dissatisfac- community violence and social support as factors. Child Psychia-
tion as a risk factor for depression in adolescent girls: Are the try Hum Dev 34:281–295
differences Black and White? J Psychosom Res 53:975–983 Petersen AC, Sarigiani PA, Kennedy RE (1991) Adolescent depression:
Gibbs JT (1988) Young Black males in America: Endangered, embit- Why more girls? J Youth Adolesc 20:247–271
tered, and embattled. In Gibbs JT (Ed) Young, Black, and male in Phinney JS (1990) Ethnic identity in adolescence and adulthood: A
America: An endangered species Auburn House, Massachusetts, review and integration. Psychol Bull 108:499–514
pp 1–36 Phinney JS (1992) The Multigroup Ethnic Identity Measure: A new
Gilligan C (1982) In a different voice: Psychological theory and scale for use with adolescents and young adults from diverse
women’s development. Harvard University Press, Cambridge, groups. J Adolesc Res 7:156–176
MA Phinney JS, Cantu CL, Kurtz DA (1997) Ethnic and American identity
Gonzales NA, Kim LS (1997) Stress and coping in an ethnic minority as predictors of self-esteem among African American, Latino, and
context: Children’s cultural ecologies. In Wolchik, Sandler (Eds) White adolescents. J Youth Adolesc 26:165–185
Handbook of children’s coping: Linking theory and intervention. Plant CT, Yanico BJ (1991) Relationship of racial identity and gender
Plenum Press, New York, pp 481–511 role attitudes to Black women’s psychological well-being. J Couns
Hammack PL, Richards MH, Luo Z, Edlynn ES, Roy K (2004) Social Psychol 38:315–322
support factors as moderators of community violence exposure Repetto PB, Caldwell CH, Zimmerman MA (2004) Trajectories of
among inner-city African American young adolescents. J Clin depressive symptoms among high risk African-American adoles-
Child and Adolesc Psychol 33:450–462 cents. J Adolesc Health 35:468–477
Harter S (1999) The construction of the self: A developmental perspec- Roberts RE, Phinney JS, Masse LC, Chen YR, Roberts CR, Romero A
tive. Guilford Press, New York (1999) The structure of ethnic identity of young adolescents from
Hops H (1995) Age- and gender-specific effects of parental depression: diverse ethnocultural groups. J Early Adolesc 19:301–322
A commentary. Dev Psychol 31:428–431 Rosenberg M (1965) Society and the Adolescent Self-Image. Princeton.
Johnson PD, Kliewer W (1999) Family and contextual predictors of de- Princeton University Press, New Jersey
pressive symptoms in inner-city African American youth. J Child Rotheram MJ, Phinney JS (1987) Introduction: Definition and perspec-
and Family Stud 8:181–192 tives in the study of children’s ethnic socialization. In Phinney J,
Kavanagh K, Hops H (1994) Good females? Bad males? Gender and Rotheram M (Eds) Children’s ethnic socialization: Pluralism and
development as contexts for diagnosis and treatment. In Ollendick development. Sage Publications, California, pp 10–28
TH, Prinz RJ (Eds) Advances in clinical child psychology vol. 16, Rowley SJ, Sellers RM, Chavous TM, Smith MA (1998) The relation-
Plenum Press, New York, pp 45–79 ship between racial identity and self-esteem in African American
Kendall PC, Watson D (Eds) (1989) Anxiety and depression: Distinctive college and high school students. J Pers Soc Psychol 74:715–
and overlapping features. Academic Press, San Diego, CA 724
Kliewer W, Lepore SJ, Oskin D, Johnson PD (1998) The role Safren SA, Gonzalez RE, Horner KJ, Leung AW, Heimberg RG, Juster
of social and cognitive processes in children’s adjustment HR (2000) Anxiety in ethnic minority youth: Methodological
to community violence. J Consult. Clin. Psychol 66:199– and conceptual issues and review of the literature. Behav Modif
209 24:147–183
Kovacs M (1985) The Children’s Depression Inventory (CDI). Psy- Sandler IN, Miller P, Short J, Wolchik SA (1989) Social support as a
chopharmacol Bull 21:995–998 protective factor for children in stress. In Belle D (Ed) Children’s

Springer
88 J Youth Adolescence (2007) 36:77–88

social networks and social support. Wiley, New York, pp 277– Svensson R (2003) Gender differences in adolescent drug use: The im-
307 pact of parental monitoring and peer deviance. Youth Soc 34:300–
Schmeelk-Cone KH, Zimmerman MA (2003) A longitudinal analysis 329
of stress in African American youth: Predictors and outcomes of Tajfel H, Turner J (1986) The social identity theory of intergroup be-
stress trajectories. J Youth Adolesc 32:419–430 havior. In Worchel S Austin W (Eds) Psychology of intergroup
Smith EP, Walker K, Fields L, Brookins CC, Seay RC (1999) Ethnic relations. Nelson- Hall, Chicago, pp 7–24
identity and its relationship to self-esteem, perceived efficacy and Weist MD, Freedman AH, Paskewitz DA, Proescher EJ, Flaherty LT
prosocial attitudes in early adolescence. J Adolesc 22:867–880 (1995) Urban youth under stress: Empirical identification of pro-
Spielberger CD, Edwards CD, Montuori J, Lushene D (1973) How I tective factors. J Youth Adolesc 24:705–721
Feel. Questionnaire (State-Trait Anxiety Inventory for Children). Wetherington E, Kessler RC (1986) Perceived support, received sup-
Counseling Psychologists Press, Palo Alto, CA port, and adjustment to stressful life events. J Health Soc Behav
Spencer MB, Cunningham M, Swanson DP (1995) Identify as coping: 27:78–89
Adolescent African American males’ adaptive responses to high- White KS, Bruce SE, Farrell AD, Kliewer W (1998) Impact of exposure
risk environments. In Harris H, Blue H, Griffith E (Eds) Racial to community violence on anxiety: A longitudinal study of family
and ethnic identity: Psychological development and creative ex- social support as a protective factor for urban children. J Child
pression. Routledge, New York, NY, pp 31–52 Family Stud 7:187–203
Spencer MB, Dupree D (1996) African American youths’ eco-cultural Yarcheski A, Mahon NE, Yarcheski TJ (2001) Social support and well-
challenges and psychosocial opportunities: An alternative analy- being in early adolescents: The role of mediating variables. Clin
sis of problem behavior outcomes. In Cicchetti D, Toth S (Eds) Nurs Res 10:163–181
Rochester Symposium on Developmental Psychopathology, Vol- Yasui M, Dorham CL, Dishion TJ (2004) Ethnic identity and psy-
ume 7, Adolescence: Opportunities and Challenges. University of chological adjustment: A validity analysis for European Ameri-
Rochester Press, Rochester, NY, pp 259–282 can and African American adolescents. J Adolesc Res 19:807–
Spencer MB, Noll E, Stoltzfus J, Harpalani V (2001) Identity and school 825
adjustment: Revisiting the “acting White” assumption. Educ Psy- Youngstrom E, Weist MD, Albus KE (2003) Exploring violence
chol 36:21–30 exposure, stress, protective factors and behavioral problems
Steinberg L, Morris AS (2001) Adolescent development. In Fiske ST, among inner-city youth. Am J Community Psychol 32:115–
Schacter DL, Zahn-Waxler C (Eds) Annual Review of Psychology 129
(vol. 52). Annual Reviews, Palo Alto, CA, pp. 83–110 Zimmerman MA, Ramirez-Valles J, Zapert KM, Maton KI (2000) A
Stevenson HC, Reed J, Bodison P (1996) Kinship social support and longitudinal study of stress-buffering effects for urban African
adolescent racial socialization beliefs: Extending the self to family. American male adolescent problem behaviors and mental health.
J Black Psychol 22:498–508 J Community Psychol 28:17–33
Stice E, Ragan J, Randall P (2004) Prospective relations between social Zimmerman MA, Salem DA, Maton KI (1995) Family structure and
support and depression: Differential direction of effects for parent psychosocial correlates among urban African American adoles-
and peer support? J Abnorm Psychol 113:155–159 cent males. Child Dev 66:1598–1613

Springer

Вам также может понравиться