reflection questions
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=hedp20
Educational Psychologist
ISSN: 0046-1520 (Print) 1532-6985 (Online) Journal homepage: http://www.tandfonline.com/loi/hedp20
Race/Ethnicity and Social Adjustment of Adolescents: How (Not if) School Diversity Matters
Sandra Graham
To cite this article: Sandra Graham (2018) Race/Ethnicity and Social Adjustment of Adolescents: How (Not if) School Diversity Matters, Educational Psychologist, 53:2, 64-77, DOI: 10.1080/00461520.2018.1428805
To link to this article: https://doi.org/10.1080/00461520.2018.1428805
Published online: 02 Mar 2018.
Submit your article to this journal
Article views: 551
View Crossmark data
Citing articles: 3 View citing articles
Race/Ethnicity and Social Adjustment of Adolescents: How (Not if) School Diversity Matters
Sandra Graham
Department of Education, University of California, Los Angeles
In this article, I describe a program of research on the psychosocial benefits of racial/ethnic
diversity in urban middle schools. It is hypothesized that greater diversity can benefit
students’ mental health, intergroup attitudes, and school adaptation via three mediating
mechanisms: (a) the formation and maintenance of cross-ethnic friendships, (b) the
development of complex social identities, and (c) decreases in perceived vulnerability. These
hypotheses are examined in a 3-year longitudinal study of approximately 6,000 ethnically
diverse 6th-grade students as they enter one of 26 urban middle schools in California. The
overarching goal of this article is to present a strong argument that greater school ethnic
diversity can buffer many of the normative challenges of early adolescence.
Educational psychology has changed dramatically since E.
L. Thorndike founded the field more than a century ago.
Our nation, too, has undergone unparalleled transformation
during the last 100 years along multiple dimensions,
including its racial/ethnic landscape. In this article, I want
to make the case that our core constructs in educational psy-
chology, such as motivation and social adaptation, need to
pay greater attention to this changing ethnic landscape. I
outline and test parts of a conceptual model of the psycho-
social and motivational benefits of school ethnic diversity
in a changing society. I study underlying processes, or the
mechanisms by which diversity is hypothesized to have
beneficial effects.
Let me begin with the historical context. Beginning in
1920 and based on decennial Census data, Figure 1 shows
the changing demographics in the United States over the
past century in 20-year increments. In 1920 Thorndike was
at the height of his career. The Law of Effect had already
been established by 1905, his classic Animal Intelligence
appeared in 1911, and the following year he was president
of the American Psychological Association. There were
about 100 million people counted by the Census in 1920.
Of those, 90% were White and roughly 10% were African
American (the other racial/ethnic groups were too small to
disaggregate, and the Census was not yet identifying people
as Hispanic/Latino). By 2010 the U.S. population had
grown to well over 300 million people, and the forces of
immigration had redefined the racial/ethnic landscape. The
White population had declined to less than 65%, African
Americans remained steady at about 12%, Latinos had
grown to 16% of the total, and Asians had become a fast-
growing racial group. The last two bar graphs in Figure 1
project what the population will look like in 2040 and
2060. It is evident that Whites will have relinquished their
numerical majority status, and Latinos, at higher than 30%,
will be well on their way to becoming the largest racial/eth-
nic group. Thus, within a just a few generations the United
States will have shifted from being primarily biracial with a
large White majority and a small African American minor-
ity to a fully multiracial/multiethnic society. We will soon
become a nation of numerical minorities.
These dramatic demographic changes are occurring
more rapidly in the under-18 population because of an
immigration that is largely non-White and different birth
rates in different racial/ethnic groups. A K-12 public school
population that was 80% White 40 years ago is now just
barely 50% White (NCES, 2015; Wells, Fox, & Cordova-
Cobo, 2016). If these trends continue, by 2040 White stu-
dents will no longer be the majority in our nation’s public
schools, Latinos will be the largest ethnic minority, and
public schools as an institution will be without a majority
of any one racial/ethnic group.
But at a time when the school-age population is becom-
ing more ethnically diverse, public schools are becoming
more ethnically segregated than they have been in the past
40 years (Orfield, 2014). For example, the typical White
student attends school where almost 80% of the students
are White, and the typical African American or Latino
Correspondence should be addressed to Sandra Graham, Department of
Education, University of California, Los Angeles, 405 Hilgard Avenue,
Los Angeles, CA 90095-1521. E-mail: shgraham@ucla.edu
EDUCATIONAL PSYCHOLOGIST, 53(2), 64–77, 2018
Copyright � Division 15, American Psychological Association ISSN: 0046-1520 print / 1532-6985 online
DOI: 10.1080/00461520.2018.1428805
student attends school where at least two thirds of the stu-
dents are from their racial/ethnic group. Moreover, the great
majority of highly segregated ethnic minority schools
are located in urban pockets of concentrated poverty, which
puts their students at greater risk for poor academic
outcomes.
This is a critical time for studying issues about school
diversity, because our courts continue to roll back the prog-
ress made in the decades following Brown v. Board of Edu-
cation in 1954. Although psychological research played a
pivotal role in the Brown decision, much of the research on
the psychosocial benefits of school diversity in K-12 public
schools has been portrayed as outdated, methodologically
weak, too focused on Black–White comparisons, and theo-
retically impoverished—all of which can make it easier for
critics of race-conscious policies to dismiss that evidence
(Linn & Welner, 2007; Wells et al., 2016). I believe that
the best counterargument to such criticisms will be rigorous
programs of research with theory-driven and testable
hypotheses about how ethnic diversity leads to better out-
comes rather than if it does.
Figure 2 displays a conceptual model that my colleague
Jaana Juvonen and I have been testing on how school diver-
sity matters for psychosocial outcomes (e.g., Graham,
Munniskma, & Juvonen, 2014; Juvonen, Kogachi, &
Graham, 2017). We begin with the setting: urban middle
schools. Our research is on public schools, while excluding
private, religious, or charter schools, mostly because the
legal discourse around racial segregation and resegregation
has focused on the public school arena (Orfield, 2014).
With increasing ethnic diversity, these public schools
become contexts that allow opportunities for cross-ethnic
friendships, exposure to multiple ethnic groups, and a
greater numerical balance of power among different ethnic
groups. Context gives rise to process, or the mediating
mechanisms that might explain the relation between school
ethnic diversity and critical developmental outcomes. We
propose three mediating mechanisms, or psychosocial pro-
cesses, capitalizing on the opportunities provided by more
diverse schools: the formation of cross-ethnic friendships,
development of complex social identities, and less per-
ceived vulnerability. Through these processes, ethnic diver-
sity can lead to improved intergroup attitudes and better
mental health, which in turn promote better school adapta-
tion. Many researchers have written about the academic
advantages of attending ethnically diverse schools (e.g.,
Mickelson, Bottia, Larrimore, & Lambert, 2016). We pro-
pose that at least some of those advantages might partly be
explained by the intrapersonal and interpersonal benefits
that diversity affords.
There are two main reasons why I focus on middle
schools and these particular psychosocial processes. First,
they are developmentally significant. We study cross-ethnic
friendships because during early adolescence, peers in gen-
eral and friendships in particular take on heightened impor-
tance (Brechwald & Prinstein, 2011). We explore complex
social identities because identity negotiation—figuring out
who I am and who we are—is a core developmental task of
adolescence (Erikson, 1968). We investigate processes
related to perceived vulnerability inasmuch feeling vulnera-
ble is largely about coping with peer victimization, which
we know peaks during the middle school years (Rivara &
LeMenestral, 2016).
The second reason is that they draw on my expertise as a
developmental social psychologist. As articulated more
next, the theory underlying each process comes from core
research in social psychology. Contact theory (Allport,
1954) is the framework for studying relations between
cross-ethnic friendships and intergroup attitudes, social
identity theory (Tajfel & Turner, 1979) is the conceptual
framework for exploring the consequences of complex
social identities, and attribution theory (Weiner, 1986) is
the organizing model for examining how attributions about
peer victimization affect mental health.
90 88 89 79 75
64 51
43
10 10 11
12 12
12
13 13
6 8 16
25 31
6 7 8 5
1920 1940 1960 1980 2000 2010 2040 2060
White Black Latino Asian Other
Year Pop. in Mil. 106 132 178 227 275 300 380 420
100%
75%
50%
25%
0%
Pe rc
en ta
g e
of th
e Po
p ul
a tio
n
FIGURE 1 Changing racial/ethnic diversity in the United States. Pop. D population; Mil. D millions. Source: U.S. Census Bureau.
ADOLESCENTS AND SCHOOL ETHNIC DIVERSITY 65
In the remainder of this article, I describe a program
of research that we have been conducting on some of the
core processes depicted in Figure 2. This is an ongoing
program of research, and not every path in the model is
discussed. As a developmental social psychologist, I
focus on psychosocial processes in the model—improved
intergroup attitudes and mental health—as primary out-
comes of greater diversity. We have yet to fully unpack
the school adjustment outcomes in our model, although
we do have evidence that feeling a sense of belonging in
diverse middle schools is related to better academic
achievement (Morales-Chicas & Graham, 2017). I also
acknowledge that a number of moderators of the pro-
cesses examined here are not discussed. For example,
there are characteristics of schools other than ethnic
diversity, such as size, academic standing, and the neigh-
borhoods in which they reside, that surely will influence
any number of paths in our model. Moreover, there are
instructional practices such as academic tracking that
limit the mixing opportunities of students even in the
most diverse schools (e.g., Moody, 2001). I made a con-
scious choice to start at the process level, documenting a
set of theory-guided psychosocial consequences of ethnic
diversity that have not previously been studied together
in one conceptual model. Theoretical and empirical
refinements to our model with the inclusion of academic
outcomes and key moderating influences will occur in
the next phase of our research.
We recruited a large and very ethnically diverse sample
of about 6,000 sixth-grade students who self-identified as
Latino, African American, Asian, White, multiracial, or
Other (groups too small to be categorized) (see Juvonen,
Kogachi, & Graham, 2017, for a more detailed description
of the sample). The students were recruited from 26 middle
schools that systematically varied in ethnic diversity as
measured by the number of different ethnic groups present
in the school and their relative representation. Some
schools were ethnically diverse such that no single ethnic
group represented a numerical majority in the population,
and members of each of the four major pan-ethnic groups
(i.e., African American, Asian, Latino, and White) were
present in the student population; some schools had two
large and relatively equal ethnic groups (e.g., Latino and
Asian) with very few members of other ethnic groups; and
other less diverse schools had a clear numerical majority
ethnic group (either African American, Asian, Latino, or
White) with a smaller number of members from each of the
other ethnic groups. We used well-established indices to
measure the diversity of these schools, which confirmed
that our schools represented a good range of diversity (see
Echols & Graham, 2016; Graham, 2016). Data on the con-
structs depicted in Figure 2 were gathered twice in sixth
grade (fall and spring) and once in the spring of seventh
and eighth grades, for a total of four waves of data in this
longitudinal study of psychosocial adjustment in middle
schools that varied in ethnic diversity.
CONTEXT SOCIAL PROCESSES OUTCOMES
6th grade 8th grade7th grade
Opportunities for Cross-
Ethnic Friendships
ETHNIC DIVERSITY Exposure toMultiple Ethnic
Groups
Complex Social Identities
• cross-cutting group memberships
Balance of Power among Ethnic
Groups
Formation of Cross- Ethnic Friendships
• frequency • quality • mutuality • stability • residential proximity
Less Social Vulnerability • peer victimization • attributional ambiguity • school safety
Mental Health • self worth • anxiety • physical symptoms
• depression
School Adaptation GPA
• engagement • attendance • discipline • school climate • sense of belonging
Intergroup Attitudes
• cognitive • affective • social acceptance
and rejection
FIGURE 2 Conceptual model of the psychosocial benefits of ethnic diversity in urban middle schools.
66 GRAHAM
SCHOOL DIVERSITY, CROSS-ETHNIC FRIENDSHIPS, AND INTERGROUPATTITUDES
The first set of processes depicted in Figure 2 involve
cross-ethnic friendships and intergroup attitudes. Cross-
ethnic friendships have long been of interest to educa-
tional psychologists. Beginning with the school desegre-
gation literature of the 1970s, researchers have been
asking whether (to what extent) children prefer friends
from their same ethnic group and the conditions under
which they become receptive to cross-ethnic friendships
(Graham & Kogachi, in press). Homophily and propin-
quity, the two major determinants of friendship, provide
answers to these questions (McPherson, Smith-Lovin, &
Cook, 2001). Homophily is the tendency to form friend-
ships with similar others, with race being one of the
most salient similarity characteristics. Propinquity is the
tendency to form friendships with others who share the
same space, such as being in the same classroom or
school. Regarding homophily, a robust finding in the
cross-ethnic friendship literature is that students show a
preference for same-ethnicity peers over different-eth-
nicity peers at every age group, but especially by ado-
lescence (Graham & Echols, in press). But propinquity
matters too. Increasing racial/ethnic diversity at the
school level does lead to more cross-ethnic friendships,
although the rate of increase does not keep pace with
the opportunity for such friendships based on the racial/
ethnic composition of schools (Moody, 2001).
Once cross-ethnic friendships are formed, our interest is
in how friendships relate to intergroup attitudes. A core
feature of the social scientist’s argument in support of
Brown v. Board of Education in 1954 was that closer con-
tact between members of different racial groups (via
school desegregation) promotes positive racial attitudes
(Pettigrew, 2004). The development of friendships is pre-
sumed to be one mechanism for sustained contact. The idea
is that cross-ethnic friendships promote more extended, vol-
untary, and intimate contact and therefore fulfill many of
the conditions of the initial contact hypothesis such as
cooperation and equal status (Allport, 1954). Based on a
meta-analysis of more than 100 studies, Pettigrew and
Tropp (2006) concluded that contact reduces prejudice,
with an average effect size of about ¡.21; on the subset of studies that involved friendship, the effect size was ¡.25. A more recent meta-analysis focusing on cross-group (not just
ethnic) friendships and intergroup attitudes found similar
effect sizes (Davies, Tropp, Aron, Pettigrew, & Wright,
2011). However, only a few of the racial contact studies
actually measured cross-ethnic friendships with different
ethnic groups; most focused on White respondents’ atti-
tudes toward ethnic minorities, which limits our under-
standing of the role of friendships endorsed by societal
minority groups, and even fewer were carried out with chil-
dren and adolescents.
To address these limitations, we focused on actual
friendships between middle school students of different eth-
nic groups and whether having cross-ethnic friendships
influenced attitudes about the ethnic groups to which those
friends belonged. Using grade rosters as a guide, students
in our sample selected the names of their good friends.
Because we had very high participation rates at sixth-grade
recruitment, we were able to determine the self-reported
ethnicity of most of the nominated friends. In a separate
part of the survey, participants reported on their affective,
cognitive, and behavioral attitudes toward the four largest
pan-ethnic groups: African Americans, Latinos, Asians,
and Whites (see Chen & Graham, 2015).
For all ethnic groups in our sample, as school diversity
increased and more ethnic groups were available to
befriend, students were more likely to form friendships
with peers from those groups, and those friendships, in
turn, predicted more positive intergroup attitudes. There
were also distinctive patterns to the data depending on
which ethnic groups composed a friendship dyad. For
example, Chen and Graham (2015) studied the cross-ethnic
friendships of sixth-grade Asian students with White,
Latino, and African American grade mates. After account-
ing for availability (propinquity), Asian students were most
likely to befriend White peers and least likely to befriend
African American peers. The effect of friendships on inter-
group attitudes was weakest for the Asian–Black dyads.
Such asymmetries underscore status differences between
groups and the general racial hierarchy in U.S. society,
with Whites (the societal privileged group) being the most
desired friends, and friendships with lower status peers
lacking the power to change negative attitudes about that
group. Power dynamics associated with cross-ethnic friend-
ships have not been sufficiently studied in the friendship
literature.
The Chen and Graham (2015) findings are cross-sec-
tional, which means that we cannot rule out selection
effects. Perhaps individuals with more tolerant intergroup
attitudes seek out cross-ethnic friends such that the direc-
tion of effects is attitudes ! friendships rather than friend- ships ! attitudes. Capitalizing on our longitudinal design, we used cross-lagged analysis across the four waves of mid-
dle school data to examine directional effects (Graham &
Kogachi, 2017). Figure 3 shows the results of these analy-
ses for Whites students’ friendships with African American
peers (top panel) and African American students’ friend-
ships with Whites (bottom panel). The friendship measure
takes into account availability (the extent to which the
number of cross-ethnic friendships exceeds chance in a par-
ticular school), and the attitude measure assesses desire for
behavioral closeness, for example, to eat lunch together or
sit on a school bus with members of a befriended peer’s eth-
nic group.
The data for White students’ friendships with African
Americans show the predicted effect of contact at the
ADOLESCENTS AND SCHOOL ETHNIC DIVERSITY 67
beginning of middle school. More African American
friendships in the fall of sixth grade (T1) predicted better
attitudes toward that ethnic group by the end of sixth grade
(T2). Thereafter, friendships were not predictive of
subsequent attitudes, in part because both friendships and
attitudes were relatively stable over time. For African
American students’ friendships with Whites (bottom panel),
there was evidence of selection effects. More positive atti-
tudes toward Whites at the end of sixth grade predicted a
greater likelihood that African American students would
befriend grade mates from that ethnic group the following
year (T3), and those friendships then predicted improved
attitudes by the end of middle school (T4). Although not
shown in Figure 3, other analyses documented that African
American students’ grade point average was a significant
predictor of intergroup attitudes in sixth grade. Perhaps
high academic status fosters the positive attitudes (and con-
fidence) necessary for African American youth to initiate a
friendship with White grade mates.
In summary, more school diversity provides opportuni-
ties for crossing ethnic boundaries to form friendships, and
these friendships resulted in more positive intergroup
attitudes. The documented patterns depended, in part, on
who was befriending whom, and there was evidence of
selection as well as influence effects. Interracial friendships
with Black classmates influenced White students’ attitudes
toward (willingness to interact with) Black students at the
beginning of sixth grade, but that influence faded over time.
High-achieving Black students who had positive intergroup
attitudes were more open to befriending White classmates
by seventh grade, which then predicted even better attitudes
toward Whites. Had we limited ourselves to cross-sectional
analyses, like most friendship-to-attitude research, we
would not have documented these interactional effects over
time.
Although I focus on the benefits of cross-ethnic friend-
ships for intergroup relations, I believe that there are aca-
demic benefits as well. Research indicates, for example,
that some ethnic minority adolescents do better in school
when they have cross-ethnic friends (e.g., Lewis et al.,
2017; Newgent, Lee, & Daniel, 2007; Stanton-Salazar &
Dornbusch, 1995). Such friends can function like social
capital (cf. Crosnoe, Cavanagh, & Elder, 2003), facilitating
the flow of important information across ethnic boundaries
T1 Cross- Race
Friendships with Whites
.56***
.54***
.06
.10
T1 Behavioral Attitudes toward Whites
T2 Cross- Race
Friendships with Whites
T2 Behavioral Attitudes toward Whites
T3 Cross- Race
Friendships with Whites
.16*
-.07
T3 Behavioral Attitudes toward Whites
T4 Cross- Race
Friendships with Whites
T4 Behavioral Attitudes toward Whites
.11*
-.05
.64***
-.13**
χ2 (52) = 285.18, p < .001, RMSEA = .05, 90% CI [.00, .08], CFI = .96
χ2 (37) = 87.54, p < .001, RMSEA = .04, 90% CI [.03, .05], CFI = .94
T1 Cross- Race
Friendships with African Americans
.54***
.53***
.05*
.06
T1 Behavioral Attitudes toward African
Americans
T2 Cross- Race
Friendships with African Americans
T2 Behavioral Attitudes toward African
Americans
T3 Cross- Race
Friendships with African Americans
.47***
.45***
-.01
.01
T3 Behavioral Attitudes toward African
Americans
T4 Cross- Race
Friendships with African Americans
T4 Behavioral Attitudes toward African
Americans
.01
.04
.42***
.52***
FIGURE 3 Cross-ethnic friendships and intergroup attitudes over time. Note: Top panel: White students’ friendships with and attitudes toward African
Americans. Bottom panel: African American students’ friendships with and attitudes toward Whites. T1 D the fall of sixth grade; T2 D end of sixth grade; T3 D the following year; T4 D end of middle school; RMSEA D root mean square error of approximation; CI D confidence interval; CFI D comparative fit index. Source: Unpublished data from author’s research.
68 GRAHAM
about what it takes to be successful in school. By forming a
relationship with an academically oriented cross-ethnic
classmate, students may be gaining access to new resour-
ces: their friends’ knowledge and skills related to achieve-
ment and possibly exposure to a larger academically
oriented social network (e.g., Goza & Ryabov, 2009;
Kawabata & Crick, 2015). The topic of the academic bene-
fits of school diversity as mediated by cross-ethnic friends
is understudied in educational psychology research.
DIVERSITY AND SOCIAL IDENTITY COMPLEXITY
What diversity-related opportunities in addition to cross-
ethnic friendships might be related to improved intergroup
attitudes? Let me turn to our second set of relations in
Figure 2 to argue that exposure to practices that promote
youths’ developing social identities might be one such
opportunity.
We define social identity as a person’s sense of belong-
ing to a social group and the meaning attached to that group
membership (Tajfel, 1981; Tajfel & Turner, 1979). Accord-
ing to social identity theory, belonging to groups helps peo-
ple to position themselves in their social world (find their
niche), feel good about themselves and their group, and
draw clear boundaries between ingroups and outgroups.
Although gender and race/ethnicity remain the most salient
social identities during adolescence, categories based on
other characteristics such as academic standing, social rep-
utation, or activities begin to emerge (e.g., Tanti, Stukas,
Halloran, & Foddy, 2011) Developmental psychologists
agree in principle that children and adolescents can simulta-
neously have multiple social identities, but there is rela-
tively little research on how these identities fit together in a
person’s self-schema and how they guide behavior.
One particularly promising approach from adult social
psychology is the model developed by Brewer and col-
leagues, labeled social identity complexity (Roccas &
Brewer, 2002). These authors defined social identity com-
plexity as the degree of perceived overlap between the
social groups with which one identifies. To illustrate, when
asked to describe himself, consider an African American
adolescent named Jordan who says, “I am Black, a good
student, a serious soccer player, and a member of the popu-
lar clique at school.” In other words, Jordan identifies with
schoolmates in his own racial group, who do well in school;
with the popular crowd; and with the good soccer players.
How does he combine these different social identities? If
Jordan believes that these four salient social groups contain
the same members (in his mind, most high achievers are
also African American, and most African Americans are
the cool kids and good soccer players), then the overlap
between his social groups is high and identity complexity is
low. On the other hand, if Jordan perceives that many high
achievers are not African American or only a few cool kids
are also good soccer players, and so on, then the overlap
between his identities is low: Not all of the members of one
identity (race) are the same as the members of the other
identities (e.g., sports, academics). In this case of partially
overlapping—or what has been called cross-cutting—group
memberships, social identity complexity is high. Group
members can simultaneously be ingroups or outgroups
depending on which social identity is most salient. Thus, it
is not so much how many social groups an individual iden-
tifies with as it is how those groups are subjectively com-
bined to represent a person’s ingroup identity. The more a
person believes that the groups to which he or she belongs
contain the same individuals, the less complex his or her
social identity.
Social identity complexity has implications for inter-
group attitudes. Ingroup–outgroup distinctions are more
likely to become polarized when membership in multiple
groups is largely overlapping (the same people or low
social identity complexity). For example, if most honors
students are thought to be Asian or most athletes are
thought to be African American, then stereotypes about
these racial groups are heightened, in part because within-
group homogeneity is emphasized. However, when individ-
uals identify with multiple groups with only partially over-
lapping memberships, distinctions between ingroups and
outgroups are diffused. In studies with adults, high social
identity complexity was associated with more positive
affect toward different racial/ethnic groups and more toler-
ant race-related attitudes (e.g., Brewer, Gonsalkorale, &
van Dommelen, 2012; Brewer & Pierce, 2005).
Complex social identities also appear to be enhanced by
living in a multicultural society because there are more
examples of social categories that do not completely over-
lap (Miller, Brewer, & Arbuckle, 2009; Schmid, Hewstone,
& Al Ramiah, 2013). These findings with adults led us to
hypothesize that greater school ethnic diversity promotes
the development of more complex social identities. More
diversity provides exposure to multiple ethnic groups and
other social groups whose characteristics are only partially
overlapping. We further predicted that increases in social
identity complexity would be related to improvements in
intergroup attitudes.
In studies with adults, social identity complexity is mea-
sured by first asking participants to list the groups that are
most important to them and then to answer questions about
the perceived overlap in group membership (i.e., how many
members of Group X are also members of Group Y?;
Brewer & Pierce, 2005). Adapting the adult measures for
adolescents proceeded in two steps (Knifsend & Juvonen,
2014, 2017). First, we asked participants to imagine that
they were filling out a Facebook profile and to write down
the three groups that best describe them. We assumed that
this was a good way to elicit the groups with whom each
participant identifies. Once the groups were listed, partici-
pants were asked a series of questions about the overlap in
ADOLESCENTS AND SCHOOL ETHNIC DIVERSITY 69
membership between all pairings of these three groups and
the individual’s self-reported ethnic group, for a total of
four groups. The questions about overlap among these
groups’ members were asked in both directions. Our hypo-
thetical participant Jordan would be asked, for example,
“How many African Americans are the cool kids?” and
“How many cool kids are African American?” Jordan’s
answers are recorded on 5-point scales, from 1 (almost all)
to 5 (hardly any). There are six questions in each direction,
for a total of 12 ratings. An index of social identity com-
plexity is calculated as the proportion of overlap between
groups, where low numbers indicate much overlap, and
therefore low social identity complexity and high numbers
indicate little overlap and therefore high social identity
complexity (see Knifsend & Juvonen, 2014). This measure
has proven to be very reliable with middle school students.
Broadening our measurement of diversity and the oppor-
tunities it affords, Knifsend and Juvonen (2017) examined
social identity complexity in the context of the diversity
of one’s extracurricular activities. We had information
on everyone’s participation in extracurricular activities,
including the number of reported activities. From this infor-
mation, we were able to calculate how diverse those activi-
ties were, based on “rosters” we created of participants who
reported participation in a particular activity (see Knifsend
& Juvonen, 2017, for details). Like in the measurement of
school diversity, the ethnic diversity of extracurricular
activities took into account the number of different ethnic
group members reporting participation and the relative
representation of those groups. Using the friendship mea-
sure described previously, we were also able to determine
how many of each participant’s cross-ethnic friends shared
those extracurricular activities.
Figure 4 shows how extracurricular activity diversity in
sixth grade predicted intergroup attitudes at seventh grade
(the same behavioral outgroup distance measure just
described) as mediated by cross-ethnic friends in those
activities and social identity complexity in seventh grade.
Moving from left to right in Figure 4, extracurricular activ-
ity diversity was related to both more cross-ethnic friends
in those activities and greater social identity complexity.
These two mediators then predicted less outgroup distance
(i.e., better intergroup attitudes). Thus, the effect of extra-
curricular diversity on outgroup distance was largely
explained by whether there were friends from different eth-
nic groups in one’s extracurricular activities and having a
more complex social identity. These are the first studies
reporting that how adolescents judge the overlap of the
identity groups to which they belong can affect their inten-
tion to interact with members of different ethnic groups.
Ethnicity of participants did not moderate these path coeffi-
cients, suggesting that the patterns we documented were
robust.
These findings also highlight the motivational significance
of extracurricular activity participation as a context for
studying exposure to ethnically diverse peers in schools that
vary in that diversity. Unlike academic programs that typi-
cally offer little choice in middle school and can often limit
the mixing opportunities of students if academic tracking is
widely practiced, extracurricular activities in principle pro-
vide free choice and should better represent the overall diver-
sity of a school (Moody, 2001). Extracurricular selections can
also satisfy some of the conditions identified by contact theory
(e.g., equal status, working toward a common goal) as prereq-
uisites for improved intergroup attitudes (Allport, 1954). On
the other hand, extracurricular activities can also limit expo-
sure even in diverse schools if they become racialized. In a
provocative analysis of yearbooks from about 200 high
schools that ranged in ethnic diversity, Clotfelter (2002,
2004) reported that ethnic minorities (primarily Black stu-
dents) participated proportionately less than their White class-
mates in all organized extracurricular activities except
basketball and track, which they tended to dominate. In some
situations, therefore, patterns of participation in extracurricu-
lar activities may actually reinforce racial/ethnic segregation
of students. Even in our sample, ethnically diverse extracur-
ricular activities were most evident for visual and performing
arts compared to sports and academic clubs (Knifend & Juvo-
nen, 2017). Thus, especially during nonacademic times, it is
important that school officials organize student activity
choices to maximize their opportunities to cross ethnic bound-
aries (Moody, 2001).
In summary, the middle school transition provides a con-
text in which early adolescents begin to calibrate and orga-
nize their social identities to find their niche and fit in. In
addition to ethnicity, other social categories tapping devel-
opmentally salient attributes (e.g., academic achievement,
popularity) increase the availability of multiple social iden-
tities. The construct of social identity complexity takes into
account not only the number of such identities that become
more salient during adolescence but also the ways in which
these identities are subjectively combined and integrated.
We documented that when extracurricular activities are
more ethnically diverse, middle school students in those
activities develop more complex social identities, defined
as less perceived overlap among important groups with
which an adolescent identifies. Complex social identities,
in turn, were related to better outgroup attitudes.
We emphasize the consequences of social identity com-
plexity for intergroup attitudes. As depicted in Figure 2, the
construct has also been linked to mental health outcomes.
For example, Roccas and Brewer (2002) demonstrated that
social identity complexity was associated with decreased
stress-related mood among college students. A more com-
plex identity is hypothesized to predict better mental health
inasmuch as failure in one domain (e.g., rejection by same-
ethnicity peers) can be buffered by another nonoverlapping
social category (e.g., identifying with popular peers in other
racial/ethnic groups). We are just beginning to explore the
mental health correlates of social identity complexity in our
70 GRAHAM
middle school sample. Early findings are documenting a
buffering effect: When ethnic minority students report
unfair race-based treatment by their peers, they feel less
anxious if they have a more complex social identity (Saafir
& Graham, 2017). Perceived assaults on one’s racial iden-
tity need not be as emotionally stressful if there are other
nonoverlapping identity groups to fall back on.
ETHNIC DIVERSITY AND PERCEIVED VULNERABILITY
Now let me turn to mental health more centrally as an out-
come to examine our third set of relations depicted in Fig-
ure 2. Here the hypothesis is that ethnic diversity is related to
less perceived vulnerability, with vulnerability operational-
ized as peer victimization and feeling unsafe at school as well
as the mental health outcomes of loneliness, social anxiety,
and depressive symptoms. During early adolescence, youth
are likely to experience heightened vulnerability, in part
because peer victimization (bullying) is thought to peak dur-
ing the middle school years (Rivara & Le Menestral, 2016).
Why should school ethnic diversity be related to less
perceived vulnerability? Our working hypothesis is that
diversity promotes less vulnerability because there is a
greater numerical balance of power between different eth-
nic groups. We based this hypothesis on the definition of
peer victimization as conflict that involves an imbalance of
power between perpetrator and victim (Olweus, 1993).
Asymmetric power relations take many forms, as when
stronger youth harass weaker classmates or when older
students pick on younger peers. At the group level, an
imbalance of power can also exist when members of major-
ity ethnic groups (more powerful in the numerical sense)
harass members of minority ethnic groups (less powerful in
the numerical sense). More ethnic groups of relatively
equal size—the hallmark of school diversity—may be pro-
tective because the numerical balance of power is less
likely to be tipped in favor of one or more large ethnic
groups exerting their influence over one or more small eth-
nic groups.
In an earlier study, we reported evidence in support of
the balance of power hypothesis with a much smaller sam-
ple that included only African American and Latino youth
(Juvonen, Nishina, & Graham, 2006). Replicating and
extending that prior work with the current much larger and
more diverse sample, we measured vulnerability as
students’ self-reported peer victimization, feeling unsafe,
and loneliness at school during sixth grade (Juvonen,
Kogachi, & Graham, 2017). Figure 5 shows the results of
these analyses. Plotted here are the significant slopes pre-
dicting levels of vulnerability at high and low levels of
school diversity. As diversity increased, self-reported vic-
timization and loneliness decreased, whereas perceived
school safety increased. Thus, when there was a shared
balance of power, students felt less vulnerable at school.
There were no moderating effects of ethnicity in these anal-
yses, suggesting that the benefits of a numerical balance of
power were true for both societally more powerful (e.g.,
Whites and East Asians) and less powerful (Africans and
Latinos) groups in our sample. Other researchers using sim-
ilar measures of school diversity have also documented
Cross-ethnic Availability
Extracurricular Activity (6th)
.61** -.13* Social Identity
Complexity (7th)
Ethnic Outgroup Distance (7th)
Cross-ethnic Friends in
Activity (6th) .80*** -.20*
.24/.02
Friendship Indirect effect: Z = -3.38*** SIC Indirect Effect: Z = -2.32*
*p <.05 **p <.01 ***p<.001
FIGURE 4 Cross-ethnic friendships and social identity complexity as mediators of the relation between diversity of extracurricular activities and intergroup
attitudes. Note: SIC D Social Identity Complexity. Source: Data from Knifsend and Juvonen (2017) and presented by permission of the authors.
ADOLESCENTS AND SCHOOL ETHNIC DIVERSITY 71
relations between increasing diversity and less peer victimi-
zation (Felix & You, 2011), as well as lower levels of
aggression (Hoglund & Hosan, 2012). Thus we consider
the findings linking school diversity to less vulnerability to
be quite robust.
An Attributional Analysis
Attribution theory (Weiner, 1986) provides insight into the
mechanisms that might explain why school diversity and a
greater balance of power are related to less vulnerability.
Attributions are answers to “why” questions such as “Why
did I fail a test?” or “Why doesn’t anyone like me?” Focus-
ing on peer victimization as a proxy for vulnerability, a his-
tory of harassment by classmates or even one particularly
painful episode might lead the victim to ask, “Why me?” In
the absence of disconfirming evidence, such an individual
might come to blame themselves for their peer relationship
problems. Such a youth might conclude, for example, “I’m
the kind of kid who deserves to be picked on.” Self-blame
can then lead to many negative outcomes, including low
self-esteem and depression. In the adult literature on causal
explanations for rape (another form of victimization),
attributions that imply personal deservingness, labeled
characterological self-blame, are especially detrimental
(Janoff-Bulman, 1979). From an attributional perspective,
characterological self-blame is internal and therefore
reflects on the self; it is stable and therefore leads to an
expectation that victimization will be chronic; and it is
uncontrollable, suggesting an inability to prevent future
harassment. Attributions for social failure to internal, sta-
ble, and uncontrollable causes lead individuals to feel both
hopeless and helpless (Weiner, 1986). In earlier research,
we documented that middle school students who make self-
blaming attributions of the characterological type feel espe-
cially bad (Graham, Bellmore, Nishina, & Juvonen, 2009;
Graham & Juvonen, 1998).
How does school diversity relate to attributional prefer-
ences? Our reasoning is as follows: In nondiverse urban
schools there are typically both a large majority ethnic
group and one or more small ethnic minority groups (e.g.,
few schools are 100% one racially homogeneous group).
Consider what it must be like to be a victim and a member
of a numerical majority group. Having a reputation as a vic-
tim when one’s ethnic group holds the numerical balance of
power might be especially painful because that person devi-
ates from what is perceived as normative for his or her
group: to be powerful and of high status. We hypothesized
that majority group victims who deviated from local norms
would be particularly vulnerable to self-blaming attribu-
tions (“It must be me”). As the number of same-ethnicity
peers increases in one’s social milieu, self-esteem buffering
external attributions become less plausible. Being a victim
and a member of the minority group should facilitate exter-
nal attributions to the prejudice of others (“It could be
them”), although such external attributions also come with
psychological costs (Major, Quinton, & McCoy, 2002).
However, in ethnically diverse contexts with a greater bal-
ance of power, we predicted the most attributional ambigu-
ity (“It might be me, but it could be them or some other
momentary cause”). Attributional ambiguity might be espe-
cially adaptive because it allows for a greater repertoire of
perceived causes of victimization (Graham, Bellmore,
Nishina, & Juvonen, 2009).
To examine whether attributions varied as a function of
school diversity, participants were presented with a sce-
nario that depicted them as the target of peer victimization
at school:
Imagine that you’ve just bought your lunch after waiting in
line for a long time. There is a group of kids in the line and
one of them sticks out their foot and trips you, as you are
walking away. You’re not hurt, but most of your food spills
on your clothes. The other kids in the line start laughing at
you.
Respondents then rated how much they agreed with 17
statements that captured their thoughts about the causes of
the victimizing incident. The items included attributions
designed to capture characterological self-blame (e.g., “If I
were a cooler kid, I wouldn’t get picked on”), context-spe-
cific behavior (e.g., “I was in the wrong place at the wrong
time”), and luck (e.g., “It was bad luck”), as well as exter-
nal attributions pertaining to others (e.g., “These kinds of
kids pick on everybody”) and the school environment (e.g.,
“Nobody is safe in this school anymore”). Each statement
was rated on a 5-point scale from 1 (definitely would think
that) to 5 (definitely would not think that) (see Graham &
Taylor, 2017).
We used latent profile analysis, a person-centered approach,
to examine patterns of attributions within respondents (Lanza
& Cooper, 2016). Latent profile analysis assumes that there are
School Ethnic Diversity
1
1.5
2
2.5
3
3.5
4
4.5
2 SD Below 1 SD Below Mean 1 SD Above 2 SD Above
School Safety
Loneliness
Perceived Victimization
FIGURE 5 Middle school ethnic diversity and perceived vulnerability as
measured by self-reports of peer victimization, perceived school safety,
and feelings of loneliness. Note: *p < .05, **p < .01, ***p < .001. Source:
Data from Juvonen, Kogachi, and Graham (2017) and presented by permis-
sion of the authors.
72 GRAHAM
subpopulations in the sample with distinct profiles comprised
of multiple indicators—in this case, distinct attribution pat-
terns. The four profiles that emerged are shown in Figure 6.
It is evident that the profiles are distinguished by the
degree to which respondents endorsed characterological
self-blame. From bottom to top, the first profile contained
members who were significantly less likely to endorse self-
blame relative to external and luck attributions. We labeled
this profile “It’s not me! It might be them.” The second pro-
file was similar to the first in that members tended to
endorse external and luck attributions rather than self-
blame, but they also endorsed specific behaviors as causes
of victimization. We labeled this profile “It’s not me—it’s
them and bad luck.” From an attributional perspective, with
low self-blaming attributions, these are the most adaptive
profiles. In the third profile, members were more likely to
endorse characterological self-blame and external attribu-
tions at moderate levels. We labeled this moderately stable
profile “It might be me but it’s definitely them.” Members
composing the fourth profile were significantly more likely
to endorse characterological self-blame and external attri-
butions; we labeled this profile “It’s me and them too!”
With the greatest probability of endorsing self-blame attri-
butions, we viewed the fourth profile as the least adaptive.
Our vulnerability indicators of feeling unsafe, socially anx-
ious, lonely, and depressed were significantly higher in this
fourth profile compared to the other three.
Next we examined predictors of profile membership
using hierarchical multinomial logistic regression.
Designating the most maladaptive fourth profile as the
reference group, the analysis can be interpreted as the
odds of membership in one of the other profiles com-
pared to “It’s me and them too!” We also had peer nom-
ination data on which students had reputations as
victims. As hypothesized, students who were both vic-
tims and members of the numerical majority in their
school (i.e., nondiverse schools) had significantly lower
odds of being in one of the more adaptive profiles com-
pared to the most maladaptive pattern. This pattern was
true for African American, Latino, Asian, and White
majority group victims. There was also a significant
Level 2 effect of school diversity. As school diversity
increased, students had greater odds of being in the
more adaptive profiles compared to the least adaptive.
In fact, the odds doubled across increasing levels of
school diversity. I suggest as a working hypothesis that
ethnic diversity creates enough attributional ambiguity
to ward off self-blaming tendencies, thereby allowing
for attributions that have fewer psychological costs. In
social contexts where multiple social cues are present
and multiple causal appraisals of social predicaments
are possible, attributional ambiguity can be particularly
adaptive if it allows the perceiver to draw from a larger
repertoire of causal schemes.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
It em
E nd
or se
m en
t ( Pr
op or
� on
)
It's me and them too!
It might be me but it's definitely them
It's not me -- it's them and bad luck
It's not me! It might be them
Characterological Self Blame Specific Behavior Luck External/Contextual Factors
I w on
't ge
t th
em in
tr ou
bl e
I'm n
ot c
oo l
A lw
ay s
ha pp
en s
to m
e
W ill
h ap
pe n
to m
e ag
ai n
O th
er k
id s
al so
d o
th is
to m
e
It 's
li ke
ly to
h ap
pe n
to m
e
I w as
in th
e w
ro ng
p la
ce
I s ho
ul dn
't ha
ve b
ee n
he re
I s ho
ul d
be c
ar ef
ul
St uff
h ap
pe ns
It w
as a
b ad
d ay
It w
as b
ad lu
ck
Ki ds
w an
na b
e to
ug h
N o
on e
ca re
s
Ki ds
a re
m ea
n
Th er
e ar
e no
r ul
es
Ev er
yo ne
g et
s pi
ck ed
o n
FIGURE 6 Profiles of attributions for peer victimization. Source: Unpublished data from author’s research.
ADOLESCENTS AND SCHOOL ETHNIC DIVERSITY 73
A good deal of peer victimization research is conducted in
urban schools where multiple ethnic groups are represented,
but very little of that research has systematically examined eth-
nicity-related context variables. In our research, we have been
attempting to bring the ethnic context to the study of peer vic-
timization (Graham, 2006). We do this by making a case for
the importance of attributions as a theoretical framework and
school ethnic composition as a central context variable, both of
which can aid our understanding of the dynamics of peer vic-
timization as well as the benefits of school diversity. I do not
think the key variable is ethnicity per se. There is no strong evi-
dence that any one ethnic group is more (or less) vulnerable to
peer victimization (see Graham & Echols, in press; Vitoroulis
& Vaillancourt, 2015). Rather, I believe that the critical vari-
able is ethnicity in context—in this case, whether one’s ethnic
group is a numerical majority, a numerical minority, or one of
several groups residing in a diverse school without highly visi-
ble majorities and minorities.
SOME CHALLENGES TO DIVERSITY
All of the mechanisms depicted in our conceptual model
(Figure 2) that explain the how of diversity treat school
diversity as a static structural variable. Although well mea-
sured, this structural approach implicitly assumes that the
experience of diversity is the same for all of the students
who walk through the school doors every morning. When
viewed as a more dynamic construct that can be different
for individuals across time and space, some of the chal-
lenges to our approach become evident. Let me provide
two illustrations of these challenges.
School Versus Classroom Diversity Mismatch
In middle school, students move from class to class through-
out their day. The average ethnic diversity of their particular
classes will vary and may differ from the overall ethnic diver-
sity of the school. Such mismatches—being in more or less
diverse courses compared to one’s school—could have
important effects on psychosocial adjustment.
We had information about each sixth-grade participant’s
unique course schedule as well as the class roster in each of
those courses. The class roster identified the ethnicity of
each student in the course. Using those rosters, we were
able to calculate a diversity index for each student’s four
academic courses. Averaging these indices yields a measure
of diversity exposure across the school day that is different
for every student. We then compared this more dynamic
and individualized ethnic diversity measure to structural
school diversity.
In a perfectly integrated school, individual exposure to
diversity in courses should match the overall diversity of the
school. What happens when there is a mismatch? We found
that when classroom diversity exposure was significantly
less than school ethnic diversity, students of all ethnic
groups perceived the racial climate of the school as more
oppressive, and they reported more negative attitudes toward
ethnic groups other than their own (Juvonen,
Kogachi, & Graham, 2017). The effects of diversity
mismatch on racial attitudes were stronger as school diver-
sity increased. A mismatch in the direction of less diversity
at the course level compared to school can signal de facto
segregation, or the sorting of students for instruction along
racial lines as in academic tracking (e.g., Mickelson, 2015).
Educational psychologists who study school diversity
should pay special attention to whether and how often stu-
dents of different ethnic groups are exposed to one another
during the school day even in very diverse schools. This will
require measurement approaches that capture dynamic as
well as structural ethnic diversity in classrooms and schools.
Changing Diversity Across a School Transition
Another kind of ethnic mismatch we study is that between
the departing and receiving school at the time of a school
transition. For example, does it make a difference whether
the numerical representation of one’s ethnic group
increases or decreases from elementary school to middle
school? One might hypothesize that a declining representa-
tion across transition settings will be associated with poorer
adjustment because there is a mismatch between the social
context of the departing and receiving school (French,
Seidman, Allen, & Aber, 2000). Developmental and educa-
tional psychologists have long argued that school transi-
tions can be challenging, especially during adolescence,
because of the mismatch between the developmental needs
of early adolescents and the school environment to which
they are transitioning (Eccles & Roeser, 2011). In this case,
mismatch involves fewer same-ethnic peers at a time when
ethnicity takes on heightened significance. In support of
this mismatch hypothesis, we found that Latino youth in
our sample who encountered a significant drop in the repre-
sentation of their ethnic group from elementary school to
middle school experienced decreases in feelings of belong-
ing and declining academic performance when they transi-
tioned to a middle school with significantly fewer members
of their own ethnic group (Morales-Chicas & Graham,
2017). Similar findings have been reported for Latino and
African American youth students across the transition from
middle to high school (Benner & Graham, 2009). These
findings suggest that it is not only absolute levels of ethnic
group representation that need to be considered when
studying school diversity but also changes in those levels
across critical school transitions.
The findings on ethnic mismatch across school transi-
tions call our attention to an important yet understudied
issue in research on school ethnic diversity. That issue is
the presence of a critical mass of same-ethnicity peers. In
the legal discourse about diversity on college campuses,
74 GRAHAM
critical mass is discussed as “meaningful numbers” or
“meaningful representation” of ethnic minorities to ensure
a diverse educational environment while encouraging
underrepresented students to participate in college life and
not feel isolated or marginalized (Garces & Jayakumar,
2014). No ethnic group is likely to benefit from an ethni-
cally diverse college or K-12 campus if their numbers are
too small to combat feelings of isolation or marginalization.
SUMMARY AND CONCLUSIONS
I began this article with the observation that the population
has become much more racially and ethnically diverse in the
century following Thorndike’s founding of educational psy-
chology, but our schools are not keeping pace with that
growing diversity. In fact, our schools appear to be resegre-
gating at an alarming rate (Orfield, 2014). I made the case
that we need strong programs of theory-driven rigorous
research that document how rather than if there are benefits
of greater school ethnic diversity. I provided empirical
examples of a conceptual model that we are testing on the
psychosocial benefits of ethnic diversity in urban middle
schools. Drawing on social psychological theory and
research that is particularly relevant to the developmental
stage of early adolescence, I tested hypotheses that greater
diversity can benefit students’ intergroup attitudes and men-
tal health via three mediating mechanisms: (a) the formation
and maintenance of cross-ethnic friendships, (b) the devel-
opment of complex social identities, and (c) decreases in
perceived vulnerability. With our large sample of multieth-
nic youth recruited from middle schools that vary in ethnic
diversity, we are building our case that school diversity can
buffer many of the challenges of early adolescence for all
youth, not just those who are members of historically mar-
ginalized groups. A growing literature, mainly about organi-
zational behavior and with adult samples, indicates that
exposure to ethnic diversity can make us smarter (i.e., more
creative, better problem solvers; Phillips, 2014). I suggest
that exposure to ethnic diversity in the everyday ecology of
school can help us be more tolerant, identity integrated, and
less socially vulnerable. I look forward to the day when it is
commonplace in our school communities to see cross-ethnic
friends, complex social identities, and students who utilize
adaptive attributions to navigate social challenges.
The absence of ethnic difference in most of the psychoso-
cial processes examined in this article underscores the robust-
ness of our findings. We had strong tests of ethnicity effects
with our large sample and because we selected our schools so
that all of the four pan-racial/ethnic groups were well distrib-
uted across schools that varied in ethnic diversity (i.e., each
ethnic group was represented in schools where their group
was the numerical majority, a minority, or one of several
groups in diverse schools). But we are not making the case
that the experiences of diversity are the same for ethnic groups
with such disparate status in American society. As the socie-
tally most privileged group, for example, White youth may
find the declining representation of their group (cf. Figure 1)
most challenging and most threatening (e.g., Kumar, Seay, &
Karabenick, 2011; see also Craig & Richeson, 2014). Our
task for the future is to better uncover what is similar across
race/ethnic groups and what is different based on the unique
historical, economic, and cultural forces that shape each eth-
nic group’s experiences in American society.
The African American scholar W.E.B. Du Bois propheti-
cally reminded us that the problem of the 20th century was
the problem of the color line, with that color line defined as
the Black–White divide (Du Bois, 1903). I believe that one of
the great social problems of the 21st century will be about
transformative changes in racial/ethnic diversity and at least
two challenges that transformation brings. First, how can we
ensure equal engagement of all ethnic groups with society’s
institutions and equal access to its resources, including good
schools? Second, when people of different ethnic groups
come in contact with one another both voluntarily and invol-
untarily, how can we reduce interethnic conflict and promote
interethnic harmony? To study these challenges about equal
access and tolerance, as developmental social and educational
psychologists we need new ways to think about the meaning
of race and ethnicity, innovative methods for measuring eth-
nic diversity as a dynamic and fluid construct that changes
across time and space, creative thinking about issues of criti-
cal mass, and a willingness to embrace all racial and ethnic
groups in our analyses, sometimes together, sometimes sepa-
rately, despite the messiness and complexity that comes with
that inquiry. My long-term goal is to help integrate the study
of motivation and social adjustment in educational psychol-
ogy with the emerging field of diversity science (see Plaut,
2010) to address a set of diversity-related school challenges
of great social significance. If I succeed, we may one day see
a world in which lists of eminent scholars such as Division
150s Thorndike Award recipients are as diverse as our nation has so rapidly become.
ACKNOWLEDGMENTS
An earlier version of this article was based on an invited
address for the E. L. Thorndike Award for Distinguished
Psychological Contributions to Education, presented to the
author at the American Psychological Association Conven-
tion in Washington, DC, August 2014.
FUNDING
The research reported in this article was supported by
grants from the National Science Foundation and the
National Institute of Child Health and Human
Development.
ADOLESCENTS AND SCHOOL ETHNIC DIVERSITY 75
REFERENCES
Allport, G. (1954). The nature of prejudice. Reading, MA: Addison-
Wesley. Benner, A., & Graham, S. (2009). The transition to high school as a devel-
opmental process among multiethnic urban youth. Child Development,
80, 365–367. doi:10.1111/j.1467-8624.2009.01265.x
Brechwald, W. A., & Prinstein, M. J. (2011). Beyond homophily: A
decade of advances in understanding peer influence processes.
Journal of Research on Adolescence, 21, 166–179. doi:10.1111/
j.1532-7795.2010.00721.x
Brewer, M., Gonsalkorale, K., & van Dommelen, A. (2012). Social identity
complexity: Comparing majority and minority ethnic group members in
a multicultural society. Group Processes & Intergroup Relations, 16,
529–544. doi:10.1177/1368430212468622
Brewer, M., & Pierce, K. (2005). Social identity complexity and outgroup
tolerance. Personality and Social Psychology Bulletin, 31, 428–437.
doi:10.1177/0146167204271710
Chen, X., & Graham, S. (2015). Cross-ethnic friendships and intergroup
attitudes among Asian American adolescents. Child Development, 86,
749–764. doi:10.1111/cdev.12339
Clotfelter, C. (2002). Interracial contact in high school extracurricular
activities. Urban Review, 34, 25–46. doi:10.1023/A:1014493127609
Clotfelter, C. (2004). After Brown: The rise and retreat of school segrega-
tion. Princeton, NJ: Princeton University Press.
Craig, M., & Richeson, J. (2014). More diverse yet less tolerant? How the
increasingly diverse racial landscape affects White Americans’ racial
attitudes. Personality and Social Psychology Bulletin, 40, 750–761.
doi:10.1177/0146167214524993
Crosnoe, R., Cavanagh, S., & Elder, G. H. (2003). Adolescent friendships as
academic resources: The intersection offriendship, race, and school disadvan-
tage. Sociological Perspectives, 46, 331–352. doi:10.1525/sop.2003.46.3.331
Davies, K., Tropp, L. R., Aron, A., Pettigrew, T. F., & Wright, S. C.
(2011). Cross-group friendships and intergroup attitudes a meta-analytic
review. Personality and Social Psychology Review, 15, 332–351.
doi:10.1177/1088868311411103
Du Bois, W. E. B. (1903). The souls of Black folk. New York, NY: Bantam
Classic.
Eccles, J. S., & Roeser, R. W. (2011). Schools as developmental contexts
during adolescence. Journal of Research on Adolescence, 21, 225–241.
doi:10.1111/j.1532-7795.2010.00725.x
Echols, L., & Graham, S. (2016). For better or worse: Friendship choices
and peer victimization among ethnically diverse youth in the first year
of middle school. Journal of Youth and Adolescence, 45, 1862–1876.
doi:10.1007/s10964-016-0516-0
Erikson, E. (1968). Identity: Youth and crisis. New York, NY: Norton.
Felix, E., & You, S. (2011). Peer victimization within the ethnic context of
high school. Journal of Community Psychology, 39, 860–875.
doi:10.1002/jcop.20465
French, S., Seidman, E., Allen, L., & Aber, J. (2000). Racial/ethnic iden-
tity, congruence with the social context, and the transition to high
school. Journal of Adolescent Research, 15, 587–602. doi:10.1177/
0743558400155004
Garces, L., & Jayakumar, U. (2014). Dynamic diversity: Toward a contex-
tual understand of critical mass. Educational Researcher, 43, 115–124.
doi:10.3102/0013189X14529814
Goza, F., & Ryabov, I. (2009). Adolescents’ educational outcomes: Racial
and ethnic variation in peer network importance. Journal of Youth and
Adolescence, 38, 1264–1279. doi:10.1007/s10964-009-9418-8
Graham, S. (2006). Peer victimization in school: Exploring the ethnic con-
text. Current Directions in Psychological Science, 15, 317–320.
doi:10.1111/j.1467-8721.2006.00460.x
Graham, S. (2016). The role of race/ethnicity in a developmental science of
equity and justice. Child Development, 87, 1493–1504. doi:10.1111/
cdev.12602
Graham, S., Bellmore, A., Nishina, A., & Juvonen, J. (2009). “It must be
me”: Ethnic diversity and attributions for victimization in middle school.
Journal of Youth and Adolescence, 38, 487–499. doi:10.1007/s10964-
008-9386-4
Graham, S., & Echols, L. (in press). Race and ethnicity in peer relations
research. In K. Rubin, W. Bukowski, & B. Laursen (Eds.), Handboook
of peer interactions, relationships, and groups (2nd ed.). New York,
NY: Guilford Press.
Graham, S., & Juvonen, J. (1998). Self-blame and peer victimization in
middle school: An attributional analysis. Developmental Psychology,
34, 587–599. doi:10.1037/0012-1649.34.3.587
Graham, S., & Kogachi, K. (2017). Interracial friendships and attitudes of
Black and White youth in middle school: A cross-lagged analysis. Manu-
script in preparation.
Graham, S., & Kogachi, K. (in press). Cross race/ethnic friendships in school.
In F. Worrell & T. Hughes (Eds.), Cambridge handbook of applied school
psychology. New York, NY: Cambridge University Press.
Graham, S., Munniskma, A., & Juvonen, J. (2014). Psychosocial benefits
of cross-ethnic friendships in urban middle schools. Child Development,
85, 469–483. doi:10.1111/cdev.12159
Graham, S., & Taylor, A. (2017). Attributions for peer victimization and
social-emotional adjustment in ethnically diverse middle schools. Manu-
script in preparation.
Hoglund, W., & Hosan, N. (2012). The context of ethnicity: Peer victimi-
zation and adjustment problems in early adolescence. Journal of Early
Adolescence, 33, 585–609. doi:10.1177/0272431612451925
Janoff-Bulman, R. (1979). Characterological and behavioral self-blame:
Inquires into depression and rape. Journal of Personality and Social
Psychology, 37, 1798–1809. doi:10.1037/0022-3514.37.10.1798
Juvonen, J., Kogachi, K., & Graham, S. (2017). When and how do students
benefit from ethnic diversity in middle school? Child Development.
Advance online publication. doi:10.1111/cdev.12834
Juvonen, J., Nishina, A., & Graham, S. (2006). Ethnic diversity and per-
ceptions of safety in urban middle schools. Psychological Science, 17,
393–400. doi:10.1111/j.1467-9280.2006.01718.x
Kawabata, Y., & Crick, N. (2015). Direct and indirective interactive
links between cross-ethnic friendships and peer rejection, internaliz-
ing symptoms, and academic engagement among ethnically diverse
children. Cultural Diversity & Ethnic Minority Psychology, 21,
191–200. doi:10.1037/a0038451
Knifsend, C., & Juvonen, J. (2014). Social identity complexity, cross-eth-
nic friendships, and intergroup attitudes in urban middle schools. Child
Development, 85, 709–721. doi:10.1111/cdev.12157
Knifsend, C., & Juvonen, J. (2017). Extracurricular activities in middle
schools: Ideal context for positive intergroup attitudes? Journal of
Research on Adolescence, 27, 407–422. doi:10.1111/jora.12278
Kumar, R., Seay, N., & Karabenick, S. (2011). Shades of white: Identity
status, stereotypes, prejudice, and xenophobia. Educational Studies, 47,
347–378. doi:10.1080/00131946.2011.587922
Lanza, S., & Cooper, B. (2016). Latent class analysis for developmental
research. Child Development Perspectives, 10, 59–64. doi:10.1111/
cdep.12163
Lewis, J., Nishina, A., Hall, A., Cain, S., Bellmore, A., & Witkow, M. (2017).
Early adolescents’ peer experiences with ethnic diversity in middle school.:
Implications for academic outcomes. Journal of Youth and Adolescence.
Advance online publication. doi:10.1007/s10964-017-0697-1
Linn, R. L., & Welner, K. G. (2007). Race-conscious policies for assigning
students to schools: Social science research and the Supreme Court
cases (Report prepared by the National Academy of Education Commit-
tee on Social Science Research Evidence on Racial Diversity in
Schools). Washington, DC: National Academy of Education.
Major, B., Quinton, W., & McCoy, S. (2002). Antecedents and consequen-
ces of attributions to discrimination: Theoretical and empirical advan-
ces. In M. Zanna (Ed.), Advances in experimental social psychology
(vol. 34, pp. 252–330). New York, NY: Academic Press.
76 GRAHAM
McPherson, M., Smith-Lovin, L., & Cook, J. (2001). Birds of a feather:
Homophily in social networks. Annual Review of Sociology, 27, 415–
444. doi:10.1146/annurev.soc.27.1.415
Mickelson, R. (2015). The cumulative disadvantages of first-and
second-generation segregation for middle school achievement.
American Educational Research Journal, 52, 657–692. doi:10.3102/
0002831215587933
Mickelson, R., Bottia, M., Larrimore, S., & Lambert, R. (2016). The
effects of school composition on K-12 reading and math achievement.
In E. Frankenberg, L. Garces, & M. Hopkins (Eds.), School Integration
matters: Research-based strategies to advance equity (pp. 56–72). New
York, NY: Teachers College Press.
Miller, K., Brewer, M., & Arbuckle, N. (2009). Social identity complexity:
Its correlates and antecedents. Group Processes & Intergroup Relations,
12, 79–94. doi:10.1177/1368430208098778
Moody, J. (2001). Race, school integration, and friendship segregation in
America. American Journal of Sociology, 107, 679–716. doi:10.1086/
338954
Morales-Chicas, J., & Graham, S. (2017). Latinos’ changing ethnic group
representation from elementary school to middle school: Perceived
belonging and academic achievement. Journal of Research on Adoles-
cence, 27, 537–549. doi:10.1111/jora.12292
National Center for Education Statistics. (2015). National elementary and
secondary enrollment by race/ethnicity projection model, 1972 through
2025. Digest of Education Statistics, 2015. Retrieved from http://nces.
ed.gov/programs/digest/d15/tables/dt15_203.60.asp
Newgent, R., Lee, S., & Daniel, A. (2007). Interracial best friendships:
Relationship with 10th graders’ academic achievement level. Profes-
sional School Counseling, 11, 98–104. doi:10.5330/PSC.n.2010-11.98
Olweus, D. (1993). Bullying at school. Malden, MA: Blackwell.
Orfield, G. (2014). Tenth annual Brown lecture in education research: A
new civil rights agenda for American education. Educational
Researcher, 43, 273–292. doi:10.3102/0013189X14547874
Pettigrew, T. (2004). Justice deferred a half century after Brown v Board of
Education. American Psychologist, 59, 521–529. doi:10.1037/0003-
066X.59.6.521
Pettigrew, T., & Tropp, L. (2006). A meta-analytic test of intergroup con-
tact theory. Journal of Personality and Social Psychology, 90, 751–783.
doi:10.1037/0022-3514.90.5.751
Phillips, K. (2014). How diversity works. Scientific American, 103, 42–47.
doi:10.1038/scientificamerican1014-42
Plaut, V. (2010). Diversity science: Why and how difference makes a differ-
ence. Psychological Inquiry, 21, 77–99. doi:10.1080/10478401003676501
Rivara, F., & Le Menestrel, S. (2016). (Eds.). Preventing bullying through
science, policy, and practice. Washington, DC: National Academies Press.
Roccas, S., & Brewer, M. (2002). Social identity complexity. Personality and
Social Psychology Review, 6, 88–106. doi:10.1207/S15327957PSPR0602_01
Saafir, A., & Graham, S. (2017). Social identity complexity as a buffer of
the effects of racial discrimination on multiethnic middle school stu-
dents. Manuscript in preparation.
Schmid, K., Hewstone, M., & Al Ramiah, A. (2013). Neighborhood diver-
sity and social identity complexity: Implications for intergroup relations.
Social Psychological and Personality Science, 4, 135–142. doi:10.1177/
1948550612446972
Stanton-Salazar, R. D., & Dornbusch, S. M. (1995). Social capital and the repro-
duction of inequality: Information networks among Mexican-origin high
school students. Sociology of Education, 68, 116–135. doi:10.2307/2112778
Tajfel, H. (1981). Human groups and social categories: Studies in social
psychology. Cambridge, England: Cambridge University Press.
Tajfel, H., & Turner, J. (1979). An integrative theory of intergroup rela-
tions. In W. Austin & S. Worchel (Eds.), The social psychology of inter-
group relations (pp. 33–48). Monterey, CA: Brooks/Cole.
Tanti, C., Stukas, A., Halloran, M., & Foddy, M. (2011). Social identity
change: Shifts in social identity during adolescence. Journal of Adoles-
cence, 34, 555–567. doi:10.1016/j.adolescence.2010.05.012
Vitoroulis, I., & Vaillancourt, T. (2015). Meta-analytic results of ethnic
group differences in peer victimization. Aggressive Behavior, 41, 149–
170. doi:10.1002/ab.21564
Weiner, B. (1986). An attributional theory of motivation and emotion. New
York, NY: Springer.
Wells, A., Fox, L., & Cordova-Cobo, D. (2016). How racially diverse schools
and classrooms can benefit all students. Education Digest, 82, 17–24.
ADOLESCENTS AND SCHOOL ETHNIC DIVERSITY 77