ArticlePDF Available

Teacher and Principal Diversity and the Representation of Students of Color in Gifted Programs: Evidence from National Data

Authors:

Abstract and Figures

Students of color are significantly underrepresented in gifted programs relative to their White peers. Drawing on political science research suggesting that public organizations more equitably distribute policy outputs when service providers share characteristics with their client populations, we investigate whether representation of students of color in gifted programs is higher in schools with racially/ethnically diverse teachers and principals. In a nationally representative sample of elementary schools created by merging two waves of data from the Civil Rights Data Collection and the Schools and Staffing Survey, we find that schools with larger numbers of Black teachers or a Black principal have greater representation of Black students in gifted programs. We find a similar relationship for Hispanic teachers and representation of Hispanic students. Further evidence suggests that a critical mass of teachers of color is necessary for teacher race/ ethnicity to be associated with higher representation of students of color in gifted programs.
Content may be subject to copyright.
TEACHER AND PRINCIPAL DIVERSITY
AND THE REPRESENTATION OF
STUDENTS OF COLOR IN
GIFTED PROGRAMS
Evidence from National Data
abstract
Students of color are signicantly underrepresented in
gifted programs relative to their White peers. Drawing
on political science research suggesting that public orga-
nizations more equitably distribute policy outputs when
service providers share characteristics with their client
populations, we investigate whether representation of
students of color in gifted programs is higher in schools
with racially/ethnically diverse teachers and principals. In
a nationally representative sample of elementary schools
created by merging two waves of data from the Civil
Rights Data Collection and the Schools and Stafng Sur-
vey, we nd that schools with larger numbers of Black
teachers or a Black principal have greater representation
of Black students in gifted programs. We nd a similar
relationship for Hispanic teachers and representation of
Hispanic students. Further evidence suggests that a critical
mass of teachers of color is necessary for teacher race/
ethnicity to be associated with higher representation of
students of color in gifted programs.
Jason A. Grissom
Luis A. Rodriguez
Emily C. Kern
van de rb ilt
university
the elementary school journal
Volume 117, Number 3. Published online February 10,2017
©2017 by The University of Chicago. All rights reserved. 0013-5984/2017/11703-0005$10.00
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
Since at least the late 1960s, research has consistently documented the sub-
stantial underrepresentation of students of color in gifted programs (Ford,
1998). Recent data show, for example, that Black students are only 59%as
likely to receive gifted services as would be predicted if their gifted partic-
ipation was proportionate to their presence in the broader student population.
1
To
receive gifted services, students must go through multiple steps, including identi-
cation as potentially gifted, referral for evaluation, and the evaluation itself, and
research suggests that students of color are less likely to pass through each of these
stages than their White peers (McBee, 2006; National Research Council, 2002).
Reasons for these disparities are complex but likely include unequal teacher per-
ceptions of student giftedness across student groups (Ford, Grantham, & Whiting,
2008; Hargrove & Seay, 2011) and the use of single, potentially culturally biased
tests to assess giftedness as a unidimensional construct (Ford et al., 2008;Harris,
Brown, Ford, & Richardson, 2004). Less often discussed is the fact that gifted rep-
resentation among students of color can vary markedly from school to school, even
among schools with similar student demographic compositions. Some differences
are attributable to state-to-state differences in gifted denitions and identication
processes, but even within states (and even districts), variation in implementation
of policy can lead to considerable disparity in rates of gifted identication (National
Research Council, 2002).
Relatively few studies, however, have examined the school-level factors inu-
encing the rates of placement in gifted programs across different student racial
and ethnic groups (Ford, 1998;McBee,2006; McBee, Shaunessy, & Matthews, 2012).
Motivated by the long literature on what is termed bureaucratic representation
or the representative bureaucracy in political science, this study focuses on one par-
ticular set of potentially important factors: the demographic characteristics of the
schools teachers and principal. Bureaucratic representation theory suggests that
more descriptively representative public organizationsthat is, those whose em-
ployees share demographic characteristics with client populationstend to more
equitably distribute policy outputs among client groups (see Grissom, Kern, & Rod-
riguez, 2015; Kennedy, 2014; Meier, 1993). The potential mechanisms are varied and
include behavioral responses to descriptive representation by both bureaucratsin
this case, educatorsand the clients (students and parents) themselves, such as
greater sensitivity of minority bureaucrats from historically marginalized groups
to minority client needs, advocacy by minority bureaucrats for organizational pol-
icies that ameliorate past disparities between minority and nonminority clients, and
increased likelihood that minority clients seek out organizational services in the
presence of bureaucrats like them(Lim, 2006). Research suggests that often these
mechanisms only surface when a critical massof bureaucrats from a minority
group is present in the organization (Henderson, 1979; Kanter, 1977; Thompson,
1976). Note that in this literature, minorityis a shorthand for groups historically
disadvantaged in public policy processes, and can include populations dened by
such characteristics as gender, sexual identity, and religion, though most often stud-
ies have dened it in terms of race and ethnicity.
Assignment to gifted services may be a particularly likely place to observe the
effects of representation because of the roles that subjectivity and discretion play
in the assignment process. Except in the case of universal screening, which is em-
teacher and principal diversity 397
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
ployed in some systems (e.g., Card & Giuliano, 2014), gifted assignment begins
with referral for evaluation, typically by a classroom teacher, on the basis of the
teachers perception of the students potential for giftedness (McBee, 2006). We
may expect teacher race or ethnicity to matter for which students are referred for
testing because, for example, teachers of color may be more attuned to giftedness
among racially and ethnically diverse students due to heightened sensitivity to cul-
tural differences among students from different groups. This consideration may be
especially important given evidence that teachers are less likely to perceive gifted-
ness among students of color than among White students even when presented
with evidence about the students that is otherwise similar (Elhoweris, Mutua,
Alsheikh, & Holloway, 2005). At the same time, parents from diverse populations
may feel more comfortable communicating with teachers from similar backgrounds
and thus more likely to request that students be evaluated for gifted services. Once
referred, students are formally tested using standardized assessments and other met-
rics; then often a placement committee, again consisting of teachers and other per-
sonnel, such as school psychologists, reviews the information gathered during eval-
uation and makes a nal determination of whether the student should be ofcially
designated as gifted (National Research Council, 2002). Evaluation and placement
decisions may present additional opportunities for teacher representation effects
for students of color. For example, teachers of color may be more likely to express
concern about evaluation procedures that disadvantage non-White students (e.g.,
IQ tests) and advocate for the use of multiple measures of giftedness or other changes
to evaluation and placement procedures that improve the likelihood of the recog-
nition of giftedness among students of color (Ford et al., 2008; Joseph & Ford,
2006).
Representative bureaucracy research has also found evidence that the race and
ethnicity of managers in a public organization can affect the distribution of policy
outputs (Grissom & Keiser, 2011; Meier & Stewart, 1992), suggesting that principal
race/ethnicity may have effects on the representation of students of color in a
schools gifted programs as well. Although principals are less likely than teachers
to be directly involved in gifted referral, evaluation, and placement decisions, rec-
ognition of inequities in assignment patterns among White and non-White stu-
dents may make non-White principals more likely to shape school assignment
practices to increase gifted representation among diverse populations. By imple-
menting policies such as allowing gifted nominations from non-teachers, using
multiple measures to evaluation giftedness, and providing multicultural training
to teachers to help circumvent biases in the identication process, principals are
positioned to inuence minority representation in their schoolsgifted programs
(Harris et al., 2004; Joseph & Ford, 2006; Matthews & Shaunessy, 2010; McBee
et al., 2012).
Investigation of connections between the diversity of a schools educators and
the representation of students of color in gifted programs is particularly important
in light of the growing mismatch between teacher and student demographics in
American schools (Boser, 2014; Grissom et al., 2015). Some earlier bureaucratic rep-
resentation research has linked teacher workforce diversity to greater rates of gifted
participation among Black and Hispanic students, though this research has been
published outside of education and remains unfamiliar to education scholars (e.g.,
398 the elementary school journal march 2017
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
Nicholson-Crotty, Grissom, & Nicholson-Crotty, 2011; Rocha & Hawes, 2009). The
purpose of this study is to build on and extend this prior work in several key ways.
First, we investigate connections between teacher racial/ethnic diversity and gifted
representation using the most recently available national data (from the 20112012
school year) to test whether associations between teacher demographic character-
istics and gifted composition observed with earlier data remain relevant and how
they may have changed. More specically, we ask: to what extent are the propor-
tions of Black and Hispanic teachers in a school associated with the racial compo-
sition of students assigned to gifted programs? Second, we investigate the critical
mass phenomenon observed in representation studies in other organizational set-
tings. That is, we ask: is there evidence that a critical number of teachers of color
must be present in a school before teacher race/ethnicity becomes associated with
the racial/ethnic composition of the schools gifted program? Third, we explore
whether the race/ethnicity of the schools principal is associated with gifted repre-
sentation, even after controlling for the characteristics of the teacher workforce, a
possibility not considered in prior empirical work. Specically, we ask whether
schools with Black or Hispanic principals have gifted programs that are more rep-
resentative of students from those racial/ethnic backgrounds and, furthermore,
whether principal race/ethnicity moderates the connection between teacher diver-
sity and the composition of the schools gifted programs.
We begin by discussing bureaucratic representation theory as a framework for
understanding the connection between the demographic diversity of a schools
faculty and differential gifted placement rates for White students and students
of color. We then discuss our data and methods before presenting the studys re-
sults. The nal section concludes with a discussion of the implications of the nd-
ings for policy and practice in the area of gifted education and efforts to increase
equity in the provision of school services more generally. We also discuss study
limitations and ideas for future work.
Bureaucratic Representation Theory and Application
to Gifted Placement
We draw on bureaucratic representation theory, a well-established theoretical
framework in the elds of political science and public administration literature,
to guide our investigation into how teacher and principal roles in the identica-
tion, referral, and evaluation processes for gifted programs may connect educator
diversity to disparate rates of gifted placement across different racial groups. Rep-
resentation scholarship begins from the observation that the composition of the
public bureaucracy inuences the implementation of public policy; that is, who
the providers of government services are matters for how policy outputs are dis-
tributed. In particular, scholars examining a variety of contexts have repeatedly
made empirical connections between descriptive representation in the bureaucratic
workforcethat is, the degree to which public sector workers share demographic
characteristics such as race, ethnicity, or gender with the populations they serve
and greater access to policy outputs for traditionally disadvantaged groups (Gris-
som et al., 2015). As an example, in one study, Equal Employment Opportunity Com-
teacher and principal diversity 399
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
mission eld ofces employing more Black and Hispanic investigators were shown
to bring a greater number of charges on behalf of Black and Hispanic complainants
(Hindera, 1993).
These connections have also been made in prior research on representation in
schools (see Grissom et al., 2015, for a review). Bureaucracy scholars have long
looked toward teachers as quintessential street-level bureaucratsgovernment pro-
fessionals who, like police ofcers or social workers, work directly with client pop-
ulations (i.e., students and their families) in roles with substantial discretion and
autonomy (Lipsky, 1980)and thus viewed schools as verdant ground for testing
many areas of bureaucratic theory. Work in this tradition has linked a larger pres-
ence of Black and Hispanic teachers to improved treatment or outcomes for Black
and Hispanic students along a variety of dimensions, including lower rates of exclu-
sionary discipline (Grissom, Nicholson-Crotty, & Nicholson-Crotty, 2009; Meier &
Stewart, 1992), lower likelihood of placement in special education (Fraga, Meier, &
England, 1986; Rocha & Hawes, 2009), and higher pass rates on standardized tests
(Meier, Wrinkle, & Polinard, 1999;Weiher,2000).
Researchers also have linked teacher racial and ethnic diversity to placement
rates for non-White students in gifted programs (Grissom et al., 2009; Rocha &
Hawes, 2009), particularly in schools where non-White students are assigned in-
frequently relative to their proportion in the overall school population (Nicholson-
Crotty et al., 2011). Combining district-level data from several sources from 2000
and 2001, Rocha and Hawes (2009)nd that larger proportions of both Black and
Hispanic teachers correlate with more equitable placement of both Black and His-
panic students into gifted programs. This nding suggests that racial/ethnic minor-
ity street-level bureaucrats may positively inuence outcomes for more than just
co-ethnic clients. In a similar study, Grissom et al. (2009) examined nationally rep-
resentative school-level data from the 20032004 school year to test for relationships
between the proportion of Black teachers in a school and the proportion of gifted
students who were Black, and how that relationship is moderated by region. The
study nds that larger proportions of Black teachers had positive associations with
Black gifted placements across all schools and particularly in the South. In more re-
cent work, Grissom and Redding (2016) and Nicholson-Crotty, Grissom, Nicholson-
Crotty, and Redding (2016) examine data from a cohortstudy of elementary students
who began kindergarten in 1998 to model gifted assignment at the individual student
level. Those studies nd evidence that Black students are assigned to gifted services
more often when their classroom teacher is Black, controlling for a variety of stu-
dent, classroom, and school characteristics.
Two other studies have drawn on representative bureaucracy theory to examine
other aspects of the relationship between educator characteristics and gifted place-
ment. Using a nationally representative dataset, Nicholson-Crotty et al. (2011) in-
vestigate whether increased non-White representation and placement into gifted
programs is also associated both with proportionately higher placement of non-
White students and lower placement of White students into gifted programs,
which would be expected if gifted placements were zero-sum.That is, if the num-
ber of slots in gifted services is xed and scarce, then an increase in the number of
non-White students placed into gifted programs due to representation by non-
White teachers would necessarily displace White students.
2
The authors report
400 the elementary school journal march 2017
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
that an increase in non-White teachers is signicantly positively related to the pro-
portion of gifted students who are non-White and negatively associated with the
placement rate of White students, but only in schools where non-White students
are underrepresented in gifted programs relative to their proportion in the schools
overall population. In schools where Black and Hispanic students are overrepre-
sented in gifted programs, an increase in teachers from those groups is not related
to increases in Black or Hispanic gifted placement, providing evidence that active
representation occurs only until equity in outcomes is reached. In the only prior
study to consider principals, Meier and Stewart (1992) explore the inuence of
management-level representation by including principal race as a covariate. Using
district-level data from 67 Florida school districts, they nd that an increase in
Black teachers is positively related to Black studentslikelihood of being placed
in a gifted program, but that gifted representation among Black students is not as-
sociated with the race of the principal.
Researchers have put forth a number of mechanisms through which descriptive
representation provides substantive benets for underserved client populations,
including both direct actions by minority bureaucrats and indirect effects minority
bureaucrats may have on the behaviors of their colleagues or the client population
itself (Lim, 2006). We illustrate the main hypothesized mechanisms in the context
of assignment to gifted services.
The mechanism most commonly set forth in the literature is that descriptive
representation benets minority clients because minority bureaucrats exercise dis-
cretion toward them in benecial ways. Some scholars have expressed concern that
this benecial exercise of discretion simply reects bias (Mosher, 1968), as would
be the case if an African American teacher, presented with a White and an African
American student of similar capacities, was more inclined to refer the African
American student for gifted evaluation. Others have suggested instead that shared
demographic characteristics proxy for shared background, values, beliefs, or un-
derstanding, which may lead to discretionary actions by minority bureaucrats that
benet minority clients (Lim, 2006). As an example, a Hispanic teacher may be
more likely to refer a Hispanic student to gifted services because linguistic or cul-
tural sensitivity better equips her to recognize giftedness in Hispanic students.
Other mechanisms linking bureaucratic diversity to improved outcomes for mi-
nority client populations are indirect, operating through changes in the behaviors
of others. A minority teacher may be more attuned to practices within the organi-
zation that disadvantage minority students and thus advocate for changes to those
practices, either informally or formally. An example of the informal case is the
non-White teacher who pushes her White colleagues to look closely at non-White
students for signs of giftedness or provides them with some tips or ideas for assess-
ing giftedness in culturally diverse students. In the more formal case, a non-White
teacher may be more likely to recognize bias toward non-White students in tests
used for gifted evaluation and advocate for the school to use a different test (Ford
et al., 2008). In both instances, the presence of non-White teachers in the school
beneted non-White students by nudging the environment in a direction that in-
creased their probability of success.
Descriptive representation may also benet minority clients indirectly by
changing the behaviors of the clients themselves. A Hispanic parent may be more
teacher and principal diversity 401
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
likely to approach her childs Hispanic teacher to request that the student be re-
ferred for gifted services, for example, if shared language or culture increases the
parents comfort in making the request. Students with teachers of similar demo-
graphic backgrounds may also perform better on assessments that make gifted
identication more likely. This demographic similarity may improve testing out-
comes via a role-modeling effect, wherein students work harder to gain approval
from teachers like them (Lim, 2006), or by reducing stereotype threat, a psycho-
logical impediment to performance based on anxiety around stereotypes, which
may be more salient in the presence of other-race teachers (Dee, 2005).
Other research on representation in organizations suggests that the impact of
descriptive representation on outcomes for diverse clients is unlikely in the ab-
sence of a critical mass, or numerical threshold, of bureaucrats from the minority
group (Henderson, 1979; Kanter, 1977; Thompson, 1976). These scholars focus on
the impact minority bureaucrats can have on organizational policies or practices,
suggesting that only when minorities have enough of a presence can they build in-
ternal support to effect change (Henderson, 1979; Thompson, 1976). Representa-
tion effects will thus be nonlinear, and in fact a pattern consistent with a critical
mass condition was observed in Hinderas study of EEOC complaints (see Hind-
era & Young, 1998). The threshold for when descriptive representation is likely to
matter is unclear. Kanter (1977) proposed that the minority group must comprise
15% of the organizational workforce before descriptive representation would pro-
duce substantive effects. Critical mass effects have largely gone ignored in studies
of schools, with the exception of one study by Meier (1993), which found some ev-
idence that the presence of Hispanic principals (but not teachers) begins to have a
positive association with student disciplinary and achievement outcomes only
once they comprise between 16% and 26% of school leaders in a district. No studies
of which we are aware have examined the critical mass idea in the context of gifted
assignment.
Studies similarly have overlooked the potential for representation effects on
gifted outcomes for school principals. Principals may inuence gifted assignments
in their schools by implementing referral and evaluation policies that may increase
rates of placement for underrepresented groups. For example, students in Florida
who attended schools with alternative policies in place to increase gifted represen-
tation by non-White students had nearly twice the probability of being identied
as gifted as students attending schools without these plans (McBee et al., 2012), il-
lustrating a school-level policy change via which principals might affect the gifted
participation of students of color. It is also possible that non-White principals
could encourage referral of students of color by providing professional develop-
ment to teachers on recognizing giftedness for underserved groups or implement-
ing a systematic screening process to increase opportunities for identication. A
non-White principals presence could also make the parent of a student of color
more comfortable in requesting a referral form for their child. Some limited em-
pirical evidence suggests that non-White principals are associated with improved
schooling outcomes for non-White students in such areas as referral to special ed-
ucation or graduation rates (Meier, 1993; Meier & Stewart, 1992; Pitts, 2005). This
work suggests that attention to possible principal representation effects in gifted
assignments is warranted as well.
402 the elementary school journal march 2017
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
Overall, the literature on bureaucratic representation in education provides a
useful framework to analyze student placement into gifted programs. There never-
theless remain gaps in the current research base surrounding both gifted programs
and bureaucratic representation that we aim to address. First, we use a theoretical
framework normally conned to the political science and public administration
literature to examine factors that inuence the prevalence of students of color in
gifted programs, which has been examined sparingly within the education research
literature. Second, we consider principal race as a possible explanatory factor in
gifted composition, which has been largely overlooked in the literature. Third, we
examine the possible role of critical mass in the relationship between teacher race
and student outcomes, which has received little empirical attention and none in
the context of gifted placement. Fourth, we make use of nationally representative
data at the school level, which helps overcome potential issues of limited generaliz-
ability or aggregation bias from single-state or single-district studies or studies that
have used district-level data to examine representation in an education context (e.g.,
Meier, 1993; Meier & Stewart, 1992; Meier et al., 1999; Pitts, 2005; Weiher, 2000).
Finally, the data we draw on are the most recently available on a national scale,
which allow for assessment of whether patterns identied in earlier research are still
present.
Data
We pair 2years of nationally representative data from two sources: the Schools and
Stafng Survey (SASS) and survey data collected by the Ofce for Civil Rights
(OCR) in the 20032004 school year and again in 20112012, referred to in the re-
mainder of the study as 2004 and 2012, respectively. We supplement these data with
additional district- and school-level information from the Common Core of Data
(CCD). SASS uses a stratied sampling method to gather information on demo-
graphic characteristics, organizational processes, and attitudes of principals and a
random selection of teachers for each school. We merge the SASS data with survey
data administered by the Department of EducationsOfce for Civil Rights (OCR),
which also uses a stratied random sampling method to collect information on ac-
ademic grouping, discipline, and educational attainment disaggregated by gender
and race. Given that students are substantially more likely to be identied for gifted
services in elementary school, we restrict our sample to U.S. public elementary
schools with gifted programs that can be matched between the two samples, then
further restrict the analytic sample to noncharter, nonmagnet schools. The nal
sample size is 2,170 schools. Approximately 6.2% of elementary students were des-
ignated as gifted in the matched sample in both waves of data. Table 1provides de-
scriptive statistics for the pooled sample and for each wave, with asterisks in the
rightmost column indicating statistically signicant differences between the 2004
and 2012 waves from two-sided t-tests.
Dependent Variables
To measure the presence of students from different racial and ethnic back-
grounds in gifted programs, we follow prior disproportionality (e.g., McBee,
teacher and principal diversity 403
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
Table 1. Descriptive Statistics, Full Sample and by Year
(1)
Full Sample
(2)
2004 Sample
(3)
2012 Sample
Students who are gifted (%) 6.22 6.19 6.26
(7.32)(8.07)(6.27)
Gifted students who are Hispanic (%) 10.09 9.59 10.72
(21.29)(21.32)(21.24)
Gifted students who are Black (%) 9.41 10.73 7.78***
(21.39)(23.36)(18.56)
Gifted students who are White (%) 71.71 70.91 72.71
(31.48)(32.30)(30.41)
Teachers who are Hispanic (%) 3.91 3.82 4.03
(11.54)(11.00)(12.19)
Teachers who are Black (%) 6.22 6.87 5.41**
(14.94)(15.82)(13.74)
Teachers who are White (%) 87.76 86.75 89.03***
(20.39)(21.21)(19.26)
Principal is Hispanic .05 .05 .05
(.21)(.21)(.21)
Principal is Black .10 .11 .09
(.30)(.31)(.29)
Principal is White .83 .82 .85
(.37)(.38)(.36)
Students who are Hispanic (%) 15.39 14.50 16.50**
(22.24)(22.41)(21.98)
Students who are Black (%) 14.49 16.20 12.37***
(22.63)(24.41)(20.01)
Students who are White (%) 62.37 61.56 63.38
(30.42)(31.15)(29.48)
Students who are eligible for FRPL (%) 48.01 46.02 50.49***
(27.76)(28.25)(26.94)
Locale type:
City .27 .33 .20***
(.44)(.47)(.
40)
Suburb .32 .32 .33
(.47)(.47)(.47)
Small town .13 .12 .13
(.33)(.33)(.34)
Rural .28 .23 .33***
(.45)(.42)(.47)
School size 491.24 488.20 495.02
(222.85)(219.85)(226.57)
District size 29,339.19 35,328.03 21,907.33***
(72,421.14)(78,063.65)(64,003.83)
Per-pupil expenditures (in $1,000s)
a
11.15 10.85 11.53***
(3.01)(2.71)(3.31)
Observations 2,170 1,200 970
Note.Standard deviations are in parentheses. Results from two-sided t-test shown. Reported sample sizes rounded to the
nearest 10 per NCES nondisclosure rules.
a
Per-pupil expenditures are in 2012 constant dollars.
*p!.10.
** p!.05.
*** p!.01.
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
2006; National Research Council, 2002) and representation (e.g., Hindera, 1993;
Nicholson-Crotty et al., 2011) research and calculate the gifted composition for
each racial/ethnic group as the number of students from that group in the schools
gifted program divided by the total number of gifted students. The gifted compo-
sition of Hispanic students, for example, is simply the number of Hispanic gifted
students in the school divided by the total number of gifted students.
3
Using OCR
data, we calculate the gifted composition of White, Black, and Hispanic students.
As Table 1shows, the composition of the average elementary school gifted popu-
lation within the pooled sample is approximately 72% White, 9% Black, and 10%
Hispanic, in contrast to the composition of the average elementary school popu-
lation at large, which in these data is 62% White, 14% Black, and 15% Hispanic.
4
These pooled averages, however, mask important changes between the samples
over the two waves. In particular, the fraction of the average schools gifted pop-
ulation that is Black fell from nearly 11%in2004 to only about 8%in2012 (p!.01),
perhaps reecting a similar decline in the average schools overall student popula-
tion, which was 16% Black in 2004 but only 12% Black in 2012. At the same time, the
percentage of elementary students who are Hispanic signicantly increased from
14.5%in2004 to 16.5%in2012 (p!.05), although the percentage of gifted students
who are Hispanic did not change signicantly. Both the proportion of elementary
students who are White and the proportion of gifted students who are White re-
mained constant over the two waves.
An alternative means of illustrating the underrepresentation of Black and His-
panic students in gifted programs appears in Figure 1, which shows the risk index,
or the percentage of each racial/ethnic group in gifted programs separately for
2004 and 2012. The gure demonstrates that whereas nearly 8% of White students
receive gifted services, only 3%4% of Black and Hispanic students dopercent-
ages that remained very stable over the two waves of data. This stability suggests
Figure 1. Percent of student population in the gifted program, by race and year.
teacher and principal diversity 405
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
that the changes in the proportions of students in gifted programs from Black and
Hispanic groups between 2004 and 2012 shown in Table 1indeed reect changes in
school composition rather than changes in the allocation of slots in gifted pro-
grams across different groups of students.
Independent Variables
The main independent variables for this study capture the racial/ethnic compo-
sition of the teacher workforce and the race/ethnicity of the principal in each
school in the pooled sample. In particular, we use SASS school questionnaire data
to calculate the percentage of teachers in each school who are Hispanic, Black, or
White. These percentages are, on average, 4%, 6%, and 88%, respectively. The per-
centage of teachers who are Black dropped from approximately 7%in2004 to
about 5%in2012 (p!.05). The percentage of Hispanic teachers remained similar
between the two waves, but the percentage of teachers who are White increased
from approximately 87%in2004 to 89%in2012 (p!.01). From SASS principal
questionnaire data, we also create two separate binary indicator variables for whether
the schools principal is Hispanic, Black, or White. In contrast, principal race and
ethnicity stayed relatively constant across the two waves of data. Approximately
5% of principals in the pooled sample are Hispanic, 10% are Black, and 83% are
White.
Control Variables
Models include control variables to account for other factors that might explain
variation in assignment to gifted programs. Our choice of control variables gener-
ally follows Nicholson-Crotty et al. (2011), which used data similar to those utilized
here. Aside from the fraction of all students in the school assigned to gifted pro-
gramswhich comes from OCR dataand locale (e.g., urban) and expenditure
informationwhich are from the CCDcontrol variables are primarily taken
from SASS. We include the percentageof all students who areBlack and the percent-
age who are Hispanic because the proportion of these populations should be highly
correlated with the proportion of students in gifted programs. We also include the
percentage of students within a school that are eligible for free or reduced-price lunch
(FRPL) because these students are less likely to be placed in gifted programs (McBee,
2006). The percentage of students who are eligible for FRPL increased signicantly
from approximately 46%in2004 to 50%in2012 (p!.01). We control for school size,
district size, and locale type because these factors may be associated with standardi-
zation of gifted identication, referral, and testing processes and different levels of
discretionary academic grouping. The percentage of students who attended schools
in various locales changed signicantly between the two waves of data. The percent-
age of students attending schools in cities declined from 33%to20%(p!.01), while
the percentage of students in rural areas increased from 23%to33%(p!.01). The
percentage of students attending schools in suburban areas or small towns remained
relatively constant across both waves of data. The observable differences likely stem
from changes in urban locale codes within the CCD across 2004 and 2012.
5
In addi-
tion, the average district size decreased signicantly between the two waves, from
406 the elementary school journal march 2017
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
more than 35,000 students to just less than 22,000 (p!.01). Per-pupil expenditures in
constant 2012 dollars is included as a control for school resources, which may inu-
ence the size of a schools gifted program; this value signicantly increased from ap-
proximately $10,850 in 2004 to $11,530 in 2012 (p!.01).
Method
We estimate a series of ordinary least-squares regression models where the depen-
dent variable is the percentage of gifted students in one of three different race or
ethnicity groups (Hispanic, Black, or White). The main independent variables of
interest are the proportion of teachers who are Hispanic, Black, or White, and
two binary variables representing whether the principal is Hispanic or Black.
Each model incorporates state and year xed effects in order to account for dif-
ferences in state procedures related to gifted program funding and assignment and
for differences across the two waves of data. We cluster standard errors at the dis-
trict level to correct for correlated errors within districts. Samples used for estima-
tion for each dependent variable are limited to schools with a student race group
between 1% and 99% of the total population.
If gifted assignment is approximately zero-sum, as Nicholson-Crotty et al. (2011)
argue, then a larger percentage of teachers from one racial group should be associ-
ated with a decrease in the gifted composition for other racial groups. Therefore, to
observe the zero-sum trade-off, we also model the relationship between Black and
Hispanic teachers and the composition of White students in gifted programs.
Additionally, we test for the moderating inuence of the principals race, inves-
tigating the hypothesis that teachers of a certain race will have a stronger inuence
over the racial and ethnic composition of the schools gifted program when the
principal is also from a non-White group. To do so, interactions between the per-
centage of teachers who are of a certain race (Black or Hispanic) and the dummy
for whether the principal is of that same race or (Black or Hispanic) are included in
the model.
Finally, we test for critical mass effects by entering teacher race/ethnicity per-
centages as a series of categorical variables dened over discrete ranges (e.g., 1%
5%, 6%10%, and so forth). These dummy variables allow us to examine whether
the association between the percentage of teachers who are of a certain race and
the placement of students of color into gifted programs is nonlinear. A critical mass
hypothesis would predict little or no association when the fraction of teachers of
color in a school is very low, with an association only becoming apparent beyond
some critical threshold of non-White teachers in the school.
Results
Teacher Race/Ethnicity and Gifted Assignments
We begin by estimating the racial/ethnic composition of the schools gifted pro-
gram in the pooled sample as a function of the fraction of Black and Hispanic
teachers in the school, plus controls. For consistency with prior research (e.g.,
teacher and principal diversity 407
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
Nicholson-Crotty et al., 2011), initially principal race/ethnicity is not included.
Models for Hispanic, Black, and White gifted student composition were run sep-
arately. Results are shown in Table 2. The rst column shows the results for the
model with the percentage of gifted students who are Hispanic as the dependent
variable. The second and third columns show results for Black students and White
students, respectively.
Our results conrm ndings from these earlier studies. The percentage of His-
panic teachers is positively related to the percentage of gifted students who are His-
panic. As shown in column 1,thecoefcient on percent of teachers who are Hispanic
(bp0.31,p!.01) means that a 10% increase in Hispanic teachers is associated with
Table 2. Results from OLS Model Testing Race of Teachers, DV: Percent of Gifted
Teacher Race/Ethnicity Only
(1)
Hispanic Students
(2)
Black Students
(3)
White Students
Teachers who are Hispanic (%) .31*** .01 .38***
(.059)(.034)(.065)
Teachers who are Black (%) .07 .32*** .30***
(.042)(.053)(.058)
Students who are gifted (%) .09*** .04 .07
(.028)(.024)(.045)
Students who are Hispanic (%) .73*** .00 .56***
(.040)(.027)(.048)
Students who are Black (%) .03 .68*** .68***
(.035)(.043)(.043)
Students who are eligible for FRPL (%) .00 .03 .08***
(.019)(.017)(.026)
Suburb .85 .11 1.29
(.813)(.839)(1.160)
Small town 3.21*** .87 4.27***
(1.158)(1.018)(1.516)
Rural .93 1.09 1.78
(.912)(.986)(1.252)
School size (in 100s) .04 .22 .23
(.162)(.156)(.202)
District size (logged) .10 .66*1.39***
(.313)(.376 )(.473)
Per-pupil expenditures (in $1,000s)
a
.07 .15 .40*
(.120)(.147)(.206)
Year .01 .04 .29***
(.076)(.078)(.101)
Constant 11.55 87.23 695.61***
(153.490)(157.138)(202.581)
Observations 1,830 1,660 2,050
Adjusted R
2
.610 .706 .587
Note.Standard errors clustered at the district level are shown in parentheses. Models include state xed effects. Controls
not shown include percent of students who are gifted, percent of students who are Black, percent of students who are Hispanic,
percent of students eligible for free and reduced-price lunch, locale type, school size, district size, and per-pupil expenditure. Samples
limited to schools containing between 1% and 99% of students in each DV group. Reported sample sizes rounded to the nearest 10 per
NCES nondisclosure rules.
a
Per-pupil expenditures are in 2012 constant dollars.
*p!.10.
** p!.05.
*** p!.01.
408 the elementary school journal march 2017
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
a3.1% increase in Hispanic gifted students. This increase is meaningful, given that
the sample average of gifted students who are Hispanic is just 10%. The percentage
of Black teachers has no detectable relationship with the percentage of gifted
students who are Hispanic. The relationship between Hispanic teachers and His-
panic students exists when controlling for the student body makeup, locale type,
per-pupil expenditures, and year.
6
The association between percent Hispanic teachers and the percentage of gifted
students who are Hispanic is shown in Figure 2. The solid line represents the re-
lationship between Hispanic teachers and Hispanic students in 2004, and the dot-
ted line represents the same relationship in 2012. Light gray lines represent the 95%
condence intervals for the predicted margins. Figure 2shows that the percentage
of gifted students who are Hispanic is predicted to increase as the percent of His-
panic teachers increases. The nearly overlapping year lines in this gure show that
the relationship between Hispanic teachers and Hispanic gifted students remained
constant between the two waves of data.
Column 2of Table 2reports the results from the model with Black gifted com-
position as the dependent variable. These results show that, even after controlling
for various school and district characteristics, a 10% increase in the percentage of
Black teachers in a school is associated with an increase in the representation of
Black students in gifted programs of about 3.2%(p!.01). Considering that the
sample average percent of gifted students who are Black is 9.4%, a 3.2percentage
point change is substantively signicant, representing an increase of 34%, on aver-
age. The proportion of Hispanic teachers has no detectable relationship with the
gifted composition of Black gifted students. Figure 3shows predicted composition
values for Black students according to the percentage of Black teachers in the
Figure 2. Predicted percent of gifted students who are Hispanic by percent of teachers who are
Hispanic, by year.
teacher and principal diversity 409
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
school, again separately for 2004 and 2012 (the two are statistically indistinguish-
able).
The results from the model of the percentage of gifted students who are White
are shown in column 3of Table 2. Schools with larger proportions of non-White
teachers are associated with less White representation in gifted programs (p!.01).
A10% increase in the proportion of either Hispanic or Black teachers is related to
approximately 4%or3% drop in the percent of gifted students who are White, re-
spectively. These results are consistent with the idea that assignment of a student to
gifted services approximates a zero-sum game (Nicholson-Crotty et al., 2011); in-
creases in the percentages of gifted students who are Hispanic or Black are asso-
ciated with a decrease in the percentage of gifted students who are White, holding
other factors constant.
Although not our main focus, before moving on it is worth pointing out a few
notable patterns from the control variables in Table 2. The percentage of students
in the school who are Hispanic or Black is positively associated with the percentage
of gifted students in the school that are Hispanic or Black, respectively, and neg-
atively related to the percentage of gifted White students. Also, the total percentage
of gifted students in a school is positively related to the gifted composition of His-
panic students (bp0.09,p!.01). Additionally, as the percentage of student el-
igible for FRPL increases by 10%, the percent of gifted students who are White
is predicted to decrease by 0.8%(p!.01). Small-town schools have lower Hispanic
gifted composition, on average (bp3.21,p!.01), but a higher percentage of gifted
students who are White (bp4.27,p!.01). District size (logged) is negatively cor-
related with the percentage of gifted students who are White (bp1.39,p!.01)but
positively correlated with the percentage of gifted students who are Black (bp0.66,
Figure 3. Predicted percent of gifted students who are Black by percent of teachers who are Black,
by year.
410 the elementary school journal march 2017
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
p!.10). Finally, the percentage of gifted students who are White was smaller in 2012
than in 2004 (bp0.29,p!.01).
Critical Mass Analysis
To test for evidence that a critical mass of teachers from a racial/ethnic group is
required in a school before there is an association between teacher race/ethnicity
and the composition of the schools gifted program, we reestimated Table 2with a
series of indicators representing varying percentages of teachers who are Hispanic
and Black. Results are shown in Table 3. As in Table 2, the three columns represent
the results from three models run with the dependent variable as the composition
of the gifted program for each of Hispanic, Black, and White students.
Coefcients in Table 3generally show evidence of a critical mass requirement.
Model 1indicates a large jump in the percentage of students in gifted programs
who are Hispanic once the fraction of teachers who are Hispanic reaches between
20% and 30%. This evidence is easier to see in Figure 4, which graphs the results
from this model. The gure shows that Hispanic representation in gifted programs
hovers around 10% when the school has fewer than 20%30% Hispanic teachers
but jumps to around 25% once that threshold is reached, with no evidence that
the fraction continues to climb as the percentage of Hispanic teachers increases
beyond that point.
The results from the model with Black students as the dependent variable are
shown in column 2of Table 3. Turning to the coefcients lower in the table that
categorize the percentage of Black teachers into bins, we see a relationship that
is similar to the Hispanic studentHispanic teacher result. Figure 5, which graphs
the predictions from these coefcients, shows the same jump at around 20%30%
Black teachers, with some evidence of further increases for higher values.
The results for the model with White students as the dependent variable are
shown in column 3of Table 3. The results indicate that there is a trade-off with
the percent of gifted students who are White with those who are Hispanic or Black.
The signicantly positive relationships for non-White gifted students and non-
White teachers are inversely related to the relationships between non-White teach-
ers and White students. As the percent of Hispanic or Black teachers reaches 20%
30%, the percent of gifted White students decreases at nearly the same magnitude
as the increases for students of color.
Examining Principal Race/Ethnicity
The next set of models, reported in the rst three columns of Table 4, adds prin-
cipal race and ethnicity measures as covariates. As in Tables 2and 3, each of these
columns represents a different race/ethnicity composition variable as the depen-
dent variable (Hispanic, Black, and White). The rst column shows that the coef-
cients for the percentage of Hispanic and Black teachers are similar in magnitude
to those shown for Hispanic students in Table 2. The race/ethnicity of the princi-
pal, however, shows no evidence of a relationship with the proportion of gifted stu-
dents who are Hispanic.
teacher and principal diversity 411
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
Column 2shows results for Black students, and again the coefcients for His-
panic and Black teachers are similar to those in the related model in Table 2. Here,
the presence of a Black principal has a signicantly positive association with the
percent of gifted students who are Black (bp3.76,p!.05), meaning that the pres-
ence of a Black principal is associated with approximately a 3.8percentage point
increase in the share of gifted students who are Black. This shift is roughly equiv-
Table 3. Results from OLS Model Testing Critical Mass Theory, DV:
Percent of Gifted Students from Each Race Group
(1)
Hispanic Students
(2)
Black Students
(3)
White Students
Teachers who are Hispanic (%):
10% and 5%2.58*** .61 .51
(.689)(.717)(1.070)
15% and 10%1.99 .79 1.91
(1.238)(1.121)(2.117)
110% and 20%1.73 2.80 5.65*
(2.241)(1.902)(3.029)
120% and 30%13.66*** .30 12.33***
(3.877)(2.701)(4.247)
130% and 50%12.13*** 1.37 16.24***
(3.847)(2.079)(4.681)
150% and 75%18.16*** .40 23.05***
(5.015)(2.008)(5.248)
175% and 100%16.62*** 3.14 25.16***
(5.695)(2.471)(7.255)
Teachers who are Black (%):
10% and 5%1.84** 1.60** 1.71
(.840)(.718)(1.209)
15% and 10%.81 2.30*.29
(1.193)(1.175)(1.771)
110% and 20%3.47** 4.69*** .63
(1.550)(1.438)(
1.976)
120% and 30%2.54 5.29** 6.45**
(1.969)(2.537)(2.590)
130% and 50%4.39 11.14*** 13.57***
(3.230)(3.386)(3.881)
150% and 75%1.97 19.84*** 18.25***
(3.355)(4.911)(5.295)
175% and 100%4.89 20.94*** 20.67***
(4.242)(4.523)(5.017)
Year p2012 .03 .01 .29***
(.074)(.074)(.101)
Constant 54.44 26.75 693.35***
(148.516)(149.374)(201.567)
Observations 1,830 1,660 2,050
Adjusted R
2
.620 .716 .586
Note.Standard errors clustered at the district level are shown in parentheses. Models include state xed effects. Controls
not shown include percent of students who are gifted, percent of students who are Black, percent of students who are Hispanic,
percent of students eligible for free and reduced-price lunch, locale type, school size, district size, and per-pupil expenditures. Samples
limited to schools containing between 1% and 99% of students in each DV group. Reported sample sizes rounded to the nearest 10 per
NCES nondisclosure rules.
*p!.10.
** p!.05.
*** p!.01.
412 the elementary school journal march 2017
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
alent to the one associated with a 10% increase in the percentage of teachers who
are Black in the school.
The results for White students are shown in column 3. The magnitude of the
coefcients for percent Hispanic and percent Black teachers are nearly identical
to those in Table 2(with both coefcients statistically signicant at the .01 level).
The presence of a Hispanic or Black principal is not statistically associated with the
proportion of gifted students who are White at conventional levels, though the co-
Figure 5. Critical mass of Black teachers and predicted percent of gifted students who are Black.
Figure 4. Critical mass of Hispanic teachers and predicted percent of gifted students who are
Hispanic.
teacher and principal diversity 413
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
efcient on the Black principal variable is relatively large in magnitude and nega-
tively signed.
Columns 4through 6of Table 4report the results of adding interactions be-
tween (a) the binary indicator for having a Black principal and percent of Black
teachers and (b) the binary indicator for having a Hispanic principal and percent
of Hispanic teachers. Column 4represents the results when the dependent variable
is the percent of gifted students who are Hispanic. The takeaway from this column
is that neither of the interaction terms is statistically signicant. The predictive
margins from this model are shown in Figure 6. The dotted line represents the pre-
dicted percentage of gifted students who are Hispanic across increasing percent-
ages of teachers who are Hispanic for schools with a Hispanic principal. The solid
line represents that relationship for non-Hispanic principals. The lighter-colored
lines represent the 95% condence intervals for those predictions. The gure shows
that while the predicted percentage of gifted students who are Hispanic is predicted
to increase as the percent of Hispanic teachers increases, there is no moderating in-
uence of a Hispanic principal on that relationship, evidenced by the substantially
overlapping condence intervals.
Table 4. Results from OLS Model Including Principal and Teacher Race, DV:
Percent of Gifted Students from Each Race Group
Teacher and Principal Teacher and Principal Interaction
(1)
Hispanic
Students
(2)
Black
Students
(3)
White
Students
(4)
Hispanic
Students
(5)
Black
Students
(6)
White
Students
Teachers who are Hispanic (%) .31*** .00 .38*** .26*** .02 .40***
(.061)(.034)(.067)(.080)(.038)(.083)
Teachers who are Black (%) .07*.29*** .27*** .11*.14** .19**
(.045)(.054)(.060)(.062)(.064)(.076)
Principal is Hispanic .20 1.10 .06 2.12 1.07 .76
(2.120)(1.123)(2.368)(2.190)(1.270)(2.682)
Principal is Black .98 3.76** 2.78 .03 .08 .39
(1.571)(1.476)(1.887)(1.920)(1.494)(2.378)
Principal is Hispanic #%of
teachers who are Hispanic –– – .10 .01 .05
(.083)(.038)(.099)
Principal is Black #%of
teachers who are Black –– –
.06 .24*** .15*
(.059)(.062)(.086)
Year p2012 .01 .03 .29*** .01 .03 .28***
(.076)(.077)(.101)(.076)(.077)(.102)
Constant 14.93 78.69 687.73*** 13.40 76.06 684.15***
(153.071)(155.812)(203.145)(152.833)(154.494)(203.272)
Observations 1,830 1,660 2,050 1,830 1,660 2,050
Adjusted R
2
.610 .709 .587 .610 .714 .588
Note.Standard errors clustered at the district level are shown in parentheses. Models include state xed effects. Controls
not shown include percent of students who are gifted, percent of students who are Black, percent of students who are Hispanic,
percent of students eligible for free and reduced-price lunch, locale type, school size, district size, and per-pupil expenditures. Samples
limited to schools containing between 1% and 99% of students in each DV group. Reported sample sizes rounded to the nearest 10 per
NCES nondisclosure rules.
*p!.10.
** p!.05.
*** p!.01.
414 the elementary school journal march 2017
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
The results from the model for Black students as the dependent variable are
shown in column 5. For Black students, there is evidence of an important interac-
tion. The coefcient on the interaction of Black principal with Black teachers is
quite large (bp0.24,p!.01), suggesting that the association between the propor-
tion of Black teachers in the school and the assignment of Black students to gifted
programs is magnied in the presence of a Black principal. Figure 7illustrates this
Figure 6. Hispanic principals, percent of Hispanic teachers, and predicted percent of gifted stu-
dents who are Hispanic.
Figure 7. Black principals, percent of Black teachers, and predicted percent of gifted students
who are Black.
teacher and principal diversity 415
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
result by plotting the predictive margins. The dotted line represents the predicted
percent of Black gifted students across increasing percentages of Black teachers in
schools with a Black principal, and the solid line represents the same relationship
for schools with a non-Black principal. The light gray lines represent the 95% con-
dence intervals. At low percentages of Black teachers in a school, the predicted
gifted composition for Black students is similar for schools with Black and non-
Black principals. The slopes of the two lines are different, however, so that increas-
ing proportions of Black teachers have a larger positive effect in schools with Black
principals. When the percentage of teachers who are Black is 30%, schools with
Black principals have approximately 20% of gifted students who are Black, com-
pared to only 13% under non-Black principals. When the percentage of teachers
who are Black is 80%, the difference is even larger: 40% of gifted students are pre-
dicted to be Black under Black principals, compared to only 20% under non-Black
principals.
Discussion and Conclusions
Traditionally disadvantaged students who often are excluded from gifted programs
perhaps benet most academically from receiving gifted services (Card & Giuliano,
2014), so identifying why some schools are more likely than others to provide stu-
dents of color with these services is an important endeavor for educational re-
search. Consistent with prior research outside education (e.g., Rocha & Hawes,
2009), our analysis shows that descriptive representation among an elementary
schools faculty and leadership is associated with greater access to gifted programs
for Black and Hispanic students. In a large, national data set spanning two time
points, larger percentages of Black teachers in the school correlate to increased
gifted representation among Black students. The association between Hispanic
teacher representation and Hispanic student presence in the schools gifted pro-
gram is of an almost identical magnitude.
At the same time, gains for Blacks and Hispanics in schools with more diverse
teaching faculties appear to come at the expense of White students, though prior
research suggests that this trade-off only occurs in schools where non-White stu-
dents are substantially underrepresented (Nicholson-Crotty et al., 2011). However,
unlike studies showing evidence of minority group competition in zero-sum con-
texts (e.g., Meier, McClain, Polinard, & Wrinkle, 2004), we do not nd that Black
or Hispanic representation at the teacher level is associated with a decrease (or in-
crease) in gifted placement for students of the other group. We also uncover evi-
dence of nonlinearities; a critical mass of racial/ethnic minority teachersin the
range of 20%to30%may be necessary before descriptive representation trans-
lates into differences in outcomes for students from the same racial/ethnic back-
ground in assignment to gifted programs.
Our results also demonstrate an association between principal race and the
composition of gifted programs. Schools with Black principals have signicantly
higher gifted representation among Black students. The presence of a Black prin-
cipal is associated with a 3.8% increase in Black representation in the schools gifted
program, equivalent to the gain associated with increasing the schools cadre of
Black teachers by 13%. Associations between teacher representation and assign-
ment of Black students to gifted programs are stronger in these schools as well.
416 the elementary school journal march 2017
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
However, we uncovered no parallel results for Hispanic principals and students,
though we note that only about 5% of the schools in our sample were led by a His-
panic principal, potentially limiting the power to distinguish such relationships.
These results point toward a greater need to understand the implications of
teacher workforce diversity in American education, particularly in light of shifting
demographics of the U.S. student population. Among all elementary schools sam-
pled by SASS, for example, the average schools Hispanic student population grew
by approximately 4percentage points between the 20032004 and 20112012 ad-
ministrations. Yet over that same time period, teacher and principal diversity
changed much more slowly, with the fraction of Hispanic teachers in the average
elementary school increasing by less than 1%, and the percentage of Hispanic prin-
cipals actually slightly decreasing. In other words, the different pace of these trends
suggests that, for Hispanic students, descriptive representation in the educational
workforce is, in fact, declining. An implication of this study is that failure of the
public school system to recruit Hispanic teachers and principals at increasing rates
may have consequences for the educational services provided to Hispanic students,
at least in the area of gifted programs.
Of course, diversication of the educator workforce is not the only means for
ensuring greater access to gifted services for students of color. Universal screening
and the use of multiple measures of giftedness are examples of strategies that can
help increase equity in access to gifted programs (Ford, 1998). Such strategies are
unlikely to eliminate the role of teacher discretion in the gifted assignment, eval-
uation, and placement processes, however, which means that training and profes-
sional development aimed at breaking the connection between teacher race/ethnic-
ity and differential access for students by, for example, helping teachers recognize
giftedness among students from diverse cultural backgrounds, may be necessary for
combating gifted underrepresentation among students of color.
This last point is speculative, however, highlighting the need for additional re-
search to understand why teacher and principal race are associated with differen-
tial assignment. Certainly different capacities for teachers to identify giftedness
among students of the same race or ethnicity is a potential mechanism, but there
are many others, including different assignment practices employed by schools
with more diverse teachers, role-modeling effects that elicit greater evidence of
giftedness among students of color, and greater propensities for parents to engage
with the school around gifted identication and evaluation when connected to
teachers or principals of the same demographic background (Grissom et al.,
2015; Lim, 2006). Moreover, we cannot be sure whether teacher or principal race
becomes a salient characteristic during the identication, referral, evaluation, or
placement stage, or some combination. Future research employing more detailed
data on gifted placement processes within and across schools can further elucidate
the empirical linkage to which this study draws attention.
The study faces limitations beyond the depth of the data. Most importantly, the
analysis relies on regression methods that do not warrant causal conclusions about
the relationship between educator demographic characteristics and the represen-
tation of students of color in gifted programs. We cannot be sure that increasing
the numbers of Black teachers in a school, for example, will affect the schools
gifted population or over what time frame, only that elementary schools with higher
percentages of Black teachers have higher rates of gifted participation among Black
teacher and principal diversity 417
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
students, conditional on other variables in our models. These results could be driven
by other factors we cannot observe. For example, schools in more progressive dis-
tricts may place a priority on ensuring diversity both in teaching and in student pro-
grams, or schools that emphasize equity among students may have an easier time
attracting teachers and principals of color. Although a large body of research on rep-
resentation suggests that educator diversity can lead to differential schooling out-
comes for students of color, and quasi-experimental research examining other out-
comes nds evidence consistent with this idea (e.g., Dee, 2005), additional research is
necessary to substantiate a causal relationship.
Even in the absence of a causal linkage, the patterns we document in this article
should raise concerns among advocates for equity, in gifted services and beyond. A
correlation between the racial or ethnic composition of a schools faculty and the
racial and ethnic composition of its gifted program suggests that a childs access to
gifted services is a function of a school characteristic he or she does not control and
that bears little apparent relationship to whether or not the child is indeed gifted.
This study lays a foundation for education researchers to dig more into this impor-
tant empirical connection and for policymakers to begin to consider steps that
might be taken to ensure that gifted students receive gifted services regardless of
such school contextual variables.
Appendix
Table A1. Results from OLS Regression Testing Year Interactions with Teacher
Variables, DV: Percent of Gifted Students from Each Race Group
(1)
Hispanic
Students
(2)
Black
Students
(3)
White
Students
Year p2012 .19 .89 1.85**
(.626)(.607)(.903)
Teachers who are Hispanic (%) .29*** .04 .33***
(.074)(.042)(.083)
Teachers who are Hispanic #year p2012 (%) .03 .05 .10
(.076)(.041)(.089 )
Teachers who are Black (%) .06 .34*** .29***
(.050)(.059)(.067)
Teachers who are Black #year p2012 (%) .02 .05 .02
(.043)(.052)(.065)
Students who are gifted (%) .09*** .03 .07
(.029)(.025)(.045)
Students who are Hispanic (%) .73*** .00 .56***
(.040)(.027)(.047)
Students who are Black (%) .03 .69*** .68***
(.035)(.043)(.043)
Students who are eligible for FRPL (%) .00 .03 .08***
(.019)(.017)(.026)
Suburb .83 .09 1.24
(.818)(.837)(1.159 )
Small town 3.19*** .91 4.23***
(1.161)(1.018)(1.513)
Rural .89 1.00 1.69
(.913)(.983 )(1.245)
418 the elementary school journal march 2017
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
Table A1.(Continued)
(1)
Hispanic
Students
(2)
Black
Students
(3)
White
Students
School size (in 100s) .04 .21 .23
(.162)(.156)(.202)
District size (logged) .09 .64*1.40***
(.315)(.377)(.473)
Per-pupil expenditures (in $1,000s)
a
.07 .15 .41**
(.120)(.147)(.206)
Constant 1.11 12.01*** 115.91***
(3.628)(4.020)(5.124)
Observations 1,830 1,660 2,050
Adjusted R
2
.609 .707 .587
Note.Standard errors clustered at the district level are shown in parentheses. Models include state xed effects. Samples
limited to schools containing between 1% and 99% of students in each DV group. Reported sample sizes rounded to the nearest 10 per
NCES nondisclosure rules.
a
Per-pupil expenditures are in 2012 constant dollars.
*p!.10.
** p!.05.
*** p!.01.
Table A2. Results from OLS Regression Testing Year Interactions with Teacher and Principal
Variables, DV: Percent of Gifted Students from Each Race Group
(1)
Hispanic
Students
(2)
Black
Students
(3)
White
Students
Year p2012 .22 .90 1.76*
(.626)(.591)(.903)
Teachers who are Hispanic (%) .28*** .00 .31***
(.076)(.044)(.087)
Teachers who are Hispanic #year p2012 (%) .08 .01 .15
(.087)(.046)(.095)
Teachers who are Black (%) .09 .30*** .29***
(.058)(.062 )(.073)
Teachers who are Black #year p2012 (%) .04 .04 .04
(.055)(.064)(.084)
Principal is Hispanic 2.26 2.79 2.46
(2.959)(1.750)(2.880)
Principal is Hispanic #year p2012 5.98 4.00*6.24
(3.953)(2.105)(4.276)
Principal is Black (%) 3.04 4.35** .68
(2.359)(2.096)(2.577)
Principal is Black #year p2012 4.59 1.48 4.91
(3.195)(2.756)(3.754)
Constant .77 11.01*** 115.07***
(3.617)(3.914)(5.158)
Observations 1,830 1,660 2,050
Adjusted R
2
.611 .709 .588
Note.Standard errors clustered at the district level are shown in parentheses. Models include state xed effects. Controls
not shown include percent of students who are gifted, percent of students who are Black, percent of students who are Hispanic,
percent of students eligible for free and reduced-price lunch, locale type, school size, district size, and per-pupil expenditures. Samples
limited to schools containing between 1% and 99% of students in each DV group. Reported sample sizes rounded to the nearest
10 per NCES nondisclosure rules.
*p!.10.
** p!.05.
*** p!.01.
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
Notes
Jason A. Grissom is associate professor of public policy and education, Luis A. Rodriguez is a
doctoral candidate, and Emily C. Kern is a doctoral candidate, all in the Peabody College of Ed-
ucation and Human Development, Vanderbilt University, in Nashville, Tennessee.
1. Authorscalculations are from the 20092010 Civil Rights Data Collection, which can be
accessed at http://ocrdata.ed.gov/StateNationalEstimations/Projections_2009_10.
2. This representative bureaucracy trade-off has been an issue of debate in the literature.
For plausibly zero-sum outcomes such as gifted placement, evidence suggests that the improve-
ments in outcomes for minority client groups linked to minority representation indeed come at
the expense of nonminority outcomes (e.g., Nicholson-Crotty et al., 2011; Rocha & Hawes, 2009).
For outcomes that are not necessarily zero-sum, such as student achievement, there is evidence
that outcomes for both racial/ethnic minority and nonminority students are higher with a more
representative workforce (Meier et al., 1999; though see Nielsen & Wolf, 2001).
3. Other studies in the representation literature have used alternative indices, such as the risk
index or the odds, to incorporate the racial/ethnic composition of the student body as a whole
directly in the dependent variable (e.g., Meier & Stewart, 1992; Rocha & Hawes, 2009). We in-
stead use the more straightforward composition measure and control for the racial/ethnic com-
position of the student body in the regression models. See Skiba et al. (2008) for a more detailed
discussion on measurement issues in disproportionality research.
4. Note that these percentages are not weighted by the number of students in the school and
thus are not the same as the percentages in the gifted or total student population.
5. CCD uses locale codes as dened by the National Center for Education Statistics (NCES).
In 2005 and 2006, NCES supported work by the Census Bureau to redesign the original locale
codes in light of changes to the U.S. population and the denition of key geographic concepts.
As a result, locale codes range from an 8-value system to a more detailed 12-value system, with
some schools designated to different values across both locale code schemes. We chose to col-
lapse locale codes across both waves into larger categories representing city, suburban, small
town, and rural locales. Where possible, we use the 2004 locale codes for all schools within
our sample and 2012 locale codes for schools that do not appear in the 2004 CCD; however, re-
gression results are not sensitive to alternate methods of construction for our locale-type control.
6. We interacted the year 2012 with both the teacher percentage (Hispanic and Black) vari-
ables and the principal (Hispanic and Black) variables to explore whether the relationship be-
tween percentage of teachers of a minority racial/ethnic group and percentage of gifted students
who share the same racial/ethnic group remains relatively constant across the 2years, control-
ling for district and school factors. The results for these separate regressions are shown in the
Appendix in Table A1(teachers only with covariates) and Table A2(teachers and principals with
covariates). We used an F-test to examine the joint statistical signicance of these interaction
terms. The null hypothesis that all were 0could generally not be rejected for either set of regres-
sions at conventional levels (with the exception of the White student model that included both
teacher and principal interactions, which was marginally signicant (pp.09). Overall, these
results point to little evidence that the associations we examine differ between 2004 and 2012
(see also Figs. 2and 3). Therefore, we chose to proceed with a pooled sample for our analyses
and include year xed effects in our main models.
References
Boser, U. (2014). Teacher diversity revisited: A new state-by-state analysis. Washington, DC: Cen-
ter for American Progress.
Card, D., & Giuliano, L. (2014). Does gifted education work? For which students? (NBER Work-
ing Paper No. 20453). Cambridge, MA: National Bureau of Economic Research.
Dee, T. S. (2005). A teacher like me: Does race, ethnicity, or gender matter? American Economic
Review,95(2), 158165.
420 the elementary school journal march 2017
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
Elhoweris, H., Mutua, K., Alsheikh, N., & Holloway, P. (2005). Effect of childrens ethnicity on
teachersreferral and recommendation decisions in gifted and talented programs. Remedial
and Special Education,26(1), 2531.
Ford, D. Y. (1998). The underrepresentation of minority students in gifted education: Problems
and promises in recruitment and retention. Journal of Special Education,32(1), 414.
Ford, D. Y., Grantham, T. C., & Whiting, G. W. (2008). Culturally and linguistically diverse stu-
dents in gifted education: Recruitment and retention issues. Exceptional Children,74(3), 289
306.
Fraga, L. R., Meier, K. J., & England, R. E. (1986). Hispanic Americans and educational policy:
Limits to equal access. Journal of Politics,48(4), 850876.
Grissom, J. A., & Keiser, L. R. (2011). A supervisor like me: Race, representation, and the satis-
faction and turnover decisions of public sector employees. Journal of Policy Analysis and
Management,30(3), 557580.
Grissom, J. A., Kern, E. C., & Rodriguez, L. A. (2015). The representative bureaucracyin edu-
cation: Educator workforce diversity, policy outputs, and outcomes for disadvantaged stu-
dents. Educational Researcher,44(3), 185192.
Grissom, J. A., Nicholson-Crotty, J., & Nicholson-Crotty, S. (2009). Race, region, and represen-
tative bureaucracy. Public Administration Review,69(5), 911919.
Grissom, J. A., & Redding, C. (2016). Discretion and disproportionality: Explaining the under-
representation of high-achieving students of color in gifted programs. AERA Open,2(1), 125.
Hargrove, B. H., & Seay, S. E. (2011). School teacher perceptions of barriers that limit the partic-
ipation of African American males in public school gifted programs. Journal for the Educa-
tion of the Gifted,34(3), 434467.
Harris, J. J., Brown, E. L., Ford, D. Y., & Richardson, J. W. (2004). African Americans and mul-
ticultural education: A proposed remedy for disproportionate special education placement
and underinclusion in gifted education. Education and Urban Society,36(3), 304341.
Henderson, L. J. (1979). Administrative advocacy: Black administrators in urban bureaucracy.
Palo Alto, CA: R&E Research Associates.
Hindera, J. J. (1993). Representative bureaucracy: Further evidence of active representation in the
EEOC district ofces. Journal of Public Administration Research and Theory,3(4), 415429.
Hindera, J. J., & Young, C. D. (1998). Representative bureaucracy: The theoretical implications of
statistical interaction. Political Research Quarterly,51(3), 655671.
Joseph, L. M., & Ford, D. Y. (2006). Nondiscriminatory assessment: Considerations for gifted
education. Gifted Child Quarterly,50(1), 4251.
Kanter, R. (1977). Men and women of the corporation. New York: Basic.
Kennedy, B. (2014). Unraveling representative bureaucracy: A systematic analysis of the litera-
ture. Administration & Society,46(4), 395421.
Lim, H. (2006). Representative bureaucracy: Rethinking substantive effects and active represen-
tation. Public Administration Review,66(2), 193204.
Lipsky, M. (1980). Street-level bureaucracy: Dilemmas of the individual in public services. New
York: Russell Sage.
Matthews, M. S., & Shaunessy, E. (2010). Putting standards into practice: Evaluating the utility of
the NAGC Pre-KGrade 12 Gifted Program standards. Gifted Child Quarterly,54(3), 159167.
McBee, M. T. (2006). A descriptive analysis of referral sources for gifted identication screening
by race and socioeconomic status. Journal of Secondary Gifted Education,17(2), 103111.
McBee, M. T., Shaunessy, E., & Matthews, M. S. (2012). Policy matters: An analysis of district-
level efforts to increase the identication of underrepresented learners. Journal of Advanced
Academics,23(4), 326344.
Meier, K. J. (1993). Representative bureaucracy: A theoretical and empirical exposition. In J.
Perry (Ed.), Research in public administration (pp. 135). Greenwich, CT: JAI.
Meier, K. J., McClain, P. D., Polinard, J. L., & Wrinkle, R. D. (2004). Divided or together? Con-
ict and cooperation between African Americans and Latinos. Political Research Quarterly,
57(3), 399409.
Meier, K. J., & Stewart, J. (1992). The impact of representative bureaucracies: Educational sys-
tems and public policies. American Review of Public Administration,22(3), 151171.
teacher and principal diversity 421
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
Meier, K. J., Wrinkle, R. D., & Polinard, J. L. (1999). Representative bureaucracy and distribu-
tional equity: Addressing the hard question. Journal of Politics,61(4), 10251039.
Mosher, F. (1968). Democracy and the public service. New York: Oxford University Press.
National Research Council. (2002). Minority students in special and gifted education: Committee
on Minority Representation in Special Education, M. Suzanne Donovan & C. T. Cross (Eds.),
Division of Behavioral and Social Sciences and Education. Washington, DC: National Acad-
emy Press.
Nicholson-Crotty, J., Grissom, J. A., & Nicholson-Crotty, S. (2011). Bureaucratic representation,
distributional equity, and democratic values in the administration of public programs. Jour-
nal of Politics,73(2), 582596.
Nicholson-Crotty, S., Grissom, J. A., Nicholson-Crotty, J., & Redding, C. (2016). Disentangling
the causal mechanisms of representative bureaucracy: Evidence from assignment of students
to gifted programs. Journal of Public Administration Research and Theory,26(4), 745757.
Nielsen, L. B., & Wolf, P. J. (2001). Representative bureaucracy and harder questions: A response
to Meier, Wrinkle, and Polinard. Journal of Politics,63(2), 598615.
Pitts, D. W. (2005). Diversity, representation, and performance: Evidence about race and ethnic-
ity in public organizations. Journal of Public Administration Research and Theory,15(4), 615
631.
Rocha, R. R., & Hawes, D. P. (2009). Racial diversity, representative bureaucracy, and equity in
multiracial school districts. Social Science Quarterly,90(2), 326344.
Skiba, R. J., Simmons, A. B., Ritter, S., Gibb, A. C., Rausch, M. K., Cuadrado, J., & Chung, C. G.
(2008). Achieving equity in special education: History, status, and current challenges. Excep-
tional Children,74(3), 264288.
Thompson, F. (1976). Minority groups in public bureaucracies: Are passive and active represen-
tation linked? Administration & Society,8,201226.
Weiher, G. R. (2000). Minority student achievement: Passive representation and social context
in schools. Journal of Politics,62(3), 886895.
422 the elementary school journal march 2017
This content downloaded from 140.234.253.009 on March 10, 2017 01:58:42 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
CopyrightofElementarySchoolJournalisthepropertyofUniversityofChicagoPressandits
contentmaynotbecopiedoremailedtomultiplesitesorpostedtoalistservwithoutthe
copyrightholder'sexpresswrittenpermission.However,usersmayprint,download,oremail
articlesforindividualuse.
... First, what constitutes a critical mass varies among these studies. Atkins and Wilkins (2013) found that a critical mass of 20% Black teachers was necessary to reduce the Black teen pregnancy rates, but Atkins et al. (2014) showed that 40% Latino teachers were needed for Latino students' college expectations to increase (see also Grissom et al., 2017). Second, the findings on critical mass differ at the street level and managerial level. ...
... Andrews et al. (2014) do not find critical mass effects at upper-managerial levels. For street-level representation, some studies show the need for a critical mass (Atkins & Wilkins, 2013;Grissom et al., 2017;Keiser et al., 2002), but many others find no critical mass is necessary (Meier, 1993;Meier & Rutherford, 2017). These differences imply that the need for a critical mass might vary at different levels of the organization and in some cases not be needed at all (see Andrews & Miller, 2013;Fernandez, Samuel, and Lee, 2018). ...
... Second, other studies separate organizations into high representation and low representation groups (Keiser et al., 2002;Nicholson-Crotty et al., 2011) and estimate representation in both groups expecting the representation to appear only in the high representation group. 3 Third, some studies include a set of dummy variables representing different levels of representation in the same equation as the way to determine the level of critical mass (Atkins & Wilkins, 2013;Grissom et al., 2017). Differences in results could be from differences in methods (Meier, 2019, pp. ...
Article
Full-text available
Studies of representative bureaucracy frequently reference the need for a critical mass before an underrepresented group can influence policy outputs or outcomes, but the empirical literature is modest and presents mixed findings. This article presents a theoretical exploration to illustrate how critical mass can link individual behavior to organizational contexts. By examining both active representation and symbolic representation at both the individual and organizational level and specifying the different microprocesses of active representation, a series of testable hypotheses are presented. The logic suggests that the concept of critical mass might also be a useful contextual variable to examine other aspects of administrative behavior. Based on the theory, two empirical illustrations from China are presented.
... Research by Sun, drawing on multiple years of quantitative data finds that black teachers were more likely to be retained when they were exposed to stronger school leadership and higher-quality professional development (Sun, 2018). Grissom, Rodriguez, and Kern (2017) investigated advanced curricular course offerings and enrollments, finding that a critical mass of teachers of color to be associated with increases in gifted and talented enrollment, especially for students of color. Thus, there is evidence that a greater proportion of teachers of color within a campus is influenced by teacher colleague and leader characteristics, with the potential to impact important student outcomes. ...
... Moreover, increasing the diversity of the teacher workforce has been shown to benefit all students, as minority teachers bring a varied set of experiences and perspectives beyond those of homogenous teaching populations (Irizarry, 2007). This qualitative study used a perspective that posits three benefits of teacher diversity: Students with teachers of the same race are 1) less likely to be suspended , 2) more likely to be referred to gifted programs (Grissom, Rodriguez, & Kern, 2017), and 3) more likely to complete high school and go to college (Gershenson, Hart, Lindsay, & Papageorge, 2017). ...
Article
Background: Student retention is the most critical challenge facing most traditional 4- and 2-year institutions. Challenges related to student belonging and retention are among many that colleges must contend with daily. Once students arrive, they are met with an academic environment that may greatly differ from their high school and home experiences. Discussion: There is evidence that classroom success is influenced by faculty tone setting and the overall campus climate. Other studies have shown that creating a sense of belonging on college campuses is essential. Classroom success can be moderated by relationships with faculty, especially with faculty of color and other minoritized groups. Aims: Given the predicted shortages of minority dental healthcare providers in the United States and the expanding diversity of the general population, it is important to recruit and retain an ethnically and culturally diverse allied dental workforce. Conclusions: The objectives of this research were to identify strategies to recruit underrepresented (underrepresented minority [URM]) students into health professions, understand the significance of underrepresented (URM) faculty's influence on the number of URM student recruitment, retention, and graduation rates, and how to implement best practices for overall minority student success.
Article
In this policy forum commentary, we call for improved national and state-level data collection and access relevant to the principal pipeline. We focus specifically on how access to quality data can inform, has informed, and is critical to policy and practice across three segments of the principal pipeline—principal preparation, licensure, and labor markets—as well as the essential issue of diversity itself and across each pipeline segment. We also outline and explain the types of data collection, storage, and access that would enable more informed and relevant principal pipeline research, policy, and preparation. We then call for action at the state and the federal levels, including reauthorization of the Higher Education Act to expand Title II educator data collection to include principal preparation programs as it does teacher preparation programs. Finally, we conclude with cautions against potential data misuse.
Article
The literature on representative bureaucracy posits that increased representation at the managerial level leads to improved outcomes for minority clients. These managers, however, must work within organizational constraints and during times of environmental turbulence to sustain organizational performance. We forward the theory that contextual factors, such as environmental turbulence, could moderate the effects of representation on organizational performance. Utilizing a Texas school-level dataset of K-12 education from 2011, we examine how the race of the school principal influences student standardized test performance in a time of widespread financial resource cuts. Our findings suggest that same-race school principal representation improves the academic performance of both African American and Latino students, but the positive effects diminish as budgetary cuts become more widespread in the school. This means that environmental turbulence can decrease the impact of representation.
Article
Extensive research exists on students who are the first in their families to pursue post-secondary education, i.e. first-generation college students. But what about first-generation students who pursue careers in education? This study highlights the experiences of first-generation college students, then graduate students, turned teachers and may shed light on how teacher preparation programs could improve persistence among first-generation college and graduate students and support this population as they embark on careers in K-12 settings. First-generation college students who become teachers frequently share demographic characteristics with students in high-need schools, i.e., many are of low socioeconomic status and/or are people of color. Thus, these teachers may help to fill a critical need as they may be more likely to persist as teachers in high-need schools, which are typically hardest to staff and have high turnover.
Article
A large body of research shows that clients of government services benefit from the presence of bureaucrats with whom they share race or ethnicity. These benefits arise from active or symbolic representation, which scholars argue are grounded in the shared backgrounds, language, and values that race and ethnicity proxy. We suggest that these shared connections are likely to be even more salient for clients and bureaucrats who share not just the same ethnicity but the same country of origin, and we look for evidence of representation based on country of origin in the context of public schools. Leveraging administrative and survey data from Miami-Dade County Public Schools, the fourth-largest school district in the United States, we employ regression models with school-by-year fixed effects to test for differences in test scores for students taught by a teacher with the same country of origin relative to similar students taught by other-origin teachers in the same school in the same year. We find that immigrant students with origin-matched teachers score modestly higher than their non-matched peers in both math and reading. These increases are most apparent among low-income students and those who are English learners. Patterns vary by immigrant students’ origin country.
Article
Gifted education programs identify students with exceptional ability or potential and provide specialized educational interventions to develop general academic and domain-specific talent. Research and legal actions have found access to gifted education riddled with inequity. Indicators of inequity in selection practices based on disproportionality have been developed in the context of employment and gifted education; however, these indicators have not been compared side by side. We evaluate combinations of disproportionality-based metrics and thresholds on their ability to flag for inequity those selection practices that are indeed inequitable while not flagging for inequity those that are not. Specifically, the risk ratio and a generalized form of the Inequity Allowance Formula/Inequity Index, called here the inequity allowance ratio, are evaluated with thresholds between 0.1 and 0.9. We evaluate false-positive rates and true-positive rates when equity or inequity in selection practices exists. Furthermore, we explore the effect of program selectivity on the performance of these indicators. Ultimately, we find that recommendations regarding which disproportionality-based metric and threshold to use depends on factors that we cannot know, including the distribution of qualification of students and the quantifiable selectivity of a program. Therefore, the use of a disproportionality-based indicator when evaluating programs for potential inequity is not recommended.
Article
Full-text available
Students of color are underrepresented in gifted programs relative to White students, but the reasons for this underrepresentation are poorly understood. We investigate the predictors of gifted assignment using nationally representative, longitudinal data on elementary students. We document that even among students with high standardized test scores, Black students are less likely to be assigned to gifted services in both math and reading, a pattern that persists when controlling for other background factors, such as health and socioeconomic status, and characteristics of classrooms and schools. We then investigate the role of teacher discretion, leveraging research from political science suggesting that clients of government services from traditionally underrepresented groups benefit from diversity in the providers of those services, including teachers. Even after conditioning on test scores and other factors, Black students indeed are referred to gifted programs, particularly in reading, at significantly lower rates when taught by non-Black teachers, a concerning result given the relatively low incidence of assignment to own-race teachers among Black students.
Article
Full-text available
Among the most-longstanding and intransigent issues in the field, the disproportionate representation of minority students in special education programs has its roots in a long history of educational segregation and discrimination. Although national estimates of disproportionality have been consistent over time, state and local estimates may show varying patterns of disproportionality. A number of factors may contribute to disproportionality, including test bias, poverty, special education processes, inequity in general education, issues of behavior management, and cultural mismatch/cultural reproduction. This article provides a report on the history, measurement, status, and factors contributing to disproportionate representation in special education, and offers recommendations based on an understanding of racial and ethnic disparities in special education as a multiply determined phenomenon.
Article
Full-text available
Bureaucratic representation—the idea that a governmental organization is better situated to serve its clients when its employee composition reflects that of its client population—has received considerable scholarly attention in the study of public institutions in the fields of political science and public administration. In a wide variety of settings, this research has demonstrated important connections between the racial, ethnic, and gender composition of the public sector workforce and how different groups—particularly traditionally underserved groups—interact with street-level bureaucrats and benefit from public services. Although scholars in those fields long ago recognized that the public school system is a large bureaucracy with diverse street-level bureaucrats (teachers) and clients (students and parents) and thus began studying bureaucratic representation in the context of schools, the concept and the causal mechanisms it hypothesizes remain largely unfamiliar to education researchers. This article synthesizes the main ideas from the bureaucratic representation literature and demonstrates their applicability to schooling outcomes—including discipline, gifted assignment, special education, and student achievement—with the goal of opening up new avenues for education research into the mechanisms linking demographic similarity among educators and students to schooling outputs and outcomes.
Article
Full-text available
The field of gifted education has faced criticism about the underrepresentation of African American, Hispanic/Latino, and American Indian students who are culturally and linguistically diverse (CLD) in its programs. This article proposes that efforts targeting both recruitment and retention barriers are essential to remedying this disparity. Educators' deficit thinking about CLD students underlies both areas (recruitment and retention) and contributes to underrepresentation in significant, meaningful ways. The authors examine factors hindering the recruitment and retention of CLD students in gifted education, attending in particular to definitions and theories, testing, and referral issues, and offer recommendations for improving the representation of CLD students in gifted education.
Article
Full-text available
Research on representative bureaucracy has failed to deal with whether or not representative bureaucracies produce minority gains at the expense of nonminorities. Using a pooled time-series analysis of 350 school districts over six years, this study examines the relationship between representative bureaucracy and organizational outputs for minorities and nonminorities. Far from finding that representative bureaucracy produces minority gains at the expense of nonminorities, this study finds both minority and nonminority students perform better in the presence of a representative bureaucracy. This finding suggests an alternative hypothesis to guide research: that representative bureaucracies are more effective than their nonrepresentative counterparts.
Article
This research extends our knowledge of the relationship between passive representative bureaucracy and active representative bureaucracy (Mosher 1968). We utilize data from the United States Equal Employment Opportu nity Commission (EEOC) to posit and test a theory of situational thresholds. Our theory is premised on the argument that Meier's (1993b) construct, criti cal mass, constitutes a particular example of a more general concept, situ ational threshold (Campbell 1963). We present evidence that three situational thresholds exist in the EEOC bureaucratic environment which determine the nature of the relationship between passive representativeness and active rep resentation : when a critical mass is reached; when a group constitutes a plu rality ; and when a group constitutes a majority In toto, these three situational thresholds delineate four distinct social environments in which active repre sentation transpires. Our findings strongly suggest that the relationship be tween passive representation and active representation is fundamentally different for each type of bureaucratic environment.
Article
Scholars have suggested that the benefits of representative bureaucracy arise from bureaucrats acting in the interests of clients who share their characteristics, increased diversity encouraging even nonminority bureaucrats work to further the interests of minority clients, and/or the actions of clients that are more responsive to bureaucrats that share their characteristics. Despite decades of research, the literature has been unable to empirically disentangle these mechanisms, primarily because the vast majority of studies examine only organization-level data, and, at the aggregate level, they all produce identical findings. In contrast, this study makes use of data that allows us to observe the behavior of individual clients and bureaucrats, as well as the aggregate characteristics of the organizations in which they interact. Specifically, we make use of student-level data to predict differences in the probability that an elementary student is referred to gifted services by race. Our results suggest that black students are more likely to be referred to gifted services when taught by a black teacher but that increased presence of black teachers in the school other than the classroom teacher has little effect. We find some evidence that the classroom teacher effect is partially driven by teachers’ more positive views of own-race students. Our results do not suggest, however, that the positive impact of teacher-student race congruence on gifted assignment can be explained by differences in student test score performance or increased parental interaction with the teacher.
Article
Scholars have provided great theoretical insight and empirical understanding to the concept of representative bureaucracy, documenting the changing makeup of the civil service and demonstrating the importance of representative bureaucracy toward democratic governance. Yet, important questions remain. What specifically does bureaucratic representation mean? Is descriptive representation necessary for policy representation? What characteristics are important? How do we measure representation? Finally, are there any negative effects of representative bureaucracy? This article provides an overview and analysis of the literature focusing on how scholars define and measure representative bureaucracy. Efforts are made to emphasize achievements and highlight areas that need attention.
Article
Work on bureaucratic representation suggests that minority citizens benefit when the programs that serve them are administered by bureaucrats with similar characteristics. This literature has not sufficiently dealt with the long-standing concern that minority benefits may come at the expense of citizens from other groups, which some critics argue makes representative bureaucracy irreconcilable with democratic values. This article suggests distributional equity as a potential moderator of bureaucratic representation and as a potential source of reconciliation. It tests for the effects of representation under different distributional conditions in a policy area in which outcomes approach a zero-sum game. Analyses of a nationally representative sample of public organizations find a relationship between bureaucratic representation and citizen outcomes only in those instances where program benefits are being inequitably distributed to the relevant group. The article concludes with a discussion of the significance of these findings for the democratic legitimacy of representative bureaucracy.
Article
Policies delegating control of educational policy to the local level are widespread, yet there has been little examination of the effects of such distributed decision making in the area of advanced education programming. We used propensity score matching to examine the effectiveness of locally developed policies for identifying intellectually gifted children identifying themselves as Black or from low-socioeconomic backgrounds across one large U.S. state (Florida) that has a state-level gifted education mandate. Ongoing underrepresentation of traditionally marginalized groups in gifted education was evident, even among districts with policies specifically designed to ameliorate disproportional representation. However, the presence of such a policy reduced the degree of underrepresentation.