Teacher Racial Composition and Exclusion Rates Among Black or African American Students

Abstract

Expulsion and suspension rates among African American students are the highest of all racial groups across elementary, middle, and high schools. This study investigates whether a more racially diverse teaching force could alleviate exclusion rates among Black students. Using administrative data from the universe of K-12 public schools in the state of Wisconsin from 2002-2003 to 2012-2013 and employing a school fixed effects approach, results suggest that increasing the representation of Black teachers by even a single percentage point is associated with lower suspension rates among Black students at the high school level. Across levels of schooling, exclusion rates of White students are unrelated to teacher racial composition.

Keywords

school discipline suspension rate expulsion rate teacher race representative bureaucracy

Introduction

There is growing attention to safety in schools and corresponding concern about school discipline. On-site security measures intensified substantially between 2000 and 2014, when the percentage of public schools that regulated access to school buildings rose from 75% to 93%, use of security camera monitoring rose from 19% to 75%, and the percentage of schools requiring that faculty and staff wear badges or picture identifications rose from 25% to 68% (U.S. Department of Education, National Center for Education Statistics, 2015). Moreover, schools are increasingly monitoring student behavior through mandatory uniforms, strict dress codes, and drug possession checks (U.S. Department of Education, National Center for Education Statistics, 2015). Such procedures are thought to maintain order and safety in schools, both inside and outside of the classroom.

As it is difficult for effective teaching and learning to occur in disruptive or unsafe spaces, student discipline may be just as important as student achievement (Gregory, Skiba, & Noguera, 2010), if not a precondition. A suspension or expulsion is a common resolution to serious student misconduct (Brown, 2007), including possessing weapons, damaging property, and breaking other school rules. The use of discipline in school has risen somewhat—between 1999 and 2007, expulsion rates increased from 3% to 3.2% and suspension rates increased from 22% to 24.5% among high school students (Aud, KewalRamani, & Frohlich, 2011).

But the rates of expulsion and suspension among Black or African American¹ students in particular are especially high, and have sparked public concern about the distribution and effects of school discipline. Black K-12 students are suspended or expelled about 3 times as often as their White counterparts (U.S. Department of Education Office for Civil Rights, 2014). From 1999 to 2007, suspension rates among Black high school students increased from 37% to 49% yet remained stagnant for Whites (18.2% in 1999 and 17.7% in 2007; Aud et al., 2011). During the same time period, high school expulsion rates increased for Blacks from 6.5% to 10.3% but decreased for Whites from 1.8% to 1.1% (Aud et al., 2011). The discrepancy in growth of school exclusions at the high school level is of concern given that a developing literature argues that excessive use of school exclusions puts students on a school-to-prison pipeline (Payne & Welch, 2010; Skiba, Eckes, & Brown, 2010; Skiba, Michael, Nardo, & Peterson, 2002; Wald & Losen, 2003). This is partly because students repeatedly suspended or expelled become less attached to school, less responsive to school and classroom rules, and less interested in academic achievement (Gregory et al., 2010).

One frequently offered explanation for these disparities is racial discrimination, particularly as evidenced by a mismatch between students and teachers from different racial groups (Monroe, 2005; Skiba et al., 2011). A few studies suggested that teachers give same-race students higher marks on behavior and academic ability (Dee, 2005; Ouazad, 2014), raising the possibility that Black students may be disciplined more frequently because they are often taught by non-Black teachers. This article considers the empirical evidence for this hypothesis by examining whether suspension and expulsion rates among African American students are affected by the number of teachers of color employed by their school. I examine this question using an administrative dataset on all K-12 public schools in Wisconsin from 2002-2003 to 2012-2013. Results from a school fixed effects model suggest that the racial composition of the teaching workforce matters. Specifically, a higher percentage of Black teachers is associated with lower suspension rates among Black high school students. The result is particular to Black teachers rather than racial/ethnic minority teachers overall.

School Factors That Influence Student Exclusions

Many scholars have attributed racial disparities in school discipline to a misinterpretation of student behavior rather than actual wrongdoing. Students of color are penalized more often for subjective or ambiguous offenses such as disobedience and disruption rather than physical or violent conduct (Christle, Nelson, & Jolivette, 2004; Lewis, Butler, Bonner, Fred, & Joubert, 2010; Mendez, 2003; Reynolds et al., 2008; Skiba et al., 2010; Skiba & Peterson, 1999; Skiba et al., 2011; Skiba et al., 2002; Skiba, Shure, & Williams, 2012; Wallace, Goodkind, Wallace, & Bachman, 2008). As schools are embedded in broader society and reflect its social, cultural, and organizational structure, scholars have suggested that the disparity in school discipline could be a reflection of discriminatory institutional practices rather than actual disproportionate misbehavior among African American students (Alexander, 2012; Arum & Velez, 2012; Eitle & Eitle, 2004; Rocque, 2010; Wald & Losen, 2003; Wilson, 2009). Following this theory, unwarranted discipline may be due to cultural mismatch between students and teachers (Evans, 2007; Ferguson, 2001; Ladson-Billings, 1995; Monroe, 2009; Skiba et al., 2002); media depictions or general stereotypes of students based on race, class, and sex (Monroe, 2005; O’Connor, 1999; Skiba et al., 2011; Verdugo, 2002; Yosso, 2005); or adult anxieties or preoccupations with social control (Males, 1999; Noguera, 2003; Pace & Hemmings, 2007).

School discipline results from multiple decisions. According to Arum and Velez (2012), administrators and teachers establish boundaries for the expression of student attitudes, behaviors, and subcultures whereas students set additional standards as acceptable by their peer environment. Prior research suggested that racial disparities in suspension and expulsion originate at the classroom level (Gregory et al., 2010; Skiba et al., 2011), with students being sent to the office by a teacher or other school official (Rocque, 2010). Students spend the majority of the school day inside classrooms where teachers have authority on whether or not to report disciplinary problems. A reported incident may result in a school removal at the discretion of the school principal. Typically, suspensions are handled at the school level whereas expulsions must be approved by the school board after a formal hearing. Although students are expected to be treated under the same disciplinary code of conduct, teachers and school administrators may make subjective decisions regarding student behavior (Reynolds et al., 2008).

Recent studies found relationships between exclusionary practices and school characteristics including percentage of Black enrollment and principal attitudes toward discipline (Skiba et al., 2014). Using 1 year of data from a Midwestern state on all students who received an in-school suspension, out-of-school suspension, or expulsion for some incident, Skiba and colleagues (2014) applied multilevel modeling to evaluate how student- and school-level variables affect the severity of disciplinary response, controlling for type of infraction. They found that the percentage of Black students in a school, rather than the race of the student disciplined, increased the risk of having an infraction result in an out-of-school suspension or expulsion rather than an in-school suspension. Relative to student behavior and demographics, systemic school-level variables—including percentage passing math and English as well as having a principal who favors school expulsion—may better explain the overrepresentation of Black students in school discipline (Skiba et al., 2014).

Suspension rates tend to be higher among both African American and White students when attending majority–minority schools (Gregory, Cornell, & Fan, 2011). An explanation is that the leaders of these schools are more likely to support the use of detentions, automatic suspensions, and criminal justice referrals when responding to student offenses (Payne & Welch, 2010). Punitive practices such as school exclusion are more common than restorative practices such as student conferences, peer mediation, restitution, and community service (Payne & Welch, 2013). Welch and Payne (2010) consistently found support of the racial threat hypothesis, which states that schools with a larger percentage of Black students are more likely to use more severe disciplinary actions as a form of social control. At the same time, applying hierarchical linear models to data on 728 Florida public schools from the 1999-2000 school year, Eitle and Eitle (2004) observed that highly segregated school districts have less White–Black discrepancy in suspension rates, supporting the resegregation hypothesis in which non-White populations do not experience more adverse effects in segregated schools. The results did not support the racial inequality hypothesis in which segregated schools have the most racial imbalance in outcomes (Eitle & Eitle, 2004). In sum, it appears that administrators impose the school disciplinary climate based on the student population, with majority–minority schools relying on punitive practices that lead to reduced racial disparity but higher exclusion rates overall.

Finally, public schools have accountability-related incentives to remove underperforming students, many of whom are economically disadvantaged or students of color. Examining more than 40,000 suspensions from Florida school districts issued over a 4-year period and employing regression techniques with multiple interactions between variables, Figlio (2006) observed that for disciplinary events that occurred around the standardized testing window, schools varied suspension durations according to the achievement propensity of students involved. In nearly 60% of incidents involving two students, schools assigned unique suspension lengths to each student involved. Lower achieving students received longer suspensions, especially for incidents occurring within a high-stakes testing window (Figlio, 2006). The results suggest that schools may attempt to alter the testing population to gain better scores.

Teacher–Student Racial Matching

The literature reviewed above suggests that school disciplinary climate is set by administrators and shaped by teachers who confront student behavior. If schools are racialized spaces, students may be more receptive to same-race teachers. Several studies revealed that teachers may give same-race students more favorable assessments of behavior and academic ability (Beady & Hansell, 1981; Dee, 2005; Ehrenberg, Goldhaber, & Brewer, 1995; Ouazad, 2014). For example, using data from the National Educational Longitudinal Study of 1988, Ehrenberg and colleagues (1995) found that Black high school students received higher evaluations from Black male teachers than from White male teachers on whether the student would go to college, should be recommended for academic honors, and works hard. Using the same dataset, Dee (2005) showed that teacher race and gender also had large effects on teacher evaluations of eighth-grade student behavior. A student taught by a different-race teacher had a 1.22 to 1.36 times greater odds of being evaluated as disruptive, inattentive, and rarely completing homework than a student taught by a same-race teacher, even after controlling for other teacher and classroom variables. A recent study by Lindsay and Hart (2017) found Black elementary students in North Carolina were less likely to receive a detention, suspension, or expulsion when taught by Black teachers. Following individual students over time, they attributed a 12% decrease in the number of students requiring any discipline to student–teacher race congruency. However, one limitation of the study is that it does not differentiate between the three types of exclusionary outcomes, which typically correspond to a range of behaviors.

Although studies have estimated the effect of characteristics among individual teachers on student outcomes, few have evaluated the collective impact of a school’s teacher population (Croninger, Rice, Rathbun, & Nishio, 2007). Teachers with strengths in some areas may be able to support their developing colleagues, leading to a schoolwide improvement (Croninger et al., 2007). Unlike elementary students who receive instruction in multiple subjects from a single teacher, high school students receive instruction across multiple subjects from multiple teachers. After 4 years of high school, each student may have been taught by several dozen teachers. Both proximate and previous teachers could influence student achievement (Monk & King, 1994) and, perhaps, student behavior. Thus, it makes sense to analyze how student outcomes are influenced by the collective teaching force.

Related to individual teacher–student racial pairings and the concept of a collective teaching force is the racial composition of schoolteachers. A review of literature suggested that racial composition is an important feature of organizations, affecting employee interactions as well as performance (Reskin, McBrier, & Kmec, 1999). Representative bureaucracy posits that a government institution better serves its clients when the racial, ethnic, and gender composition of its employees reflects that of its clients. Well known in political science, bureaucratic representation has received some but not much attention in education policy (Grissom, Kern, & Rodriguez, 2015; Meier & Stewart, 1992). Analyses of 1 year of data on 67 districts in Florida (Meier & Stewart, 1992) and 3 years of school-level data from Georgia (Roch, Pitts, & Navarro, 2010) found that a higher proportion of Black teachers in schools was associated with lower suspension rates across all students. A 1-year sample of 3,000 public schools across the nation (Grissom, Nicholson-Crotty, & Nicholson-Crotty, 2009) exhibited a similar result regarding suspension rates among Black students.² The present study improves on the methods of previous studies on bureaucratic representation and school discipline by using a panel of administrative data from an entire state, and disaggregating results by race and school level. Given that the structure of middle and high schools greatly differs from that of elementary schools, it is possible that the impact of teacher composition varies by school level.

I seek to address the following research questions. What is the relationship between teacher racial composition and exclusion rates among Black students in elementary, middle, and high schools, and does it differ for White or all students? Second, if there exist an association between Black disciplinary outcomes and the percentage of Black teachers, is it specific to Black teachers or minority teachers in general?

Data and Descriptive Statistics

This study utilizes administrative data from the universe of public schools in the state of Wisconsin from 2002-2003 to 2012-2013. The 11-year period was chosen based on data availability—2002-2003 was the first year that teacher race could be retrieved from the Wisconsin Department of Public Instruction, and 2012-2013 was the most recent year of data available at the creation of this project.

Wisconsin is an appropriate site to research racial disparities in school discipline. A nationwide study showed that for the 2011-2012 school year, the state had the largest percentage of African American students suspended (34%) and the largest Black/White gap (30 percentage points) at the high school level (Losen et al., 2015). Among elementary schools, Wisconsin had the third largest percentage of African American students suspended at 12.2% and second highest Black/White gap at 11.2 percentage points (Losen et al., 2015).

Table 1 shows that the statewide racial composition of Wisconsin public school students was fairly stable during this time period, with the exception of a gradual increase in the percentage of Hispanic students and decrease in that of White students. On average, Wisconsin’s student population was 76.6% White, 10.2% Black, 7.7% Hispanic, 3.5% Asian, and 1.4% American Indian, with the remaining proportion identified as Other race. Beginning in 2010-2011, Wisconsin expanded its racial categories to include “two or more races” and “native Hawaiian or other Pacific Islander.” About 2% students took on the new classifications, making the last 3 years of descriptive statistics less comparable. I am unable to address any discrepancy in the summary statistics but perform a sensitivity analysis of the main results later in the article.

Table 1.

Demographics of Students Enrolled in Wisconsin Public Schools, Statewide.

School year	% White	% Black	% Hispanic	% Asian	% American Indian
2002-2003	79.53	10.29	5.37	3.35	1.46
2003-2004	78.90	10.46	5.83	3.37	1.45
2004-2005	78.35	10.43	6.30	3.45	1.46
2005-2006	77.84	10.42	6.74	3.56	1.45
2006-2007	77.31	10.45	7.2	3.58	1.46
2007-2008	76.9	10.41	7.62	3.62	1.46
2008-2009	76.42	10.42	8.01	3.67	1.48
2009-2010	76.02	10.37	8.42	3.69	1.49
2010-2011	74.42	9.89	9.27	3.51	1.33
2011-2012	73.76	9.77	9.71	3.53	1.29
2012-2013	73.12	9.7	10.1	3.56	1.26
Average	76.60	10.24	7.69	3.54	1.42

School expulsion rate and suspension rate are the two disciplinary outcomes of interest. Table 2 presents statewide exclusion rates by race across time, and Figure 1 shows corresponding graphs. Expulsions are grave disciplinary consequences. Typically, an expulsion forces a student out of a school for at least the remainder of the school year. Table 2 shows that expulsions were rare, affecting about 0.15% of Wisconsin students per year. School expulsion rate is equivalent to the number of students expelled divided by student enrollment. Black students had the highest expulsion rate at about 0.5%, 5 times that for White students.

Table 2.

Disciplinary Rates by Race, Statewide.

School year	White	Black	Hispanic	Asian	American Indian	All
Expulsion rate (in percent)
2002-2003	0.13	0.32	0.26	0.08	0.27	0.16
2003-2004	0.15	0.39	0.32	0.11	0.39	0.19
2004-2005	0.14	0.50	0.33	0.08	0.44	0.19
2005-2006	0.15	0.61	0.29	0.08	0.32	0.21
2006-2007	0.12	0.51	0.25	0.08	0.29	0.17
2007-2008	0.10	0.52	0.24	0.05	0.26	0.15
2008-2009	0.10	0.43	0.17	0.05	0.40	0.15
2009-2010	0.08	0.58	0.18	0.06	0.14	0.14
2010-2011	0.07	0.60	0.15	0.05	0.16	0.13
2011-2012	0.06	0.50	0.13	0.02	0.25	0.11
2012-2013	0.06	0.45	0.09	0.04	0.14	0.10
Average	0.11	0.49	0.22	0.06	0.28	0.15
Suspension rate (in percent)
2002-2003	3.88	23.23	10.21	2.71	14.96	6.33
2003-2004	4.03	26.05	11.07	3.08	15.09	6.87
2004-2005	3.92	27.42	10.79	2.76	15.86	6.94
2005-2006	3.88	30.48	10.56	2.68	15.00	7.22
2006-2007	3.23	28.91	9.41	2.37	11.36	6.44
2007-2008	3.19	30.95	9.43	2.37	11.39	6.64
2008-2009	3.03	30.62	8.40	2.09	10.52	6.41
2009-2010	2.97	30.12	8.07	2.01	9.28	6.28
2010-2011	2.74	26.41	6.55	1.61	9.62	5.54
2011-2012	2.50	25.57	5.99	1.44	8.69	5.21
2012-2013	2.26	21.32	4.89	1.10	7.09	4.48
Average	3.24	27.37	8.67	2.20	11.71	6.21

Figure 1.

Disciplinary rates by race, statewide.

Of all demographic groups, African Americans had the highest suspension rate, defined as the number of students suspended divided by student enrollment. From 2002-2003 to 2012-2013, 27% of Black students received a suspension, which was almost 9 times the rate for White students. The dotted line shows that about 6% of Wisconsin students were suspended each year. Taken together, Table 2 shows that African American students by far experienced the highest rates of school exclusion and, thus, are the primary group of interest for this study.

How large was the overrepresentation of Black students in school discipline? Figure 2 superimposes the percentage of students who are African American to the percentage of disciplinary outcomes received by African American students. Despite representing just more than 10% of the student population, Black students received about 45% of suspensions and 34% of expulsions.

Figure 2.

Black share of disciplinary outcomes and Black student enrollment, statewide.

Because the high rates of exclusion among Black students are of concern, schools serve as the unit of analysis. Teacher race was aggregated to the school level as follows:

R_{s t} = \frac{\sum_{j = 1}^{N_{s t}} r_{s t j}}{N_{s t}} .

(1)

The percent of teachers of a given race R in school s during year t is equivalent to the summation of an indicator variable r—which takes on value 1 if matching R, 0 otherwise—corresponding to the jth through $N_{s t}$ teacher, divided by the total number of teachers $N_{s t}$ . Percent of Black teachers and of non-Black minority teachers are of interest in this study and were calculated accordingly. Only regular, licensed teachers not involved with special education were included. According to the Department of Public Instruction, program aides or other support staff are not licensed or considered to be teachers. I assume that the primary teacher holds the authority of the classroom and that teacher aides report student disciplinary issues to the primary teacher. Future studies could evaluate whether race congruency between students and teacher aides results in improved disciplinary rates.

For simplicity, special education teachers are excluded from the analysis. In 2011-2012, I find that about 15% of all teachers taught special education. The proportion is not different for Black teachers—about 15% of Black teachers are special education teachers. Furthermore, I do not find that special education teachers are more likely to be Black—2.3% of special education teachers and 2.4% of non–special education teachers are Black. Given these similarities, I do not anticipate that results would differ if special education teachers were included in the analysis.

Finally, the small number of part-time teachers were included in this study. In 2002-2003, more than 92% of regular and licensed, non–special education teachers held a position equivalent to at least 100% full-time (FTE); more than 95% at least 75% FTE; and more than 99% at least 50% FTE. Teacher racial composition at a given school was similar regardless if part-time teachers were counted.

Table 3 shows that during the time period of interest, 95.5% of teachers in Wisconsin public schools were White, 2.2% were Black, and 2.3% were neither Black nor White. Although the percentage of Black teachers gradually declined, the percentage of non-White teachers was fairly consistent due to a small rise in the percentage of other teachers of color. My calculations of teacher racial composition are similar to those reported by Beck (2014), Boser (2011), and the U.S. Department of Education, National Center for Education Statistics (2004).

Table 3.

Demographics of Wisconsin Public School Teachers, Statewide.

School year	% White teachers	% Black teachers	% Non-Black minority teachers	% Non-White teachers
2002-2003	95.59	2.64	1.76	4.41
2003-2004	95.42	2.67	1.91	4.58
2004-2005	95.48	2.51	2.01	4.52
2005-2006	95.69	2.26	2.05	4.31
2006-2007	95.34	2.22	2.44	4.66
2007-2008	95.76	2.07	2.17	4.24
2008-2009	95.64	2.03	2.33	4.36
2009-2010	95.78	1.91	2.32	4.22
2010-2011	95.38	1.92	2.70	4.62
2011-2012	95.30	1.88	2.82	4.70
2012-2013	95.36	1.76	2.88	4.64
Average	95.52	2.17	2.31	4.48

How did teacher racial composition compare with that among students? Table 4 presents statewide pupil–teacher ratios by racial group, with the number of respective pupils as the numerator and the number of respective teachers as the denominator. On average, there was one White teacher for every 13 White students, one Black teacher for every 80 Black students, and one non-White teacher for every 86 non-White students. There was one non-White, non-Black teacher for every 92 non-White, non-Black students. The representation of non-Black minority teachers was even smaller than that of Black teachers, possibly due to the gradual increase in Hispanic student enrollment.

Table 4.

Ratio of Students to Teachers, Statewide.

School year	Overall^a (1)	White^b (2)	Black^c (3)	Non-Black minority^d (4)	Non-White^e (5)
2002-2003	16.40	13.65	63.92	94.63	76.21
2003-2004	16.48	13.63	64.65	91.92	76.03
2004-2005	16.27	13.36	67.65	90.73	77.91
2005-2006	16.65	13.54	76.73	95.59	85.68
2006-2007	16.78	13.61	79.04	84.13	81.71
2007-2008	16.83	13.51	84.45	98.49	91.62
2008-2009	16.61	13.27	85.28	94.04	89.96
2009-2010	16.77	13.31	91.09	98.53	95.17
2010-2011	16.61	13.17	86.98	88.02	87.59
2011-2012	17.00	13.42	90.13	89.35	89.66
2012-2013	16.95	13.29	95.45	89.99	92.06
Average	16.67	13.43	80.49	92.31	85.78

N Pupils: 1 Teacher.

N White Pupils: 1 White Teacher.

N Black Pupils: 1 Black Teacher.

N Non-Black Minority Pupil: 1 Non-Black Minority Teacher.

N Non-White Pupils: 1 Non-White Teacher.

Table 5 provides summary statistics of the analytic sample and by school level for variables of interest and related variables. Elementary/secondary combination schools and charter schools were excluded. Outcome means based on fewer than six students or where cell information of student groups could be calculated by subtraction were unavailable. School-by-year observations missing information on teacher racial composition or other control variables were omitted. In total, the aggregate sample was based on 20,048 observations across 2,007 schools. The average Wisconsin school expelled 0.11% and suspended 4.53% of its students. Middle and high schools had about the same average suspension rate, but expulsion rates were greater among high schools. Student racial composition varied by school, as evidenced by the large standard deviations. About half of the student population was proficient in reading. The disparate means in exclusion rates across elementary, middle, and high schools suggest that it is appropriate to address the research questions by school level.

Table 5.

Descriptive Statistics by School Level, Analytic Sample.^a

	All schools		Elementary schools		Middle school		High schools
	M	SD	M	SD	M	SD	M	SD
Outcome variables
Expulsion rate	0.11	0.35	0.01	0.08	0.16	0.44	0.30	0.53
Suspension rate	4.53	8.58	2.83	6.70	6.88	11.13	6.89	9.33
Explanatory and related variables
% White teacher	96.37	9.59	95.63	10.50	97.05	9.12	97.67	7.06
% Black teacher	1.82	7.40	2.29	8.05	1.57	7.87	0.84	4.71
% Non-White teacher	3.64	9.60	4.38	10.50	2.96	9.13	2.35	7.09
% Non-Black minority teacher	1.81	5.37	2.09	6.00	1.37	3.77	1.49	4.74
Enrollment	440.97	327.35	356.38	156.97	449.26	259.53	647.96	536.56
% White students	80.58	24.28	76.98	26.59	83.90	19.72	86.91	19.43
% Black students	7.84	18.69	9.71	21.12	5.81	14.09	4.81	14.46
% Hispanic students	6.52	11.00	7.95	12.98	5.23	7.27	3.98	6.64
% Asian students	2.90	4.78	3.32	5.41	2.71	4.06	2.00	3.25
% American Indian students	1.66	6.41	1.46	5.75	1.88	7.32	1.98	7.14
% Reading proficiently	48.37	24.20	48.03	25.42	50.01	23.83	47.88	21.13
N observations	20,048		11,631		3,817		4,600
N schools	2,007		1,172		386		449

Based on all groups of students.

Method

The following fixed effects model estimates the impact of teacher racial composition on student exclusion rates, controlling for student population characteristics:

y_{r s t} = β_{0} + β_{1} {(p b l a c k T)}_{s t} + β_{2} {(p n w n b T)}_{s t} + θ X_{r s t} + σ_{s} + τ_{t} + ε_{r s t} .

(2)

Dependent variable $y_{r s t}$ is the expulsion rate or suspension rate among student group r in school s during year t. Subscript r is all students, White students, or Black students in a given school. As some schools in Wisconsin did not have many minority students, the sample size for Whites is much larger than that for Blacks. Coefficients $β_{1}$ and $β_{2}$ are of interest and estimates the effect of increasing the percentage of Black teachers and non-Black minority teachers, respectively, on disciplinary rates. Vector $X_{r s t}$ controls for student population characteristics such as enrollment, percent of students in each minority group (African American, Hispanic, Asian, and American Indian), and percentage of students achieving proficient or advanced on the reading portion of the annual state exam. In Wisconsin schools, the percent of students in each minority group is weakly correlated to each other and can therefore be in the specification together so long as the percent of White students is omitted. The same applies to teacher racial composition. School fixed effects $σ_{s}$ control for unobserved heterogeneity among schools. Year fixed effects $τ_{t}$ capture year-specific variation in outcomes that is shared by all schools in the state. Intercept $β_{0}$ and error term $ε_{r s t}$ complete the model. Standard errors are clustered at the school level to correct for correlated errors within schools.

Fixed effects control for stable omitted variable bias across schools but have the disadvantage of discarding some of the signal in the data. Instead of exploiting variation in the entire dataset, fixed effects compare the changes that occur within schools. Still, a fixed effects model is employed in this article rather than a random effects model for several reasons. First, it is difficult to satisfy the random effects assumption that unobserved school characteristics are uncorrelated with observed covariates, and it is unclear whether the assumption generally holds in practice (Clarke, Crawford, Steele, & Vignoles, 2010; Firebaugh, Warner, & Massoglia, 2013). Second, a fixed effects model is more suitable when only one policy-relevant variable is of interest (Clarke et al., 2010), as teacher representation is in the present study. Although the data do show more variation in teacher racial composition between rather than within schools, within-school variation is not nonexistent. Later in the article, a sensitivity analysis takes this into account.

The primary limitation of this study is that the use of school-level rather than student-level data prevents a more rigorous empirical approach and does not allow for causal inferences to be drawn because observations are based on different groups of students across time. I am unable to comment on whether being taught by a Black teacher reduces a Black student’s likelihood of being suspended or expelled, but can state how disciplinary rates change with the number of teachers of color employed at a school. Second, the data do not reveal the reasons for expulsions and suspensions, preventing the analysis from controlling for student behavior. I can only make statements regarding disciplinary outcomes, which may not perfectly capture student conduct. Finally, it is unclear whether the results are generalizable to other states. Wisconsin is less racially diverse and pupil–teacher racial dynamics may vary by region (Grissom et al., 2009). But despite these limitations, the study still makes a contribution given that the quantitative research on discipline is relatively small (Noltemeyer, Ward, & Mcloughlin, 2015) and that school exclusions may result in significant social costs (Rumberger & Losen, 2016).

Results

Results by school level and student racial group are presented in Table 6. Aggregate results are driven by the large number of elementary schools in the analytic sample and are available from the author. Across levels of schooling, the explanatory power of the baseline specification is higher for suspension rates than it is for expulsion rates, suggesting that incidents leading to expulsion occur more randomly and that suspensions may be more embedded in school culture.

Table 6.

The Relationship Between Teacher Racial Composition and Disciplinary Rates.

	Expulsion rate			Suspension rate
	All students	White	Black	All students	White	Black
	(1)	(2)	(3)	(4)	(5)	(6)
Panel A. Elementary schools
% Black teachers	0.00264*	0.000842	0.00461**	0.192**	0.0590	0.182*
% Black teachers	(0.00136)	(0.00135)	(0.00226)	(0.0827)	(0.0687)	(0.100)
% Non-Black minority teachers	−0.000250	0.000203	−0.00446	0.0161	−0.00379	0.227**
% Non-Black minority teachers	(0.000298)	(0.000310)	(0.00381)	(0.0183)	(0.00926)	(0.108)
N observations	11630	7,410	2,012	11,630	7,357	1,969
N schools	1,176	1,051	416	1,176	1,051	413
adjusted R²	.206	.073	.049	.762	.641	.636
Panel B. Middle schools
% Black teachers	0.0311	−0.00489	0.00227	0.378*	0.283	0.293
% Black teachers	(0.0225)	(0.00683)	(0.0151)	(0.197)	(0.211)	(0.359)
% Non-Black minority teachers	0.000811	−0.00160	−0.00517	−0.0336	−0.0231	0.246
	(0.00245)	(0.00181)	(0.0201)	(0.0435)	(0.0252)	(0.376)
N observations	3,817	2,949	1,098	3,817	2,811	1,058
N schools	388	360	192	388	360	189
adjusted R²	.382	.217	.237	.884	.737	.730
Panel C. High schools
% Black teachers	0.00202	−0.00221	−0.0239	−0.0749	−0.169	−1.244***
% Black teachers	(0.00617)	(0.0107)	(0.0252)	(0.121)	(0.224)	(0.448)
% Non-Black minority teachers	−0.000128	−0.00335	−0.0135	0.117*	0.00535	−0.330
% Non-Black minority teachers	(0.00183)	(0.00302)	(0.0438)	(0.0686)	(0.0191)	(0.412)
N observations	4,600	2,448	596	4,600	2,255	555
N schools	448	380	112	448	379	112
adjusted R²	.281	.364	.312	.786	.698	.748

Note. All models include school and year fixed effects and control for enrollment, percent of students in each minority group (Black, Hispanic, Asian, and American Indian), and percentage of students achieving reading proficiency. Standard errors in parentheses.

Significant at the *10% level, **5% level, ***1% level.

At the elementary school level, Black teacher representation appears to be positively associated with the expulsion of Black students (p < .5) and of all students (p < .10), but not of White students. A one percentage point increase in the proportion of Black teachers is approximately related to one additional expulsion for every 220 Black elementary students or for every 380 elementary students overall, a trivial finding. Similarly, Black teacher representation has a positive relationship with suspension rates among Black students (p < .10) and of all students (p < .5), but not among White students. The positive and significant estimate on all students seems to be driven by Black students and possibly non-Black minority students.

On average and for both Black and White students, teacher racial composition seems to be unrelated to middle school expulsion rates. However, the percentage of Black teachers in a middle school shares a positive relationship with the suspension rate among all students (p < .10), but not for White or Black students alone. The association appears to be driven by non-Black minority students.

Finally, high school expulsion rates on average and for both Black and White students are uncorrelated with teacher racial representation, similar to the null finding for middle schools. However, holding everything constant, a one point increase in the percentage of Black teachers in a high school is associated with a 1.2 point reduction in the percentage of Black students who are suspended in a given year (p < .01). Suspension rates among White students again seem to be unrelated to teacher race.

Is the association between Black disciplinary outcomes and teacher representation specific to Black teachers or minority teachers overall? For elementary schools, the only coefficient for percent of non-Black minority teachers that is statistically significant is that corresponding to Black suspension rates (p < .01) and is actually larger than the estimate on the percent of Black teachers. None of the estimates regarding non-Black minority representation in middle schools are statistically meaningful. Finally, for high schools, the coefficient on the percent of non-Black minority teachers is large and negative but nonsignificant, showing that the reduction in Black suspension rates only occurs given an increase in Black teachers rather than a general increase in teachers of color.

Why does the relationship between Black teacher representation and suspension rates among Black students differ by school level? Earlier, Table 4 revealed that statewide, there was one White teacher per 13 White students and one Black teacher per 80 Black students during this time period. Based on the analytic sample, Table 7 disaggregates pupil–teacher race-congruence ratios by school level and shows that there was one White teacher for every 11, 12, and 13 White elementary, middle, and high school students, respectively. On the contrary, there was one Black teacher for every 66, 60, and 105 Black elementary, middle, and high school students. The difference is substantial. This study provides some evidence that hiring more African American teachers in high schools may be advisable. At the same time, the positive and null findings for elementary and middle schools, respectively, suggest that at a certain point, hiring more minority teachers may not improve disciplinary outcomes for Black students (though it may have other benefits). In the analytic sample, high schools with the largest representation of African American teachers did not necessarily have the lowest suspension and expulsion rates. There are certainly other factors that give rise to disciplinary problems.

Table 7.

Ratio of Students to Teachers by School Level, Analytic Sample.

	N Pupils:1 teacher	N White pupils: 1 White teacher	N Black pupils: 1 Black teacher
Elementary	13.96	11.23	65.80
Middle school	14.03	11.60	59.76
High school	15.36	13.00	105.25

Note. These numbers are slightly different from the statewide ratios in Table 4 because the analytic sample excludes elementary/secondary combination and charter schools.

For which high schools could a larger presence of Black teachers be associated with lower suspension rates among Black students? Table 8 shows that the inverse relationship is driven by high schools that suspended an average of at least 50% of their Black students. The inverse association still applies to high schools that suspended fewer than 25% or between 25% and 50% of their Black students, but it is statistically nonsignificant.

Table 8.

Suspension Rates Among Black High School Students.

	Schools that suspend less than 25% of Black students^a	Schools that suspend between 25% and 50% of Black students^a	Schools that suspend 50% or more Black students^a
	(1)	(2)	(3)
% Black teachers	−0.266	−0.538	−1.060**
% Black teachers	(0.409)	(0.445)	(0.384)
N observations	297	203	55
N schools	72	30	10
adjusted R²	.379	.237	.665

Note. All models include school and year fixed effects and control for percent of non-Black minority teachers, enrollment, percent of students in each minority group (Black, Hispanic, Asian, and American Indian), and percentage of students achieving reading proficiency. Standard errors in parentheses.

On average.

Significant at the *10% level, **5% level, ***1% level.

Discussion

African American students receive more suspensions and expulsions than their counterparts, possibly due to racial discrimination or actual rule-breaking. Studies have shown that student–teacher race congruence could lead to academic gains but have yet to evaluate its impact on school discipline. Using administrative panel data from 2002-2003 to 2012-2013 on all schools in the state of Wisconsin, this study takes a statewide, multiyear approach that is more credible than many existing studies. I find that a 1-point increase in the percentage of Black teachers reduces suspension rates among Black high school students by 1.2 percentage points, supporting the representative bureaucracy hypothesis. This magnitude is likely an underestimate of the overall impact on total suspensions because suspended students often receive more than one suspension in a given year (Cheng, 2016). The effect is attributable to an increase in African American teachers specifically, rather than a general increase in teachers of color, and is most evident in high schools that suspend more than 50% of their Black students. For a given school, having more Black teachers is associated with a rise in the expulsion of Black elementary students, but the magnitude is trivial. Last, average suspension rates in elementary schools are positively associated with Black teacher representation, but the results may be due to non-Black minority students who are beyond the scope of this study. Across levels of schooling, exclusion rates of White students are unrelated to teacher racial composition.

This study is unable to identify the mechanism, direct or indirect, in which having more Black teachers reduce suspensions among Black high school students. Perhaps, a more diverse teacher population alters the overall school environment (Reskin et al., 1999). For example, African American teachers may bring to attention disciplinary issues that their colleagues might not be aware of (see Lim, 2006, for a discussion on possible direct and indirect contributions of minority bureaucrats). Another possibility is that students behave differently when around same-race teachers. Or perhaps, Black teachers address, tolerate, or do not report incidents that would otherwise result in a suspension. It is important to note that recorded disciplinary events may not correspond perfectly to actual student behavior, and that suspension rates could decline without experiencing improvement in student conduct. Future research could explore any of these possibilities.

Finally, it is important to note that although interesting, research on racial matching is somewhat controversial due to the ethicality of its implications. For example, the policy implication of finding positive effects of teacher–student racial pairing is to make classroom assignments based on race, promoting segregation at the classroom level. The broader consequence is racial segregation of teachers and students at the school level. Even if race congruence produces higher test scores, researchers should consider whether such policy is ethical and by what means schools are willing to obtain results. Taking an alternative approach, this study treated the teaching population at a given school as a collective unit. This framework assumed that racial dynamics work in subtle ways such that students do not need to be directly taught by a same-race teacher to be positively affected. The implication of such study would encourage adjustments to the overall teaching pool without necessitating individual student–teacher pairings.

Footnotes

Acknowledgements

I would like to thank Sara Goldrick-Rab, Eric Grodsky, Erica Turner, Katie Broton, Sohyun Kang, and Rachel Feldman for helpful comments.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biography

Dorothy A. Cheng completed her Ph.D. in Educational Policy Studies at the University of Wisconsin, Madison. Her research interests include school discipline and racial disparities in student outcomes.

References

Alexander

(2012). The new Jim Crow: Mass incarceration in the age of colorblindness. New York, NY: The New Press.

Arum

Velez

(Eds.). (2012). Improving learning environments: School discipline and student achievement in comparative perspective. Stanford, CA: Stanford University Press.

Aud

KewalRamani

Frohlich

(2011). America’s youth: Transitions to adulthood (NCES 2012-026). Washington, DC: National Center for Education Statistics.

Beady

C. H.

Hansell

(1981). Teacher race and expectations for student achievement. American Educational Research Journal, 18, 191-206.

Beck

(2014, December 14). “I never had a teacher that looked like me”: Challenges exist in hiring a diverse staff. Retrieved from http://host.madison.com/news/local/education/local_schools/i-never-had-a-teacher-that-looked-like-me-challenges/article_f2435038-b105-5299-86c9-621e4724d94c.html

Boser

(2011). Teacher diversity matters: A state-by-state analysis of teachers of color. Center for American Progress. Retrieved from https://www.americanprogress.org/wp-content/uploads/issues/2011/11/pdf/teacher_diversity.pdf

Brown

T. M.

(2007). Lost and turned out academic, social, and emotional experiences of students excluded from school. Urban Education, 42, 432-455.

Cheng

D. A.

(2016). Effects of a firearm zero tolerance policy on student expulsions and suspensions (Working paper). Retrieved from https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2909104

Christle

Nelson

C. M.

Jolivette

(2004). School characteristics related to the use of suspension. Education & Treatment of Children, 27, 509-526.

10.

Clarke

Crawford

Steele

Vignoles

(2010). The choice between fixed and random effects models: Some considerations for educational research (Working Paper No. 10-10). London, England: Department of Quantitative Social Science-Institute of Education, University of London.

11.

Croninger

R. G.

Rice

J. K.

Rathbun

Nishio

(2007). Teacher qualifications and early learning: Effects of certification, degree, and experience on first-grade student achievement. Economics of Education Review, 26, 312-324.

12.

Dee

T. S.

(2005). A teacher like me: Does race, ethnicity, or gender matter? The American Economic Review, 95, 158-165.

13.

Ehrenberg

R. G.

Goldhaber

D. D.

Brewer

D. J.

(1995). Do teachers’ race, gender, and ethnicity matter? Evidence from the National Educational Longitudinal Study of 1988. Industrial & Labor Relations Review, 48, 547-561.

14.

Eitle

T. M.

Eitle

D. J.

(2004). Inequality, segregation, and the overrepresentation of African Americans in school suspensions. Sociological Perspectives, 47, 269-287.

15.

Evans

A. E.

(2007). School leaders and their sensemaking about race and demographic change. Educational Administration Quarterly, 43, 159-188.

16.

Ferguson

A. A.

(2001). Bad boys: Public schools in the making of black masculinity. Ann Arbor: University of Michigan Press.

17.

Figlio

D. N.

(2006). Testing, crime and punishment. Journal of Public Economics, 90, 837-851.

18.

Firebaugh

Warner

Massoglia

(2013). Fixed effects, random effects, and hybrid models for causal analysis. In Morgan

S. L.

(Ed.), Handbook of causal analysis for social research (pp. 113-132). Dordrecht, The Netherlands: Springer.

19.

Gregory

Cornell

Fan

(2011). The relationship of school structure and support to suspension rates for Black and White high school students. American Educational Research Journal, 48, 904-934.

20.

Gregory

Skiba

R. J.

Noguera

P. A.

(2010). The achievement gap and the discipline gap: Two sides of the same coin. Educational Researcher, 39, 59-68.

21.

Grissom

J. A.

Kern

Rodriguez

L. A.

(2015). The “representative bureaucracy” in education: Educator workforce diversity, policy outputs, and outcomes for disadvantaged students. Education Researcher, 44, 185-192.

22.

Grissom

J. A.

Nicholson-Crotty

(2009). Race, region, and representative bureaucracy. Public Administration Review, 69, 911-919.

23.

Grissom

J. A.

Rodriguez

L. A.

Kern

(2017). Teacher and principal diversity and the representation of students of color in gifted programs: Evidence from national data. The Elementary School Journal, 117, 396-422.

24.

Ladson-Billings

(1995). Toward a theory of culturally relevant pedagogy. American Educational Research Journal, 32, 465-491.

25.

Lewis

C. W.

Butler

B. R.

Bonner

I. I.

Fred

Joubert

(2010). African American male discipline patterns and school district responses resulting impact on academic achievement: Implications for urban educators and policy makers. Journal of African American Males in Education, 1(1), 7-25.

26.

Lim

H. H.

(2006). Representative bureaucracy: Rethinking substantive effects and active representation. Public Administration Review, 66, 193-204.

27.

Lindsay

Hart

(2017). Teacher race and school discipline. Education Next, 17(1), 72-78.

28.

Losen

Hodson

C. I.

Keith

I. I.

Michael

Morrison

Belway

(2015). Are we closing the school discipline gap? (K-12 Racial Disparities in School Discipline). Los Angeles: The Center for Civil Rights Remedies at The Civil Rights Project of University of California, Los Angeles.

29.

Males

(1999). Framing youth: Ten myths about the next generation. Monroe, ME: Common Courage Press.

30.

Meier

K. J.

Stewart

(1992). The impact of representative bureaucracies: Educational systems and public policies. The American Review of Public Administration, 22, 157-171.

31.

Mendez

L. M. R.

(2003). Predictors of suspension and negative school outcomes: A longitudinal investigation. New Directions for Youth Development, 2003(99), 17-33.

32.

Monk

D. H.

King

J. A.

(1994). Multilevel teacher resource effects in pupil performance in secondary mathematics and science: The case of teacher subject matter preparation. In Ethrenberg

R. G.

(Ed.), Choices and consequences: Contemporary policy issues in education (pp. 29-58). Ithaca, NY: ILR Press.

33.

Monroe

C. R.

(2005). Why are “bad boys” always Black? Causes of disproportionality in school discipline and recommendations for change. The Clearing House: A Journal of Educational Strategies, Issues and Ideas, 79, 45-50.

34.

Monroe

C. R.

(2009). Teachers closing the discipline gap in an urban middle school. Urban Education, 44, 322-347.

35.

Nicholson-Crotty

Grissom

J. A.

Nicholson-Crotty

Redding

(2016). Disentangling the causal mechanisms of representative bureaucracy: Evidence from assignment of students to gifted programs. Journal of Public Administration Research and Theory, 26, 745-757.

36.

Noguera

P. A.

(2003). Schools, prisons, and social implications of punishment: Rethinking disciplinary practices. Theory into Practice, 42, 341-350.

37.

Noltemeyer

A. L.

Ward

R. M.

Mcloughlin

(2015). Relationship between school suspension and student outcomes: A meta-analysis. School Psychology Review, 44, 224-240.

38.

O’Connor

(1999). Race, class, and gender in America: Narratives of opportunity among low-income African American Youths. Sociology of Education, 72, 137-157.

39.

Ouazad

(2014). Assessed by a teacher like me: Race and teacher assessments. Education Finance and Policy, 9, 334-372.

40.

Pace

J. L.

Hemmings

(2007). Understanding authority in classrooms: A review of theory, ideology, and research. Review of Educational Research, 77, 4-27.

41.

Payne

A. A.

Welch

(2010). Modeling the effects of racial threat on punitive and restorative school discipline practices. Criminology, 48, 1019-1062.

42.

Payne

A. A.

Welch

(2013). Restorative justice in schools: The influence of race on restorative discipline. Youth & Society, 47, 539-564.

43.

Reskin

B. F.

McBrier

D. B.

Kmec

J. A.

(1999). The determinants and consequences of workplace sex and race composition. Annual Review of Sociology, 25, 335-361.

44.

Reynolds

C. R.

Skiba

R. J.

Graham

Sheras

Conoley

J. C.

Garcia-Vazquez

(2008). Are zero tolerance policies effective in the schools? An evidentiary review and recommendations. The American Psychologist, 63, 852-862.

45.

Roch

C. H.

Pitts

D. W.

Navarro

(2010). Representative bureaucracy and policy tools ethnicity, student discipline, and representation in public schools. Administration & Society, 42, 38-65.

46.

Rocque

(2010). Office discipline and student behavior: Does race matter? American Journal of Education, 116, 557-581.

47.

Rumberger

R. W.

Losen

D. J.

(2016). The high cost of harsh discipline and its disparate impact. Los Angeles, CA: The Center for Civil Rights Remedies at The Civil Rights Project of University of California, Los Angeles.

48.

Skiba

R. J.

Chung

C. G.

Trachok

Baker

T. L.

Sheya

Hughes

R. L.

(2014). Parsing disciplinary disproportionality: Contributions of infraction, student, and school characteristics to out-of-school suspension and expulsion. American Educational Research Journal, 51, 640-670.

49.

Skiba

R. J.

Eckes

Brown

(2010). African American disproportionality in school discipline: The divide between best evidence and legal remedy. New York Law School Law Review, 54, 1071-1112.

50.

Skiba

R. J.

Horner

R. H.

Chung

C. G.

Rausch

M. K.

May

S. L.

Tobin

(2011). Race is not neutral: A national investigation of African American and Latino disproportionality in school discipline. School Psychology Review, 40(1), 85-107.

51.

Skiba

R. J.

Michael

R. S.

Nardo

A. C.

Peterson

R. L.

(2002). The color of discipline: Sources of racial and gender disproportionality in school punishment. The Urban Review, 34, 317-342.

52.

Skiba

R. J.

Peterson

(1999). The dark side of zero tolerance: Can punishment lead to safe schools. Phi Delta Kappan, 80, 372-382.

53.

Skiba

R. J.

Shure

Williams

(2012). Racial and ethnic disproportionality in suspension and expulsion. In Noltemeyer

A. L.

Mcloughlin

C. S.

(Eds.), Disproportionality in education and special education. Springfield, IL: Charles C. Thomas Publisher.

54.

U.S. Department of Education, National Center for Education Statistics. (2004). Percentage distribution of public school teachers by race/ethnicity, percentage minority, and state: 2003–04. Retrieved from https://nces.ed.gov/surveys/sass/tables/state_2004_18.asp

55.

U.S. Department of Education, National Center for Education Statistics. (2015). Percentage of public schools that used safety and security measures: Various school years, 1999-2000 through 2013-14. Retrieved from http://nces.ed.gov/programs/digest/d15/tables/dt15_233.50.asp

56.

U.S. Department of Education Office for Civil Rights. (2014). Civil rights data collection—Data snapshot: School Discipline. Retrieved from https://www2.ed.gov/about/offices/list/ocr/docs/crdc-discipline-snapshot.pdf

57.

Verdugo

R. R.

(2002). Race-ethnicity, social class, and zero-tolerance policies: The cultural and structural wars. Education and Urban Society, 35, 50-75.

58.

Wald

Losen

D. J.

(2003). Defining and redirecting a school-to-prison pipeline. New Directions for Youth Development, 2003(99), 9-15.

59.

Wallace

J. M.

Jr. Goodkind

Wallace

C. M.

Bachman

J. G.

(2008). Racial, ethnic, and gender differences in school discipline among U.S. high school students: 1991-2005. The Negro Educational Review, 59(1-2), 47-62.

60.

Welch

Payne

A. A.

(2010). Racial threat and punitive school discipline. Social Problems, 57(1), 25-48.

61.

Wilson

W. J.

(2009). More than just race: Being black and poor in the inner city (Issues of Our Time). New York, NY: W. W. Norton & Company.

62.

Yosso

(2005). Whose culture has capital? A critical race theory discussion of community cultural wealth. Race Ethnicity and Education, 8, 69-91.