Abstract
School discipline is a salient problem of educational policy and practice. Teachers play an important role in the production and disruption of racial inequities in school discipline, yet there remains a need to disentangle the relationship between teacher characteristics, their perceptions of school climate, and school discipline patterns. This study uses longitudinal data representing over one thousand public middle and high schools in New York City and linear regression methods to examine the relationship between teachers’ perceptions of school climate, teacher characteristics, and the rates of and disparities in suspensions. Overall, results indicate more positive teacher perceptions of school climate, higher years of teaching experience, and a greater share of Black teachers are independently associated with lower rates of office referrals and suspensions, overall and especially for Black and Latinx students. The study concludes by discussing implications for teacher education and on-the-job support as well as school leadership.
Keywords
Introduction
School discipline remains an urgent educational issue of the modern era. Prior research has shown that exclusionary forms of school discipline such as suspensions and expulsions have various deleterious consequences for children and youth, including exacerbating achievement disparities by disproportionately removing students of color from the classroom and disrupting their exposure to standard classroom instruction (Gregory et al., 2010; N. Hwang, 2018; Skiba et al., 2016; Welsh & Little, 2018a), and ultimately increasing the likelihood of long-term negative outcomes, such as higher rates of involvement in the juvenile and criminal justice systems as well as poorer health and socioeconomic well-being into adulthood (Bacher-Hicks et al., 2019; Davison et al., 2022; Duarte et al., 2023). Black and Brown students disproportionately experience exclusionary discipline compared to their White and Asian peers, even in response to similar subjective infractions, such as not following school dress code or “being too loud” (Anyon et al., 2016; Barrett et al., 2021; Welsh, 2021a; Welsh & Little, 2018b). Prior research documents that racial disparities in exclusionary school disciplinary outcomes are largely due to subjective offenses (Girvan et al., 2017; Legette & Anyon, 2023)—or disciplinary decisions largely at the discretion of educators in their response to perceived misbehavior (Barrett et al., 2021; Girvan et al., 2017; Welsh, 2022, 2023a). Scholars contend that racial and ethnic disparities in exclusionary discipline arising from teacher-subjective decision-making are tied to structural inequities and racism within schools (Legette & Anyon, 2023; Sobti & Welsh, 2023).
Teachers serve a central role in school discipline processes, as they often initiate the process by reporting instances of student misbehavior through an office discipline referral (ODR) to identify a student for behavioral intervention (Gregory et al., 2010; Marcucci & Elmesky, 2020; Nichols et al., 2006). More specifically, teachers’ classroom management skills (Cho et al., 2020); teacher-student racial mismatch and the lack of racial diversity within the educator workforce (Blake et al., 2016; Grissom et al., 2015; Lindsay & Hart, 2017); and teachers’ negative perceptions, beliefs, expectations, and bias of students (Hines-Datiri, 2015; Okonofua & Eberhardt, 2015; Okonofua et al., 2016) can significantly shape the prevalence of and differential processing of exclusionary discipline in schools.
While teachers are an important aspect of the production and disruption of exclusionary discipline, they are situated within a broader school climate that shapes their relationships and interactions with students and school administrators throughout the disciplinary process. Scholars and policymakers have posited that improvements to school climate may reduce racial inequality in school discipline (Gage et al., 2016; Skiba et al., 2014; U.S. Department of Education, 2017). However, relatively few empirical studies have examined the link between how teachers perceive their school climate and school discipline outcomes (Bear et al., 2014; Heilbrun et al., 2018; O’Brennan et al., 2014; Steinberg et al., 2013; Zulauf-McCurdy & Zinsser, 2021, 2022)—those that do have largely relied on cross-sectional data in elementary and middle school settings. Prior research has established a correlation between positive school climates, largely from student perspectives, and lower rates of office discipline referrals and suspensions (e.g., Gage et al., 2016; Gregory et al., 2011; Lee et al., 2011). Yet, how teachers perceive school climate and how those perceptions shape disciplinary outcomes remains poorly understood from a research perspective.
This study uses data from public middle and high schools for New York City Public Schools (NYCPS) to examine the relationship between teachers’ perceptions of school climate, school contextual factors related to teachers (demographic composition and qualifications), and outcomes arising from school discipline processes. Specifically, analyses presented herein are guided by the following research questions: Do teachers’ perceptions of school climate and characteristics of teachers within schools independently and interactively relate to the prevalence of ODRs and suspensions at the school level, particularly for Black and Latinx youth, above and beyond other school characteristics? And which dimensions of perceived school climate and characteristics of teaching staff are most notably related to ODRs and school suspensions?
This study is one of few recent studies that link teachers’ observable characteristics, teachers’ perspectives of school climate, and school discipline outcomes, and does so using a longitudinal sample of middle and high schools located in a large, racially diverse, urban multidistrict school system in NYCPS. The lion’s share of prior studies examining the link between teacher characteristics and perceptions of their school environment to school discipline outcomes have been limited to cross-sectional samples at the high school level (Bottiani et al., 2017; Gregory et al., 2011); our study extends the prior literature base by leveraging robust longitudinal administrative and teacher survey data for urban middle and high schools to provide further insight into the way teachers are associated with variation in school discipline patterns while controlling for year-specific and time-invariant school-level confounding factors.
In this study, there is a particular focus on the exclusionary discipline of Black and Latinx students. The majority of attention has been paid to Black-White differences in disciplinary experiences with relatively little attention to the disciplinary experiences of Latinx students (Welsh & Little, 2018b) and these results have been inconsistent and mixed (Anyon et al., 2014; Gage et al., 2020; Ramey, 2015; Welsh, 2022; Welch & Payne, 2018). Some studies have highlighted racial discipline disparities for Latinx students (Cruz & Rodl, 2018; Gage et al., 2020; Peguero & Shekarkhar, 2011), and others have found no significant differences between the disciplinary outcomes of Latinx and White students (Brown & DiTillio, 2013; Morgan & Wright, 2018). Although a substantial proportion of the research on school discipline has an urban focus (Welsh, 2021b; Welsh & Little, 2018b), a limitation of the literature is its consideration of the prevalence and disparity of exclusionary discipline experienced by Latinx students and their contrast with Black student disciplinary experiences within the same urban context (Anyon et al., 2016). Although the prevalence of exclusionary discipline is lower for Latinx students compared to Black students (in 2017–18, Latinx students were 27.2% of student enrolments, 23.2% of ISSs, and 21.7% of OSSs (Office for Civil Rights, 2021), it nevertheless remains important to examine the disciplinary experiences of Latinx students. Given these inconsistent findings and the demographic growth of Latinx students in the United States (Gage et al., 2020), this study intentionally examines how teachers are associated with discipline patterns for Latinx students.
Study findings have implications for education policy and practice by furthering an understanding of the central role of teachers in the disciplining of an increasingly diverse student population, with the aim of identifying which teachers or perceived schooling experiences of teachers tend to be significantly related to exclusionary and inequitable school discipline practices. More specifically, the findings generated from this study may inform revisions and upgrades to the preparation of teachers to meet contemporary school discipline challenges by situating school discipline not only as a matter of classroom management (Welsh, 2023b) but, more broadly, as a matter of school climate affecting teachers.
School Discipline, Teachers, and School Climate
Teachers, school climate, and school discipline are intricately interconnected elements within K–12 schools—as a guiding framework, Figure 1 depicts these interconnections. Teachers play a central role in shaping the school climate through their interactions with students, colleagues, and the overall learning environment they cultivate. A positive and inclusive school climate, characterized by a sense of trust, safety, respect, and a sense of belonging, is often fostered by educators who prioritize students’ social and emotional development alongside academics (Cohen et al., 2009; Thapa et al., 2013). In turn, a nurturing school climate can have a profound impact on school discipline. When students feel safe, supported, and committed to their learning, there can be a decrease in disciplinary issues (Bottiani et al., 2017; Huang & Cornell, 2018). Conversely, a negative or more authoritarian (structured but not supportive) school climate can exacerbate discipline problems (Bear et al., 2017; Bottiani et al., 2017; Fefer & Gordon, 2020; Gage et al., 2016; Gregory et al., 2011, 2012; Heilbrun et al., 2018; Huang & Cornell, 2018).
Teachers, school climate, and school discipline.
Drawing on the tenets of the Systems View of School Climate (SVSC) (Rudasill et al., 2018), we posit that teachers contribute to, function within, and are affected by school climate and school disciplinary processes affecting students. There is a lack of consensus on a universal way to conceptualize and measure school climate (Rudasill et al., 2018). Under SVSC, school climate is defined as “the affective and cognitive perceptions regarding social interactions, relationships, values, and beliefs held by students, teachers, administrators, and staff within a school” (Rudasill et al., 2018, p. 46). Rudasill et al. (2018) further distinguish between school climate, which reflects group-level perceptions of a school, and other aspects of a school such as its structure (e.g., enrollment or class size), other contextual characteristics (e.g., student and staff composition or the prevalence of bullying, absenteeism, or teacher turnover), and processes that govern school norms and rules (e.g., school discipline practices). Structural components of a school, contextual characteristics of its members, and processes within a school “may form a base from which perceptions are formed, but they are conceptually and empirically different than school climate” (Rudasill et al., 2018, et al., p. 46).
SVSC highlights how teachers, students, and other school stakeholders interact in their immediate contexts or microsystems as well as within nanosystems nested within microsystems such as classrooms (Rudasill et al., 2018). In this regard, teachers and their perceptions of school climate may shape how schools address instances of perceived misbehavior from students (Welsh, 2023a, 2023b). Given that SVSC offers a clear distinction between school climate and related concepts like school context, structures, and processes that inform and reflect school climate, this framework serves as a valuable perspective for analyzing the connection between key school stakeholders, such as teachers, the way they perceive school climate, and the results of school processes, including the prevalence of exclusionary discipline stemming from school disciplinary practices. Moreover, along with school climate, the compositional characteristics of teachers (e.g., race/ethnicity, gender, years of experience) serve as school contextual characteristics that can independently and interactively relate to the prevalence of exclusionary discipline. That is, teachers’ perceptions of their school may matter more for the disciplining of students given certain attributes of teaching staff (e.g., race or experience).
In addition to SVSC, this study also incorporates the theory of racialized organizations (Ray, 2019) to consider the role of race and racism in the interactions among students and teachers within the school discipline process. It is essential to recognize that schools are not race-neutral environments; rather, the theory of racial organizations underscores the intricate role of schools in perpetuating, reinforcing, and expanding racial disparities in the context of school discipline (Cruz & Firestone, 2023). Ray (2019) defined “ race not as a thing but as a relationship between persons mediated through things” (p. 29). Within the school discipline process, teachers act as street-level bureaucrats, wielding considerable discretion over the consequences students face (Haider-Markel et al., 2022; Marcucci & Elmesky, 2020). Notably, this level of discretion can operate on the basis of a student or teacher’s race or ethnicity. Previous research has illustrated that race at the individual and school level plays an important role in whether students are referred to the office and suspended (Rodriguez & Welsh, 2022; Vavrus & Cole, 2002; Welsh & Little, 2018b; Welsh & Rodriguez, 2024). Owens and McLanahan (2020) examined the drivers of racial disparities in suspensions and expulsions and found that the differential treatment and support of students with similar behaviors accounted for the majority of disparities in exclusionary discipline. Black students receive more frequent and harsher punishment than their White peers for similar infractions and are disproportionately removed from classrooms (Barrett et al., 2021; Rodriguez & Welsh, 2022; Skiba et al., 2014; Welsh, 2022). This emphasis on differential treatment underscores the significant role that both teachers’ and students’ racial backgrounds play in the school disciplinary process.
As such, we use the lens of racialized organizations to situate teachers’ race in relationship with students to consider how the racial/ethnic composition of teaching staff within schools may affect the prevalence of school discipline among racial or ethnic student groups who have historically been marginalized within the U.S. public education system, Black and Latinx youth (Skiba et al., 2011). Conceptualizing schools as racialized organizations provides an analytic tool for understanding how schools produce educational (in)equity and reproduce societal inequality (Ray, 2019). One of the main theoretical tenets proposed by Ray (2019) posits that racialized organizations enhance or diminish the agency of racial groups. As noted by Ray (2019), “one’s position in racialized organizations shapes agency” (p. 36) and an increase in the presence of teachers of color may mark a shift in organizational racialization. This shift in organizational racialization may enhance or diminish the agency of teachers of color, especially regarding their use of discretion in disciplinary decisions (Marcucci & Elmesky, 2020; Vavrus & Cole, 2002; Welsh, 2023a, 2023b).
The subsequent sections proceed by presenting a review of existing literature across two pivotal domains. The next section reviews research examining the link between teacher characteristics on school discipline, while the subsequent subsection focuses on research that delves into the interplay between teacher perceptions of school climate and its influence on the administration of discipline within educational settings. A deeper exploration of both areas furnishes a more comprehensive understanding of the intricate relationships between teacher attributes, their perceptions of school climate, and their relevance to school discipline outcomes.
Teachers Characteristics and Students’ Disciplinary Outcomes
Prior studies have highlighted the importance of teachers in implementing school discipline policy and the link between teacher characteristics and discipline disparities (Capers, 2019; Milner, 2010). Teacher-student racial mismatch has been highlighted as a pertinent consideration in discipline disparities (Bradshaw et al., 2010; Lindsay & Hart, 2017). Using aggregates at the school level, several studies have found that Black and Latinx students are less likely to be suspended or expelled in schools with higher proportions of Black and Latinx teachers (Blake et al., 2016; Grissom et al., 2015). Classroom-level studies of teacher racial matching with students have confirmed that Black students tend to get suspended less when assigned to Black teachers (Holt & Gershenson, 2019; Lindsay & Hart, 2017; Shirrell, 2021). The diversity of school personnel may directly influence the interpersonal relationships between students and educators (DeMatthews et al., 2017). Scholars have underlined the significance of teachers being culturally congruent with their students in the disciplinary process (Capers, 2019; Marcucci & Elmesky, 2020; Redding, 2019). Cultural congruence, which pertains to a teacher’s responsiveness to the cultural values, beliefs, practices, and ways of being of students and families in the classroom, is linked to racial congruence. Capers (2019) noted that “ teachers that share the race of their students are more likely to align culturally with the students, though other identities (i.e. gender, nationality) may also foster cultural congruence” (p. 793). Thus, in a racialized organization of schools and a racialized process of perceiving and responding to student misbehavior, Black and Latinx teachers may discipline differently in several ways to the benefit of students of color (Capers, 2019; Lewis & Diamond, 2015; Marcucci & Elmesky, 2020; Milner, 2010).
The intersection of gender and racial identities may also play an instructive yet complex role in the disciplinary process in schools (Bradshaw et al., 2010; Brockenbrough, 2015; Welsh, 2024). For instance, there is increasing attention to the burden of Black male teachers as the primary disciplinarians in schools (Brockenbrough, 2015). Bradshaw et al. (2010) examined the contributors to ODRs in 21 elementary schools and found that student-teacher racial matches did not decrease the risk of ODRs among Black students. Moreover, Black male students in classrooms with Black teachers had the greatest likelihood of receiving an ODR. In their interpretation of these findings on racial congruence, Bradshaw and colleagues postulated that “it is possible that the Black teachers may have a higher standard for the behavior of their students and thus may be more inclined to use major ODRs with their students than other teachers.” (p. 515; Bradshaw et al., 2010). Golann et al. (2019) observed in no-excuses charter schools that the inclination to be “strict on one’s own” is viewed as a preferred form of discipline, especially for Black and Latinx families. Black teachers’ higher standard of behavior in Bradshaw et al. (2010) may be an example of “strict on your own” discipline and illustrates the complexity of how race and cultural disciplinary norms and preferences of educators may shape the disciplinary process and outcomes for students (Marcucci & Elmesky, 2020; Welsh, 2023a, 2023b). Overall, there may be nuanced differences in how Black and Latinx teachers may refer students to the office relative to their teacher peers, especially for the type of infractions (subjective vs. nonsubjective offenses) that results in a higher prevalence of exclusionary discipline, for all students and particularly for Black youth.
The extant literature has emphasized the importance of teacher race and gender to school discipline but has rarely examined the influence of other observable teacher characteristics, such as experience and qualifications, on how teachers respond to perceived student misbehavior (Blake et al., 2016; Hughes et al., 2020). Prior research has emphasized classroom management as an area of growth for novice teachers (Hirsch et al., 2021), with some qualitative evidence to suggest that teachers of varying experience levels employ different approaches to classroom management (Pressley et al., 2020; Wolff et al., 2015). Moreover, recent studies have found that schools with more experienced teachers are associated with lower rates of suspension, with mixed evidence suggesting that more experienced teaching staff are associated with lower disproportionate discipline among Black students (Wang et al., 2022; Williams III et al., 2018). Considering prior evidence to suggest a potential link between teacher experience, classroom management, and school discipline, the inequitable exposure of students of color to novice teachers (Rodriguez et al., 2023) presents a cause of concern with respect to school discipline. The inequitable distribution of novice teachers among students and schools may translate to disparities in classroom discipline, negatively impacting the educational experiences and outcomes of students of color and reinforcing racialized hierarchies within and across schools.
Research on Teachers’ Perceptions
Prior studies have linked teachers’ perceptions and biases to the use of ODRs and suspensions as well as reports of student misbehavior (Gilliam et al., 2016; Marcucci & Elmesky, 2020). For instance, Santiago-Rosario et al. (2021) examined the relationship between teacher expectations of students and ODRs and found that lower expectations were associated with higher disparity in ODR between Black, Latinx, and White students. Atiles et al. (2017) examined how teachers’ values and beliefs affected their disciplinary practices and found that teachers who valued developmentally appropriate practices more frequently used nonpunitive discipline responses when addressing instances of student misbehavior. In a separate study, O’Brennan et al. (2014) found an association between a comprehensive measure of teachers’ perceptions of school climate and reported problem behaviors occurring in their classrooms. There is evidence that teachers may also implicitly believe that Black students specifically are innately less stable, are more likely to misbehave and create trouble, and are more likely to be perceived as a potential risk or threat (Gilliam et al., 2016; Welsh & Little, 2018b).
Teachers and their perceptions of school climate provide an invaluable window into the disciplinary process for a few reasons. First, teachers interact with students, parents, and school leadership throughout the disciplinary process (O’Brennan et al., 2014)—as such, teachers are a central figure among many stakeholders who contribute to a student’s experience with discipline in schools. Second, teachers’ perspectives shed light on several of the essential dimensions of school climate not captured through a student lens. Thapa et al. (2013) identify five dimensions of school climate: (a) safety, (b) relationships, (c) teaching and learning, (d) institutional environment, and (e) the school improvement process. Although SVSC categorizes relationships and social interactions as one of three key components of school climate (Rudasill et al., 2018), most research on school climate has focused on the teaching and learning or safety dimensions of school climate (Cohen et al., 2009). Positive teacher-student and teacher-parent relationships are critical yet have often been overlooked in the school discipline literature (Lindsay & Hart, 2017; Thapas et al., 2013; Zulauf-McCurdy & Zinsser, 2021, 2022). Moreover, relationships between teachers and principals function as a vital yet overlooked component of school climate that can affect teacher behavior and school discipline (Hampden-Thompson & Galindo, 2017; Rodriguez et al., 2022; Viano et al., 2023). Prior studies have highlighted trust and communication in the social interactions and relationships among these key actors in classrooms and schools (Anyon et al., 2016; Rodriguez & Welsh, 2024; Wahman et al., 2022; Zulauf-McCurdy & Zinsser, 2021, 2022). In essence, teachers have a unique vantage point of school climate that may be particularly informative given their range of interactions with other key contributors to school climate as well as their centrality in the disciplinary process in schools.
Relatively few studies have employed teachers’ perceptions to measure school climate and examine its association with students’ disciplinary outcomes (Bear et al., 2014; Heilbrun et al., 2018; O’Brennan et al., 2014; Steinberg et al., 2013; Zulauf-McCurdy & Zinsser, 2021, 2022). For instance, Steinberg et al.’s (2013) analysis, in addition to student surveys, considered teacher reports of school climate in Chicago. However, their study was more oriented toward an analysis examining school safety rather than school discipline outcomes. Heilbrun et al. (2018) used student- and teacher-reported measures of school climate in a sample of middle schools in Virginia and found that only student-reported measures were significantly associated with lower suspension rates, and higher disciplinary structure (i.e., student perceptions of fairer and stricter school discipline policies) was significant for Black but not White students. Recent studies have used teacher surveys to underscore the importance of relationships between teachers and parents—one aspect of school-family relationships—for students’ disciplinary outcomes (Zulauf-McCurdy & Zinsser, 2021, 2022). For instance, using surveys of roughly 300 early childhood teachers, Zulauf-McCurdy and Zinsser (2021) found that teachers’ perceptions of high-quality, teacher-parent relationships were associated with a lower likelihood of expulsion for children who have not yet been expelled, with a stronger association found among Black children.
Middle and High School Discipline in New York City Public Schools
Our study leverages NYCPS as an empirical case to examine the relationship between the school contextual attributes of teachers, school climate as perceived by teachers, and outcomes arising from school discipline processes. It goes without saying that NYCPS is a highly unique urban setting that would be problematic to decisively generalize to other contexts; however, we argue that the city nevertheless offers valuable insights into the issue of student discipline. The city’s public school system serves over one million students and comprises over seventy thousand teachers, far outranking all others in terms of size and diversity. Due to the scope of diversity, unlike many other settings, the study of NYCPS provides the opportunity for further inquiry on discipline patterns associated with many student racial/ethnic groups, specifically for Latinx students.
The NYCPS discipline code classifies infractions into five levels based on severity, ranging from Level 1 infractions considered “uncooperative or noncompliant behavior” (e.g., cutting or being late for class); Level 2 infractions considered “disorderly behavior” (e.g., using profane language or possession of cigarettes); Level 3 infractions considered “disruptive behavior (e.g., engaging in cheating, vandalism, or gang-related behavior); Level 4 infractions considered “aggressive or injurious/harmful behavior” (e.g., possessing controlled substances or engaging in harassing, intimidating, and/or bullying behavior); to Level 5 infractions considered “seriously dangerous or violent behavior” (e.g., starting a fire, possessing or selling weapons or controlled substances, or attempting to inflict or inflicting serious injury on students or others). Over the period observed in this study (2011–12 through 2018–19), approximately 21% of all reported infractions at the middle and high school level were Level 1 and 10% were Level 2; the most prevalent forms of infractions were at Level 3 (33%) and Level 4 (32%) while only about 4% of reported infractions were Level 5.
As per NYCPS discipline code, different levels of infractions are eligible for distinct forms of disciplinary response. Subsequent to 2012, suspensions were no longer assigned for Level 1 and 2 infractions due to previous suspension reforms taking effect (Craig & Martin, 2023). Among infractions that led to a suspension at the middle and high school level over the period observed in this study, 34% were Level 3, 54% were Level 4, and 12% Level 5. Level 3 infractions and above could result in one of two types of suspension: a “principal’s suspension,” lasting one to five school days, or a “superintendent’s suspension,” lasting more than five school days and assigned with approval from the district-wide Office of Safety and Youth Development (OSYD). Regardless of the suspension type, students are to receive some form of alternative instruction, either at a different location within their school for a “principal’s suspension” or at an alternative learning center or designated buddy school for a “superintendent’s suspension.” Thus, there is no traditional “out-of-school” suspension option in NYCPS, as students are not entirely excluded from the schooling system. Nevertheless, the quality of instruction provided during suspensions may vary and serve as an imperfect substitute for standard classroom instruction.
Beginning in 2015, the NYCPS began to promote restorative justice and social-emotional learning (SEL) approaches to address nonviolent forms of misbehavior in lieu of tactics involving exclusionary discipline (Rodriguez & Welsh, 2022). As part of this citywide initiative, the city initially allocated $47 million dollars toward staff training on restorative disciplinary practices and mental health supports. In addition, with the appointment of former Schools Chancellor Richard Carranza in 2018, antibias training for teachers and administrators as well as culturally responsive education have become further embedded in the fabric of NYCPS education policy reforms. Several other school discipline policy changes were implemented in tandem with these initiatives, including a firm requirement that principals obtain written approval from NYCPS central office to suspend a student for infractions considered of lower severity. In light of implemented suspension reforms, the number of infractions and suspensions declined precipitously between 2011–12 to 2018–19 (Rodriguez & Welsh, 2022). Nevertheless, several recent studies have mixed evidence of these reforms in reducing disparities (as opposed to the overall prevalence) in exclusionary disciplines. While some studies have shown evidence that behavioral interventions such as restorative justice programs have reduced differential processing of suspensions among Latinx students (Davison et al., 2022; Nishioka et al., 2021), others have raised concerns about whether these policy and programmatic changes are beneficial for improving the disciplinary experiences of Black students (Cruz et al., 2021; Davison et al., 2022; Nishioka et al., 2021; Welsh, 2023c).
Data and Methods
The study relies on administrative and teacher survey data for NYCPS maintained by the Research Alliance for New York City Schools (RANYCS) and represents longitudinal information across the 32 geographic school districts in NYCPS from 2011–12 to the 2018–19 school years for middle schools (grades 6 through 8), high schools (grades 9 through 12), and schools serving any combination of grades 6 through 12 (i.e., schools with grades K through 8, K through 12, and 6 through 12). We restrict our analysis to schools serving grades 6 through 12, as these are the grade levels for which discipline records are comprehensively available through RANYCS. In addition, NYCPS maintains entirely separate discipline codes that distinguish between students in grades K through 5 from those in grades 6 through 12.
Table 1 provides summary statistics for the full sample of middle and high schools in NYCPS along with conditional statistics for schools with high suspension rates (top tertile of distribution) and low suspension rates (bottom tertile). The school-level analytic sample for this study consists of 1,091 unique middle and high schools across all years—approximately 29% are traditional middle schools, 44% traditional high schools, and 27% are schools serving other combined grade configurations. Schools with higher suspension rates are staffed by teachers with lower years of teaching experience as well as a larger share of Black teachers. On average, these schools serve a higher share of students who are Black and Latinx, from low socioeconomic status backgrounds, living in temporary housing, and receiving special education services with discernibly lower proficiency in math and ELA at the middle school level and rates of graduation at the high school level.
Summary Statistics, NYCPS Middle and High Schools (2012–2019)
Note. SD = standard deviation.
Includes schools in the top tertile of distribution of school-level suspension rates in a given year.
Includes schools in the bottom tertile of distribution of school-level suspension rates in a given year.
Calculated among middle schools and combined secondary schools only.
Calculated among high schools and combined secondary schools only.
The period under study marks a shift in the city’s mitigation of exclusionary suspension policies, which is reflected by a notable drop in suspension across middle and high schools in NYCPS. As shown in Figure 2, roughly 6.7% of students were suspended in 2011–12 compared to 4.2% in 2018–19. Interestingly, the percentage of students receiving an ODR remained comparatively steady over the same period—approximately 12%—as did the gap in suspension rate between Black students and other racial/ethnic groups. By 2019, Black students still had the highest rate of exposure to suspensions (6.4%), followed by a comparable rate between Latinx students (3.6%) and White students (3.4%), and trailed by Asian students (1.6%).
Discipline rates over time, overall and by student race/ethnicity.
Measures
The key data elements for the study are drawn from several sources, including student discipline records; school administrative records of students and education staff; and teacher responses to the annual NYC School Survey. For the years under study, characteristics retrieved from administrative data sources (i.e., nonsurvey information) had negligible missing information for students, staff, and schools in the analytic sample. At the individual student-year level, except for the foreign-born indicator, which was missing for 0.8% of all student-year records, all other student characteristics were missing under 0.1% across all student-year records. At the individual staff-year level (before aggregating the data to the school level), all staff characteristics (teacher and principal) were missing under 0.1% across all staff-year records. As for the teacher survey data, 80 to 83% of teachers responded to the NYC School Survey overall in any given year while three-fourths of all schools had at least 74% of teachers respond each year. Below, we present a more detailed description of the key measures, the empirical strategy used for this study, and specification checks conducted, including those to account for survey nonresponse.
School Discipline Rates (Overall and by Student Race/Ethnicity)
The study considered two distinct forms of exclusionary discipline for its outcomes of interest: ODRs and suspensions. An ODR involves the reporting of an infraction (normally reported by a teacher) that might entail the removal of the involved student from their classroom as they await a disciplinary decision. An ODR may result in a suspension, though other disciplinary consequences aside from a suspension (or no further disciplinary outcome at all) might arise as a result of an ODR.
The available discipline data contain information on reported infractions (i.e., ODRs) and suspensions. To measure the prevalence of exclusionary discipline outcomes at the school level, we calculate the risk of exposure to exclusionary discipline as the number of unique students receiving an ODR or suspension divided by the number of students enrolled within a school during that year.
In addition to overall school-level discipline rates, we calculate subgroup-specific ODR and suspension rates by disaggregating based on student race/ethnicity. For students, race and ethnicity are reported by the parental guardian on a “Student Ethnic and Race Identification” form, which is collected by NYCPS when a student first registers for enrollment in the system as per federal law. The current version of the form allows the parental guardian to select whether the student is “Hispanic, Latino, or of Spanish origin” as well as one or more of the following five racial groups: (1) “American Indian or Alaskan Native,” (2) “Asian,” (3) “Native Hawaiian or Other Pacific Islander,” (4) “Black,” and (5) “White.” In prior years, race/ethnicity information was collected more crudely by forcing the option to select only one category (with Hispanic being included as one category along with the five other current race categories). To maintain consistency across the available panel, the study operationalizes mutually exclusive subgroups for non-Hispanic Black, Asian, and White students and Hispanic/Latinx students. American Indian or Alaskan Native and Native Hawaiian or Other Pacific Islander were combined into a single category due to the low percentage of students and staff identified with those categories in the student enrollment and staff records (under 1.9% of students and 4.0% of teachers).
For discipline rates specific to a racial/ethnic student subgroup, we calculate the ODR and suspension risk index for Black (non-Hispanic) and Latinx (Hispanic) students separately as the number of unique students receiving an ODR or suspension in a group divided by the total number of students in that group within a school year. These racial/ethnic-specific measures are able to better assess whether the prevalence of ODRs and suspensions is significant specifically for Black and Latinx students, a matter of concern due to the racialized hierarchies that have historically marginalized both student groups in the US public education system (Skiba et al., 2011)
Teacher Demographics and Qualifications
Human resources data provide demographic and background information for all full-time classroom teachers employed by NYCPS. Average years of teaching experience as well as the percentage of teachers with a particular education level (bachelor’s, master’s, or more than a master’s degree) serve as school contextual attributes pertaining to teacher qualifications. To identify the demographic composition of teaching staff within a school, we consider the percentage of teachers who are Black, Latinx, Asian, or other racial/ethnic background as well as percent of teachers who are male.
Teacher Perceptions of School Climate
Drawing on the NYC School Survey, we construct measures of teacher perceptions of school climate. The survey data we use for this study span the 2011–2012 through 2018–2019 school years and represent more than 270,000 teacher-year responses. For the period under study, the survey was administered to teachers beginning mid to late February up through late March to early April, depending on the year. Due to NYCPS policy, responses are anonymized at the individual teacher level but remain linkable to other forms of data—teacher demographics and background information, student suspensions—at the school level, which serves as the unit of analysis for this study.
The NYC School Survey is administered annually to teachers and intends to capture teachers’ opinions across six broad reporting categories within an overarching framework for school improvement: (1) rigorous instruction; (2) collaboration between teachers; (3) aspects for whether the school is a supportive environment; (4) effective school leadership; (5) ties between the school, families, and general community; (6) and trust between teachers, administrators, and parents. Each year, the NYC School Survey for teachers consists of anywhere between 20 to 30 Likert-scale questions; however, to extend the extant school discipline literature, which has primarily utilized cross-sectional teacher survey data to measure school climate, we utilize data from survey items that are consistently available throughout the period of study to support a longitudinal analysis—nine survey items in total, listed in Table 2. The nine items mapped onto four conceptually distinct elements based on NYCPS’s framework for the NYC School Survey: (1) one item for the subcategory “School Commitment” among teachers included under the broader “collaborative teachers” category, (2) one item for the subcategory “Safety” included under the broader “supportive environment” category, (3) six items for “Effective School Leadership,” and (4) one item for “Teacher-Teacher Trust.”
Results From Exploratory Factor Analysis (EFA) and Propensity Components Analysis (PCA) of NYC School Survey Items
Notes. Based on teacher survey responses to the NYC School Survey from 2012 to 2019 (n = 270,843). Questions were restricted to those that were consistently available on the survey throughout that time period.
Scholars have employed various frameworks to conceptualize the dimensions of school climate (e.g., Cornell & Huang, 2016; Huang & Anyon, 2020). Rudasill et al. (2018) argue that conceptual clarity is necessary for school climate construct validity, suggesting that school climate constructs should focus on any or all of the three components of school climate: a) perceptions of social interactions and relationships within the school, b) shared beliefs and values in the school, and c) the sense of safety within the school. As detailed later, we use a measure of “overall school climate” that captures these three components as well as four distinct elements of school climate based on teacher survey items: a) school commitment, which captures both perceptions of a host of social interactions (teacher-teacher, teacher-students, teacher-school administrator, teacher-parent relationships) as well as a component of shared beliefs and values in school; b) safety, which is aligned with a sense of safety or notions of maintaining order and discipline; c) effective school leadership, which is a blend of perception of teacher-school administrator interactions as well as shared beliefs and values; and d) teacher-teacher trust, which captures perceptions of teacher-teachers social interactions and relationships.
Our study relies on both data- and theory-driven approaches to construct measures of teacher perceptions of school climate. First, we conducted an exploratory factor analysis (EFA) as a data-driven approach, which identified one orthogonal dimension of variation from the school survey item responses that explained 75% of the total item variance. Visual examination of the scree plot aligned with the Kaiser-Guttman stopping criteria to suggest a “breaking point” after the first factor, as the items did not capture other meaningful dimensions. We interpret the retained factor as a measure of “overall school climate” that captures broad organizational features of schools, which include effective school leadership, safety, teacher trust, and teacher commitment to the school. As shown in Table 2, all items similarly loaded strongly onto the single factor, with all factor loadings exceeding 0.74.
As an alternative theory-driven approach to measuring school climate, we considered the conceptual framework developed and used by the NYCPS for the NYC School Survey so as to attempt to disaggregate its underlying dimensions. “Effective School Leadership” was the only framework element for which there was more than one available item. Therefore, we conducted a principal’s component analysis (PCA) with only the items related to school leadership. As shown in Table 2, the results of the PCA yielded one viable component that explained 83.9% of the overall variability of the school leadership items. We acknowledge that the alternate approach has its limitations, especially as the four remaining framework elements did not arise as distinct dimensions in the EFA. However, we pursue this alternate approach to better understand which of the underlying survey items may be driving the relationship between the “overall school climate” measure (generated from the EFA) and school suspension patterns.
As individual teacher survey responses were anonymized in the data and linkable at the school level only, after calculating factor and principal component scores at the teacher-year level via EFA and PCA, we then averaged these scores to the school-year level to obtain our primary predictors. In addition to the aggregated factor (“overall school climate”) and principal component (“effective school leadership”), we aggregated individual survey responses for each of the items aligned with the “safety,” “teacher trust,” and “school commitment” framework elements to calculate the percentage of agreement within a school year (e.g., for “school commitment,” percent of teachers in a school year that usually look forward to work).
Other School-Level Information
Finally, administrative records provide data on grade configuration, enrollment, principal background information from the human resources data (gender, race/ethnicity, years of experience, degree attainment), and the demographic composition of students served—race/ethnicity, gender, foreign-born status, temporary housing status, low socioeconomic status, special education status, and English language proficiency. Gender is optionally reported as “male” or “female” by staff and—for students—by parents upon initial school enrollment, with the option made available for staff and students to request to change their identified gender on record.
In addition, other school-level information includes measures of school performance—average attendance, percent proficient in English language arts (ELA) and mathematics exams for middle schools, and graduation rates for high schools. For middle schools (grades 6 through 8), we standardize math and ELA proficiency rates within each year and calculate the average among the two subjects to construct a measure of middle school–level performance. For high schools (grades 9 through 12), the graduation rate—standardized among all schools within each year—serves as the measure of school-level performance. Finally, for schools that serve grade configurations that span traditional middle school grades 6 through 8 and high school grades 9 through 12, we construct a school-level measure of performance by calculating the average of the standardized ELA proficiency rate, math proficiency rate, and graduation rate within each year.
Empirical Strategy
Our study is concerned with the relationship between organizational characteristics—namely, teacher perceptions of school climate, school contextual attributes related to teachers, and outcomes arising from school discipline processes (ODR and suspension rates). As the primary analytic strategy, we employ linear regression methods specified to isolate within-school variation over time. In our first set of analyses, we model school-level ODR and suspension rates as follows:
where
The primary coefficients of interest in (1) are
In addition to the primary analyses described, we estimate modified versions of (1) to include interactions between the measures of perceived school climate and school-level teacher characteristics to test the interactive relationship between perceived school climate and teaching staff with school discipline patterns. The interactions tested involved interactions between the measures of perceived school climate and the school teaching staff’s (a) gender composition, (b) racial/ethnic composition, (c) degree attainment, and (d) average years of teaching experience. When estimating these models, a statistically significant estimate on the interaction term would be indicative that teacher traits differentially be associated with ODR and school suspension patterns based on teachers’ perception of school climate. We hypothesize that teacher characteristics matter more in schools where teachers have more negative perceptions of school climate since more qualified and diverse teaching staff may be better equipped to navigate challenging work settings to handle instances of student misbehavior without resorting to exclusionary disciplinary consequences.
In addition to the primary aforementioned analyses, we conduct a series of sensitivity checks to assess the robustness of our findings. First, as teacher perceptions of school climate serve as key predictors for the study, we are naturally concerned about the extent to which survey nonresponse may affect the study’s results. While overall teacher response rates to the NYC School Survey are generally high from year to year at the middle and high school levels (above 80%), we reestimate model (1), adjusting for the school-level response rate as a covariate.
A second concern is that any significant relationships observed may be confined to infractions of a specific severity level. That is, perhaps teacher characteristics are only strongly associated with discipline rates for the most severe infractions (i.e., Level 5). We, therefore, explore how results change when exclusively examining ODR and suspension rates for infractions of a given level of severity as the dependent variable of interest through a series of sensitivity checks.
Lastly, another challenge is the presence of a reverse temporal relationship between measures of perceived school climate and school discipline patterns. Specifically, teachers’ ratings of safety (e.g., “order and discipline are maintained at my school”) may be a function of prior-year discipline rates, in which case we would expect both measures to covary within schools over time. Indeed, existing empirical evidence suggests student (mis)behavior would affect teacher reports of school climate and their general well-being (Spilt et al., 2011). Similar empirical challenges have been common in prior literature utilizing observational measures of teaching and school effectiveness (e.g., Kane et al., 2011). For this reason, readers should be cautious about interpreting our results as causal, given the possibility of a reverse temporal relationship. We explore this issue further when presenting a set of sensitivity checks that includes reestimation of a modified version of equation (1) but with
Results
To consider the role of teachers in shaping school discipline outcomes, we focus on examining school contextual attributes pertaining to teaching staff as well as school-level teacher perceptions of school climate. Table 3 provides regression estimates predicting the relationship between overall school climate as perceived by a school’s teaching staff, teacher demographics (gender and racial/ethnic composition), and teacher qualifications (degree level and years of teaching experience) with school ODR and suspension rates. We report results from the combined total suspensions (inclusive of both superintendent and principal suspension types) given that the overwhelming majority comprise principal suspensions (77.9%); however, we disaggregate between the two types of suspensions in auxiliary analyses, which are qualitatively similar and available upon request. The results indicate that more positive teachers’ perceptions of school climate are associated with lower rates of ODRs and suspensions. As shown in columns (1) and (2), respectively, a 1 SD increase in the “overall school climate” measure is associated with a roughly .65 percentage point decline in ODR rates and .36 percentage point decline in suspension rates when controlling for other observable school characteristics. This shift represents about a 5 percent decline in the baseline average of ODR observed among middle and high schools across the studied period and a 7 percent decline in the baseline average rate of suspensions.
Estimates Predicting Overall School Discipline Rates
Note. Standard errors reported in parentheses are clustered at the school level. ODR = Office discipline referral. “Overall school climate” derived from exploratory factor analysis (EFA) of NYC School Survey items; “effective school leadership” derived from principal components analysis (PCA) of NYC School Survey items on school leadership; “safety” represents percent of teachers within school who agreed “order and discipline are maintained”; “teacher trust” represents percent of teachers within school who agreed “teachers trust one another”; and “school commitment” represents percent of teachers within school who agreed with the statement “I look forward to coming to work every day.” Std = measure standardized within a year. Covariates not shown include the remaining characteristics listed in Table 1.
p < 0.05; **p < 0.01; ***p < 0.001.
Table 3 columns (1) and (2) also indicate that the characteristics of a school’s teaching staff are significantly associated with suspension rates. Our results are congruent with previous studies that highlighted that teacher racial/ethnic diversity is associated with lower rates of exclusionary discipline (Hughes et al., 2020). Based on model estimates, a 1 percentage-point increase in the percentage of Black teachers is associated with about a .05 percentage-point decrease in ODR rates and .04 percentage-point decrease in suspension rates. We also find evidence that more teacher experience is associated with lower suspension rates. Schools staffed by teachers that, on average, have one more year of teaching experience are predicted to have lower ODR and suspension rates by .37 and .16 percentage points, respectively, which translates to about a 9 to 10 percent reduction off baseline ODR and suspension rates for 1 SD increase in the average years of teaching experience within a school. Overall, these findings suggest that with respect to school discipline, teachers matter beyond their racial/ethnic diversity, as teaching experience appears to be a significant factor as well.
Also in Table 3 columns (3) and (4), we disaggregate the elements of overall school climate into the underlying four measurable components: effective school leadership, safety, teacher-teacher trust, and teacher school commitment. As before, the percentage of teachers who are Black and teaching experience within a school are significantly related to school discipline patterns when including the disaggregated school climate measures. With disaggregated school climate measures included, model estimates indicate that as more teachers agree that there is order and discipline in their school (“safety”) and the more they look forward to coming to work (“school commitment”), ODR and suspension rates are predicted to decline. While “teacher-teacher trust” was not significant across the models, model estimates predicted higher ODR and suspension rates when teachers rated school leaders as more effective. The “effective school leadership” measure was based on survey items related to awareness (e.g., “principal knows what’s going on in my classroom”), school vision (e.g., “principal communicates a clear vision for this school”), and instructional leadership (e.g., “principal participates in instructional planning with teachers”). The association between these indicators and elevated discipline rates prompts critical questions concerning the nuances of school leadership normatively considered effective, a point that we return to in the forthcoming discussion.
An important consideration is whether school contextual attributes pertaining to teachers and school climate, as perceived by teachers, may signal improved agency to reduce exposure to exclusionary discipline among Black and Latinx students. Table 4 presents model estimates predicting Black and Latinx ODR and suspension rates. For Black and Latinx students, an increase in teachers’ perception of “overall school climate”—columns (1) through (4)—is associated with a decline in ODR and suspension rates. However, the magnitudes are well over 50% larger when predicting Black student ODR and suspension rates, which suggests that teacher perceptions of school climate are a substantially stronger predictor of disciplinary outcomes of Black students. When we disaggregate the overall school climate into the underlying measurable elements, we find that the patterns are being driven by teachers’ perception of school safety for both Black and Latinx ODR and suspension rates as well as teachers’ commitment to their school for suspension rates for Black students. The years of experience of teaching staff within a school are also significantly related to ODR and suspension rates for Black students in particular, such that a higher average year of experience among teachers within a school is associated with lower ODR and suspension rates for Black students. The results are mixed for teacher racial/ethnic diversity; however, a higher percentage of teachers who are Black within a school is consistently associated with lower suspension rates for Black students regardless of how school climate was operationalized. By contrast, Latinx students’ suspension rates were not associated with the percentage of teachers who identify as Latinx in either model; however, a greater percentage of Black teachers was associated with lower suspension rates for Latinx students, suggesting that the presence of a greater share of Black teachers may translate to other marginalized racial/ethnic student groups.
Estimates Predicting Black and Latinx Student Discipline Rates
Note. Standard errors reported in parentheses are clustered at the school level. “Overall school climate” derived from exploratory factor analysis (EFA) of NYC School Survey items; “effective school leadership” derived from principal components analysis (PCA) of NYC School Survey items on school leadership; “safety” represents percent of teachers within school who agreed “order and discipline are maintained”; “teacher trust” represents percent of teachers within school who agreed “teachers trust one another”; and “school commitment” represents percent of teachers within school who agreed with the statement “I look forward to coming to work every day.” Std = measure standardized within a year. Covariates not shown include the remaining characteristics listed in Table 1.
p < 0.05; **p < 0.01; ***p < 0.001.
Finally, we investigate whether teachers’ perception of school climate is more strongly related to school discipline outcomes when considered in tandem with observable teacher attributes. We estimate separate modified versions of the school fixed effect models from Tables 3 and 4 to test a number of interactive relationships, including interactions between the measures of perceived school climate and (a) teaching experience, (b) teacher education, (c) teacher racial/ethnic composition, and (d) teacher gender. The estimated interactive relationships between the measures of perceived school climate and teacher characteristics were all null at the 95% confidence level (results available in Appendix Tables A.1–A.3), suggesting that teachers’ perceptions of school climate and teacher characteristics are independently and noninteractively related to school ODR and suspension patterns.
Sensitivity and Falsification Checks
We conducted a series of checks to assess the sensitivity and robustness of our results. First, we re-ran model (1) where the dependent variable was defined as the percent of students receiving an ODR for a given severity level of infraction (e.g., percent of students receiving an ODR for a Level 3 infraction) as well as the percent of students suspended for a given severity level of infraction (e.g., percent of students suspended for a Level 3 infraction) to determine whether the presented findings are confined to certain kinds of infractions. When examining suspension, we focused on the three highest infraction levels, since suspensions were no longer assigned for Level 1 and 2 infractions post-2012 (the first year observed in the studied period). For Level 1 and 2 infractions, average teacher experience was the only statistically significant factor of interest such that higher experience was associated with lower ODR rates, overall and for Black and Latinx students. For overall and Black ODR and suspension rates for Level 3 infractions and above, regression estimates were qualitatively similar in terms of direction and statistical significance compared to the main results, with the exception of the percentage of Black teachers, which was not statistically significantly related to overall and Black ODR rates for Level 3 or 5 infractions and suspension rates for Level 5 infractions. None of the key estimates were statistically significant for Latinx ODR and suspension rates for Level 3 infractions, while the estimates for school climate remained qualitatively similar for Latinx ODR and suspension rates and percent of Black teachers for Latinx ODR rates for Level 4 and 5 infractions. In terms of magnitude, statistically significant estimates were larger when predicting discipline rates for Level 3 and 4 infractions than for Level 5 infractions.
We also reestimated models adjusting for school-level teacher response rates to the NYC School Survey (results available in Appendix Tables A.4 and A.5). Estimates remained qualitatively similar in terms of direction, magnitude, and level of statistical significance.
Lastly, we conducted tests to assess the direction of the relationship between the school climate measures and school discipline, which presents a considerable challenge to the credible interpretation of the temporal order of the estimated relationships of interest. School discipline and teacher perceptions of school climate may have a complex bidirectional relationship (Huang & Cornell, 2018). More specifically, teachers’ perceptions of climate and discipline may be endogenous, as negative perceptions of school climate among teachers may arise due to high suspension or, alternatively, high suspension may arise due to poor school climate as perceived by teachers. In auxiliary analyses not presented here, we estimate models where we exchange our outcome variables, school ODR and suspension rates (overall, Black, or Latinx), and overall school climate predictor, and instead model school climate measured in time t as a function of outcome measures from time t – 1. Results from these analyses indicated that neither overall, Black, or Latinx student ODR and suspension rates in the prior year were significantly related to teacher perceptions of school climate in the subsequent year. We nevertheless estimate modified versions of model (1) with perceived school climate measured in the prior year and controlling for prior-year ODR and suspension patterns within a school (results available in Appendix Tables A.6 and A.7). Controlling for prior-year discipline patterns presents an analytic trade-off, as discipline patterns are unobserved prior to 2011–12 but doing so more precisely isolates the relationship between teachers’ perceptions of school climate and school discipline patterns in the subsequent year. A more positive perception of the overall school climate among teachers was associated with a statistically significantly lower ODR and suspension in the subsequent year, yet the magnitude of the relationship is slightly attenuated when controlling for prior-year ODR and suspension patterns. The relationship appears to be largely driven by teachers’ perception of school safety, as that remains the only measurable component of teachers’ perception of school climate that was statistically significantly associated with school ODR and suspension patterns in the subsequent year even after controlling for prior-year ODR and suspension rates. In other words, a stronger sense of safety and order among teachers within the school year appears to be significantly related to a lower prevalence of ODR and suspension in the following year. With the exception of Latinx ODR rates, higher teaching experience within a school remains statistically significantly associated with lower ODR and suspension rates, as is the percentage of Black teachers within a school but only for Black student suspension rates.
Discussion
A number of salient findings arose in our study’s interrogation of the relationship between school contextual characteristics related to teachers, school climate as perceived by teachers, and school disciplinary outcomes in middle and high schools in NYCPS. First, a more positive overall school climate, as perceived by teachers, is associated with lower overall, Black, and Latinx ODR suspension rates. Second, we find that a larger share of Black teachers is associated with lower rates of suspension, overall and specifically for Black and Latinx students, which supports the notion concluded in prior studies that teacher diversity matters for school discipline. As prior studies show, Black students tend to be suspended less when exposed to more Black teachers (Holt & Gershenson, 2019; Lindsay & Hart, 2017; Shirrell, 2021). However, findings from this study suggest the benefits of Black teaching staff may also extend to the disciplining of Latinx students—this could be reflective of a level of heterogeneity among Latinx students not captured by the racial/ethnic indicators available for this study—namely congruence among students identifying as Afro-Latinx. Third, we find that teacher experience also matters for school discipline, especially for overall and Black suspension rates, which is congruent with prior studies highlighting differences in disciplinary outcomes between novice and veteran teachers (Wang et al., 2022; Williams III et al., 2018; Welsh, 2023b). For instance, Losen et al. (2014) found that novice teachers were more likely to suspend African American students. Classroom management is a major challenge for novice teachers (Wolff et al., 2015), which may also partly explain the significance of teacher experience while culturally responsive and restorative approaches applied to disciplinary issues may be more prevalent among teachers of color (Achinstein & Ogawa, 2012). A few plausible alternative explanations may account for the significance of teacher experience and race/ethnicity, such as the sorting of more experienced teachers into schools with lower suspension rates; however, many of these patterns are robust when controlling for prior-year school discipline patterns.
We also find that a stronger sense of teacher school commitment (I usually look forward to coming to work) is also associated with lower prevalence of overall suspension and Black suspension rate, even after controlling for teachers’ perceived sense of safety within the school. Another likely significant factor in looking forward to coming to work is the working conditions of the school. Prior work has found that teacher working conditions, including a school’s culture, principal leadership, and relationships among colleagues, matter a great deal for teacher satisfaction and, in turn, teacher retention and student academic growth (Johnson et al., 2012). Moreover, teacher workload and teacher perceptions of student discipline in school are factors also previously found to be closely associated with teacher job satisfaction (Toropova et al., 2021)—our findings extend this research base by highlighting the link between the degree to which teachers are committed to and satisfied with their school environment and school discipline patterns. Similarly, our findings point to the importance of perceptions of order and discipline as safety was associated with lower suspension rates. We posit that teachers’ sense of school commitment and safety are largely a function of a host of social interactions and relationships in schools such as among teachers as well as between teachers and school leaders, teachers and students, and teachers and families. Forming professional learning communities among teachers has been emphasized as a powerful strategy for enhancing teacher school commitment in prior research (Hausman & Goldring, 2001). Additionally, trust among teachers, as well as between teachers and parents, has been found to relate to teachers’ job satisfaction (Van Meale et al., 2012), as has the degree of perceived parental involvement at home and teacher-parent contact (Li & Hung, 2012).
Surprisingly, our study finds that aspects of school leadership (as perceived by teachers) are nevertheless associated with increased suspension rates within schools. This analysis was based on survey items like “School leaders communicate a clear vision for this school,” “The principal is an effective manager who makes the school run smoothly,” and “The principal knows what’s going on in my classroom.” This raises important questions about aspects of school leadership normatively considered effective with respect to exclusionary discipline. For instance, to what extent do normative qualities of effective leadership reflect a principal’s ability to enhance teacher accountability and autonomy when handling instances of student behavior without resorting to exclusionary discipline? It is certainly possible that more effective principals tend to intervene in disciplinary issues in ways that reduce classroom-managed infractions but increase office-managed infractions, thus leading to a greater number of suspensions. Recent studies have highlighted “disciplinary tensions” that exist in the relationships and interactions between teachers and school leadership that may be reflected in teachers’ perceptions (Welsh, 2023a). While our results motivate further inquiry into the role of school leaders in the disciplinary process in schools, it is important to note the significance of estimates on the “effective school leadership” measure was not robust in models controlling for prior-year discipline patterns—this may be suggestive of strategic assignment of principals viewed as effective in the normative sense in schools with history of high exclusionary discipline activity.
The findings from this study should be interpreted in light of its limitations. First, the setting and population under study can most credibly generalize to racially diverse urban school systems and middle and high school contexts. Second, the study’s results should be interpreted as correlational, as there remains a possibility of unmeasured or unobserved confounding factors. Specifically, the study centers its analysis at the organizational or school level; thus, it is unable to discern how individual teacher perceptions and traits interplay to affect student discipline outcomes at the individual level. Third, forms of exclusionary discipline beyond ODRs and suspensions remain unobserved in the available administrative data and beyond the scope of the current study, though they certainly may manifest in schools, particularly for lower levels of infractions—a topic of concern that should be addressed in future research. Fourth, our study was restricted to teacher survey items that were consistent across the available data panel; this results in a loss of dimensionality in the school climate construct, as the measures used in this study do not pertain to various aspects of perceived school climate such as teacher-parent relationships and teacher-student relationships, which prior research highlights are salient aspects of school climate that are linked to the school discipline outcomes (e.g., Rodriguez & Welsh, 2024). Lastly, the concern remains that we may have in part identified how discipline patterns impact teachers’ perceptions of school climate and composition of teaching staff in schools rather than how teacher perceptions of school climate and teacher characteristics affect discipline outcomes. We suggest, however, that these results confirm that teachers and their perceptions are closely linked with exclusionary discipline practices in schools, which nevertheless presents a number of implications for teacher education and educational research, the most salient of which we enumerate in what follows.
Implications for Education Policy and Practice and Future Research
The findings hold several implications for education policy, practice, and research. First, the significance of teacher experience for school discipline hints that preservice experiential learning and in-service professional development may be an integral part of reducing racial inequality in suspensions. We find that the presence of more experienced teachers is associated with lower rates of suspension, which underscores the need to target teacher education and professional development focused on discipline, relationship building, classroom management, and cultural competency to bolster novice teachers’ capacities to address student behavioral issues. Experiential learning and professional development may function as crucial channels through which preservice and novice teachers can develop well-needed experience in managing classrooms and addressing disciplinary issues. Despite national standards requiring clinical and field experience in diverse environments, this experience is often neglected by teacher education programs (Allen et al., 2017; Darling-Hammond, 2010; Hollins, 2015).
The significance of teacher experience in our study’s findings also behooves educational policymakers and administrators to consider more robust on-the-job support for novice teachers to enhance their capacity to address disciplinary issues. Policymakers may consider providing coaching in tandem with professional development that place emphasis on school discipline-relevant skills and strategies to proactively address common struggles encountered in the disciplinary process. Professional development may prioritize building relationships with students given its importance for both instructional quality and school discipline (Gregory & Korth, 2016; Kwok, 2019; Welsh, 2023b) while reinforcing classroom management strategies.
Third, the findings lend support for initiatives by policymakers that have advanced improving school climate as a way to reduce school discipline disparities (U.S. Department of Education, 2017) and interventions that aim to enhance teacher self-efficacy and address emotional exhaustion (Eddy et al., 2020). In the wake of the COVID-19 pandemic, teacher support has acquired even greater significance as novice and experienced teachers navigate unprecedented challenges (Carter Andrews et al., 2021a). This points to the growing role that teacher preparation can play in developing socio-emotional skills and resilience among teachers. Indeed, Carter Andrews et al. (2021b) similarly argue, “[j]ust as teacher education programs depend upon practicing teachers for the quality and integrity of their programs, they have a responsibility for playing a role in ongoing support for these educators” (p. 269). Our results underline the importance of teachers’ school commitment as a factor that is significantly associated with the prevalence of exclusionary discipline practices within schools. In this regard, a possible avenue to reduce the use of exclusionary discipline in schools is creating a more positive school climate for both teachers and students. This may include establishing school norms that are conducive to teachers’ socio-emotional well-being so they are better able to respond to and de-escalate situations involving student misbehavior. While the issue of how to support teacher wellness is an area lamentably understudied in the field of education research, a few notable practices for districts and school leaders to consider include promoting favorable organizational structures and supports that enhance instructional autonomy and work-life balance for teachers (Dizon-Ross et al., 2019; Steiner et al., 2023), mindfulness techniques (Spilt et al., 2011), opportunities to develop enhanced teacher-student relationships (Hwang et al., 2017; Milatz et al., 2015), and teacher professional learning communities (Owen, 2016).
Our study also underscores the significant relationship between the racial/ethnic composition of teachers and school suspension patterns. The findings support the current efforts made by the NYCPS and other schooling systems to diversify the teaching workforce and retain educators of color as a way to lessen disciplinary disparities. Yet, given the immediate supply constraints of teacher candidates of color, which have been compounded by the historical displacement of Black public school educators in the wake of the U.S. Supreme Court’s 1954 Brown ruling, should not be the single lever through which districts aim to do so (Fultz, 2004; Smith et al., 2023). Moreover, the salience of teacher diversity for school discipline also compels a reexamination of teacher educators as well as the candidates entering into teacher education programs and provides a boost for “grow-your-own” programs that develop teachers who are familiar with the cultures and communities of students within their schools (Valenzuela, 2017).
Finally, several directions for future research emerge from our study’s findings. To build on findings of current and past research, it is imperative that future research explore the link between teachers and school discipline outcomes from an intersectional perspective, taking into account various student characteristics alongside race and ethnicity, such as gender and disability. As prior studies have demonstrated, examining school discipline from an intersectional lens can continue to enhance the discourse by identifying students most at risk of facing exclusionary discipline and in need of behavioral support (Annamma, Connor, & Ferri, 2013; Annamma, Anyon, et al., 2019; Morris, 2016). Secondly, there remains a need for more research at the nexus of teacher education and school discipline given the central role teachers play in the disciplinary process. It is a fitting moment to re-survey the role and prominence of classroom discipline and management in teacher education curriculum, especially as there is relatively limited empirical research addressing the efficacy of specific strategies for classroom management (Djigic & Stojiljkovic, 2011; Kwok, 2019). Future studies may examine the association between the strategies of classroom management and the prevalence of and disparities in exclusionary discipline to better inform curricular and program changes in teacher education. Findings from the current study suggest that a fruitful direction of future research is to delve deeper into how school contextual factors—namely teacher diversity and experience—may contribute to the agency in the disciplinary process that disrupts discipline disparities. Further research on the disciplinary practices of teachers of color and experienced teachers, including their interactions with students, parents, and school administrators, would be beneficial for teacher education to identify promising relational strategies and challenges encountered in the realm of addressing student behavioral issues. Thirdly, the current study’s results highlight a key area for future inquiry on the drivers of teachers’ perceptions of school climate, particularly pertaining to perceptions around school safety and discipline and the degree to which they may be a function of teacher biases versus school processes and school leadership. Finally, the analysis presented herein considers the disciplinary process as one that is encapsulated within the school environment; however, we encourage future research that considers systems and stakeholders external to the immediate school environment that affect school discipline policy and outcomes. Moving forward, researchers, policymakers, and practitioners should endeavor to develop a more comprehensive understanding of the interrelated entities, resources, and institutions situated in a broader social ecosystem comprising a child’s home and community that affect whether and how they experience exclusionary discipline in schools.
Supplemental Material
sj-docx-1-ero-10.1177_23328584241263860 – Supplemental material for School Climate, Teacher Characteristics, and School Discipline: Evidence From New York City
Supplemental material, sj-docx-1-ero-10.1177_23328584241263860 for School Climate, Teacher Characteristics, and School Discipline: Evidence From New York City by Luis A. Rodriguez, Richard O. Welsh and Chelsea Daniels in AERA Open
Footnotes
Acknowledgements
The authors gratefully acknowledge the Research Alliance for New York City Schools and New York City Public Schools for their cooperation and assistance in providing the necessary data to conduct these analyses. The views expressed herein are those of the authors and should not be attributed to their institutions or data providers. Any and all errors are attributable to the authors.
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.
Authors
LUIS A. RODRIGUEZLuis A. Rodriguez is an assistant professor of educational leadership and policy studies at New York University. His research focuses on the intersections of educator quality, retention, and diversity; school discipline; and school climate.
RICHARD O. WELSH is an associate professor of education and public policy in the Department of Leadership, Policy and Organizations at Peabody College of Vanderbilt University. Dr. Welsh’s research focuses on understanding and transforming inequality in K–12 education, and his scholarship interests include urban education, the economics of education, school discipline, student mobility, the politics of education, and research-practice partnerships.
CHELSEA DANIELS is a PhD candidate in the Department of Sociology at New York University. She uses quantitative methods to explore racial and socioeconomic disparities in education with a special focus on the factors that shape students’ and families’ decision-making and educational trajectories.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
