Identifying the Mental Health Profiles of Black Adolescents Who Experience School Policing and School Discipline: A Person-Centered Approach

Abstract

Objective:

Educational environments that are structured by race perpetuate poor mental health for Black adolescents. This empirical relationship is pronounced when it is examined through Racial Battle Fatigue theory, which provides a framework that links educational environments and poor psychological health of Black students. School police have major effect on Black adolescents educational and health experiences and trajectories. The purpose of this person-centered study was to assess the risk of Black students’ depressive symptoms who were stopped by school police, saw other students stopped by school police, or experienced school discipline.

Method:

Black student youths from the Fragile Families Study Year 15 wave (N=1,601) was used to conduct a latent class analysis to identify subgroups of school policing (i.e., being stopped by school police, seeing other students stopped by school police) and school discipline on the distal outcome of depressive symptoms. Covariates included demographic and school attachment.

Results:

We identified 4 distinct Black student subgroups: (1) unscathed [no school policing or discipline reported]; (2) school disciplined; (3) combined school policing; and (4) school policed [arrested by police]. Each subgroup had an incremental increase in mean depressive symptom scores. Compared to the unscathed subgroup each subgroup also had lower school attachment.

Conclusion:

This study disrupts the notion that education environments are a social determinant of health and a great equalizer. This study critically exposes how educational institutions, complicit with school policing, are associated with racism-related mental health conditions of Black youths.

Introduction

The literature concerning the relationship between educational environments and mental health outcomes of Black adolescents is still emerging. Empirical studies find that educational environments with disproportionate numbers of White students are related to the increased depressive symptoms of Black students; as the number of White students increases, so do the levels of depressive symptoms for Black students.^1,2 Further, these depressive symptoms, borne of inequitable educational environments, persist into adulthood.² Indeed, racially arduous experiences perpetuated by educational institutions negatively affect the mental health of Black youths, but the literature is scant. Hurst and colleagues³ provide a review that posits the adverse mental health effects of school policing towards Black students. These authors discuss the history of over-policing enforcement tactics against Black communities. Furthermore, these scholars identify an association between the legacy of egregious state-sanctioned policing against Black communities and psychological health outcomes such as fear, paranoia, depressive symptoms, and Racial Battle Fatigue of Black students. Racial Battle Fatigue (RBF) is the psychological and physiological responses in Black people when coping with institutional and structural racism replete in educational environments.⁴ Their contribution establishes a framework for this burgeoning catalog of research, as their paper intends to spur empirical work regarding the relationship between school policing and the mental health outcomes of Black people. Our study builds on the seminal work of Hurst and her colleagues³ by identifying the mental health profiles of Black adolescent high school students stopped by school police and received school discipline.

“Zero-Tolerance” education policies began in the 1980s and 1990s, which propelled increases in school discipline and school policing in campuses across U.S. schools.^5–7 “Zero-tolerance policies refer to policies that require school officials to apply predetermined consequences, such as suspension or expulsion, regardless of the situational context, mitigating circumstances, or the seriousness of the offense.”⁸ Among these zero-tolerance, ‘get tough’ policies was the Gun Free Schools Act of 1994.^9–12 These policies were adapted from the criminal justice system which called for swift and harsh punishment, usually suspension or expulsion, for students with no track record of infractions.¹³ Before zero-tolerance discipline policies of the 1990s, exclusionary school discipline was a last resort for egregious cases like drug or gun possession or physical violence.^6,13–15 A result of these harsh discipline practices of the 1980s and 1990s, children of color are suspended or expelled more than their White counterparts for the same offenses.^15–18

In addition to increases in school discipline rates, zero-tolerance policies also increased the presence of school police in U.S schools. From the 1950s to the 1990s, the number of school police rose to 17,000.¹⁹ An exact amount is unknown, but a recent study estimates that over 20,000 school police officers in the U.S. as of 2015.²⁰ Even though zero-tolerance policies and school police presence was used to ensure student safety,^21,22 this goal was largely not achieved.^6,23 Police presence in schools leads to two types of enforcement scenarios. The first scenario is where school police are the enforcers of mundane discipline issues that educators addressed in yesteryear.²⁴ The second discipline scenario is where school police are co-enforcers with educators and are present when educators interrogate students.²⁵ This passive co-enforcement scenario allows school police to gather self-incriminating information from students, bypassing their constitutional Miranda rights (for review, see U.S. Supreme Court case Miranda v. Arizona, 1966).²⁵ Both of these enforcement tactics are associated with racial inequities in education,^5,6 which increases interaction with the juvenile just system, pushing Black students into the “school-to-prison pipeline” (STPP).⁶

In aligning studies on the effects of educational inequities, the depressive symptoms of Black adolescents are high when the racial composition of the student body becomes increasingly White.^1,2 Applying the framework provided by Hurst and colleagues³ suggests that the educational inequities propelled by school policing are associated with poor mental health outcomes of Black students. Therefore, being stopped by school police may increase the odds of depressive symptoms of Black students, which is consistent with the tenets of RBF.²⁶

Theoretical Framework

Underpinning this research study is RBF theory which undergirds the relationship between educational environments, replete with racism, and racism-related health outcomes of Black students. Further, this theory explains the psychological and physiological responses in Black people when coping with institutional and structural racism commonly found in educational environments.⁴ The application of RBF theory raises understanding of how educational environments place Black students at risk of harmful psychological and physiological health outcomes.²⁷ According to RBF, Black students report psychological and physiological health problems like frustration, stress, rumination, hopelessness, anxiety, or fear.^4,26,27 Specifically, RBF suggests that racism-related stress manifests in certain mental health symptoms such as: stress, frustration, shock, anger, disappointment, resentment, anxiety, helplessness, hopelessness, and fear.⁴ Our study applies RBF theory to understand better the relationship between school policing and depressive symptoms of Black students. Thus, disrupting dominant narratives about the mental health of Black students places the onus of education and health on historically white institutions and instead of Black communities.

Purpose

Indeed, more anti-racist research is needed to center the educational and health experiences of Black adolescents as they navigate educational settings, which are historically structured by race and racism. This paper seeks to identify the nexus between school policing, school discipline, and depressive symptoms of Black adolescents. We hypothesize that there would be nuanced subgroup differences between Black adolescent high school students and levels of depressive symptoms by reported school policing (i.e., whether they were and/or saw another student on campus being stopped by school police) and school discipline. We further hypothesize that these person-centered profiles would vary by gender.

Method

Materials

To identify high school student subgroups of school policing and discipline and depressive symptoms, we used latent class analysis (LCA), a person-centered analysis on the Fragile Families dataset. Data for the present study came from Year 15 from the Fragile Families and Child Wellbeing Study (FFCWS), a large, population-based cohort study beginning in 1998 and is presently ongoing. The national random sample is from 77 large U.S. cities with populations over 200,000 people in 1994, between 1998 and 2000.^28,29 This study followed a cohort of new parents and their children, providing unparalleled data and information about conditions and contexts surrounding unwed partners and families. Data was collected from 4,700 births (3,600 non-marital and 1,100) in 75 different hospitals in various U.S. cities.²⁸ Notably, this study is not without flaw; the intent of the FFCWS was to understand unwed couples and targeting communities of color for this information stems from a deficit-based perspective about these communities.

Latent Variable Indicators

School policing had two categories. The first was the student report of being “stopped by police at school”, categorized as yes or no. The second school policed variable was if student had “ever seen someone stopped by police in your school?”, which was categorized as yes or no. School discipline was based on student report of “ever been suspended or expelled in past 2 years?” and was categorized as yes or no. Sex/gender of adolescents was categorized as male and female.

Distal continuous outcome.

Depressive symptoms were scored using a 4-point scale [strongly disagree=0, somewhat disagree=1, somewhat agree=2, and strongly agree=3] using five questions selected by the Fragile Families et al.³⁰ from the Center for Epidemiologic Studies Depression (CES-D) scale based on participant feelings in the past week: (1) you felt that you could not shake off the blues, even with help from your family and your friends; (2) you felt sad; (3) you were happy; (4) you felt life was not worth living; and (5) you felt depressed. The CES-D “you were happy” item is positively worded as such was reverse coded.

Covariates

There were three demographic covariates that indicated whether mothers were married or unmarried, level of maternal income (up to $14,999, $15,000 to $34,999, and $35,000 or more), and maternal high school education (or equivalent). There were four school-level covariates, where the last was considered an indicator of perceived discrimination:¹ (1) “I feel like I am part of my school”; (2) “I feel close to people at my school”; (3) “I am happy in school”; and (4) “Teachers in school treat the students with respect”. With a Likert scale, we also assessed “How likely [students were] to graduate from college?”. This last covariate was included to indicate to whether punitive control strategies may influence long-term academic goals and attitudes.

Data Analysis Plan

We first conducted a descriptive analysis to report the characteristics of the Black adolescent student study sample. We used LCA to identify and characterize school policed subgroups with conditional probabilities of observed school policing and discipline indicators using depressive scores as the distal continuous outcome. A comparative approach was used to select the best LCA model for interpretation. This approach compares multiple models (i.e., 1-class to 4-class solution) using the following model fit criteria: Bayesian information criterion, (BIC), sample size adjusted (SSA) BIC, Lo-Mendell-Rubin (LMR) adjusted likelihood-ratio-test (LRT), parametric bootstrapped (PB) LRT, and entropy. Entropy provides an index of reliability for separation of classes or subgroups.³¹ Once the final model was selected a multinomial logistic regression was used to examine the role of covariates in order to assess between subgroup differences. All analyses were performed in Mplus 8.4 (Muthén & Muthén).

Results

Sample Descriptives

The majority of the Black student sample had not been ever stopped by police at school (77.6%) and had never been disciplined at school (64.5%). The sample had almost equally seen others stopped by police at school. The sample was also equally male and female. The depressive symptom mean score was 2.92.

Latent Class Model

The four-class solution was selected for interpretation based on the criteria seen in Table 2. Classes or subgroups were named based on participant reports of school discipline and policing. Classes were ordered by CES-D depressive mean scores.

Table 2.

Model Fit Comparisons (N=1,601)

	Bayesian information criterion (BIC)	Sample Size Adjusted-BIC (SSA-BIC)	Entropy	Lo-Mendell-Rubin (LMR) adjusted likelihood-ratio-test (LMR LRT)	p-value	Parametric Bootstrapped (PB-LRT)	p-value
1-Class Solution	15030.3	15011.2	-	-	-	-	-
2-Class Solution	14662.9	14624.7	0.85	402.6	0.000	411.7	0.000
3-Class Solution	14630.4	14573.2	0.57	75.0	0.003	76.7	0.000
4-Class Solution	14651.3	14575.1	0.82	5.5	0.326	8.7	1.000

As seen in Table 3, Class 1 had the lowest depressive mean score and had almost equal probabilities to be male or female. Class 1 was categorized as the Not School Policed nor Disciplined subgroup as this profile had the highest conditional probability (95%) of never having been stopped by the police at school. Students in this subgroup had exclusively never been disciplined at school.

Table 3.

Four Class Solution Conditional Probabilities From the Latent Class Analysis (N=1,601)

	Class 1	Class 2	Class 3	Class 4
	Not School Policed nor Disciplined	School Disciplined	Observed School Policing and/or Policed or Disciplined	School Policed
	828	188	292	293
	51.7%	11.7%	18.2%	18.3%
Have you ever been stopped by police at school?
No	0.95	0.92	0.66	0.67
Yes	0.06	0.09	0.34	0.33
Have you ever seen others stopped by police at school?
No	0.67	1.00	0.00	0.43
Yes	0.33	0.00	1.00	0.57
Have you ever been disciplined at school?
No	1.00	0.00	0.36	0.58
Yes	0.00	1.00	0.65	0.42
Sex/Gender
Male	0.46	0.60	0.62	0.40
Female	0.54	0.40	0.39	0.60
Depressive Symptoms
Mean	1.54	2.11	2.16	7.75

Class 2 had an intermediate-low depressive mean score compared to all classes. Students had a higher conditional probability to be male in this subgroup. Students in this subgroup were categorized as School Disciplined due to having been exclusively school disciplined. The school disciplined also had the second highest conditional probability of never been stopped by police at school, as well as exclusively never seen others stopped by the police at school.

Class 3 had an intermediate-high mean depressive score compared to all classes. Students had a higher conditional probability to be male in this subgroup. Students in the Class 3 subgroup were categorized as Observed School Policing and/or Policed or Disciplined due to having been had the highest conditional probability of having been stopped by police at school, as well as exclusively seen others stopped by the police at school.

Class 4 had the highest depressive mean score and the highest conditional probability to be female. This student subgroup was categorized as School Policed due to having the second highest conditional probability of being stopped by the police at school.

The school disciplined subgroup (Class 2) had 48% lower odds of having a married mother and 61% students reporting “I am happy in school” compared to unscathed subgroup (Class 1). The combined school policed subgroup (Class 3) was found to have 60% lower odds of students reporting “I am happy in school” and 73% lower odds for students reporting “teachers in school treat the students with respect” compared to unscathed subgroup (Class 1). The school policed subgroup (Class 4) had 76% lower odds of feeling “I feel like I am part of my school,” 69% lower odds of students reporting “I am happy to be at my school,” and 74% lower odds for students reporting “teachers in school treat the students with respect” compared to unscathed subgroup (Class 1). See Table 4 for all covariates.

Table 4.

Multinomial Logistic Regression of Student Subgroups Using the Unscathed Subgroup (Class 1) as Reference (n=1,184)

	Class 2				Class 3				Class 4
		95% CI				95% CI				95% CI
	OR	Lower	Upper	p-value	OR	Lower	Upper	p-value	OR	Lower	Upper	p-value
Income	0.69	0.44	1.09	0.11	1.20	0.77	1.86	0.41	0.89	0.58	1.38	0.61
Married mother	0.52	0.28	0.98	0.04	0.67	0.37	1.22	0.19	0.66	0.38	1.17	0.16
US-born mother	0.74	0.28	1.99	0.55	2.15	0.55	8.42	0.27	2.36	0.60	9.18	0.22
Mother with high school/GED	0.86	0.56	1.32	0.49	1.46	0.95	2.24	0.08	1.14	0.76	1.72	0.53
“I feel like I am part of my school”	0.51	0.21	1.24	0.14	0.46	0.19	1.12	0.09	0.24	0.11	0.52	0.00
“I feel close to people at my school”	0.74	0.39	1.40	0.36	0.80	0.40	1.57	0.51	0.60	0.33	1.09	0.10
“I am happy in school”	0.39	0.20	0.78	0.01	0.40	0.19	0.86	0.02	0.31	0.16	0.59	0.00
“Teachers in school treat the students with respect”	0.52	0.22	1.24	0.14	0.27	0.10	0.68	0.01	0.26	0.11	0.60	0.00
Felt likely to graduate and go to college	3.27	0.25	43.09	0.37	0.76	0.18	3.23	0.71	0.48	0.13	1.74	0.27

Discussion

The purpose of the current study was to identify the depressive symptom profiles of Black high school students who experienced school policing, i.e., were stopped by school police or who observed other students being stopped by school police, as well the mental health profiles of Black high school students who experienced school discipline. Students who reported these two types of school policing had higher odds of experiencing depressive scores when compared to students that did not experience school policing or school discipline. If we uncritically assume that education is healthy, we erase essential experiences of Black students; centering the educational experiences of Black students should propel scholars to investigate countering narratives about the education-health relationship.

Our findings also revealed that all student subgroups, when compared to the unscathed students, were less likely to feel happy at school. The intermediate-high depressive score subgroup of the combined school policed in addition to not feeling happy at school also felt that teachers did not treat students with respect. The highest depressive mean score group of the school policed subgroup were found to have lower odds of feeling happy in school and felt teachers in school treated students with respect, in addition to feeling like they were part of their school. The school policed subgroup had an increased probability of being female. We gain further understanding of these mental health outcomes when making the symptomatic link to RBF, as well as “vicarious racism,” which includes students of color observing racism.³²

Although this study advances the literature, it has limitations. First, this study employs a cross-sectional design which does not allow for causal inference found in longitudinal studies. Furthermore, we recommend the need for prospective studies to ascertain and better identify causal inference. The latter studies would inform if health observation changes latent classes over time; therefore, future studies should replicate these findings using longitudinal analyses. Second, the two measures that observe being stopped by school police are binary and may provide simple information; more nuanced, comprehensive instrumentation on school policing is needed such as what educators were present, educator-police interaction, the number of referrals, time of day, the context of an altercation, geographic area of school and so on. Third, there are potential underlying mechanisms underpinning the analysis, such as institutional, systemic, and interpersonal racism—even the “mis-education” of Black people, where U.S. education is used as an apparatus to both assimilate and oppress Black students.³³

Despite the above limitations, this investigation has strengths that advance the extant literature on both RBF theory and the attributable, person-centered health profiles of Black students who are compelled to interact with school police. First, this study is nationally representative and sampled from major cities in the U.S. The second strength is that the Fragile Families Study includes both education and health measures, which allows for comprehensive, interdisciplinary health and education research. Finally, the ability of this data set to lend information to structure the education-health relationship by race and racism allows us to establish a counternarrative that exposes the deficits of the educational enterprise, rather than perpetuates the dominant deficit-based narrative of Black adolescents’ education-borne mental health conditions.

Future directions will need to focus on longitudinal data collection and analyses using both person-centered and variable-centered approaches. Not only do we need to understand if certain mental health observations change latent classes over time, but we also need to know the long-term mental health impacts of school policing, and other forms of social control strategies sanctioned by the educational enterprise, against Black people. As such, we can better inform educational and health policy towards the education, literacy, and well-being of Black youth and their communities. It will help researchers and practitioners understand whether the educational enterprise is a social determinant of health or whether it perpetuates the disenfranchisement from historically white institutions. The authors call on researchers, clinicians, and policymakers to create prospective studies or, retro-fit existing studies with the intent to better conceptualize, understand, and investigate how the relationship between the educational enterprise and health outcomes are structured by race and racism. Furthermore, the authors encourage child and adolescent psychiatrists to consider social systems, social forces, or institutions when identifying the potential causes of depressive symptoms in Black children and adolescents. Specifically, when psychiatrists examine and treat depressive symptoms of Black youth and the role that school policing has on the etiology and comorbidity of depression. We also encourage child and adolescent psychiatrists to implement, and participate in, community-based interventions to help remedy the mental health conditions that present challenges to Black adolescents.

This catalog of research cannot be ahistorical; empirical research must go above and beyond couching studies with lists of events or problems. Scholars need to couch research in historiography, for this method reflects on the histories of power and oppression. Studies often include a lacuna of properly understanding outcomes of the 1954 Supreme Court decision, Brown v. Board of Education, which has been less impactful to the educational quality of Black communities.³⁴ Finding ways to examine the quality of education and how culturally responsive and inclusive education is, with nationally representative data, is paramount to changing the narrative of Black students. Finally, the history of redlining and “de jure segregation,”³⁵ as well as “resegregation”³⁶ must be considered when investigating the education-health relationship.

Table 1.

Black Subsample Descriptives (N=1,601)

	n	%
Have you ever been stopped by police at school?
No	384	77.6
Yes	111	22.4
Have you ever seen others stopped by police at school?
No	802	50.3
Yes	792	49.7
Have you ever been disciplined at school?
No	1021	64.5
Yes	563	35.5
Sex/Gender
Male	803	50.2
Female	798	49.8
			M	SE
Depressive symptom scores (R: 0, 15)			2.92	0.07