Examining Individual and Aggregate Correlates of Police Killings of People with Mental Illness: A Special Gaze at Race and Ethnicity

Abstract

This study assesses the influences of race and ethnicity on the presence of resistant behavior during deadly police contacts with people experiencing mental illness. Drawing upon information mainly from the Mapping Police Violence Database, regression analysis showed that race/ethnicity is not a consistent predictor of resistance among people with mental illness. Hispanics with mental illness were less likely than White counterparts to attack the police before being killed. Counties with a higher percentage of the Black population have a higher likelihood of people with mental illness attacking the police.

Keywords

use of deadly force mental illness race and ethnicity public resistance attack police

Introduction

Wallace, an aspiring rapper and father of nine, had a mental illness. . .Police officers responded twice to the Wallace residence Monday before returning a third time to a report of a person with a weapon, police said. Two officers – who did not have Tasers – each fired at least seven rounds at Wallace after yelling at him to drop a knife, according to police. Wallace was hit in the shoulder and chest. One of the officers put him in a police vehicle and drove him to a hospital, where he was pronounced dead a short time later (Hauck, USA Today, October 29, 2020).

The video footage of the shooting of Walter Wallace, Jr. went viral on social media, stirring several days of public protests and violent confrontations with the Philadelphia police amid the COVID-19 pandemic. The city later settled the wrongful-death lawsuit filed by the Wallace family with a $2.5 million payment in 2021. This deadly incident elucidates the continuing difficulties and problems in police encounters with people experiencing mental illness. It is challenging for police officers to identify people with mental illness by simply observing their behavior through limited interactions (Alpert, 2015). It could also be problematic as many police officers are uncertain about the proper approaches to handle people with mental illness due to inadequate training (Wells & Schafer, 2006). Furthermore, when the non-justice response is unavailable, police officers often resort to coercive actions to handle persons with mental illness, further aggravating the problem. Given that approximately a quarter of police fatal contacts involve individuals with mental illness (Fuller et al., 2015), further research is crucial to increase our understanding of these situational contexts and to alleviate this disparity.

The killing of Walter Wallace, Jr. also represents a more significant issue of racial disparities in police use of deadly force. For example, Mapping Police Violence (2022a) reports that between 2013 and 2022, Black citizens have been 2.9 times more likely to be killed by police use of force. Looking specifically at fatal shootings, the Washington Post (2022) delineates that, between 2015 and 2022, both Black and Hispanic citizens have been disproportionately shot and killed by the police despite accounting for less than 13% and 12% of the U.S. population, respectively. Compared to the rate at which White Americans are killed (16/million), deadly force kills Black Americans at more than twice the rate (39/million) and Hispanic Americans at twice the rate (28/million). These disparities in police killings raise essential questions about the relationship between race/ethnicity and citizen resistance toward the police. Are Black and Hispanic Americans more likely to resist the police? Does an increased likelihood of resistance explain the disproportionate rates at which Black and Hispanic Americans encounter police use of deadly force?

The main purpose of this study is to examine the relationships between race and ethnicity and the resistant behavior of people with mental illness before being killed by the police. Although many studies have examined the effects of various predictors of police use of deadly force, we have limited knowledge about the interactions between the police and people with mental illness. Even rarer is research examining the behavior of people with mental illness, particularly their non-compliance or resistance, before being killed by the police. This study assesses the linkages between race and ethnicity and the presence of resistant behavior during deadly police contact with people experiencing mental illness. This study is not about how mental illness may shape the relationship between race/ethnicity and resistance. Instead, this study addresses the research question: Are race and ethnicity associated with the likelihood of resistance among people with mental illness in their deadly encounters with the police?

Citizen resistance, or any behavior or demeanor that police officers perceive as challenges to their authority or control efforts, is an integral factor in shaping officers’ decision-making as police use of force must correspond with the occurrence and level of resistance. We focus on the connection between individual race/ethnicity and community racial/ethnic minority representation and such resistance, given that racial/ethnic minorities—particularly Black Americans—face heightened risks of police killings (Mapping Police Violence, 2022a). This study contributes to the existing literature on policing individuals with mental illness by analyzing the association between race/ethnicity and citizen resistance preceding the use of deadly force, which previous studies have seldom investigated.

As we will show, the literature examining predictors of citizen resistance during police encounters and interactions between the police and citizens with mental illness sheds light on possible explanations of resistance but also presents necessary gaps to fill. Generally, findings concerning the relationship between individual-level characteristics and citizen resistance are mixed. More specifically, research has not determined if and how citizen race/ethnicity is associated with citizen resistance. Further, fewer studies have considered the role of aggerate-level characteristics, and those findings are also mixed. Most research on police encounters involving citizens with mental illness focuses specifically on how mental health status predicts the use of force and citizen resistance. While this is a necessary focus, there is also a need for studies that examine how individual-level and aggregate-level predictors are associated with citizen resistance, specifically within the context of citizens with mental illness in deadly policing encounters. Few studies have also looked to theories such as Black’s (1976) theory of the behavior of law and Sykes and Clark’s (1975) Deference Exchange Theory to explain citizen resistance as a result of asymmetrical power and social stratification between citizens and legal officials. Whether civilian race/ethnicity and neighborhood racial/ethnic composition plays a role in citizen resistance and the nature of police encounters involving citizens with mental illness needs further investigation.

The following sections first discuss the conceptualization and classification of citizen resistance and individual- and aggregate-level characteristics linked to citizen resistance. We then review the literature on policing civilians with mental illness. Following the literature review sections is the introduction of methods used in the study, including data sources, population, and measures. In the last section, we present the results from regression analyses and discuss our major findings, study limitations, and directions for future research and policy.

Conceptualizing and Predicting Resistance

Police use of force is generally understood as the amount of effort deemed necessary to resolve an incident, deter the commission of a crime, make arrests, or subdue a non-compliant citizen in situations of self-defense and the defense of others (International Association of Chiefs of Police, 2001; National Institute of Justice, 2020). Resistance, therefore, is a key predictor of police violence and killings. The policing literature conceptualizes citizen resistance as any behavior or demeanor that is perceived to challenge police officers’ authority or their understandings of the legitimacy underlying their authority to gain social control in the current context (Engel, 2003). Examining the classifications of citizen resistance across 371 police agencies, Terrill and Paoline (2013, p. 51) delineated six categories of resistance, including: “compliant, verbal (e.g., refusing verbal direction), passive (e.g., failing to respond to an officer/ignoring), physical defensive (e.g., bracing, pulling away, and fleeing), physical active (e.g., hostile and overt physical aggression toward the officer), and deadly (e.g., attempt or actual attack that could cause death).” Further, the most used citizen resistance continuum by agencies placed types of resistance in the following order ranging from the lowest to highest threat level: (1) compliant, (2) verbal/passive combined, (3) physical defensive, (4) physically active, and (5) deadly.

Though the officer use-of-force continuums adopted by agencies vary, the common expectation and standard are that force should only be used when necessary, such as in self-defense or defense of others (National Institute of Justice, 2020). Additionally, the level of force must correspond with the level of resistance so that officers should only use higher levels of force in the most severe circumstances and lower levels of force in the least severe circumstances (National Institute of Justice 2009, 2011). Thus, the most severe contexts would include a physically aggressive civilian attacking, using, or attempting to use a weapon against an officer or another person. Further, circumstances would be more severe if a civilian is armed with a deadly weapon rather than a non-deadly weapon.

Several theories have been used to explain citizen resistance toward the police, such as Black’s (1976) theory of the behavior of law and Sykes and Clark’s (1975) Deference Exchange Theory. Although Black originally developed his theory to predict the quantity and style of law, it has since been extended to explain citizens’ behavior toward the law. One key proposition is that as the level of stratification increases between a citizen and a legal official, so does the quantity of the law enforced upon a citizen. Black further argues that levels of stratification also predict other forms of social control, such as delinquent behavior in response to the law. Though a study examining citizen compliance, McCluskey et al. (1999) use Black’s theory to argue that a citizen’s social status, relative to a legal official, can influence their compliance with the demands of a legal official. They further argue that citizen compliance is more likely when the legal official is of the same or higher social status as the citizen.

Sykes and Clark’s (1975) Deference Exchange Theory is a prediction of citizens’ reactions to officers’ efforts to enforce social control. Sykes and Clark (1975) conceptualize encounters between the police and citizens as governed by an asymmetrical power norm where police occupy higher statuses than citizens due to the nature of their occupational role. Police, as they explain, “. . .[are] symbolic of the law, the ultimate Weberian rational legal basis of social authority in modern societies” (p. 588). Further, because of the asymmetrical power norm and what the police represent, police officers expect deference or obedience from all citizens but are not expected to reciprocate such courtesy. However, as Sykes and Clark (1975) further explain, this relationship between police and citizens has unique complications when it comes to minority populations, especially racial and ethnic minorities. One would expect that because racial and ethnic minorities may perceive the police as enforcers of racial/ethnic stratification, they will be more likely to resist police efforts of control than White citizens.

Both Black’s theory of the behavior of law and Sykes and Clark’s Deference Exchange Theory would predict racial/ethnic minority citizens to be more likely to resist police control than White citizens. Black’s theory emphasizes stratification and inequality (U.S. Government Accountability Office, n.d.), while Sykes and Clark (1975) include the role of asymmetrical power and citizens’ perceptions of policing as enforcing stratification and inequality (Ekins, 2016). Additionally, as citizens with mental illness experience social and economic inequality and social stigmatization (Boysen et al., 2020; Funk et al., 2012), these theories would suggest that resistance may manifest differently among this group. Therefore, research examining race/ethnicity as predictors of citizen resistance and specifically looking at policing encounters involving citizens experiencing mental illness is necessary.

Correlates of Citizen Resistance

This study focuses on the impact of race and ethnicity on resistant behavior among people with mental illness killed by the police. Besides race and ethnicity, other individual-level correlates, such as gender and age, and aggregate-level variables, such as percent poverty and region, are also included in the analysis. In the following two sections, we review the relevant literature regarding the individual- and aggregate-level non-race variables associated with citizen resistance, followed by a review of the literature examining whether and how race and ethnicity at the individual- and aggregate-level predict citizen resistance.

Individual-Level Characteristics

Previous studies have examined citizen resistance during police encounters, specifically, what factors predict resistance. Across this literature, such individual-level characteristics as civilian age, gender, and race/ethnicity are factors of common interest to researchers. Of the studies that have explored the role of civilian age, researchers have consistently elucidated that age is not a significant predictor of a civilian’s likelihood of resistance (Belvedere et al., 2005; Crawford & Burns, 1998; Engel, 2003; Kavanagh, 1997; Nix et al., 2017; Paoline et al., 2018; Thomas et al., 2021). One outlier, McCluskey et al. (1999), identified age as a significant predictor of citizen resistance, with younger civilians more likely to resist the police. Fewer studies have assessed the role of gender in predicting citizen resistance as women are underrepresented in police-citizen encounters and resisting the police (Nix et al., 2017). Among the studies that have considered gender, results are mixed. For example, one study found that females were more likely than males to resist the police (Mastrofski et al., 1996), whereas two other studies revealed no significant gender differences in resistance (Crawford & Burns, 1998; McCluskey et al., 1999). A fourth study showed that females were more likely to disrespect the police but were not more likely to be resistant (Engel, 2003).

Most relevant to the current study is the literature examining the relationship between citizen race/ethnicity and resistance. Studies examining the relationship between civilian race/ethnicity and resistance have produced mixed results. Some studies found that Black (Belvedere et al., 2005) and non-White (Sykes & Clark, 1975) civilians are more likely than White counterparts to engage in resistance. In a more recent study, Nix et al. (2017) reported no significant difference in the likelihood of resistance between Black and White people. Still, they found that Hispanic civilians are more likely to exhibit resistance than Black and White civilians. Most other studies showed a weak relationship or no racial difference in resistance toward the police (Crawford & Burns, 1998; Kavanagh, 1997; McCluskey et al., 1999; Paoline et al., 2018). Furthermore, racial differences in civilians’ likelihoods of resistance could be conditional, based on civilian age, officer race, and type of resistance. For example, Thomas et al. (2021) demonstrated that, whereas older Black men were less likely than older White men to be armed, there was no racial difference in armed status among younger age groups. Additionally, comparing resistance among non-White and White civilians, Engel (2003) found that non-White civilians were more likely to be non-compliant (i.e., refusing to answer questions or otherwise cooperate) specifically toward White officers. However, there were no racial differences in the more aggressive forms of resistance (i.e., disrespect, verbal aggression, or physical aggression).

Aggregate-Level Characteristics

Including aggregate-level factors in predicting citizen resistance proves necessary as analyses show that ecological contexts and environments influence officers’ decisions to use force (Edwards et al., 2018; Feldman, 2020; Lautenschlager & Omori, 2018; H. Lee et al., 2014; Parker et al., 2005). Further, higher levels of police use of force in a specific context might imply the existence of higher levels of citizen resistance. However, higher levels of police use of force could also be explained by factors outside of citizen behavior such as the presence of racial discrimination or racial threat. Such perspectives argue that greater police use of force may be associated with racially biased decisions to use force (e.g., Kramer & Remster, 2018) or an increased need for police force in response to an increasing minority population (e.g., Lautenschlager & Omori, 2018).

Aggregate-level characteristics, such as region (i.e., Northeast, South, Midwest, and West) and crime rates, have been considered in a few studies. Region has important implications for citizen resistance to the police force due to regional crime and violence. In 2019, the Northeast region of the U.S. had the lowest violent crime rate (292.4 per 100,000), followed by the Midwest (361.7), South (406.6), and West (413.5) (Federal Bureau of Investigation, 2019). Additionally, the South has long been associated with a “culture of violence”—the thesis that specific features of Southern culture, including the need to protect one’s reputation, self, and family from threats, explain the South’s higher rates of violence (Cohen et al., 1996; M. Lee et al., 2007). Studying the interaction between race and armed status among Black and White men fatally shot by the police, Thomas et al. (2021) found that region matters as Black men were less likely to be armed in the South but more likely to be armed in the Midwest. Nonetheless, Nix et al. (2017) reported no regional differences between the armed and attack statuses of civilians who have been shot and killed by the police. When examining the role of violent crime rates, Nix et al. (2017) did find that civilians living in moderate crime areas were more likely to be unarmed and attack officers or other civilians than those living in low crime areas.

There is consistent evidence showing that police officers are more likely to use force in areas with higher levels of crime (H. Lee et al., 2010, 2014; Terrill & Reisig, 2003) and economic disadvantage (Feldman, 2020; Parker et al., 2005; Sun et al., 2008). Research shows that police use of force varies by type of neighborhood (i.e., urban, suburban, and rural) and geographic areas (i.e., regions and divisions). Edwards et al. (2018) analyzed police killings of Black, Latino, and White men and found that the risk of death by police violence was highest in large urban areas and Mountain states (i.e., Montana, Idaho, Wyoming, Nevada, Utah, Colorado, Arizona, and New Mexico), while suburban areas and Middle Atlantic states (i.e., New Jersey, New York, and Pennsylvania) had the lowest observed rates.

Among the aggregate-level characteristics, the most relevant variable to this study is neighborhood racial/ethnic composition. Evidence concerning the influence of racial/ethnic composition is mixed. Some studies indicated that police officers are more likely to use force in communities of color (Feldman, 2020; Hoekstra & Sloan, 2020; Parker et al., 2005), while others suggested a disconnect between the two (Lawton, 2007). Interestingly, while Lautenschlager and Omori’s (2018) study found police use of force to be primarily concentrated and more severe in Black neighborhoods, they also found that use of force incidents decreased in number but increased in severity in communities with higher racial and ethnic heterogeneity. As scholars have shown, aggregate-level characteristics such as crime, poverty, geography, and racial/ethnic composition must be considered for an understanding of not only police decision to use force but citizen resistance to such force.

Policing Citizens With Mental Illness

While some studies have investigated the influence of civilian mental health status on police use of force, little attention has focused on understanding the nature of actual encounters between civilians experiencing mental illness and the police. Researchers have presented mixed findings on whether civilian mental health status influences police officers’ use of force. While some studies showed officers as more likely to use force on civilians experiencing mental illness (Alpert & Dunham, 1997; Lawton, 2007; Morgan et al., 2020; Sun & Payne, 2004), others showed no significant relationship between mental health status and force (Johnson, 2011; Kaminski et al., 2004; Terrill & Mastrosfki, 2002). Further, of the few studies examining the association between mental health status and citizen resistance, there is a consensus that civilians experiencing mental illness are more likely than those not experiencing mental illness to be resistant to police use of force (Engel & Silver, 2001; Johnson, 2011; Mulvey & White, 2014; Novak & Engel, 2005; Rossler & Terrill, 2017).

However, researchers have also shown that the relationship between mental illness and resistance may vary by specific types of resistance and particular police departments. For example, Nix et al. (2017) study of threat perception failure in fatal police shootings examined whether civilians were not attacking and were unarmed at the time of their death. In terms of mental illness, their findings revealed that civilians were more likely not to be attacking and no more or less likely to be unarmed if they had a mental illness. Morabito et al. (2012) also studied police encounters with individuals experiencing mental illness in four Chicago Police Department districts. They found that officers were more likely to encounter civilians with mental illness who were cooperative/compliant than resistant.

Most relevant to the current study is the literature examining the relationship between race and resistance among people with mental illness. To the best of our knowledge, only one study has examined the intersections between civilian race/ethnicity, resistance, and mental health status to understand the nuances influencing the racial disparities that persist in police killings (Thomas et al., 2021). In their study, Thomas and colleagues assessed whether race and armed status interacted in the fatal police shootings of Black and White males. Using a case-only design, they analyzed potential three-way interactions between race and armed status with civilian age, U.S. region, and mental illness status. They found that Black men who were perceived as mentally ill were less likely to be armed when the police fatally shot them, showing a three-way interaction between race, armed status, and mental illness status. Ultimately, the authors argued, these findings indicate that anti-Black bias, compared to armed status, may be a stronger contributing factor in influencing deadly officer use of force on people with mental illness.

This Current Study

Considering the persistent racial/ethnic and mental health disparities in police killings, it is imperative to critically evaluate the range of predictors contributing to the vulnerability that civilians who are Black, Hispanic, and living with mental illness face during their encounters with police. While much research has examined various individual-, situational-, aggregate-, and organizational-level predictors of police use of force, little attention has been given to the influence of civilian behavior, the nature of police encounters involving civilians with mental illness, and the linkages between race/ethnicity and resistance. As shown throughout the literature review, citizen resistance, or any behavior or demeanor perceived to challenge police officers’ authority and legitimacy to gain control in the current situation, is an integral factor in considering police actions. Scholars have used Black’s (1976) theory of the behavior of law and Sykes and Clark’s (1975) Deference Exchange Theory to argue that civilians experiencing social inequality (e.g., racial/ethnic minorities) are more likely to resist police control. The theories attribute a higher likelihood of resistance to differences in social statuses between racial/ethnic minorities and law enforcement and perceptions of law enforcement as oppressive. In terms of predictors of resistance, research thus far has presented mixed findings, questioning whether and how civilian race/ethnicity is connected to resistance.

This study extends the current literature by analyzing situational and community/regional contexts related to police killings of people with mental illness. The focus of this study is not to assess how mental illness may influence the connection between race/ethnicity and resistance. Instead, we focus on how individual racial/ethnic backgrounds and community racial/ethnic compositions may be linked to resistant behavior toward the police among persons with mental illness. Using information primarily collected from the Mapping Police Violence Database, this study examines the associations between individual- and aggregate-level racial/ethnic characteristics and two forms of resistant behaviors among 775 people with mental illness killed by the police over 5 years. This study is important as it seeks to explore further the role of race/ethnicity in predicting various measures of citizen resistance and, more specifically, during encounters with individuals experiencing mental illness.

Based on our research question and literature review, two hypotheses were formulated and tested in the study:

H1: Blacks and Hispanics with mental illness are more likely than their White counterparts to be armed and attacking during their deadly encounters with the police, controlling for other individual- and aggregate-level variables.

H2: People with mental illness living in areas with higher percentages of Black and Hispanic populations are more likely than those living in areas with low Black and Hispanic populations to be armed and attacking during their deadly encounters with the police, controlling for individual- and other aggregate-level variables.

Methodology

Police killings of citizens with mental illness across the U.S. were analyzed to examine the relationship between race/ethnicity and citizen resistance. This study focuses on fatal police encounters involving three racial/ethnic groups: Blacks, Whites, and Hispanics. Using several data sources, we conducted regression analyses to examine whether and how race/ethnicity predicts two types of citizen resistance: whether a citizen was armed or attacking during the fatal encounter with police.

Data Sources

The data used for this study originates from the Mapping Police Violence Database (MPVD) as well as the U.S. Census Bureau’s 2019 American Community Survey (ACS) 5-Year Estimates and Small Area Income and Poverty Estimates (SAIPE). MPVD is a compilation of all police use of force incidents resulting in civilian deaths since 2013. The period for this study was January 1, 2016, to December 31, 2020, during which there were 5,529 total police killings. The MPVD database is considered the most comprehensive account of people killed by the police since 2013. It includes not only incidents of police shootings but also incidents involving any use of force by police that killed a civilian such as tasers, physical force, and police vehicles. The database combines data from the three largest crowdsourced police use of force databases: FatalEncounters.org, KilledbyPolice.net, and the U.S. Police Shootings Database. Although researchers have used MPVD to investigate various topics related to police use of deadly force (e.g., Bor et al., 2018; Gray & Parker, 2019; Nix et al., 2017; Nix & Lozada, 2021), researchers should acknowledge some classification issues associated with the dataset. For example, MPVD classifies individuals with toys or BB guns as “unarmed.” Such a classification is problematic as it contradicts the legal standards of self-defense, the principles of justification, and the reasonable person standard commonly considered by legal judgments (Nix & Lozada, 2021).

MPVD reports many variables relating to the individual-, encounter-, aggregate-, and agency-level characteristics associated with each police killing. The following variables were used for this study: Victim’s Age, Victim’s Race, Victim’s Gender, whether or not the victim exhibited Symptoms of Mental Illness during the encounter, Armed/Unarmed Status, Alleged Weapon, Alleged Threat Level, and the State, City, County, and Community Type where a police killing took place.

We introduced additional variables into the dataset to explore further the aggregate-level characteristics of the violent contexts connected to such police killings. Using the U.S. Census Bureau’s 2019 American Community Survey (ACS) 5-Year Estimates, racial/ethnic composition data was gathered to determine each county’s percent Black and Hispanic population. Additionally, we used the U.S. Census Bureau’s Small Area Income and Poverty Estimates (SAIPE) to gather county-level estimates of the percentage of people living in poverty for each year from 2016 to 2019. Lastly, according to Census regions, each state was linked to its corresponding region: Northeast, South, Midwest, and West.

Population

Of the 5,529 total police killings occurring from 2016 to 2020, 1,038 victims had a mental illness. We removed the following cases where variables were missing information or documented as “unknown,” “unclear,” or “undetermined”: individuals’ age (36 cases), armed/unarmed status (52 cases), alleged weapon (0 cases), and alleged threat level (33 cases), as well as the aggregates’ community type (9 cases), county racial composition (6 cases), and the county poverty estimates (1 case). In addition, we removed three cases because the individuals’ gender identity was missing or identified as transgender, which was too few to consider in the analyses. Lastly, 123 cases were removed because the individuals’ racial/ethnic identity was missing or identified as Pacific Islander, Native American, or Asian. These racial/ethnic groups were outside our study scope as we focused solely on Black, Hispanic, and White individuals. After removing all of these cases, our final population consisted of 775 individuals. As shown in Table 1, among those killed by the police, 67% were White, 19% were Black, and 14% were Hispanic. Most of them (93%) were male, and their average age was about 40.

Table 1.

Descriptive Statistics of All Variables (n = 775).

Variables	M	SD	Min	Max
Dependent variables
Armed (MPVD)	0.86	0.35	0.00	1.00
Armed (reclassified)	0.91	0.28	0.00	1.00
Attacking	0.62	0.49	0.00	1.00
Independent variables
Individual-level
White^a	0.67	0.47	0.00	1.00
Black	0.19	0.39	0.00	1.00
Hispanic	0.14	0.35	0.00	1.00
Age	39.26	13.82	10.00	88.00
Male	0.93	0.26	0.00	1.00
Aggregate-level
Percent Black	11.50	11.77	0.10	75.60
Percent Hispanic	17.53	17.13	0.30	95.50
Percent poverty	13.71	4.26	2.98	34.30
Rural^a	0.22	0.41	0.00	1.00
Suburban	0.56	0.50	0.00	1.00
Urban	0.22	0.41	0.00	1.00
South^a	0.42	0.49	0.00	1.00
Northeast	0.10	0.30	0.00	1.00
Midwest	0.18	0.39	0.00	1.00
West	0.30	0.46	0.00	1.00

Note. M = mean; SD = standard deviation; Min = minimum; Max = maximum.

Reference group in regression analysis.

Measures

The MPVD and ACS data were used to create two indicators of citizen resistance as the dependent variables and individual- and aggregate-level predictors, including race and non-race correlates, as independent variables. Based on previous literature examining citizen resistance (e.g., Engel, 2003), civilian physical resistance was operationalized using three dichotomous variables constructed by combining the MPVD variables: “armed (original),” “armed (reclassified),” and “attacking.” According to the MPVD’s definitions for Armed/Unarmed Status, a person was considered to be Unarmed/Did Not Have a Weapon if they were one or more of the following:

not holding any objects or weapons when killed, holding household/personal items that were not used to attack others (cellphone, video game controller, cane, etc.), holding a toy weapon (BB gun, pellet gun, air rifle, toy sword), an innocent bystander or hostage killed by a police shooting or other police use of force, a person or motorist killed after being intentionally hit by a police car or as a result of hitting police stop sticks during a pursuit (Mapping Police Violence, 2022b)

A person was considered Allegedly Armed if they were alleged to have objects or weapons in circumstances other than those previously stated. Using the original MPVD’s data on whether a civilian was (a) armed or (b) unarmed, a dichotomous measure “armed (MPVD)” was created (0 = unarmed, 1 = armed). Acknowledging the classification issues associated with MPVD, we created a second “armed” variable by reassigning incidents involving a “toy weapon” and “BB gun” (46 cases in total) to the “armed (original)” category. The variable “armed (reclassified)” is a dummy variable, with one representing armed. Also, MPVD included whether or not a civilian was (a) attacking or (b) other. Incidents were coded as non-attacking if civilians were documented in the dataset as other. Thus, “attacking” is a dichotomous measure of whether the civilian was attacking or not (0 = non-attack, 1 = attack).

The key predictors for this study are race and ethnicity. At the individual-level, we created three dummy variables to represent Black, Hispanic, and White, with the last group (White) serving as the reference category in data analysis. At the aggregate-level, Percent Black and Percent Hispanic were constructed to reflect the 5-year averages for the percentages of county-level Black and Hispanic populations.

Other theoretically important predictor variables were included in the analyses as independent variables. For individual-level characteristics, we included civilian age and gender. Age was measured on a continuous scale. A dummy variable, Male, was also created to capture gender. We included community type, county poverty estimates, and region for aggregate-level characteristics. Three dummy variables were created to capture community types: Urban, Suburban, and Rural (reference category). One variable captured the 4-year average percentage of people living in poverty at the county level. Lastly, four dummy variables were created to capture regions: South (reference category), Northeast, Midwest, and West. Table 1 displays descriptive statistics of all variables used in the study.

Results

Table 2 reports the results from binary logistic regression models predicting each of the three measures of citizen resistance. Looking at the first model predicting “armed (MPVD)” persons with mental illness, the two racial/ethnic variables fail to achieve statistical significance, suggesting no difference between White and Black and between White and Hispanic in terms of being armed before the police killed them. Further analysis using Black as the reference group also reveals that Hispanics and Blacks experiencing mental illness do not differ in predicting whether they were armed or not. At the aggregate-level, percent Black and percent Hispanic are also not significantly linked to the dependent variable of “armed (MPVD).” These results are inconsistent with our hypotheses (H1 and H2). The only significant predictive in the model is age. Older people experiencing mental illness were more likely than their younger counterparts to be armed during their deadly encounters with the police (B = 0.02; odds ratio = 1.02). With every 1-year increase in age, the odds of being armed rise by 2%. None of the aggregate-level variables are predictive of whether or not persons with mental illness killed by the police were armed.

Table 2.

Summary of Binary Logistic Regression Results (n = 775).

Variables	Armed (MPVD)			Armed (reclassified)			Attacking
Variables	B	OR	SE	B	OR	SE	B	OR	SE
Individual-level
Black	−0.46	0.64	0.27	−0.80*	0.45	0.32	−0.14	0.87	0.21
Hispanic	0.29	1.34	0.36	0.06	1.06	0.43	−0.49*	0.61	0.25
Male	0.67	1.96	0.36	0.10	1.11	0.50	0.12	1.13	0.30
Age	0.02**	1.02	0.01	0.02	1.02	0.01	0.01	1.01	0.01
Aggregate-level
Percent Black	−0.00	1.00	0.01	0.00	1.00	0.02	0.02*	1.02	0.01
Percent Hispanic	−0.00	1.00	0.01	−0.01	0.99	0.01	0.01	1.01	0.01
Percent poverty	0.04	1.04	0.03	0.02	1.02	0.04	0.02	1.02	0.02
Suburban	−0.15	0.86	0.31	−0.40	0.67	0.41	−0.28	0.76	0.22
Urban	−0.40	0.67	0.36	−0.49	0.61	0.47	−0.77**	0.46	0.26
Northeast	−0.16	0.86	0.37	0.02	1.02	0.47	0.28	1.33	0.28
Midwest	0.06	1.06	0.33	−0.10	0.91	0.40	0.36	1.43	0.23
West	−0.11	0.90	0.31	0.19	1.21	0.37	0.31	1.36	0.22

p < .05. **p < .01.

Since our data comes from a population rather than a sample, evaluating substantive effect size in addition to statistical significance is reasonable due to its non-random nature (see Hirschi & Selvin, 1973). Despite not reaching statistical significance, the odds ratios associated with citizen race/ethnicity and other variables reveal interesting substantive effects. For instance, compared to White citizens with mental illness, Black citizens with mental illness were 36% less likely to be armed (B = −0.46; odds ratio = 0.64), and Hispanic civilians with mental illness were 34% more likely to be armed (B = 0.29; odds ratio = 1.34). Additionally, males with mental illness were nearly twice (96%) more likely to be armed than females with mental illness (B = 0.67; odds ratio = 1.96). Compared to citizens with mental illness living in rural areas, those living in urban areas were 33% less likely to be armed (B = −0.40; odds ratio = 0.67).

Switching to the second model in Table 2, where the dependent variable was reclassified by considering individuals with toy weapons/BB guns as “armed,” the race/ethnicity effect surfaces, whereas the age connection disappears. Compared to their White counterparts, Blacks with mental illness were 55% less likely to be armed before being killed by the police, which contradicts our hypothesis. The aggregate-level racial/ethnic indicators and other variables remain ineffective in predicting the armed status.

Finally, in the third model, three predictors, Hispanic, percent Black, and urban, reached statistical significance in the “attacking” model. Contradictory to our hypothesis (H1), being Hispanic lowered the odds of attacking the police by 49% (B = −0.48; odds ratio = 0.61) compared to a White person experiencing mental illness. Although the relationship between “Black” and “attacking” is not statistically significant (B = −0.14; odds ratio = 0.87), the odds ratio indicates that being Black reduced the odds of attacking the police by 13%. At the aggregate-level, percent Black matters (B = 0.02; odds ratio = 1.02). With every one-percentage increase in the Black population, the odds of a person with mental illness attacking the police raise by 2%. Also, living in an urban area reduced the odds of assaulting the police by 54% (B = −0.77; odds ratio = 0.46) compared to living in a rural area. Finally, albeit not statistically significant, the effect sizes of suburban, Northeast, Midwest, and West also show their importance as predictors of attacking the police. Compared to citizens with mental illness living in rural areas, those living in suburban areas were 24% less likely to be attacking the police (B = 0.02; odds ratio = 1.02). Further, compared to those living in the South, citizens with mental illness living in the Northeast were 33% more likely to be attacking (B = 0.28; odds ratio = 1.33). The odds of attacking the police also increase by 43% when citizens live in the Midwest (B = 0.36; odds ratio = 1.43) and 36% when citizens live in the West (B = 0.31; odds ratio = 1.36).

Discussion

Recent deadly encounters between the police and people with mental illness have crystallized the persistent and challenging nature of such police-public contacts. Complicating matters further, racial/ethnic inequality in deadly police incidents has been widely reported, with Black Americans consistently over-represented in police use of deadly force. The intertwined nature of race and ethnicity and fatal police force has led us to investigate the potential relationships between race and ethnicity and resistance that arguably pre-conditioned police use of deadly against people with mental illness. We analyzed two types of citizen resistance (i.e., armed and attacking). Our findings do not reveal a consistent association between race/ethnicity and resistant behavior in the context of police killings of people experiencing mental illness. We further elaborate on our major findings below.

First, our findings show mixed results regarding the relationship between citizens’ race/ethnicity and their resistance among people with mental illness before being killed by the police. Our findings contradict Black’s (1976) and Sykes and Clark’s (1975) predictions that racial/ethnic minority citizens would be more resistant toward the police due to racial/ethnic stratification and asymmetrical power between the police and racial/ethnic minority citizens. We found that Blacks and Hispanics with mental illness were not statistically different from their White counterparts in the armed status when using the MPVD’s original classifications to operationalize the variable. Our reclassification of the armed variable reveals more evidence contradicting the expectation of high resistant behavior among racial/ethnic minorities. Blacks with mental illness were less likely than their White counterparts to be armed before being killed by the police.

Similarly, contradictory to earlier predictions, we also found that Hispanic citizens with mental illness were less likely to be attacking the police compared to their White counterparts. Such a result does support the idea that the relationship between race and ethnicity and resistance could vary across resistant types (Engel, 2003) and racial/ethnic groups. Furthermore, we found a limited relationship between racial/ethnic composition and citizen resistance. People with mental illness who resided in predominately Black counties were more inclined to attack the police than those in predominately White counties. (They were not, however, more likely to be armed.) We cannot directly link our findings to the results showing the greater likelihood of police deadly force in Black communities (Feldman, 2020; Hoekstra & Sloan, 2020; Lautenschlager & Omori, 2018; Parker et al., 2005). Nonetheless, one may speculate that persistent higher levels of police violence in Black communities could have bred a street culture of resistance, symbolizing the psychological resilience needed to survive adverse environments shaped by social injustice (Y. A. Payne, 2011). In other words, a higher possibility of attacking the police in predominately Black communities could be an adaptive response to police violence embedded firmly within community values and expectations. Our data do not allow us to test such speculation. Perhaps drawing upon in-depth interview data, future studies can shed a stronger light on this possibility.

Second, our findings show that only two non-racial/ethnic characteristics are linked to resistance among people with mental illness. In the “armed (MPVD)” model, our findings reveal the relevance of age in predicting resistant behavior toward the police among people with mental illness. Although previous studies failed to establish a consistent link between age and citizen resistance (Belvedere et al., 2005; Crawford & Burns, 1998; Engel, 2003; Kavanagh, 1997; Nix et al., 2017; Paoline et al., 2018; Thomas et al., 2021), we found that the likelihood of being armed rose when the age of the person with mental illness increased. The average age of our population is close to 40, which falls within the most common age category (i.e., 30–44) for people shot to death by the U.S. police between 2015 and 2022 (Washington Post, 2022), which may explain a positive relationship between age and being armed found in this study. Future research should further investigate the significant connection between age and citizen resistance among individuals with mental illness.

Finally, besides age, we also found that rural residents with mental illness were more likely than their urban counterparts to attack when encountering the police. This finding challenges the common wisdom about greater violence embedded within police-citizen encounters in urban than rural areas. Our finding further points to the lesser-known feature of rural policing, going beyond the traditional police activities involved in dogs, drunk, disorder, and dysfunction (B. Payne et al., 2005). One may suspect that high weapon ownership and inadequate mental health services in rural areas could have exacerbated confrontational situations defining police actions toward resistance displayed by people with mental illness. More studies are needed to assess the validity of this speculation.

Before discussing policy implications, several limitations associated with using secondary data should be acknowledged. First, although scholars have distinguished among different types of resistance (Hollander & Einwohner, 2004), our data do not make such conceptual distinctions across our cases. For example, it would be interesting to see whether varying factors may be linked to different types of resistance due to differences in the civilian’s intention to resist. Further examination could focus on overt resistance (i.e., intentional behavior that is visibly recognized as an act of resistance by the target and observers) and unwitting resistance (i.e., an intentional act by actors yet is recognized as threatening by the target and observers) (Hollander & Einwohner, 2004). Similarly, the dataset does not contain any information on the police, making the control of key officer background and work characteristics such as race/ethnicity, gender, experience, and assignment impossible.

Second, county was used as the unit of analysis for the aggregate level because it is the lowest level of data available for measuring racial/ethnic composition, poverty, and community type. However, county-level estimates do not account for the nuances in community characteristics within smaller geographic areas such as cities and neighborhood census tracts. Therefore, our estimates of racial composition and poverty within counties could be overlooking essential differences between and within various cities and neighborhoods. A relevant issue is that using aggregate-level variables at the individual-level analysis violates the assumption of the variables’ independence. Although prior studies found that, compared to hierarchical modeling techniques, the employment of police encounter-level predictors at the citizen level analysis yielded unbiased outcomes (e.g., Mastrofski et al., 1995), the multilevel analysis should be considered in future research if data permits.

Finally, though this study sought to examine the relationship between race/ethnicity and citizen resistance in contexts where police use deadly force, the data only includes incidents where deadly force was used and resulted in a civilian’s death. This focus limited the scope of analysis as it did not include all deadly use-of-force incidents and overlooked all other incidents where civilians survived. More insight can be gained into the association between civilian race/ethnicity and resistance using a more representative sample of incidents capturing all deadly use of force incidents. Considering these limitations, future research should diversify data collection and analysis methods. Unique approaches could include utilizing video clips captured by officers’ body-worn cameras and conducting systematic social observations with patrol officers to gather data on police use of force against individuals with mental illness. Future research would particularly benefit from using qualitative methods with survivors of deadly police force to unveil further what influences racial/ethnic minority citizens and those experiencing mental illness to resist or comply with police control.

Our findings suggest a few directions for policymakers, police administrators, and researchers to improve police encounters with Black and Hispanic citizens experiencing mental illness. First, the key finding that there is no concrete relationship between citizen race/ethnicity and resistance among those with mental illness calls into question the racial/ethnic disparities in police use of deadly force that many have linked to subject resistance. The focus should be shifted beyond citizens’ behavior during these encounters to include how police officers accurately or inaccurately perceive the threat of a civilian with mental illness and the larger role that bias could play in those perceptions. As Black and Hispanic citizens have historically and contemporarily been associated with stereotypes of criminality and dangerousness (Eitle & Taylor, 2008; Taylor et al., 2019), we must consider how biases—which can manifest consciously or outside of conscious awareness—can affect police behavior.

Second, our findings relating to age, gender, and rural neighborhoods suggest ways police encounters with citizens experiencing mental illness can be improved. Police non-enforcement and service activities need to be geared toward citizens with mental illness who are older, male, and residing in rural neighborhoods to alleviate greater citizen resistance to police interventions. If possible, police officers should take advantage of their routine patrol and calls-for-service opportunities to become acquainted with those with a mental illness, emphasizing males and older citizens. Establishing connections and building trust can go a long way when crises arise around those with mental illness. Rural police departments should also be aware of the potential high violence in responding to distress calls from individuals with mental illness and their family members. Working knowledge about people with mental illness in their jurisdictions could be a promising approach for police officers to minimize possible violent resistance. Police academy and in-service training should improve officers’ educational experiences in dealing with people in mental distress. Police departments can also consider sending officers to advanced professional training to enhance their ability to recognize and help people with mental health needs.

Footnotes

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.

ORCID iD

Ivan Y. Sun

Author Biographies

Krishanna J. Prince is a sociology doctoral student and Equal Justice Fellow in the Department of Sociology and Criminal Justice at the University of Delaware. Her research interests include the intersections of race/ethnicity, inequality, and policing encounters, with a particular focus on experiences within the Black community.

Ivan Y. Sun is a Professor in the Department of Sociology and Criminal Justice at the University of Delaware. His research interests include police attitudes and behavior, public opinion on legal authorities, and crime and justice in Asian societies. His most recent publications on policing have appeared in Criminal Justice and Behavior, Journal of Research in Crime and Delinquency, Police Quarterly, and British Journal of Criminology.

References

Alpert

G. P.

(2015). Police use of force and the suspect with mental illness: A methodological conundrum. Criminology & Public Policy, 14(2), 277–283.

Alpert

G. P.

Dunham

R. G.

(1997). The force factor: Measuring police use of force relative to suspect resistance. Police Executive Research Forum.

Belvedere

Worrall

Tibbetts

(2005). Explaining suspect resistance in police-citizen encounters. Criminal Justice Review, 30(1), 30–44.

Black

(1976). The behavior of law. Academic Press.

Bor

Venkataramani

Williams

Tsai

(2018). Police killings and their spillover effects on the mental health of black Americans: A population-based, quasi-experimental study. The Lancet, 392(10144), 302–310.

Boysen

G. A.

Isaacs

R. A.

Tretter

Markowski

(2020). Evidence for blatant dehumanization of mental illness and its relation to stigma. The Journal of Social Psychology, 160(3), 346–356.

Cohen

Nisbett

Bowdle

Schwarz

(1996). Insult, aggression, and the southern culture of honor: An experimental ethnography. Journal of Personality and Social Psychology, 70(5), 945–960.

Crawford

Burns

(1998). Predictors of the police use of force: The application of a continuum perspective in Phoenix. Police Quarterly, 1(4), 41–63.

Edwards

Esposito

Lee

(2018). Risk of police-involved death by race/ethnicity and place, United States, 2012–2018. American Journal of Public Health, 108(9), 1241–1248.

10.

Eitle

Taylor

(2008). Are Hispanics the new ‘threat’? Minority group threat and fear of crime in Miami-Dade County. Social Science Research, 37(4), 1102–1115.

11.

Ekins

(2016). Policing in America: Understanding public attitudes toward the police. Results from a national survey. Cato Institute. https://www.cato.org/survey-reports/policing-america-understanding-public-attitudes-toward-police-results-national

12.

Engel

(2003). Explaining suspects’ resistance and disrespect toward police. Journal of Criminal Justice, 31(5), 475–492.

13.

Engel

Silver

(2001). Policing mentally disordered suspects: A reexamination of the criminalization hypothesis. Criminology, 39(2), 225–252.

14.

Federal Bureau of Investigation. (2019). 2019 crime in the United States. https://ucr.fbi.gov/crime-in-the-u.s/2019/crime-in-the-u.s.-2019/topic-pages/tables/table-4

15.

Feldman

(2020). Police killings in the US: Inequalities by race/ethnicity and socioeconomic position. People’s Policy Project. https://www.peoplespolicyproject.org/wp-content/uploads/2020/06/PoliceKillings.pdf

16.

Fuller

Lamb

H. R.

Biasotti

Snook

(2015). Overlooked in the undercounted: The role of mental illness in fatal law enforcement encounters. A Report from the Office of Research and Public Affairs. Treatment Advocacy Center. https://www.treatmentadvocacycenter.org/storage/documents/overlooked-in-the-undercounted.pdf

17.

Funk

Drew

Knapp

(2012). Mental health, poverty and development. Journal of Public Mental Health, 11(4), 166–185.

18.

Gray

Parker

(2019). Race, structural predictors, and police shootings: Are there differences across official and “unofficial” accounts of lethal force? Crime and Delinquency, 65(1), 26–45.

19.

Hauck

(2020, October 28). Walter Wallace had a mental illness, his family says. Why did police respond? USA Today. https://www.usatoday.com/story/news/nation/2020/10/28/walter-wallace-philadelphia-shooting-mental-illness/6053831002/

20.

Hirschi

Selvin

(1973). Principles of survey analysis. The Free Press.

21.

Hoekstra

Sloan

(2020). Does race matter for police use of force? Evidence from 911 calls (Working Paper No. w26774). National Bureau of Economic Research.

22.

Hollander

Einwohner

(2004). Conceptualizing resistance. Sociological Forum, 19(4), 533–554.

23.

International Association of Chiefs of Police. (2001). Police use of force in America. Author. https://www.theiacp.org/sites/default/files/2018-08/2001useofforce.pdf

24.

Johnson

(2011). Suspect mental disorder and police use of force. Criminal Justice and Behavior, 38(2), 127–145.

25.

Kaminski

Digiovanni

Downs

(2004). The use of force between the police and persons with impaired judgment. Police Quarterly, 7(3), 311–338.

26.

Kavanagh

(1997). The occurrence of resisting arrest in arrest encounters: A study of police-citizen violence. Criminal Justice Review, 22(1), 16–33.

27.

Kramer

Remster

(2018). Stop, frisk, and assault? Racial disparities in police use of force during investigatory stops. Law & Society Review, 52(4), 960–993.

28.

Lautenschlager

Omori

(2018). Racial threat, social (dis) organization, and the ecology of police: Towards a macro-level understanding of police use-of-force in communities of color. Justice Quarterly, 36(6), 1050–1071.

29.

Lawton

(2007). Levels of nonlethal force: An examination of individual, situational, and contextual Factors. Journal of Research in Crime and Delinquency, 44(2), 63–184.

30.

Lee

Jang

Yun

Lim

Tushaus

(2010). An examination of police use of force utilizing police training and neighborhood contextual factors. Policing: An International Journal of Police Strategies & Management, 33(4), 682–702.

31.

Lee

Vaughn

Lim

(2014). The impact of neighborhood crime levels on police use of force: An examination at micro and meso levels. Journal of Criminal Justice, 42(6), 491–499.

32.

Lee

Bankston

Hayes

Thomas

(2007). Revisiting the southern culture of violence. The Sociological Quarterly, 48(2), 253–275.

33.

Mapping Police Violence. (2022a). Mapping Police Violence. Campaign Zero. https://mappingpoliceviolence.org

34.

Mapping Police Violence. (2022b). Data and Methodology. Campaign Zero. https://mappingpoliceviolence.org/methodology

35.

Mastrofski

Snipes

Supina

(1996). Compliance on demand: The public’s response to specific police requests. Journal of Research in Crime and Delinquency, 33(3), 269–305.

36.

Mastrofski

Worden

Snipes

(1995). Law enforcement in a time of community policing. Criminology, 33(4), 539–563.

37.

McCluskey

Mastrofski

Parks

(1999). To acquiesce or rebel: Predicting citizen compliance with police requests. Police Quarterly, 2(4), 389–416.

38.

Morabito

Kerr

Watson

Draine

Ottati

Angell

(2012). Crisis intervention teams and people with mental illness: Exploring the factors that influence the use of Force. Crime & Delinquency, 58(1), 57–77.

39.

Morgan

Logan

Olma

(2020). Police use of force and suspect behavior: An inmate perspective. Journal of Criminal Justice, 67, 101673.

40.

Mulvey

White

(2014). The potential for violence in arrests of persons with mental illness. Policing: An International Journal of Police Strategies & Management, 37(2), 404–419

41.

National Institute of Justice. (2009). The use-of-force continuum. National Institute of Justice. https://nij.ojp.gov/topics/articles/use-force-continuum

42.

National Institute of Justice. (2011). Overview of less-lethal technologies. National Institute of Justice. https://nij.ojp.gov/topics/articles/overview-less-lethal-technologies

43.

National Institute of Justice. (2020). Overview of police use of force. National Institute of Justice. https://nij.ojp.gov/topics/articles/overview-police-use-force

44.

Nix

Campbell

Byers

Alpert

(2017). A bird’s eye view of civilians killed by police in 2015: Further evidence of implicit bias. Criminology & Public Policy, 16(1), 309–340.

45.

Nix

Lozada

(2021). Police killings of unarmed black Americans: A reassessment of community mental health spillover effects. Police Practice and Research, 22(3), 1330–1339.

46.

Novak

Engel

(2005). Disentangling the influence of suspects’ demeanor and mental disorder on arrest. Policing: An International Journal of Police Strategies & Management, 28(3), 493–512.

47.

Paoline

Gau

Terrill

(2018). Race and the police use of force encounter in the United States. British Journal of Criminology, 58(1), 54–74.

48.

Parker

MacDonald

Jennings

Alpert

(2005). Racial threat, urban conditions and police use of force: Assessing the direct and indirect linkages across multiple urban areas. Justice Research and Policy, 7(1), 53–79.

49.

Payne

Berg

Sun

(2005). Policing in small town America: Dogs, drunks, disorder and dysfunction. Journal of Criminal Justice, 33(1), 31–41.

50.

Payne

Y. A.

(2011). Site of resilience: A reconceptualization of resiliency and resilience in street life oriented Black men. Journal of Black Psychology, 37(4), 425–451.

51.

Rossler

Terrill

(2017). Mental illness, police use of force, and citizen injury. Police Quarterly, 20(2), 189–212.

52.

Sun

Payne

(2004). Racial differences in resolving conflicts: A comparison between black and white police officers. Crime & Delinquency, 50(4), 516–541.

53.

Sun

Payne

(2008). The impact of situational factors, officer characteristics, and neighborhood context on police behavior: A multilevel analysis. Journal of Criminal Justice, 36(1), 22–32.

54.

Sykes

Clark

. (1975). A theory of deference exchange in police-civilian encounters. American Journal of Sociology, 81(3), 584–600.

55.

Taylor

Guy-Walls

Wilkerson

Addae

(2019). The historical perspectives of stereotypes on African-American males. Journal of Human Rights and Social Work, 4(3), 213–225.

56.

Terrill

Mastrofski

(2002). Situational and officer-based determinants of police coercion. Justice Quarterly, 19(2), 215–248.

57.

Terrill

Paoline

E. A.

III . (2013). Examining less lethal force policy and the force continuum: Results from a national use-of-force study. Police Quarterly, 16(1), 38–65.

58.

Terrill

Reisig

(2003). Neighborhood context and police use of force. Journal of Research in Crime and Delinquency, 40(3), 291–321.

59.

Thomas

Jewell

Allen

(2021). Black and unarmed: Statistical interaction between age, perceived mental illness, and geographic region among males fatally shot by police using case-only design. Annals of Epidemiology, 53, 42–49.

60.

U.S. Government Accountability Office. (N.d.). Race in America. https://www.gao.gov/race-america

61.

Washington Post. (2022). Police shooting database, 2015–2021. https://www.washingtonpost.com/graphics/investigations/police-shootings-database/

62.

Wells

Schafer

(2006). Ofﬁcer perceptions of police responses to persons with a mental illness. Policing: An International Journal of Police Strategies and Management, 29(4), 578–601.