Black Bodies on the Ground II: Examining Impactful Policing Policies and Analysis of Policing Disparities Affecting Unarmed African American Men (2015-2020)

ABSTRACT

This article critically analyzed impactful policies and practices governing the policing of Black men historically, and investigated the role these policies played in contributing to the untimely deaths of unarmed Black men by police. The researchers examined cases of police-involved shootings as reported by the Washington Post, with a particular focus on the killing of unarmed Black men over a six-year time frame. It was found that unarmed Black men were disproportionately killed by police and this finding is consistent with prior research that has considered long-term trends in the occurrence of police killings based on race. The researchers conclude with policy recommendations at the local, state, and federal levels to help stem the crisis.

KEYWORDS:

• Unarmed shooting
• African American men
• criminal justice
• police
• policy

Introduction

It has been five (5) years since the publication of Black Bodies on the Ground: Policing Disparities in the African American Community (M. A. Robinson, 2017). While the purpose of this paper is to continue the research on police violence against unarmed African American men, it is essential that one understand how we arrived at this tenuous circumstance. In this paper, the researchers briefly present significant policies affecting the policing of Black men from 1619 through 2020, particularly focusing on where the previous paper left off in the 1970s, to present an argument that policing today is a result of its past. To further our understanding on lethal use of police force against unarmed African American men, we conduct a multivariate analysis of cases involving police shootings of civilians (n = 5,949). The paper concludes with community, state, and federal policy recommendations focusing on improving police relations with the African American community and stemming the rise in police killings of unarmed African American men.

Historical underpinning of laws and policies affecting enslaved and newly freed Blacks (1619-1960ʹs)

To set the context for this paper, we need to go back and examine the policies around policing African Americans that led us to where we are today. Not to rehash what was previously discussed in the first installment of Black Bodies on the Ground, we will review and take a fresh look at policies written and unwritten that directly or indirectly affect the killings of unarmed black men. In 1619, the Dutch Man of War ship docked in Point Comfort, Virginia and unloaded the first enslaved Africans to set foot in the colonies (M. A. Robinson, 2017). These enslaved Africans were brought here to work the fields for planters, but before that could happen, they needed to be indoctrinated into a culture that considered them as subhuman or chattel. To justify this treatment, laws, like the Slave Codes, were passed to govern the enslaved Africans and to protect the planters. These codes had roots in Barbados 1661, Jamaica in 1664, and arrived in the colony of South Carolina in 1691 (Rugemer, 2013). By the end of the eighteenth century, every slave holding state adopted a version of the Slave Codes which were enforced by Slave Patrols. In 1793, the Second United States (U.S.) Congress enacted the Fugitive Slave Laws. The Fugitive Slave Laws were designed to return runaway slaves to their owners and punish anyone aiding in their escape (Finkelman, 1990).

Slavery was abolished after the Civil War and so the laws governing the former enslaved population were no longer deemed just. Five days after the end of the Civil War, President Lincoln was assassinated, and Andrew Johnson assumed the presidency. He quickly moved to readmit all the southern states back into the Union. Former Confederate soldiers were placed in positions of authority until a government could be formed. This led to many southern states passing laws to restrict the movement of the former enslaved population. These laws were known as the Black Codes and remained in effect until 1866. The Black Codes controlled the movements of the newly freed Blacks by arresting them for something as inconsequential as walking down the street (Walking while Black), standing for an extended period of time in one place (Standing while Black), and young Blacks who were accused of the same infractions were oftentimes forced into unpaid apprenticeships with the planters (M. A. Robinson, 2017). Political powers shifted and Congress passed the Fourteenth Amendment and the Civil Rights Act of 1866 (Williams, 2017). The end of the Black Codes gave rise to the repressive and cruel Jim Crow Laws.

Jim Crow Laws were in effect for approximately 100 years after the Civil War. These were written and unwritten laws governing etiquette as a means of controlling the Negro population. Essentially, Jim Crow Laws legalized racial segregation and were enforced by police and the justice system. These tyrannical laws prevented Blacks from drinking from “Whites only fountains,” and public places such as parks, restaurants, and theaters were all segregated. Even jails were segregated which were now being filled by Blacks at record numbers. Jim Crow Laws were challenged by the Civil Rights movement of the 1960ʹs and later abandoned. However, centuries of discriminatory policies consistently devalued the lives of Black Americans, which often utilized law enforcement as a mechanism for punitive control, leading to many tense encounters between officers and the Black community. Moreover, Blacks would continue to face the wrath of a society that still deemed them subhuman and a judicial system that supported the notion that Blacks were socially undesirable as inherent criminals.

Criminal policies and policing practices impacting Blacks in the U.S. (1970-2020)

Modern policy developments and policing practices from 1970 to 2020 further illustrate how policing on the ground has been shaped to treat Black Americans as a potential threat. It is our view that policing today is a result of its past, and the long history of oppressive policies and practices places Black males in disproportionate contact with law enforcement where potential lethal encounters can occur (Barnum & Perfetti, 2010; Tolliver et al., 2016; Warren & Tomaskovic–Devey, 2009). This section briefly highlights several significant policies during this time period and connects them to the killings of unarmed Black males by police.

The controversial “War on Drugs” began during the Nixon administration and found its legal basis in the Controlled Substances Act (CSA), which is a part of the larger Comprehensive Drug Abuse Prevention and Control Act of 1970 (Sirin, 2011). The CSA established drug “schedules” based on medical use and potential for abuse of a given substance. Marijuana and heroin were classified as Schedule I, considered to be the most dangerous, which resulted in heavy criminalization of both substances. It was later stated by President Nixon’s former aide, John Ehrlichman, that the administration had deliberately aimed at getting the public to associate heroin with Blacks in order to target and disrupt the Black community under the new basis of the CSA (Drug Policy Alliance, 2020). Soon after, Nixon formed the Drug Enforcement Administration (DEA) in 1973, which was closely tied to the practice of racial profiling and developed the profiling of drug couriers during the mid-1980s to combat interstate drug trafficking (Farrell & McDevitt, 2010).

Going into the 1980s, expanded efforts in the drug war marked staggeringly high rates of incarceration. The number of people incarcerated for nonviolent drug offenses grew from 50,000 in 1980 to over 400,000 by 1997 (Drug Policy Alliance, 2020). The Anti-Drug Abuse Act of 1986 created mandatory minimum sentencing for federal drug and firearm offenses. Mandatory minimums for drug offenses meant that unique circumstances of a case were dismissed, and that judges must hand down a sentence in accordance with the established minimum. Typically ranging from at least 5–20 years, drug offense minimums were based solely on the type and weight of the drug involved (FAMM, 2021). Most notorious, the shocking disparity between crack and powder cocaine where each gram of crack was considered to be equivalent to 100 grams of powder cocaine in calculating sentences (Fischman & Schanzenbach, 2012; McDonald & Carlson, 1994).

The “crack epidemic” during the 1980s had a devastating impact on many inner-city Black communities (Dunlap et al., 2006; Foundation, 2021; Thompson, 2002). As a result, mandatory minimums for crack cocaine impacted the Black community in nearly every way. Indeed, relative to White offenders, Black offenders were 21% more likely to be sentenced at or above the minimum (Meierhoefer, 1992), and federal prosecutors were twice as likely to charge Blacks with offenses that carried a mandatory minimum sentence than comparable Whites (Starr & Rehavi, 2013). In 2010, the Fair Sentencing Act reduced the disparate ratio between powder and crack cocaine from 100 to 1 to 18 to 1, but the damage of mandatory minimums diminished trust and strained relations between law enforcement and the Black community.

The “tough on crime” political mentality of the 1970s and 1980s continued into the new decade and ushered in the most sweeping crime bill in U.S. history, the Violent Crime Control and Law Enforcement Act of 1994, commonly referred to simply as the crime bill. Most detrimental, the crime bill of 1994 is best known for implementing the “Three Strikes, You’re Out” provision at the federal level with many states adopting the same policy soon after. While the original intention of the law was to focus on repeated serious or violent offenders, the arbitrary inclusion of other property or nonviolent offenses can mean a mandatory life sentence for a defendant with two prior convictions (Shinbein, 1996; Stanford Law School, 2020). The most notorious example is California where over 45% of inmates serving life sentences under the Three Strikes law are Black according to the California Department of Corrections (Stanford Law School, 2020). These federal policies not only impacted sentencing structures and the severity of punishment, but greatly influenced state policing practices, particularly what and where law enforcement targeted, with profound disparities among the Black community (Lynch et al., 2013; Mauer, 2009a).

Over the last three decades, several policing practices consistent with the “tough on crime” policy approach were implemented across the country, shaping everyday interactions between law enforcement officers and the Black community. “Broken windows policing,” originating from broken windows theory (Wilson & Kelling, 1982), focuses on social or physical signs of disorder in communities. Disorder is thought to be a visual cue for motivated offenders to target an area, which leads to increased fear, crime, and withdrawn informal social controls from neighborhood residents (Piscitelli & Doherty, 2018; Wilson & Kelling, 1982). Police were encouraged to intervene on visual cues of disorder, no matter how minor, thereby preventing bigger crimes. In practice, broken windows policing was essentially “zero tolerance” policing, which requires arrests and summons as opposed to informal responses to minor infractions (Howell, 2015). Zero tolerance policing was particularly popular in New York City where this form of policing was conducted primarily in communities of color (Howell, 2015). Broken windows and zero tolerance policing exponentially increased the number of police-civilian encounters for minor offenses, which increased the potential risk for discretionary use-of-force by police to end in injury or death (Howell, 2015).

Another contributor to the risk of lethal use-of-force is “stop and frisk,” a policing tactic where officers briefly stop suspicious individuals to see if that person may be involved in a crime or in possession of a weapon. If reasonably suspected of having a weapon, the officer is permitted to pat down the individual, which was ruled constitutional during the 1968 Supreme Court case, Terry v. Ohio. A recent study on stop-and-frisk practices in Philadelphia found that predominantly Black neighborhoods had 70% more frisks than non-Black areas, but these frisks generated less weapons (Hannon, 2020). Furthermore, in 2013, a New York City judge ruled that stop-and-frisk policies as they were conducted by the New York Police Department were evidence of “indirect racial profiling,” possibly due to unconscious bias, and violated the constitutional rights of minorities, but this decision did not end the practice of stop-and-frisk (Goldstein, 2013).

More recently, following the highly controversial killings of Michael Brown and Eric Garner in 2014, the Obama administration established the President’s Task Force on 21^st Century Policing (Hineman, 2020; President’s Task Force, 2015). Along with recommendations for fundamental reform toward community-oriented policing practices and altering police culture, the task force led to a sharp increase in body-worn cameras by police officers to increase accountability and transparency (Koslicki et al., 2021). The task force’s final report made 59 recommendations (President’s Task Force, 2015), with one survey reporting that approximately 40% of police departments made changes to their use of force policies and increased de-escalation training within the first two years after the final report (Stephens, 2019).

However, in 2016, the Trump administration championed a police-first position and vowed to change what he called an anti-police atmosphere (DeVylder, 2017). Despite his rhetoric, in June 2020, President Trump signed an executive order targeting police reform following the high-profile police killing of George Floyd (Kelly & Naylor, 2020). The order called for the creation of a database to track officers with repeated incidents of misconduct as well as the use of federal funds to incentivize police departments to maintain higher certifications for protocols on use of force (Kelly & Naylor, 2020). While the long-term impact of this order remains to be seen, skeptics continue to question the efficacy of efforts to reduce these tragedies in the wake of the deaths of George Floyd, Breonna Taylor, Jacob Blake, and others (Sullivan, 2020).

Lethal use of force from police

The publicized killings in 2020 have added to the already tenuous relationship between police and the Black community, sparked national outrage, and demonstrated that action is still desperately needed. The history of written and unwritten laws, policies, and policing procedures have greatly contributed to the current social justice crisis of police-related deaths of Black males, particularly in creating and sustaining a culture that systematically denigrates the Black community in the U.S. (Tolliver, 2016). Many researchers have found that police officers disproportionately intervene with racial and ethnic minorities in a variety of ways including arrests, searches, being stopped, and use of force (Eith & Durose, 2011; Higgins et al., 2011; Langton & Durose, 2013; Walker et al., 2016). Further, scholars have attempted to try and understand these disparities as well as why law enforcement officers make these decisions with the use of theory, beyond any explanations of greater offending among the Black community (Engel et al., 2002; Fridell & Lim, 2016; Leiber, 2008).

A number of theories have been posited to explain disparities in instances of lethal use of police force. One theory attempting to explain the killings of unarmed Black men by police is Critical Race Theory (CRT), which displays how racism is embedded into social institutions, including law enforcement entities, and how White leadership can reinforce this discrimination (Chaney & Robertson, 2013). A part of CRT and related legislation is the prosecution rate of police officers: only 33% of misconduct cases by police officers were convicted. In addition, “the average incarceration, post-conviction, for officers was only 14 months compared to 49 for the general public” (Chaney & Robertson, 2013). Another theory aimed at explaining this phenomenon is Contact Theory (CT), and the notion behind this theory is that intergroup interactions can decrease prejudice, which extends to the police and Black Americans (Davis, 2018). In order for there to be a positive change, Allport requires 4 conditions to be met: “equal status between groups in the situation, common goals, intergroup cooperation, and the support of authority and law” (Allport, 1954). This can relate to deadly police force against Black men because the in-group’s greater knowledge and understanding of the out-group can reduce stereotyping and prejudice, as well as the out-group’s overall mental functioning, including anxiety and perceived threats (Davis, 2018).

In addition, Fundamental Attribution Error (FAE) has been applied to explain the interactions between Black males and law enforcement. FAE underscores that “people tend to make internal (personal) attributions about others, rather than external ones, even when there are clear signs of potential situational causes,” which is especially relevant in individualistic cultures like the United States (Davis, 2018). This is mainly due to biased information processing but can also be impacted by quick decision-making in high-stress environments (Davis, 2018). Davis actually proposed an equation to predict shootings of Black males: “intense pressure + racial stereotyping x dispositional attributions of the actor = shooting of unarmed African Americans” (Davis, 2018); however, this formula needs to be studied more comprehensively in the literature. Similarly, other cognitive bias theories, like “implicit bias,” suggest that people unconsciously link persons with stereotypes of their groups, which can influence behavior and occur even when people hold conscious non-prejudiced beliefs (Correll et al., 2007; Fridell & Lim, 2016; Hardin & Banaji, 2013)

Recall the New York judge that cited the role of implicit bias resulting in more Black people being stopped and frisked (Goldstein, 2013). The history of policies and practices that have facilitated a social environment that categorizes Black males with aggressiveness or criminal behavior could play into the potential for implicit cognitive biases to arise in the lethal use of police force (Correll et al., 2007; Fridell & Lim, 2016). The authors of this study make the supposition that the aforementioned policies were written to protect a society from the perceived threat of African American men. These policies directly and/or indirectly lead to increased contact between police and African Americans, and oftentimes, resulted in the premature death of these men. Moreover, these policies protected police officers and in some instances, armed civilians who committed homicide using these laws as justification. Just as policies and practices have contributed to the social issue, changes in policies and practices can help reorient the potential and severity of these situations to occur. However, not enough is known about the prevalence and characteristics of lethal use of police force against unarmed Black males among national samples. Building on the work of our previous study, we will examine lethal use of force by police committed against African American males from January 2015 through December of 2020 using a national database.

According to the Washington Post database, 1,636 Black males have been fatally killed by police since 2015 and of those, 138 were unarmed. In comparison, 3,015 White males have been killed by police forces since 2015, and of those, only 164 were unarmed (The Washington Post, 2022). Given these numbers, 8.4% of Black men who were shot were unarmed compared to 5.4% of White unarmed men. It is also important to account for population demographics in considering these statistics: “[Black Americans] account for less than 13% of the U.S. population, but are killed by police at more than twice the rate of White Americans” (The Washington Post, 2022). We are particularly interested in understanding the incident-level correlates of police-related deaths of unarmed Black males, as these individuals likely posed little, if any, threat to the life of the officer given their unarmed status.

Current study

The historical, policy, and theoretical contexts lay a foundation for understanding how African Americans are disproportionately encountering police officers, but research is still needed on national prevalence and characteristics surrounding lethal use of police force against unarmed African American men. Thus, the current study examines incidents of police-related civilian deaths over a six-year period (2015–2020) to assess trends and factors associated with unarmed African American male deaths. Using national data from the Washington Post, this study seeks to answer the following research questions: 1) Which characteristics are more likely to be present in police shooting deaths of unarmed individuals? 2) Are rates of police shootings higher among African American individuals compared to their representation in the United States population?

Methods

Sample

The data analyzed in this study was retrieved from the Washington Post (WaPo), which was founded in 1877 and is widely read and respected as a credible source of news. The Washington Post began compiling a database on police killings of civilians starting from January 1, 2015. The Washington Post retrieved the data by reviewing news sources, internet accounts, civilian accounts, and public records. The researchers of this study triangulated 25% of the WaPo data by researching local news sources where the killings occurred. For example, Mr. Robert Lawrence White was in the WaPo database and verified by an internet source which located the story The Daily Record, a publication in the state of Maryland, and in the Crime and Safety section of the Patch, a publication in Silver Springs, Maryland. Additionally, several other studies have begun using the WaPo database to examine police-related fatalities (Nix et al., 2017; Pinchevsky & Nix, 2018; Shane et al., 2017).

Measures

Dependent variable

The dependent variable for the current study was whether victims were armed or unarmed at the time they were killed by police. This variable has been dummy coded for the current analyses as 0 = armed and 1 = unarmed. Individuals were grouped into one of two categories based on whether they were unarmed, or whether they were in possession of a gun, knife, or other weapon at the time of their death according to the WaPo database. Individuals were coded in this manner given the main focus of this research was on individuals who were unarmed, as recent police shootings publicized in the media have been against unarmed African American individuals, which was the impetus for the Black Lives Matter movement (Adedoyin et al., 2018). However, both armed and unarmed individuals were analyzed.

Independent variables

A key variable for the current study was the race/ethnicity of civilians killed by the police. This is an important consideration, given recent and historical incidents of police misconduct and implicit biases on the basis of race/ethnicity. For the current analyses, race is broken down into the following categories: White, African American, Hispanic, and Other. Each race variable was dummy coded for the current analyses. Of note, when combining Native American, Hawaiian, and other ethnic groups in the “other” category, this group still constituted the smallest category in terms of race/ethnicity.

Another variable included in the current analyses indicate whether the state in which the incident occurred was historically a slave state or a free state. States were identified as being a slave state or free state based on their status prior to the Civil War in 1860. This variable was coded as 0 = free state and 1 = slave state. This variable was included as prior research has indicated that former slave-holding states represent greater numbers of unarmed individuals killed by the police annually in the United States (M. A. Robinson, 2017).

Other variables included were the presence of signs of mental illness, which was dummy coded as 0 = no and 1 = yes, and whether or not the victim was fleeing at the time of the incident. Furthermore, if they were fleeing, how were they fleeing? This variable was operationalized as 0 = not fleeing, 1 = car, 2 = foot, 3 = other. Threat level indicates whether the officer faced an immediate threat at the time the civilian was killed. This was operationalized as 0 = attack, 1 = other, and 2 = undetermined. For the threat level to be classified as an attack, the individual had fired shots, brandished a weapon, or physically attacked a person. Other threats were those that did not involve firearms or physical attacks, but involved some other serious threat. Meanwhile, undetermined threats were those cases with limited information or were undergoing further investigation (Brittain, 2015). Finally, civilian age is measured as a continuous variable.

Analytic strategy

For the purposes of this study, we reviewed cases that involved police killings of African American men, non-Hispanic White American men, and LatinX men which occurred between January 1 ^t, 2015, and December 31, 2020. This allowed us to analyze 5,949 reported cases of police shootings of civilians that occurred during this six-year time period. We begin by presenting descriptive information for civilians killed by the police, and the breakdown of characteristics present in these shootings. We then present the racial breakdown of unarmed civilian shootings as both raw counts and as incidence rates based on U.S. racial representation per 100,000 as of 2019. Next, we will present bivariate results in the form of chi-square analyses, which will then lead to the presentation of our logistic regression model. Analyses were conducted using SPSS 27.0.

Results

Table 1 provides the frequencies and descriptive information for all civilians killed by the police from 2015 through 2020 as reported in the media (n = 5,949). The majority of these civilians were shot by police, which resulted in their death. For each year, most civilians that were killed by police were White, although as will be explained momentarily, this count differs based on their representation within the U.S. population. Most civilians killed by the police did not demonstrate signs of mental illness upon their death. Across the six-year time span, the threat level toward the police in many cases was classified as other, meaning that civilians were engaging in certain behaviors toward the police or other citizens that did not constitute an imminent threat toward either party (see, Nix et al., 2017 for a detailed explanation).

Table 1. Descriptive information for civilians killed by the police, 2015–2020.

	2015		2016		2017		2018		2019		2020
	n	%	n	%	n	%	n	%	n	%	n	%
Total shootings	993		960		986		990		999		1021
Civilian
Armed	852	90.0	837	92.9	835	92.2	912	94.0	945	94.6	960	94.1
Unarmed	95	10.0	64	7.1	71	7.8	58	6.0	54	5.4	60	5.9
Manner of death
Shot	924	93.1	898	93.5	935	94.8	959	96.9	961	96.2	972	95.2
Shot and Tasered	69	6.9	62	6.5	51	5.2	31	3.1	38	3.8	49	4.8
Age (mean)	36.7		36.6		36.7		37.5		38.0		37.4
Gender
Male	951	95.8	920	95.8	940	95.4	938	94.7	956	95.7	983	96.3
Female	42	4.2	40	4.2	45	4.6	52	5.3	43	4.3	38	3.7
Race
White	503	50.7	467	48.6	461	46.8	462	46.7	424	42.4	457	44.8
African American	258	26.0	235	24.5	224	22.7	232	23.4	252	25.2	243	23.8
Hispanic	172	17.3	161	16.8	181	18.4	167	16.9	167	16.7	170	16.7
Other	60	6.0	97	10.1	120	12.2	129	13.0	156	15.6	151	14.8
Signs of Mental Illness
No	733	73.8	713	74.3	747	75.8	770	77.8	784	78.5	808	79.1
Yes	260	26.2	247	25.7	239	24.2	220	22.2	215	21.5	213	20.9
Threat Level
Attack	732	73.7	547	57.0	598	60.6	615	62.1	689	69.0	685	67.1
Other	227	22.9	386	40.2	368	37.3	339	34.2	278	27.8	293	28.7
Undetermined	34	3.4	27	2.8	20	2.0	36	3.6	32	3.2	43	4.2
Fleeing? If so, how?
Not fleeing	716	72.1	629	66.6	598	63.9	615	65.6	593	64.1	533	59.3
Car	143	14.4	143	15.1	182	19.4	162	17.3	173	18.7	177	19.7
Foot	108	10.9	121	12.8	123	13.1	137	14.3	138	14.9	142	15.8
Other	19	1.9	51	5.4	33	3.5	24	2.6	21	2.3	47	5.2
Was officer wearing body camera?
No	918	92.4	814	84.8	878	89.0	872	88.1	863	86.4	850	83.3
Yes	75	7.6	146	15.2	108	11.0	118	11.9	136	13.6	171	16.7
Free State or Slave State?
Free state	475	48.0	448	46.9	481	48.9	473	47.8	435	43.6	476	46.7
Slave State	514	52.0	507	53.1	503	51.1	516	52.2	563	56.4	544	53.3

Some values may not equal the total number of shootings due to missing information.

While it appears that the numbers of unarmed individuals killed by police have decreased over time, these differences are more accurately depicted when considering their representation within the U.S. population broken down by race, which is depicted in Figure 1. When considering their representation within the U.S. population per 100,000, African American civilians demonstrated to have the highest rates of being killed by the police when unarmed. Population data were derived from the United States Census’ Annual Estimates of the Resident Population by Sex, Race, and Hispanic Origin for the USA: April 1, 2010 to July 1, 2019 (table NC-EST2019-SR11H), where Census data were aggregated by race and ethnicity. Census data from 2020 were obtained from the 2020 Decennial Census Redistricting Supplementary tables.

Figure 1. Rates of shootings of unarmed civilians by the police per 100,000.

Chi-square analyses and t-tests were conducted to determine the presence and extent of significant relationships with civilian armed/unarmed status at the bivariate level. Chi square analyses indicated that officer use of body camera (X² (1, N = 5,743) = 15.834, p < .001), whether or not the citizen was fleeing (X² (3, N = 5,743) = 38.984, p < .001), threat level (X² (2, N = 5,743) = 185.397, p < .001), race (X² (3, N = 5,270) = 14.027, p = .003), and signs of mental illness (X² (1, N = 5,743) = 5.247, p = .022) were all significant at the bivariate level. A t-test was conducted for age and armed/unarmed status, which indicated that age was significant at the bivariate level. Unarmed citizens killed by police were younger (M = 32.50, SD = 11.05) than armed citizens (M = 37.65, SD = 13.14) killed by the police, on average, t (490.320) = 8.839, p < .001. The occurrence of shootings in a slave state vs. free state was not found to be significant at the bivariate level.

Prior to conducting the multivariate analyses, we sought to determine if these results would be affected by multicollinearity. In this case, VIF scores were used to check for multicollinearity; scores ranged from 1.009 to 1.081, indicating that multicollinearity is not a problem with the current analyses. Multivariate logistic regression analyses were then conducted to determine the influence of the above variables on whether police killed armed or unarmed citizens (Table 2). Within the model, manner of death (OR = 2.475, 95% CI [1.722–3.556]), age (OR = 0.973, 95% CI [0.963–0.983]); presence of body camera (OR = 1.362, 95% CI [1.025–1.808]); facing an attack, other (OR = 2.929, 95% CI [2.331–3.681]) or undetermined threat (OR = 9.990, 95% CI [6.381–15.642]), both as compared to an attack; and various levels of fleeing as compared to not fleeing–fleeing by car (OR = 1.539, 95% CI [1.146–2.036]) and fleeing by foot (OR = 1.763, 95% CI [1.325–2.345]) were all significant. If an individual was shot and tasered, they were approximately 2.5 times more likely to be unarmed compared to those who were shot. For each year increase in the individual’s age, there was a 2.7% decrease in the likelihood of being armed. Unarmed individuals were 36% more likely than armed individuals to be killed by officers with body worn cameras compared to officers without body worn cameras. Unarmed individuals facing an “other” or undetermined threat were more likely than those facing a direct attack to be killed by officers; they were three and ten times more likely, respectively, to be killed. Finally, unarmed individuals who were fleeing by car or on foot were 54% and 76% more likely to be fatally shot by officers compared to individuals who were not fleeing, respectively.

Table 2. Logistic regression for variables contributing to police shootings of unarmed citizens.

Variables	b	SE	OR		95% CI
Manner of death	0.906***	0.185	2.475		1.722–3.556
Age	−0.028***	0.005	0.973		0.963–0.983
Gender (male reference)	0.296	0.240	1.345		0.839–2.155
Signs of mental illness	0.252	0.145	1.287		0.969–1.710
Body camera	0.309*	0.145	1.362		1.025–1.808
Slave owning state	−0.018	0.118	0.982		0.779–1.237
Race (White as reference)	-
African American	0.096	0.137	1.100		0.841–1.440
Hispanic	−0.017	0.155	0.983		0.726–1.331
Other	0.081	0.272	1.084		0.636–1.848
Threat level (Attack as reference)	***
Other threat	1.075***	0.117	2.929		2.331–3.681
Undetermined	2.302***	0.229	9.990		6.381–15.642
Fleeing? If so, how? (Not fleeing as reference)	***
Fleeing by car	0.431**	0.150	1.539		1.146–2.036
Fleeing on foot	0.567***	0.146	1.763		1.325–2.345
Fleeing by other means	0.000	0.311	1.000		0.544–1.839
Constant	−2.597	0.260	0.074
	Nagelkerke R^2 = 0.120

* p ≤ .05; ** p ≤ .01; *** p ≤ .001.

Discussion

The Washington Post police shootings database was developed in 2015 after their investigation discovered that the FBI had been underreporting these occurrences in their official counts (Tate et al., 2016). Since its development, stories of police shootings in the media have seemingly increased during this same time period, despite analyses of long-term trends of police shootings indicating the contrary (Campbell et al. (2018). It was the attention from these high-profile shootings, however, that raised the question of police legitimacy, particularly with regard to race (Wallace et al., 2018).

The current study focused on unarmed civilians killed by the police between 2015 and 2020. Descriptive statistics found that in general, specific information surrounding these shootings varied widely across years, which may be partially attributed to the details provided for each incident within the WaPo database. While the findings indicated that the number of killings of unarmed citizens did not vary greatly across years, when considering racial representation per 100,000 within the U.S. population, there was a greater proportion of unarmed African Americans killed by police during this period. This finding is consistent with prior research that has considered long-term trends in the occurrence of police killings on the basis of race (Krieger et al., 2015; Price & Payton, 2017). While prior research has considered the race of the officer in addition to that of the civilian, this data is currently not available in the WaPo database. Furthermore, without the availability of such comprehensive data, it is difficult to make the conclusion that police disproportionately use force against minorities on a national scale, as many agencies have demonstrated reluctance in publicizing use of force incidents (Nix et al., 2017).

Other characteristics of police killings of unarmed civilians were also taken into consideration in this study. For instance, facing an attack, other, or undetermined threat, presence of a body worn camera, not fleeing, fleeing by car, and fleeing by foot all increased the odds that the citizen was unarmed at the time of their death, while only age decreased the odds of the citizen being unarmed at the time of their death.

Facing an attack or other imminent threat and the association with unarmed citizen deaths takes into consideration the “objectively reasonable” use of force standard, and its basis in the Fourth Amendment. In Graham v. Connor (1989), the Court ruled that the “objective reasonableness” of a use of force standard should be based on the perspective of the officer at the scene, while considering the severity of the offense, the threat posed by the citizen, and any attempts at evading or resisting arrest. Because of the broad definition of “objectively reasonable,” agencies have further clarified their use of force policies by developing parameters in applying such force (Walker, 2007), typically by providing a continuum outlining the progression of force ranging from least restrictive/verbal to most restrictive/deadly force (Terrill & Paoline, 2013). However, the extent to which this progression is carried out varies widely.

Other significant variables – year of incident, being African American, not fleeing, fleeing by car, and fleeing on foot – decreased the odds that a citizen would be unarmed at the time of their death. In other words, these variables were associated with armed citizens. Given the increased media attention to these high-profile shootings, tensions between the police and the public remain elevated, as does the presence of firearms, especially in disadvantaged areas where officer-citizen encounters are more prevalent (Fridel et al., 2020). Similarly, being African American was associated with being armed at the time of death. This has also been indicated in previous research, which has found that officers respond more quickly in identifying guns or gun-like objects with African Americans than with other races (Lee, 2004; Payne, 2001). Reasons why African American males may be more likely to carry firearms during these incidents are challenging to completely disentangle. However, results from qualitative studies have suggested that concepts similar to Anderson’s (2000) The Code of the Streets, such as lack of faith in the police, fear of the police, perceived lack of police accountability, and beliefs that police officers are often aggressive or abusive, may play a role in these individuals feeling a need to carry a firearm for self-protection (Calvert et al., 2020; O’Neill et al., 2021; Rich & Grey, 2005).

Those citizens killed by the police who were not fleeing were more likely to be armed. Not fleeing may indicate a number of scenarios with an officer; for example, the citizen might have threatened death or imminent bodily harm to the officer, or the citizen (as an innocent bystander) might have been killed as they were attempting to use deadly force on another fleeing citizen. With respect to fleeing by car or fleeing on foot, the vehicle or other available objects could potentially be used as weapons in addition to the firearm, thereby increasing the likelihood that an officer would resort to deadly force.

Our findings suggest that while the number of unarmed men killed by police has decreased from 2015–2020, African American men are still being killed at disproportionate rates. Further, African Americans having higher odds of being armed in these encounters highlights the enduring tumultuous relations between law enforcement and the African American community. Policy reform that accounts for the unique adverse impact of police-related fatalities among this community may serve to reduce the disproportionate rates, tensions, and likelihood of encountering armed civilians. Based on our findings and our account of major U.S. policies and practices, we provide several recommendations to address ongoing racial disparities and lethal use-of-force toward the Black community.

Future recommendations

Community recommendations

As we contend in our first paper, Black Bodies on the Ground (M. A. Robinson, 2017), increasing community involvement and relationship building from police officers in the areas they patrol can lead to reduced fear amongst Black civilians, which aids in facilitating trust and cooperation (Brown & Benedict, 2002; Stoutland, 2001). Additionally, public forums with community members and law enforcement officers provide a safe, productive environment for forming positive relationships (M. A. Robinson, 2017). Although these two recommendations were first discussed in greater detail during our previous installment, the same issues have persisted since its publication, suggesting these strategies are still needed.

Turning toward alternative initiatives, instituting civilian review boards comprised of citizens and stakeholders living in the community to review all cases of police misconduct and complaints (i.e., unwarranted stops, use of excessive force, etc.) would increase police accountability to its citizens. Community policing, as both a philosophy and practice, is predicated on “co-production” between police officers and community members, which necessitates clear standards for police accountability and “adequate citizen voice in holding to account” (Somerville, 2009, p. 268). This lateral measure of accountability utilizes peer review and collective self-regulation to ensure community members are included in forming shared goals for community safety and fair law enforcement. Achieving adequate citizen voice cannot be understated, and for Black community members, ensuring sufficient representation in review boards is paramount.

State and local policy recommendations

For states and localities, there is a large deal of control over day-to-day law enforcement operations. As such, states should adopt trailblazing policies that target racial disparities as well as improve officer training and accountability. Requiring racial impact statements, an evaluative tool for lawmakers, provides legislators with a report that examines potential unforeseen policy consequences that could worsen racial disparities (Mauer, 2009b). Racial impact statements aim to evaluate disparities of proposed legislation prior to implementation (Mauer, 2009b; Porter, 2021). At the time of this research, there are nine states (Colorado, Connecticut, Florida, Iowa, Oregon, Maine, Maryland, New Jersey, and Virginia) that have adopted the use of racial impact statements (Porter, 2021), indicating that there is substantial room for improvement in this area.

While the efforts from the President’s Task Force on 21^st Century Policing during the Obama administration saw an increase in body-worn cameras by police officers (Koslicki et al., 2021), there is still a need for consistent and transparent use of this technology. Body cameras should be common practice in all police departments and should be activated the entire shift, except for personal time, and lunch/breaks. If they are not activated, an alert could be sent to the shift commander as this may cut down on the number of cases where the body camera was not activated during an encounter with a civilian. This suggestion is supported by data, as our findings show an overwhelming majority of the incidents involving police killing of unarmed African Americans where the officer did not have an activated body camera or did not have a body camera.

States could also increase police accountability by implementing policies that promote self-regulation and lateral responsibility among officers. Like the Himmel Rule (Gosbee, 1991) for attorneys, there should be a similar rule governing police officers (Chauvin Rule) who do not intervene and/or report violations of other officers for inappropriate behavior; for instance, when excessive force is used, improper or unwarranted police stops, or any other inappropriate behavior by fellow officers. Finally, in response to contemporary policing practices that emphasize punitive measures for minor infractions, such as “zero tolerance” and “stop and frisk,” law enforcement agencies should emphasize more commensurate controls (i.e., warnings or citations) for minor offenses. State and local officials have sufficient power to reduce arrests for minor infractions, such as low-level drug offenses, and shift protocols for handling unarmed offenders. Investing in evidence-based implicit bias training (Kang & Lane, 2010), prioritizing appropriate referrals to social services for minor disturbances related to mental health or poverty, and diversion to treatment for minor drug-related offenses are all strategies to ameliorate racial disparities and reduce the incarcerated population.

Federal policy recommendations

At the federal level, we begin with a recommendation that was made previously (M. A. Robinson, 2017), but has yet to be widely adopted. Legislation is needed requiring police departments to report all use-of-force incidents that result in the deaths of civilians to a national database. To ensure agencies comply, all federal funds should be withheld until all incidents are reported. While the Federal Bureau of Investigation (FBI) began collecting national use of force data in 2019, participation from law enforcement agencies is entirely voluntary (FBI Website, 2020). Further, their inaugural release of national use-of-force data in 2019 revealed that only 41% of law enforcement officers were represented in the data (FBI Website, 2020). Much like the challenges faced while conducting this research, improved data collection and sharing on this issue may serve to enhance evaluation of tailored policies and practices.

A national use-of-force policy could be instituted, which would be developed by an ensemble of policing experts, medical personnel, community members, and mental health experts. While each law enforcement agency has a use-of-force standard in place, these standards and regulations generally vary depending on the jurisdiction (D’Antonio, 2020). Thus, federal legislation that sets a singular use-of-force standard nationwide is needed to curb the failures in use-of-force regulation. This national standard should clearly emphasize de-escalation, preservation of life, and reducing excessive violence. As a starting reference for this legislative push, the International Association of Chiefs of Police (IACP) has released a discussion paper on a guided national consensus from 11 law enforcement leadership/labor organizations on use-of-force (See National Consensus Discussion Paper on Use of Force and Consensus Policy, IACP Website, 2020).

To boost collaborative emergency response, the federal government should work to provide funding or reallocate current funding for developing and training joint police and social worker mental health response teams. These teams can be dispatched for domestic violence, citizens in distress, or an apparent mental health crisis. Specialized response teams that include both police officers and social workers may be better suited to correctly identify and provide service-based solutions that address the underlying causes for minor crimes such as mental illness, poverty, or substance use. In addition to improving law enforcement responses to reduce the incarcerated population, the federal government should look to effectively end the War on Drugs by reducing the number of low-level drug offenders prosecuted in federal courts. Moreover, federal legislation should largely eliminate mandatory minimum sentences to improve accounting of individual case circumstances by judges. As previously mentioned, there is some unclear discretion given to federal prosecutors when deciding who would be charged with mandatory minimum sentences, which disproportionately impacts Blacks (Starr & Rehavi, 2013). Ending mandatory minimums would reduce this racial disparity in the federal prosecutorial process.

All the recommendations at the community, state/local, and federal levels would work toward breaking the cycle of unarmed Black men unjustifiably losing their lives at the hands of law enforcement officers. Indeed, the history of major criminal policies and practices highlighted in this paper provides contextual understanding for the ongoing incidents that have occurred since our first paper from 2017. Most distressing, the culmination of police violence and civil unrest during 2020 has undoubtedly demonstrated the need for reform. This period in the U.S. should be seen as a historical turning point leading to unified efforts and the passing of effective, just policies.

Limitations

The database used in this study is maintained by an unofficial source, the Washington Post; this is because police departments are not required to report deaths of civilians by police to any national database (M. A. Robinson, 2017). The researchers are reasonably sure that the incidents in the database are accurate; however, if a death by police was not reported in a reputable news source, there is a chance it may not be included in the database. Another limitation is the reporting of the mental health status of the men in the database, as it may not have been known at the time that the case was input in the dataset, and research shows that African American men are less likely to seek mental health support, and more likely to be misdiagnosed (Hampton, 2007; Holden et al., 2012). A third limitation is related to information that is available within the database for more recent occurrences. While the story is accounted for in the database, specific information may be lacking regarding details of these shootings, such as the type of weapon that was used, or nuanced details regarding the victim. Over time, these details are added to the database as they are triangulated with other sources. A final limitation is related to the availability of officer characteristics within the database, including gender, race/ethnicity, and number of officers involved in the incident. Without the availability of these characteristics, we were unable to determine potential racial disparities between the officer(s) and the victim.

Future directions for research and practice

Given the nationwide recognition and action for police reform, particularly following the events of 2020, there are substantial avenues for future research and practice in this area. One important study that is warranted is collecting data on the race/ethnicity of the police officer as this may illuminate whether this is a policing culture issue, racial bias issue, or both. In addition, it would be compelling to investigate the police officer’s employment history analyzing complaints and/or disciplinary procedures levied against the officer, as it may provide insight into policies that govern that police department. With continuous changes at both the state/local and federal levels, future research needs to examine the impact of novel policy changes. Moreover, it would be beneficial to investigate police killings of African Americans with a particular focus on cities with co-responding police and mental health teams, compared to cities without the teams. Case study comparisons in this way could help ascertain what makes a difference and what circumstances shape those changes. Finally, for social work practitioners, it is essential to establish larger roles within police departments. Social workers should assist officers with service-related solutions and increase knowledge on evidence-based interventions for social issues that plague many minor offenders.

Disclosure statement

No potential conflict of interest was reported by the author(s).