The dark footprint of state violence: A synthetic approach to the American crime decline

Abstract

This project combines the conversation on the national crime rate with emerging discussions on the violence that the state perpetrates against civilians. To measure US lethal violence holistically, we reconceptualize the traditional definitional boundaries of violence to erase arbitrary distinctions between state- and civilian-caused crime and violence. Discussions of the “crime decline” focus specifically on civilian crime, positioning civilians as the sole danger to the health, wealth, and safety of individuals. Violence committed by the state—from police homicide to deaths in custody to in-prison sexual assault—is not found in the traditionally reported crime rate. These absences belie real dangers posed to individuals which are historical and contemporary, nonnegligible, and possibly rising. We present Uniform Crime Report data side-by-side with data on police killings, deaths in custody, and executions from sources such as Fatal Encounters, the Washington Post, the Guardian, and the Center for Disease Control to produce a robust discussion of deaths produced through the criminal legal system. We ground this empirical analysis in a broader conceptual framework that situates state violence squarely within the realm of US crime, and explore the implications of this more holistic view of crime for future analyses.

Keywords

crime control crimes of the state crime trends critical criminology definitions of crime homicide police and policing police homicide state violence violent crime

Introduction

Mainstream criminology relies for disciplinary coherence on a number of assumptions. In criminological theory, bright lines separate criminals from non-criminals, conformers from nonconformers, and law abiders from legal transgressors. Methodologically, violence recorded by official statistics provides information about crime and crime trends. Violence which escapes—or which is misclassified by—these counts is acknowledged but typically assumed to be missing nonsystematically. These assumptions structure criminology’s canon and motivate its core concerns. Consider, for instance, the apparently simple criminological truth that crime dropped precipitously in the 1990s, returning to the relatively low levels of the 1970s and prior. Combined with evidence of soaring incarceration rates and the proliferation of law enforcement personnel and infrastructure, the Great American Crime Decline provided scholars with a rich body of empirical data and a clear disciplinary thrust: the examination and furtherance of crime prevention by state action. Consumers of such research are guided to the unavoidable conclusions that the state has succeeded in making US society quite safe and it is getting safer.

Here, we argue that the crime decline (and thus social safety) has not only been mismeasured, but more importantly, misconceptualized. Criminology’s myopic focus on civilian crime and violence obscures actions by the state and results in a substantial underestimation of the prevalence of violence. The state is a major purveyor of violence and has a violent footprint larger than the collective civilian violence of most US states and territories. This underestimation is systematic, beginning with criminological definitions and measurement of crime, and largely invisible.

The illusion that public safety is synonymous with the official civilian crime rate may be showing some cracks under the Trump administration. In 2018 journalists uncovered a massive ring of kidnappers, traffickers, and torturers. Organized, armed groups kidnapped nearly 2000 families, coercively holding parents and separating them from their children, perhaps permanently. Some kidnap victims were secretly drugged, subjected to torturous terms of confinement, and sexually assaulted; some died as a result (Ellis et al., 2018; Reilly, 2018; Speri, 2018). Much of the US public was horrified and took moral (and sometimes physical) stands against this violence, which continues unabated. Yet none of this—save actions of those protesting the violence (Ravani, 2018; Shepherd, 2018)—was considered “crime” in any official sense and none appears in national crime figures.

This example describes only a portion of the current federal policy of “zero tolerance” toward undocumented immigrant and refugee families (violations of civil, not criminal, law). The kidnappers and torturers mentioned above are federal agencies and contractors—Immigration and Customs Enforcement, Border Patrol, Homeland Security, Federal Protective Services, and detention facilities—performing their prescribed tasks. Our reformulation here is a reversal in the classic style of Bursik and Grasmick (1993) who identified the double standard separating urban gangs from college fraternities. Both use specific signs/signals, commit crimes against persons disproportionately, abuse substances, are geographically tied, and so on, yet are ideologically separated by arbitrary and inflated distinctions. Real differences are obscured: one group consists primarily of impoverished racial minorities facing “multiple marginality” (Vigil, 2002), while the other is predominantly wealthy, white, and connected to powerful institutions. We apply this logic here: violence enacted by the state is still violence, but these actions will escape categorization as crime for the purposes of official statistics. Regardless of future determinations of the legality of the current federal immigration policy, no agents will be prosecuted for kidnapping or child trafficking.

While concerns of safety from the state may have risen to new prominence, they are not new. Crimes by the state—including genocide and slavery—define American history (Zinn, 2005). In considering violence and public safety, the state is the unnamed elephant in the room. This oversight is reified and perpetuated by an empirical discipline which relies upon data created, provided, and vetted by the state and whose analyses defer to state definitions. Official crime trends, touted now to be as low as in periods prior to the 1970s, might look different if the tally included the full scope of social violence.

Importantly, even this comparison to earlier eras is itself a distortion. Rather than a tranquil period of American yeomanry, the pre-Civil Rights Movement period was incredibly violent; much of this violence was enacted or sanctioned by the state. Police particularly were leaders or co-conspirators of such violence, which often took the form of lynching (Ginzburg, 1988). The Tuskegee Institute reported that the vast campaign of racial terrorism lynched nearly 5000 persons in the late 1800s and early 1900s—although the vast majority were Black, Latin@s, Asians, and some white labor leaders were also victims (Carrigan and Clive, 2003; Carroll, 2016; Escobar, 1999). Significantly, lynching achieved widespread attention not through state agents (who were often involved) but by radical Black activist-journalist Ida B Wells. Other disruptions of the neat crime decline imaginary include the pervasiveness of in-prison sexual assault by prison staff, massacres of Native peoples, or the billions of civil asset forfeiture dollars seized from civilians by police without due process, none of which have ever officially counted as crime, violence, or theft. In sum, crime trends not only suffer from significant distortions by omitting broad swathes of data and making illusory historical comparisons, but also compound intersectionalities that place certain populations—particularly Black women, as Richie (2012) demonstrates—at elevated risk for violence from civilians and the state.

Including the state as a significant and perpetual violent perpetrator gets us further in assessing social hazard than does a traditional civilian-only conceptualization of crime. The time is ripe to adjust our terms and definitions. Traditional notions of crime, violence, and justice struggle to make sense of de facto permanent immigrant family separation policies, to continue the earlier example. Because the perpetrator is the state itself, the entity to which criminologists turn for their categories and data of crime and violence, scholars are left with neither data nor appropriate interpretive frameworks. Contemporarily, as federal agents and the military deploy to “pacify” rebellions in major US cities (Miroff and Berman, 2020), the painful irony of being in the trough of the Great American Crime Decline demonstrates criminology’s disciplinary inability to cope with the complete spectrum of social violence. We suggest that the first steps to remedying this problem are to define it, place it in context, and attempt preliminary measurement of its scope.

First, we lay out a new conceptual framework for understanding crime that includes both civilian and state crime. Our overarching research question is conceptual: Can we reconceptualize how crime is counted in a way that includes the dark figure of state violence? From this emerge three specific and empirical subquestions: (1) What are the best available data to measure this concept? (2) What is the annual US homicide rate when adjusted to include state sources of violent death? (3) For how much of this adjusted homicide rate does state violence account?

To answer these questions, we unpack and define state violence and homicide statistics, craft an inclusive definition of violence, and operationalize it for the case of homicide. We confront the challenges inherent in our method and present empirical data estimating homicide beginning in 2000, decomposed into civilian and state homicide. Our findings suggest that state homicide may conservatively account for as much as 16% of the corrected rate and this percentage may be rising. We conclude by discussing the implications of our framework for the trajectory of crime rates, crime control, and the complex relationship between them.

Theoretical background

Countering criminological transcendentalism with “synthetic” scholarship

An attitude of transcendental change characterizes scientific reasoning, beginning in the Enlightenment and continuing through late modernity. By applying scientific method and reasoning, social change proceeds on a generally uninterrupted upward trajectory making life better for all. This attitude suffuses academic disciplines and is popular with the general public. Indeed, psychologist Steven Pinker (2011) argues that a combination of rationalism, liberal political schemata, religious and political ideologies, and respect for life has produced violence rates lower than any in human history. The particular transcendentalism of criminology is shaped by the institutional bent of the field. While Pinker uses the crime decline narrative explicitly as evidence of a pacifying world, the implicit premise of criminology is similar—society transcends crime through state intervention. Frank Zimring (2007) and other commentators, including former Los Angeles and New York police chief William Bratton (Bratton and Knobler, 1998), argue that while some of the decline may be idiosyncratic or “cyclical”, much is due to the deployment of governmental resources, particularly police. The crime decline thus confirms the effectiveness of rational governance strategies. Together with other critics (e.g. Herman and Peterson, 2012), we argue that this conclusion reflects selective data analysis which ignores and defines out of existence state violence.

Criminology is replete with studies on juvenile delinquency and street crime, but has been noticeably slow to address larger issues such as white-collar crime and state crime, whether domestic or international (Green and Ward, 2000; Rothe and Friedrichs, 2006). Past presidents of the American Society of Criminology have urged the field to do exactly this (e.g. Chambliss, 1989; Peterson, 2017). To this end, Ward (2015) has crafted an understanding of the “slow violence” of state-organized racial victimization—a term that refers less to spectacular police homicides, and more to the subtler complicity of courts, historians, and archivists in erasing, misclassifying, or downplaying such events in the public record. For criminology as a field to more fully assess the threat of violence from the state alongside that of civilians requires both reconceptualization of what crime counting might mean and new ways to measure the problem.

To remedy this, we follow Hagan et al. (2006) in approaching such scholarship “synthetically”. These authors faced similar bridging concerns in considering “complex humanitarian emergencies”—by which they mean genocide and then the ensuing death from public health concerns. To synthesize the siloed fields of demography, public health, criminology, and socio-legal studies, the researchers begin with basic measurements of mortality, specifically the crude mortality rate. To enable diverse perspectives to coalesce, they proffer a definition:

Complex humanitarian emergencies . . . include situations in which efforts to drastically restructure a state, society, or social group lead[s] to civil or international war, resulting in the violent death of large civilian populations and in their substantial displacement to detrimental living conditions that produce disease, dehydration, starvation, malnutrition, and other sources of extraviolent deaths.

(Hagan et al., 2006: 331)

This definition defines its actors by their actions, such that both state and non-state actors can induce large scale violence and related death. Further, it relies on an articulated causal chain (violence leads to displacement) rather than third party adjudication (a finding of guilt in a court setting) to enable its comparisons. This does not make such measurement easy—“detrimental living conditions” alone produces debate—but it does make it possible.

Hagan and colleagues do not begin with mortality arbitrarily. Citing Berlin’s (1969) and Sen’s (1981) discussions on freedom and human rights, they locate the “hard” human right against coercion (including death) as the bedrock upon which “softer” human rights (e.g. material conditions for civil society) rest. Consequently, they suggest that, although the notion of synthesizing these human rights-related fields of study should not end with the study of mortality, it must begin there.

The synthetic work of conceptualizing and measuring state violence alongside civilian violence must begin with a similar “hard” human right against coercion (see also Green and Ward, 2000). We do not depart significantly from commonsense and legal definitions of violence and assault—intentionally harming or putting another in reasonable apprehension of physical harm—but rather unfetter this to include the state and state actors. We propose a broad standard for measuring such violence which includes events directed by the state, the actions of state agents, and violence and self-harm that occur under the direct supervision of the state—in sum, a measure of the footprint of state violence upon the populace. Importantly, we do not allow categories such as “justified” or “legal” to influence our considerations—as history reminds us through spectacular examples of the enslavement of Black Americans or the Second World War Holocaust (of Jews, LGBTQ+ people, Romani, other nonWhites, and leftists of all stripes), legality is temporal, partial, reversible, and fundamentally a matter of power relations. Harm can easily hide within such categories.

Our overarching conceptual concern is to include the harm brought about by the state directly or through specific neglect of its incapacitated charges. We purposely include those within the USA’s geographic borders, such as recent migrants, rather than restrict consideration to citizenship. We also include those for whom the state is legally responsible who may not reside within these borders—the assassination of US citizen Anwar al-Awlaki remains a pointed example (Miller, 2014)—although measurement is difficult. Inspired by Hagan et al.’s (2006) “hard” right against coercion and guided by the doctrine of parens patriae, we include those in state custody or for whom the state retains primary responsibility (either permanently or in an instant encounter) under the notion that police power so arrogated corresponds to an equivalent degree of responsibility. In depriving people of liberty, the state assumes responsibility for their welfare, a point made forcefully by the 2011 US Supreme Court decision Brown v. Plata.¹ We extend that reasoning from the carceral context to law enforcement encounters of all kinds. Police, by training and by law, establish a comparable “command and control” situation that precludes civilian autonomy. This renders street encounters a targeted version of martial law, suspends the human right against coercion, and places civilian welfare into the hands of the state. Deaths which occur at the hands of state actors or as the result of the conditions of confinement in the context of street encounters or within carceral facilities are a form of state violence.

Current conceptualization and measurement of homicide

Before we operationalize state homicide below, we briefly revisit the current criminological (and official Justice Department) conceptualization and measurement of both civilian and state homicide.

Civilian homicide

The current conceptual and empirical boundaries of homicide are not straightforward. In the aptly named report, The nation’s two measures of homicide (Regoeczi and Banks, 2014: 3), the Bureau of Justice Statistics (BJS) defines homicide for reporting purposes as “the willful killing of one human being by another, [which] includes murder [and] nonnegligent manslaughter”. BJS relies on two different dimensions—intent and justifiability—to determine which deaths count as homicide.

First, the degree of intent splits the figure: homicide is “first degree murder” combined with “second degree murder” or nonnegligent manslaughter. This is designed to exclude accidental or unintended deaths (“negligent manslaughter”). As discussed above, negligent manslaughter by the state is eligible under our parens patriae standard.² Next, justifiability introduces the idea that some intentional killings are defensible and acceptable. Justifiability is not made explicit in BJS’s conceptual definition of homicide, but rather is revealed by the existence of a separate “justifiable homicide” category. “Homicide”, assumed a priori to be without grounds, stands in contrast to “justifiable homicide”, or the killing of a person for a legally supported reason. Justifiable homicides comprise two subcategories: law enforcement and civilian killings. While civilian justifiable homicides are relegated to extreme situations such as self-defense, the definition of police justifiable homicide is “the killing of a felon by a peace officer in the line of duty” (Criminal Justice Information Services Division, n.d.). That is, when police kill civilians (“felons”) the killings are assumed to be justified, and therefore do not count as homicide. Indeed, the act of police killing civilians converts civilians into “felons” as a matter of definition.

Given these empirical boundaries, we refer to the conventional BJS “homicide” category as “civilian homicide”. This terminology clarifies what is actually included in the category, and evokes the necessary taxonomic distinction from the more robust category of “homicide” which includes justifiable homicides by civilians and law enforcement.

In criminology, civilian homicide data are culled largely from a single source: the Uniform Crime Report (UCR; Bureau of Justice Statistics, n.d.), which is voluntarily reported by local law enforcement. The Federal Bureau of Investigation (FBI) requests these figures, but has no enforcement mechanism to ensure they arrive. Analyses of UCR civilian homicide data over time reveal the consequences of this methodology: zeros for entire states in some years, for instance, compared with large numbers in other years and uneven coverage for smaller jurisdictions across time. Maltz’s (2006) comprehensive study of missingness in UCR homicide data argues both that large departments produce the best data and that more than half the country’s population lives in an under- or nonreporting jurisdiction. Larger departments are more reliable due to institutionalized data collection and constant dialogue with federal investigators; smaller departments inconsistently report due to a lack of personnel, crime, interest, and stable federal relationships. For locales with persistent data gaps, BJS uses statistical extrapolation to estimate crime figures. Together, this renders federal crime statistics a patchwork of voluntarily reported counts and statistical guesswork using misleadingly narrow definitions (MD Maltz, UCR data and justifiable homicides, personal communication (email), 8 May 2018).

State homicide

Despite these limitations, criminology tends to prefer that its analyses and range of empirical questions conform to official data. Consequently, the preferred source for measuring state homicide is the police justifiable homicide measure drawn from the UCR’s Supplementary Homicide Report (SHR), a footnote to the already-voluntary survey.

A fundamental conflict of interest is revealed when we shift focus from civilian to state violence (such as police killings): culpable subjects produce the official data necessary to determine the scope (or even the existence) of the suspicion. One of the authors observed the mundanity of this conflict firsthand when he coded raw police data—homicide files known as “murder books”—on site with the Los Angeles Police Department (LAPD). The physical process of accessing murder books revealed that files on “officer involved shootings” were stored in a separate geographic location from civilian homicides. These were never made available to researchers, not considered comparable in any sense, and thus did not count as murders for research purposes.

Indeed, a number of government officials and national commentators identified deficiencies in the official methods of counting police killings “based solely on police investigation” (Bureau of Justice Statistics, n.d.). Opinions in the nation’s largest newspapers lament the lack of solid, government-collected data covering police killings of civilians (Berman, 2015; Lowery, 2014). A former FBI Director described federal data on justifiable killings as unreliable due to voluntary reporting requirements, describing journalists instead as the “lead sources” in determining the scope of the problem (Tran, 2015) and a former Attorney General made similar comments (Richardson et al., 2015).

The undercounting of civilian homicide (Maltz, 2006) compounds when we shift to police homicide, shifting the vectors of concern as well. Richardson et al. (2015) report that in 2012 only 4% of jurisdictions (about 750 of 18,000) reported police killings to the FBI. Directly contrary to civilian homicides, this extreme undercounting stemmed in part from “large jurisdictions that do not routinely report police-involved deaths” including one jurisdiction of more than 1300 officers which “did not view justifiable homicides by law enforcement officers as reportable data”. Similarly, The New York City Police Department (NYPD) reported zero police homicides between 2007 and 2012 (Richardson et al., 2015: S11). Maltz (personal communication, 8 May 2018) suggests in summary that some jurisdictions report those killed by police as justifiable homicides, others consider them part of the regular homicide rate directly (contra BJS), while others omit them completely. The end result is both simple and disturbing: some, maybe many, who die violent deaths at the hands of state agents are not counted in any official justice statistics whatsoever.

Data and method

Operationalizing an adjusted homicide rate

Given the official data limitations, we consulted a diverse set of data sources to operationalize civilian and state violence. We compare government data with crowd-sourced and journalistic data to yield an accounting of violent US deaths, both by civilians and the state—our operationalization of the true homicide rate.³ We begin by using traditional UCR civilian homicide cases, which exclude all justifiable homicides. While civilian justifiable homicide cases are more closely adjudicated than police homicides—the latter is presumed legitimate—we then expand our count to include these. We rely on UCR data here for two reasons: (1) unlike police homicides, there are no alternative source data, and (2) there is less inherent conflict of interest in the state’s reporting of civilian justified homicides. We isolate UCR cases of civilian justifiable homicide and include them in our adjusted homicide count calculations. Importantly, UCR-SHR law enforcement killings are substantially limited and we remove them completely in favor of other data as described below.

Next, we include in our calculations deaths in state custody. The Bureau of Justice Statistics “Mortality in Local Jails and State Prisons” files are the only comprehensive accounting of deaths in jails and prisons known to us. Incarceration records are closely gatekept; independent accountings arise mainly through public records requests. Fortunately, causes of death are broken out, comprising generally disease-related deaths (e.g. heart disease, “illness”, AIDS-related), suicide, homicide, accidents, drug and alcohol intoxication, “other”, and missing. Our inclusion decisions here were guided by a cautious conservatism together with our general parens patriae doctrine. A more liberal inclusion criteria would include nearly all deaths occurring in carceral facilities—as the Plata decision pointed out for California prisons, “a preventable or possibly preventable death occurred once every five to six days”, many from inadequate healthcare. Further, suicide rates in prison were 80% higher than the general population, most of which were caused by inadequate treatment. While a strong case could be made for many AIDS-related or even drug and alcohol deaths, the uncertainties surrounding these led us to restrict inclusion to deaths unmediated by direct individual medical concerns. Consequently, we include only the substantive categories of homicide, suicide, and accident, alongside data categorized as “missing” (not “other”—a mystery category), under the more conservative notion that while the state is responsible its wards’ well-being, it cannot completely eradicate disease or easily adjust preexisting individual levels of health.

State executions are perhaps the most straightforward cases of state violence. Despite the well-documented correlates of poverty, inadequate representation, and other social and legal pathology that determine executions independent of case facts, many states and the federal government continue to commit this form of homicide, so we include this as well.

The bulk of our state violence data relates to justifiable homicides by law enforcement. Because official sources misrepresent the scope of the problem, we turn to more comprehensive, extra-official sources. The Malcolm X Grassroots Movement (MXGM, 2012) which began after the killing of Trayvon Martin, originated a host of nonofficial efforts to independently count these killings with their publication Report on Black people executed without trial by police, security guards and self-appointed law enforcers. This report provides a rough count of Black extra-judicial killings and establishes a general media search methodology which other efforts have elaborated upon. We discuss these next-generation efforts below alongside more traditional sources.

Uniform Crime Report’s Supplementary Homicide Report (UCR-SHR)

Many studies of police violence have relied on the UCR’s Supplementary Homicide Reports (SHR) to document police killings, even though estimates suggest that the SHR report only between 30 and 45% of all such incidents (Campbell et al., 2018; Planty et al., 2015; Williams et al., 2019). Such deficiencies have produced public demands for increased data collection, in response to highly visible and controversial civilian deaths at police hands (see Ruane, 2017). The reporting of police justifiable homicide specifically remains under the control of the jurisdictions themselves, which have every reason to underreport such data and few carrots to encourage compliance or accuracy. An interview with police killing journalist and data collection expert D Brian Burghart (personal communication (phone interview), 20 March 2018) suggests also that killings by teams such as Special Weapons and Tactics (SWAT) are not always included, and that no UCR data include killings by federal officers.

National Violent Death Reporting System (CDC)

The Center for Disease Control (CDC) implemented the National Violent Death Reporting System (NVDRS) in 2000. The NVDRS links information from several sources, including law enforcement, medical examiners and coroners, toxicology reports, and death certificates to record violent deaths in participating states. Its death category classifications include: legal intervention (including deaths by law enforcement and other persons with legal authority to use deadly force), homicides, suicides, accidental firearm, and undetermined intent. State participation in NVDRS is voluntary—not until 2018 did all 50 states contribute data. By the CDC’s own account, even their expansion to all 50 states, Puerto Rico and the District of Columbia, only brings them “closer to the goal of providing a complete picture of violent deaths across the nation” (National Center for Injury Prevention and Control, n.d., emphasis added).

Killedbypolice.net

Killedbypolice.net was created anonymously in 2013 and utilizes exclusively crowd-sourced media reports. Since then, the creator of the website has backed away from the project which is maintained by an anonymous colleague. The database’s curation methods are unknown, the interface is clunky, and the data are produced through a substantial amount of crowd-sourced labor. The database is current from 2013 to 2018.

Fatal Force

In 2015, the Washington Post began tracking police shootings in a series known as Fatal Force (Tate et al., 2016). Fatal Force attempted a partial look at the problem, including killings similar to the homicide of Ferguson, MO’s Michael Brown in 2014—shootings of civilians in the line of duty only. Fatal Force uses media sources, law enforcement websites and statements, social media tracking, open records requests, and independent reporting. While shootings comprise the vast majority of police killings, there are many other ways in which state agents cause death in civilians and these are explicitly excluded from the archive which is current from 2015 to 2019.

The Counted

Similarly, in 2015, UK newspaper the Guardian began tracking police homicide in a series known as The Counted. Ascribing to a comparable journalistic rigor as the Washington Post, The Counted aspired to a comprehensive record of police killings, rather than shootings only, and tracked deaths for two years, ending systematic data collection in 2016. According to public statements, the team suspended tracking police homicide due to Department of Justice (DOJ) reforms. Announced by the Obama administration, these reforms attempted to implement better tracking of officer-involved deaths, partly in response to such investigations; their current status is unclear.

Fatal Encounters

D Brian Burghart, a newspaper editor, publisher, and journalist, organized Fatal Encounters (FE) in 2013. FE collected data on police killings moving forward, but also engages in retrospective research. In 2014, FE received 501(c)3 status and began paying researchers to aid Burghart. FE is the most inclusive tracking effort to date, providing a set of inclusion criteria that cover chain of custody problems ignored by other data (such as post-booking, interview room, and other pre-jail deaths). FE began by using public records requests, but due to poor cooperation and high cost⁴ quickly expanded to include crowd-sourced input, local news, and other online resources. Each entry has supporting documentation and Burghart’s team fact-checks each before publication, allowing user transparency (DB Burghart, personal communication (phone interview), 20 March 2018). The database is comprehensive and current from 2000.

Bias and underreporting in extra-official sources

The main sources of bias for journalistic and crowd-sourced efforts at counting police homicide are related to (1) the method of data provision and (2) the ubiquity of internet media coverage. First, although one could plausibly refer to some of these data efforts (particularly FE) as “all source”, media reports predominate. Media are a capitalist market currently experiencing reductions in local journalism expenditures. An obvious source of bias is that under-resourced and profit-driven media simply ignore police killings of racial minorities, the houseless, and the poor that lack an amplifying protest movement or hashtag. Relatedly, media sources tend to be overly credulous in accepting law enforcement narratives. Officers and administrators sometimes collude to obfuscate the role of police in the killing of civilians. To the extent that understaffed media sources accept these questionable narratives, data may be missing.

Second, the ubiquity of internet media is fairly new. Although the availability of information in recent years has probably reached a saturation point, this is less the case the further back one goes. Burghart (personal communication (phone interview), 20 March 2018) notes that 2005–2006 was a transitional period for data storage. Before this, a lack of online storage space meant media reports were routinely deleted after an interval. Some are no doubt available through other means, but others are likely unrecoverable without a massive campaign of microfiche analysis that will inevitably miss smaller markets and geographies. Since 2006, most media information has been stored on “the cloud” and losing such media reports is no longer an issue. Because this methodology has inherent longitudinal performance difficulties, FE only researches retrospective data back until 2000—to delve further back would court severe undercounting.

Our empirical case: An adjusted US homicide rate

Our first research question was conceptual: can we reconceptualize the counting of crime in a way that includes the dark figure of state-originated violence? We argue above that we can, and that homicide is the place to begin. Our remaining research questions are empirical and pertain to homicide specifically: (2) What are the best available data to measure state homicide? (3) What is the homicide rate when adjusted to include state sources of violent death? (4) For how much (adjusted) homicide does state violence account? Following Hagan and colleagues (2006), we begin by counting violent deaths, regardless of the actor who caused them or the legality of the context. We refer to our revised conception as the adjusted US homicide rate. This concept rounds out the traditional homicide rate (first degree murder and nonnegligent manslaughter) by adding in civilian justifiable homicides and state violence. To account for state violence, we include police justifiable homicides and deaths in jails and prisons that are explicitly homicides (including executions), otherwise preventable, and those for which officials cannot account.

Police homicide

To answer the first empirical question, we argue that, for police homicide, extra-official sources of data are superior to official sources. Figure 1 compares official sources (DOJ and CDC) to the four unofficial sources discussed above for a year for which all sources produced data, 2015. The figure is not a strict apples-to-apples comparison. The boundaries of inclusion for each data source shift somewhat, as described above, which accounts for some, but not all, of the discrepancies. Even the most conservative extra-official source (Fatal Force), which includes only police shooting homicides, reported an 87% higher homicide count than CDC and 119% higher than DOJ. Fatal Encounters’ more comprehensive numbers report a count of police homicide 246% higher than that reported by DOJ. Moreover, each and every entry in Fatal Encounter’s database includes key incident details and linked news stories, while the official sources are anonymized. Due to its longitudinal scope and similarity of inclusion criteria to our conceptual definition, we use Fatal Encounters for the adjusted US homicide accounting below.

Figure 1.

Counts of police homicide by source in 2015.

Incarceration deaths

Figure 2 depicts the official DOJ rates of jail and prison deaths between 2000 and 2015 (state prison figures were not available for 2000) and their accompanying trendlines. We rely on official figures from the Bureau of Justice Statistics, but undercounts here are also likely. Both numbers remain relatively flat across the timeframe analyzed, jail deaths declining slightly and prison deaths increasing slightly, while the national rate of execution continues to fall slowly. The preponderance of available commentary suggests that immigration detention facilities are not included in this counting (e.g. Granski et al., 2015).

Figure 2.

Local jail, state prison, and executions death rates (BJS and US Census).

Adjusting the US homicide count

Figure 3 depicts the traditional US homicide rate from 1977 through 2015 (first degree murder and nonnegligent manslaughter). Depending on one’s definition (see Zimring, 2007), the homicide decline begins around 1991. The figure also displays our extra-official data availability, which begins near the trough of the crime decline (notwithstanding the 2001 World Trade Center attack spike). It is this traditional measurement which we will round out with additional measures of homicide.

Figure 3.

Traditional UCR homicide rate, including time range availability for crowdsourced estimates of police homicide.

Our second research question addresses the dark figure of state violence—that is, what is the homicide rate when adjusted to reflect state sources? Our first task is to group the included forms of state homicide together and assess their general character. Figure 4 depicts our categories of homicide together from 2001 to 2013—the years for which data are available in the important categories. Jail, prison, and execution deaths retain their relatively flat trajectories. Police homicide exhibits a clear upward trend. If the media transition to the cloud in 2005–2006 exerted an effect on these data, it is subtle and part of a general upward trend. In terms of gross counts, the number of police homicides from 2001 (816) nearly doubled in just over a decade (1781). More recent years are not depicted in the absence of comparable jail and prison figures, but the partial counts without jail and prison figures stand at 1712 (2014), 1596 (2015), 1586 (2016), and 1749 (2017).

Figure 4.

Different forms of state homicide (expressed in rates), 2001–2013.

We address the final research question—for how much of the overall homicide rate does state violence account?—in Figures 5 and 6. In Figure 5, we construct our new US homicide rate, plotting the constituent components together for clarity. At the top, the space between the traditional rate and the revised rate represents the “dark figure” of state homicide, real violence occurring in society excised from traditional consideration. This is depicted separately at the bottom. We draw attention to the relative contribution of state homicide to the overall adjusted homicide rate as well as the trajectory of each. Civilian homicide remains on a decline—just above this line is the adjustment for civilian justifiable homicide, which is not usually counted. This affects the intercept slightly but not the general character of the line. Visually, the contribution of state homicide to the overall rate is considerable. To use a comparison year from the middle of the graph, in 2007, California led the nation in civilian homicides with 2260, while Texas, number two, produced 1419. Positioning state violence as its own unit ranks it just behind California with 2036 killings. Note that the rate of state violence alone appears to be climbing, while the complete US homicide rate at the top of Figure 5 retains most of its original character. Although the addition of state violence flattens the decline somewhat, the complete rate is elevated significantly from the traditional measurement, underlining the significant contribution of state violence to the overall prevalence of homicide.

Figure 5.

Proposed adjusted homicide rates alongside constituent measures (2001–2013).

Figure 6.

State violence plotted against total adjusted homicides.

Figure 6 depicts our complete homicide rate plotted against the percentage of that count represented by state violence (note the two y axes).⁵ Acknowledging both the potentially deflated figures on the left side and the 3000+ 11 September 2001 deaths, even this conservative figure averages just under 10%. State homicide in the graph’s last five years averages about 12.5% of the total homicide count and rises—the final, and likely most accurate, year of the analysis places the state violence contribution at nearly 16%, larger than any single US state. The steady Figure 4 climb in state homicide is magnified by calculating these figures as a percentage of an overall declining count, as the great American homicide decline continues among civilians.

Conclusion

This article makes the case for revising traditional violence definitions using the case of homicide as a baseline. Including the state as a significant, perpetual perpetrator gets us much further in assessing social violence than a civilian-only conceptualization. The time is ripe to adjust our terms and definitions. State agents legally execute orders which comport with no coherent standard of justice or decency. The state is responsible for the resulting trauma and death but criminology has no definitional framework to comprehensively address this.

In recent years between 11 and 16% of US homicides connect directly to state action and are obscured from criminological or public accounting. Our estimation is conservative—for example, fatal conditions of confinement and media undercounts of police killings suggest that contemporary figures could be much higher. The federal immigration system also recently received a considerable shock yet comprehensive data on violence in such facilities are difficult to obtain (Granski et al., 2015).

Our most compelling findings include the strong upward trajectory of state homicide. There are several competing explanations which need not be mutually exclusive: expanding carceral and law enforcement apparatuses, shifts in journalism, the ubiquity of cell phone cameras incident captures, widespread cultural shifts in resisting police brutality, and the pushback to these by police. George Holliday’s 1991 video of LAPD’s brutalization of Rodney King began a cultural shift in which civilians fill the gap left by the state and journalists in publicizing police brutality. In addition to the ubiquity of online journalism, increased media attention to police killings perhaps has corrected media inattention in earlier eras. This likely renders recent assessments more accurate, while calling into question the general trajectory. Over the analytic timeframe, cell phones went from luxury item to mundane necessity. Recording and publicly posting police killings is easier now than ever, which, interestingly, provokes journalistic coverage. Simultaneously, the reach of the criminal legal system continues to expand, rendering more official guns and less oversight. As Balko (2013) documents, police have increasingly militarized, using sophisticated operations units armed with military weaponry to perform basic tasks. While less-lethal weaponry is also more widely available, there is little evidence that this has reduced state violence.

Our approach reveals a number of ideological implications which plague criminology’s understanding of “public safety”. As Ward (2015) suggests, the slow violence of state-organized racial victimization distorts the early years of the crime curve by erasing institutional violence against Black, Native, and other historically marginalized peoples. Our homicide curve adjustment is a necessary beginning, but is far from the only correction required. The earlier record—not just the 1800s, but also the 1990s—must be corrected as well. The state and its institutional handmaidens systematically conceal state violence and bolster state legitimacy by unclassifying and erasing such figures. In its uncritical reliance on official figures, criminology is complicit.

Empirical methods must demonstrate the risk of state violence, much as traditional research has determined correlates that predict risk of civilian homicide. If the historical record elides slow genocidal violence, contemporary research suggests that race and class continue to predict police deployments, stop engagements, and subsequent decision making that results in homicide and other brutality (Blasi and Stuart, 2008; Correll et al., 2002, 2007; Fagan et al., 2009; Gelman et al., 2012; Greenwald et al., 2003; Krieger et al., 2015; Lynch et al., 2013; Nix et al., 2017; Peterson et al., 2006; Sadler et al., 2012). Future research must consider comprehensive violence definitions to assess the multiple social risks of violence on minority communities, rather than assuming civilian risk and state protection.

We touch upon two possible arenas for future research. Carceral facilities represent a logical point of inclusion, as prisoners suffer fully and intimately state strictures and responsibilities. Crime in carceral facilities is not “incapacitated” but rather continues undetected by traditional measures. About 4% of state and federal prisoners and 3.2% of those jailed reported sexual victimization between 2011 and 2012; 2.4% reported victimization by facility staff (Beck et al., 2013). Other reports put sexual assault victimization by correctional staff at about half of all incidents (Rantala et al., 2014). Correctional staff both perpetrate violence on prisoners and fail to prevent violence by other prisoners. In considering indirect violence and criminal neglect in the facilities (including immigration detention), the torture of solitary confinement, and disease exposure, confinement experts could easily make use of the state violence framework (Aleaziz, 2019; Haney, 2018; Kerwin, 2020; Merkl, 2020; Patel and Jawetz, n.d.).

Identifying state violence outside carceral facilities is also challenging. While the 1992 uprisings head the list of most devastating US “riots” to date, several competitors remain uninterrogated due to myriad historical slow violence factors. For example, rampaging whites completely destroyed the Greenwood district, 35 Black middle-class blocks in Tulsa, Oklahoma, in 1921. The official count was 36 massacred residents, but recent reevaluations increased that estimate 10 times. Local law enforcement deputized many rioters to commit this violence (Luckerson, 2018; Oklahoma Commission, 2001; Tulsa Historical Society, 2010). Similarly, the 1943 Los Angeles “Zoot Suit Riot”—or the “Government Riot” (Mirandé, 1987)—involved the collusion and participation of state agents. Several hundred servicemen and off-duty police invaded LA’s Mexican barrios to attack Latin@ and Black youth. They beat 600 victims bloody, stripped them of their zoot suits, and left them unconscious in the streets for arrest. The crime rate of these years—and historical rankings of “worst” mass violence—remain wildly incomplete, “laundering” in Ward’s (2015: 309) terminology, “racial violence with legal variables”.

These historical accounts are spectacular, but hint at the more mundane sanitizing of the myriad forms of violence committed against marginalized US populations. Systematic data on these incidents are not officially collected and remain unconceptualized without a framework that takes seriously state violence. The scale of this violence, a dark figure whose surface we have only begun to scratch, demands such a reconceptualization.

Footnotes

Acknowledgements

We would like to thank Thythy Bui and Amanda Zuniga for their research assistance. The authors remain solely responsible for the analyses, interpretations, and opinions presented herein.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Aaron Roussell

Notes

Author biographies

Aaron Roussell is an associate professor of sociology at Portland State University. He studies policing, drug policy, racial capitalism, and state violence. He is the co-author of The Limits of Community Policing: Civilian Power and Police Accountability in Black and Brown Los Angeles (2019).

Lori Sexton is an associate professor of Criminal Justice and Criminology at the University of Missouri-Kansas City. She studies incarceration and other sanctions with an emphasis on the lived experience of punishment. Her work has been funded by the National Science Foundation and the National Institute of Justice, and is published in Law & Social Inquiry, Punishment & Society, Criminology, Justice Quarterly, and Critical Criminology.

Paul Deppen is an alum of Bloomsburg University of Pennsylvania (BA) and Portland State University (MS). He currently works as a Clinical Research Coordinator at the Children’s Hospital of Philadelphia.

Marisa Omori is an assistant professor of Criminology and Criminal Justice at the University of Missouri-St. Louis. Her work focus on racial inequality in the criminal justice system, courts and sentencing, and punishment and social control.

Esther Scheibler is a doctoral student in the School of Criminal Justice at the University of Cincinnati. Her research interests are centered on crime & victimization in Jewish communities, faith-based prisoner rehabilitation, and prisoner reentry.

References

Aleaziz

(2019) More than 52,000 people are now being detained by ICE, an apparent all-time high. Buzzfeed News, 20 May. Available at: buzzfeednews.com/article/hamedaleaziz/ice-detention-record-immigrants-border.

Balko

(2013) Rise of the Warrior Cop: The Militarization of America’s Police Forces. New York: PublicAffairs.

Beck

Berzofsky

Caspar

, et al. (2013) Sexual victimization in prisons and jails reported by inmates, 2011–12. NCJ 241399. US Department of Justice, Bureau of Justice Statistics.

Berlin

(1969) Four Essays on Liberty. Oxford: Oxford University Press.

Berman

(2015) Holder: We need better data on police shootings and officer deaths. Washington Post, 15 January. Available at: washingtonpost.com/news/post-nation/wp/2015/01/15/holder-we-need-better-data-on-police-shootings-and-officer-deaths.

Blasi

Stuart

(2008) Has the Safer Cities Initiative in Skid Row reduced serious crime? Los Angeles, CA: UCLA School of Law.

Bratton

Knobler

(1998) The Turnaround: How America’s Top Cop Reversed the Crime Epidemic. New York: Random House.

Bureau of Justice Statistics (n.d.) Homicide. Office of Justice Programs, US Department of Justice. Available at: bjs.gov/index.cfm.

Bursik

Grasmick

(1993) Neighborhoods and Crime. Lexington, KY: Lexington Books.

10.

Campbell

Nix

Maguire

(2018) Is the number of citizens fatally shot by police increasing in the post-Ferguson era? Crime & Delinquency 64(3): 398–420.

11.

Carrigan

Clive

(2003) The lynching of persons of Mexican origin or descent in the United States, 1848 to 1928. Journal of Social History 37(2): 411–438.

12.

Carroll

(2016) The mysterious lynching of Frank Little: Activist who fought inequality and lost. Guardian, 24 September. Available at: theguardian.com/us-news/2016/sep/21/mysterious-lynching-of-frank-little-equality-activist.

13.

Chambliss

(1989) State-organized crime: The American Society of Criminology, 1988 presidential address. Criminology 27(2): 183–208.

14.

Correll

Park

Judd

, et al. (2002) The police officer’s dilemma: Using ethnicity to disambiguate potentially threatening individuals. Journal of Personality and Social Psychology 83: 1314–1329.

15.

Correll

Park

Judd

, et al. (2007) Across the thin blue line: Police officers and racial bias in the decision to shoot. Journal of Personality and Social Psychology 92: 1006–1023.

16.

Criminal Justice Information Services Division (n.d.) Expanded homicide. Federal Bureau of Investigation, US Department of Justice. Available at: ucr.fbi.gov/crime-in-the-u.s/2016/crime-in-the-u.s.-2016/topic-pages/expanded-homicide.

17.

Ellis

Hicken

Ortega

(2018) Handcuffs, assaults, and drugs called “vitamins”: Children allege grave abuse at migrant detention facilities. CNN Investigates, 21 June. Available at: cnn.com/2018/06/21/us/undocumented-migrant-children-detention-facilities-abuse-invs/index.html.

18.

Escobar

(1999) Race, Police, and the Making of a Political Identity. Berkeley, CA: University of California Press.

19.

Fagan

Geller

Davies

, et al. (2009) Street stops and broken windows revisited: The demography and logic of proactive policing in a safe and changing city. Columbia Law School Public Law & Legal Theory Working Paper Group.

20.

Gelman

Fagan

Kiss

(2012) An analysis of the New York City Police Department’s “stop-and-frisk” policy in the context of claims of racial bias. Journal of the American Statistical Association 102: 813–823.

21.

Ginzburg

(1988) 100 Years of Lynching. Baltimore, MD: Black Classic Press.

22.

Granski

Keller

Venters

(2015) Death rates among detained immigrants in the United States. International Journal of Environmental Research and Public Health 12(11): 14414–14419.

23.

Green

Ward

(2000) State crime, human rights and the limits of criminology. Social Justice 27(1): 101–115.

24.

Greenwald

Oakes

Hoffman

(2003) Targets of discrimination: Effects of race on responses to weapons holders. Journal of Experimental Social Psychology 39(4): 399–405.

25.

Hagan

Schoenfeld

Palloni

(2006) The science of human rights, war crimes, and humanitarian emergencies. Annual Review of Sociology 32: 329–349.

26.

Haney

(2018) The psychological effects of solitary confinement: A systematic critique. Crime and Justice 47(1): 365–416.

27.

Herman

Peterson

(2012) Reality denial: Steven Pinker’s apologetics for western-imperial violence. ColdType. Available at: coldtype.net/Assets.12/PDFs/0812.PinkerCrit.pdf.

28.

Kerwin

(2020) Immigrant detention and COVID-19: How the US detention system became a vector for the spread of the pandemic. Center for Migration Studies of New York. Available at: cmsny.org/publications/immigrant-detention-covid.

29.

Krieger

Kiang

Chen

, et al. (2015) Trends in US deaths due to legal intervention among black and white men, age 15–34 years, by county income level: 1960–2010. Harvard Public Health Review 3: 1–5.

30.

Lowery

(2014) How many police shootings a year? No one knows. Washington Post, 8 September. Available at: washingtonpost.com/news/post-nation/wp/2014/09/08/how-many-police-shootings-a-year-no-one-knows.

31.

Luckerson

(2018) Black Wall Street: The African American haven that burned and then rose from the ashes. The Ringer, 28 June. Available at: theringer.com/2018/6/28/17511818/black-wall-street-oklahoma-greenwood-destruction-tulsa.

32.

Lynch

Omori

Roussell

, et al. (2013) Policing the “progressive” city: The racialized geography of drug law enforcement. Theoretical Criminology 17: 335–357.

33.

Malcolm X Grassroots Movement (2012) Report on Black people executed without trial by police, security guards and self-appointed law enforcers. Malcolm X Grassroots Movement. Available at: finalcall.com/reports/MXGM.PDF.

34.

Maltz

(2006) Analysis of missingness in UCR crime data. No. 215343. Criminal Justice Research Center, Ohio State University. Available at: ncjrs.gov/pdffiles1/nij/grants/215343.pdf.

35.

Merkl

(2020) What’s keeping thousands in prison during covid-19. Brennan Center for Justice. Available at: brennancenter.org/our-work/research-reports/whats-keeping-thousands-prison-during-covid-19.

36.

Miller

(2014) Here’s the secret memo that justified Anwar al-Awlaki killing. Time, 23 June. Available at: time.com/2912137/memo-anwar-al-awlaki-doj-drone.

37.

Mirandé

(1987) Gringo Justice. Notre Dame, IN: University of Notre Dame Press.

38.

Miroff

Berman

(2020) Trump threatens to deploy federal agents to Chicago and other U.S. cities led by Democrats. Washington Post, 20 July. Available at: washingtonpost.com/national/defending-portland-crackdown-trump-pledges-to-deploys-feds-to-chicago-and-other-us-cities-led-by-democrats/2020/07/20/fda42b8a-caaa-11ea-89ce-ac7d5e4a5a38_story.html.

39.

National Center for Injury Prevention and Control (n.d.) CDC’s National Violent Death Reporting System (NVDRS). Center for Disease Control. Available at: cdc.gov/violenceprevention/pdf/NVDRS-factsheet508.pdf.

40.

Nix

Campbell

Byers

, et al. (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.

41.

Oklahoma Commission (2001) Report on Tulsa race riot of 1921. Available at: archive.org/stream/ReportOnTulsaRaceRiotOf1921/TulsaRaceRiot1921_djvu.txt.

42.

Patel

Jawetz

(n.d.) Condition of confinement in immigration detention facilities. American Civil Liberties Union. Available at: aclu.org/other/conditions-confinement-immigrant-detention-facilities.

43.

Peterson

(2017) Interrogating race, crime, and justice in a time of unease and racial tension. Criminology 55(2): 245–272.

44.

Peterson

Krivo

Hagan

(eds) (2006) The Many Colors of Crime: Inequalities of Race, Ethnicity, and Crime in America. New York: New York University Press.

45.

Pinker

(2011) Better Angels of Our Nature: Why Violence Has Declined. New York: Viking Books.

46.

Planty

Burch

Banks

, et al. (2015) Arrest-related deaths program: Data quality profile. NCJ 248544. US Department of Justice, Bureau of Justice Statistics.

47.

Rantala

Rexroat

Beck

(2014) Survey of sexual violence in adult correctional facilities, 2009–11: Statistical tables. NCJ 244227. US Department of Justice, Bureau of Justice Statistics.

48.

Ravani

(2018) 39 Occupy ICE protesters in San Francisco arrested. San Francisco Chronicle, 9 July. Available at: sfchronicle.com/bayarea/article/Occupy-ICE-protesters-in-San-Francisco-report-13059692.php.

49.

Regoeczi

Banks

(2014) The nation’s two measures of homicide. NCJ 247060. US Department of Justice, Bureau of Justice Statistics. Available at: bjs.gov/content/pub/pdf/ntmh.pdf.

50.

Reilly

(2018) Nearly 2,000 children have been separated from their families during Trump border crackdown. Time, 16 June. Available at: time.com/5314128/trump-immigration-family-separation-2000-children.

51.

Richardson

St Vil

Cooper

(2015) Who shot ya? How emergency departments can collect reliable police shooting data. Journal of Urban Health 93(1): S8–S31.

52.

Richie

(2012) Arrested Justice: Black Women, Violence, and America’s Prison Nation. New York: New York University Press.

53.

Rothe

Friedrichs

(2006) The state of criminology and crimes of the state. Social Justice 33(1): 147–161.

54.

Ruane

(2017) Re(searching) the truth about our criminal justice system: Some challenges. Sociological Forum 32(S1): 1127–1139.

55.

Sadler

Correll

Park

, et al. (2012) The world is not Black and White: Racial bias in the decision to shoot in a multiethnic context. Journal of Social Issues 68(2): 286–313.

56.

Sen

(1981) Poverty and Famines: An Essay on Entitlement and Deprivation. Oxford: Clarendon.

57.

Shepherd

(2018) Federal agents fire pepper balls and make arrests at Occupy ICE camp after protesters allegedly assaulted an officer. Willamette Week, 9 July. Available at: wweek.com/news/courts/2018/07/09/federal-agents-fire-pepper-balls-and-make-arrests-at-occupy-ice-camp-after-protesters-allegedly-assaulted-an-officer.

58.

Speri

(2018) Detained, then violated. The Intercept, 11 April. Available at: theintercept.com/2018/04/11/immigration-detention-sexual-abuse-ice-dhs.

59.

Tate

Jenkins

Rich

, et al. (2016) How The Washington Post is examining police shootings in the United States. Washington Post, 7 July. Available at: washingtonpost.com/national/how-the-washington-post-is-examining-police-shootings-in-the-united-states/2016/07/07/d9c52238-43ad-11e6-8856-f26de2537a9d_story.html.

60.

Tran

(2015) FBI chief: “Unacceptable” that Guardian has better data on police violence. Guardian, 8 October. Available at: theguardian.com/us-news/2015/oct/08/fbi-chief-says-ridiculous-guardian-washington-post-better-information-police-shootings.

61.

Tulsa Historical Society (2010) The Tulsa race riot. Available at: web.archive.org/web/20101015171251/http://www.tulsahistory.org/learn/online-exhibits/the-tulsa-race-riot.

62.

Vigil

(2002) A Rainbow of Gangs: Street Cultures in the Mega-City. Austin, TX: University of Texas Press.

63.

Ward

(2015) The slow violence of state organized race crime. Theoretical Criminology 19(3): 299–314.

64.

Williams

Bowman

Jung

(2019) The limitations of government databases for analyzing fatal officer-involved shootings in the United States. Criminal Justice Policy Review 30(2): 201–222.

65.

Zimring

(2007) The Great American Crime Decline. Oxford: Oxford University Press.

66.

Zinn

(2005) A People’s History of the United States. New York: Harper Perennial.