An Analysis of Homicides by Off-Duty Police Officers

Abstract

The current study examined 119 incidents of off-duty police shootings resulting in 123 deaths. Results show that the majority (69%) of off-duty deaths were the cumulative result of the officer being a target of a crime or interjecting themselves to assist others who were targeted. Those who intercede on behalf of others were not subjected to criminal proceedings. Other findings of note revealed how off-duty fatal police shootings frequently involved people of color. The various approaches of states, police departments, and district attorney offices in reporting of these cases was problematic. A greater transparency within the investigation is discussed.

Keywords

off-duty police police shootings race/ethnicity

When a law enforcement officer (LEO) is on-duty, the potential for a deadly incident is an understood part of the profession. However, officers may choose to continue enforcing the law, take a secondary employment security position, or have a need to defend themselves/others after their shift ends. Being off-duty is a function of the department’s timeclock. Off-duty officers can be in or out of uniform, driving a marked unit or their personal car, and/or presenting themselves verbally or not as police officers. If the officer is not on his or her regular rotation or overtime service, then any time outside of normally scheduled shift is off-duty. Research on the law, training, and circumstances surrounding these choices and/or professional responsibilities is lacking. Sometimes in situations where an off-duty officer acts, particularly under color of the law, deadly force may be used (Fyfe, 1980). While police use of deadly force is considered rare (Fridel et al., 2020), increasingly these incidents have garnered much media and public attention (Culhane et al., 2016), while also representing the subject of great scholarly inquiry (Phillips & Kim, 2021). However, there has been very limited research in the field when it comes specifically to off-duty officers and their use of deadly force over the last two decades.

White (2000) summarized police shooting research in the major cities of Philadelphia, New York, Chicago, and Los Angeles. A notable percentage, between 17% and 24%, of shootings occurred while off-duty. White (2000) observed that the officer’s home department disapproved of the off-duty shooting. Off-duty officers who discharged firearms in these situations, did so in violation of the administrative policy and/or criminal law. However, these vast majority of studies were conducted using data from periods of time (1970s) in which the rules of police shootings were different than today. After the cases of Tennessee v. Garner (1985) and Graham v. Connor (1989), police departments had to codify changes that only a few already had with respect to police shootings. Durán’s (2016) data found slightly fewer shootings to take place while off-duty or working a second job as security. Only around 10% of cases in his review were off-duty shootings, but it was also noted that officers “often” worked in private security roles (p. 76).

Beyond the specific topic of deadly force and duty status, research on the general topic of off-duty officers is generally limited. Despite this, there are a small assortment of studies that have explored a number of salient issues related to off-duty officer behavior. Studies exploring fatal encounters between on and off-duty officers’ interactions, also known as blue-on-blue, are among the most prevalent area of study on off-duty officer fatalities (Charbonneau et al., 2017; D. A. O’Neill et al., 2021; J. O’Neill et al., 2018). D. A. O’Neill et al. (2021) presented a scenario and survey to American and Canadian LEOs asking for current practices in identifying fellow officers who are off-duty and out of uniform. In their scenario, an on-duty officer would approach a scene with two other individuals, one of whom is holding a gun. Results suggest a lack of training on blue-on-blue encounters in the field, as most officers have never received any training on the subject. Officers reported acceptable practices of off-duty officers as a simple verbal alert or displaying a badge as verification of their status as a LEO. This research followed earlier studies of the more serious cases involving accidental discharges resulting in blue-on-blue injury/death (J. O’Neill et al., 2018) and fatal use of force by on-duty officers on off-duty officers (Charbonneau et al., 2017).

J. O’Neill et al.’s (2018) other project involving off-duty officers examined unintentional discharges (UD) of their firearm. More than one-third of all UDs accounted for in the dataset occurred when the officer was off-duty. These UDs resulted in a small number of deaths. The authors did not specifically compare the on- and off-duty officers, so it is difficult to determine what deaths occurred with respect to their employment status. Charbonneau et al. (2017) investigated racial disparities in blue-on-blue fatalities when one officer was on-duty and the other was either undercover or off-duty. When examining just the off-duty officers, their findings suggested a large racial gap in the deaths of off-duty officers shot by one of their fellow LEOs. Whereas most police officers in the United States are White, the majority of those killed in Charbonneau et al.’s analysis were Black. All of these deaths were the result of a crime or dispute in progress, thus leading to the intervention of an on-duty LEO.

It is important to note that a decision to intercede in assisting another civilian or attempting to stop a crime while off-duty could have legal ramifications as well. The International Association of Chiefs of Police’s (IACP) Law Enforcement Policy Center promotes a Standard of Conduct recommendation for departments to adopt. In the most recent iteration (IACP, 2019) it is noteworthy that officers are expected to uphold the same standards regardless of duty status. The failure to act is considered a neglect of duty while on the job, and as the document reads, could also be considered a failure while off-duty. Fyfe (1980) noted that off-duty officers were often required to be armed, with an anecdotal expectation that the citizenry was safer because of such policies. LEOs, whether on- or off-duty, have no legal obligation to protect a particular citizen from harm (DeShaney v. Winnebago County Department of Social Services, 1989; Castle Rock v. Gonzales, 2005). However, when off-duty, the potential loss of qualified immunity could result in substantial civil penalties. Cameron (2014) argues that essentially off-duty LEOs are no different legally than a private citizen with a few exceptions. However, the standards put forth by the IACP would suggest otherwise. One such exception is the choice to act in a capacity of an on-duty police officer, thus assuming the responsibilities and repercussions. The officer in question may or may not, however, have the legal protection of qualified immunity from civil or criminal prosecution. Still, Cameron’s (2014) assertions may not hold true given the circumstances surrounding any particular off-duty incident.

Criminal proceedings against an officer (both on- and off-duty) are possible. One large study (Stinson et al., 2012) reviewed a dataset of news reports centering on arrests made of law enforcement officers from across the United States. Over 1,100 officers were identified as committing criminal infractions while off-duty. The majority of these off-duty officers committed some act of violence (with simple and aggravated assault being the most common) leading to the arrest of the LEO. Other crimes committed by off-duty LEOs included driving under the influence (DUI), statutory rape, pornography/obscenity crimes, disorderly conduct, property destruction/vandalism, and other alcohol-related offenses. One crime that did not appear in Stinson and colleagues’ results was murder, nor were there any other homicide related offenses reported. Boateng et al. (2021) reviewed over 6,000 cases of officers who were arrested in a seven-year period. The largest category precipitating arrest was violence at nearly 48%. In subsequent analysis, such arrests were more likely to be perpetrated when the officer was off-duty. It is important to note that not all crime was found to be more likely to occur when off-duty, but violence advances the likelihood of a fatal encounter.

Recent research has shown that mock jurors find officers guilty more often when presented with use of force scenarios in which the LEO is off-duty (Pica et al., 2020). Duty status had several interactions with other manipulated variables in the study, such as weapon usage. Pica and her colleagues concluded “When police are off-duty, it is possible that any use of force would be perceived as unnecessary and violent” (p. 355). Their study is the only one known to look at public perceptions of the off-duty officer’s use of force actions.

Brunet (2008) examined the implications behind the decision of off-duty officers [in North Carolina] to become employed as security for a privately-owned company. Officers in this capacity are often looking to supplement their income through these opportunities. In some agencies the participation rate was “almost universal” among the force (p. 167). Further, in these departments, the officers would default back to law enforcement status upon the witnessing of criminal activity, and the officers were under the same expectations of use of force guidelines when employed as a private security contractor. As these officers are participating in crime prevention functions, the potential for deadly shootings while off-duty may rise. Other potential negatives can translate to on-duty performance. For example, officers may be tired and not as sharp on the shift; officers may be less productive as a mechanism of avoiding court or other obligations which would cut into their outside employment opportunities; and/or, officers may use public resources (cars, firearms, training, etc.) for the benefit of the private company and at the expense of the maintenance of these assets.

Stoughton (2017) noted that no states, nor the District of Columbia, have statutes prohibiting the moonlighting of officer employment as private security while off-duty. Most states have no statutory laws at all regulating the practice, and the remaining states either authorize or regulate such practices. When statutes do exist regarding the practice, the state can be insulated from potential liability. States may simply bar the private company from holding the state liable, may require liability insurance to be carried by the company, or both. Much of the policy formation has been handled at the department level, as Stoughton found over 94% of surveyed departments had written policies governing the private employment of officers when off-duty.

Race and Police Shootings

The death of minorities at the hands of police officers has a long and fraught history. Research has pointed to a long-standing finding of disproportionality of police-minority shootings (Geller, 1982). In particular, Black citizens are shot at a much higher rate “compared to the racial makeup of the community” (Geller, 1982, p. 163). Multiple cases have become high profile in the media and public conscience. The Mississippi Burning deaths of Civil Rights workers in 1964 was coordinated in conjunction with local law enforcement (FBI, 2022a). The riots of 1989 in Los Angeles, California after the televised beating of Rodney King, the 2014 protests in Ferguson, Missouri touched off following the shooting death of Michael Brown, and the dramatic rise of the Black Lives Matter movement across the country after the death of George Floyd are all examples of minorities being abused or killed by law enforcement. Further, these deaths had a galvanizing quality of the citizen population leading to a reckoning that police agencies and the country had to face (Smith et al., 2022).

Studies on the shooting deaths of minorities at the hands of police are mixed in their findings of potential racial biases. Phillips and Kim (2021) found that White officers did not disproportionally target citizens of color. Instead, they concluded that situational factors were more responsible for the decision to shoot. Likewise, Worrall et al. (2018) noted that Black suspects were no more likely to be shot by LEOs than Whites, further concluding that officers were not implicitly biased to violence toward Black citizens.

Contrary to the two aforementioned studies, Scott et al. (2017) determined that LEOs were more likely to shoot Black suspects when controlling for criminal activity in metropolitan areas across a 21-year span. Similarly, studies using the Washington Post Police Shooting Database have shown that Black civilians were more likely to be shot than White civilians when unarmed and shot more often regardless of the status of physical confrontation (Nix et al., 2017). That is, both groups (White and Black) equally attacked the officer, yet Blacks were more likely to be killed. Ross (2015) reported that Black individuals were at greater risk of being shot by police than White suspects. In all racial groups, unarmed suspects were less likely to be shot than armed ones. However, a Black person was nearly 3.5 times more likely than a White person to be shot when armed. Furthermore, the likelihood of being shot while unarmed and Black, was nearly equal to being armed and White. Ross (2015) argues the patters are “consistent with police being more discerning of armed/unarmed status before shooting a white suspect than shooting a black or hispanic suspect” (p. 10). Ross et al. (2021) produced similar findings of an anti-Black bias in police shooting situations. Unarmed Blacks were more likely to be killed by police when considering the criminal population of an area.

Given the contradictory research and the attention paid to this area, further tests are warranted. The topic of racial bias in police use of deadly force has elicited the most research in the subfield of police shootings over the last 10 years (Oramas Mora et al., 2022). Furthermore, a close examination of off-duty police interactions and the race of the suspect/victim is needed.

Community and Police Shootings

Characteristics of the community could serve as factors in the understanding of shooting incidents. For example, areas of lower income have seen greater rates of criminal activity, thus accounting for an increase in the likelihood of shootings (Zare et al., 2022). Furthermore, criminal activity could be a function of the density of the population area, as more people mean more opportunity for crime. Neighborhoods with higher criminal violence are more likely to have victims of an officer involved shooting (Sheppard & Tucker, 2022; Zare et al., 2022). As crime is more prevalent in urban settings, particularly poor ones, the community indices need to be controlled for in analyses of police shooting factors.

Current Study

This research sought to test available data on off-duty police shootings across the nation, the first such examination of its kind. Previous research has generally been limited to a single city or state sample. The aim of the paper has a multitude of research questions to answer. First, what are the most common reasons an off-duty officer shooting is likely to occur? An exploratory analysis surrounding the reasons off-duty police shootings occur via the coding of the incidents should shed light on this question. Second, to test off-duty shootings as a function of race/ethnicity, as this issue is at the forefront of the public attention. More specifically, are Persons of Color more likely, less likely, or similar to Whites to engage in criminal actions likely to result in a fatal encounter? To test this question, statistical analyses of whether either group was armed in an encounter, whether either group was committing a crime against person, and whether either group was actively attacking the officer at the time of the incident were conducted. Finally, additional analysis focused on the legal implications of off-duty deadly force. The likelihood of facing criminal charges was limited, but included where feasible. The results of this work increase our knowledge of the causes and outcomes of these types of shootings.

Method

Cases

Utilizing a complete database of LEO involved shootings is a challenge (Alpert, 2016). Researchers have continuously noted the lack of a Federal database of officer-involved shootings (Jennings et al., 2020; Lim, 2017; Shane et al., 2017). For this project, three sources of officer-involved fatalities were examined and collapsed. The first data source was obtained from Mapping Police Violence (https://mappingpoliceviolence.org/). The second source was the Washington Post Police Shooting Database (https://www.washingtonpost.com/graphics/investigations/police-shootings-database/). The third was obtained from Fatal Encounters (https://fatalencounters.org/). These sources of data represent a “step in the right direction to police accountability” (Gray & Parker, 2019, p. 42), but also limit the potential biases of relying on self-reporting by police agencies. Furthermore, the merging of multiple datasets does provide additional checks of the accuracy of the reported details surrounding the shootings.

The databases initially identified 125 potential off-duty officer involved deaths from a five-year period starting January 1, 2015 and ending December 31, 2019. These cases were subsequently subjected to greater scrutiny. Specifically, extensive internet searches and review of the news coverage of the cases was undertaken on each case. Several of these cases had not been updated for new information with regard to the variables already included in the datasets. News media provides additional insight into the cases, as well as follow-up information not presented in the databases (Durán & Loza, 2017). Some cases had missing information that was not available via news reporting. Indeed, multiple news agencies simply reported the initial incident and no follow-up information about the case thereafter. Cases from two states required contacting state agencies and filing a Freedom of Information Act (FOIA) request. Additional information from two cases was not available due to the open investigation of the case.

Variables

The aforementioned datasets included variables on the race and gender of the suspect/victim, the age of the suspect/victim, on- or off-duty status, reason/type of crime causing the initial encounter between the officer and victim/suspect, legal outcomes (prosecutorial decisions), whether there was a direct attack on the officer, and whether the suspect/victim was armed or unarmed. The last two variables have been found to be significantly high in police shootings. Fagan and Campbell (2020) found that nearly 93% of suspects killed by police were armed in some fashion at the time of the shooting. In the same study, attacks on the officer was the second most common factor at nearly two-thirds of cases.

Control variables in all models included the age of the deceased, the median household income according to the 5-Year American Community Survey (ACS) 2019 data for the zip code of the death location, and the National Center for Health Statistics (NCHS) Urban-Rural Classification Scheme for Counties (Ingram & Franco, 2014). The ACS maintains a litany of data estimates for the entire country on hundreds of social, economic, and health variables. The median income of the area was chosen as an indication of the affluence/poverty of the neighborhood where the death took place. The mean income of the area for all shootings was $55,877 (SD = 19,654; min = $20,909; max = $111,169). The NCHS classification scheme measures the population criteria of counties and divides them into six classifications: Large central metro, Large fringe metro, Medium metro, Small metro, Micropolitan, and Non-Core. This variable was included as a continuous variable indicating less density and population size as the numbers progressed.

New variables were added for the purposes of this research project. More specifically, the cases were coded for whether or not there was an intentional homicide on the part of the individual, who just so happened to also be a LEO. Intentional homicides were cases in which the officer committed a criminal offense, whether purposely or recklessly, on another individual (i.e., murder, manslaughter, negligent homicide). Likewise, the cases were studied for possible relationships (personal, blood, friendship, acquaintance, or none) among the officer and the deceased. Cases were coded whether or not the officer was the initial target of the interaction. That is, was the officer being targeted by the deceased as a start to the interaction, which was mostly a crime perpetrated against the officer. Finally, the shootings were coded for whether the officer was coming to the aid of someone else or not.

Data Analysis

Initial analyses of the dataset centered on a series of descriptive statistics, as well as crosstabulations of the variables of interests. A series of binomial logistic regressions, modeled in part after Nix et al. (2017), examined variables of criminal activity. The following variables were tested in individual logistic regression models: the deceased was armed (Yes/No), active attacking of the officer (Yes/No), and the initial encounter was the result of a crime against person (Yes/No). When not the focus of the test as the dependent variable, these variables served as controls in the other tests. Other control variables in all models included a log transformation of the ACS median household income and the NCHS Urban-Rural Classification Scheme for Counties (Ingram & Franco, 2014).

To further examine the variable of race in off-duty police shootings, cases were coded with a dichotomous Person of Color (POC) or White suspect/victim. The race of the suspect/victim was entered into the model as a potential predictor of the three dependent variables.

Results

Two cases were removed from the data set because the shooting was miscoded as off-duty and was actually carried out by an on-duty officer. Of the remaining 123 deaths, four were noted as incidents with multiple victims (2, 2, 2, and 3). Therefore, the final count was 119 off-duty officer(s) involved incidents which lead to the death of the 123 individuals.

The majority of those killed were males (n = 104). The age of the deceased was available in all but two cases. The victims/suspects had an average age of 32.26 (SD = 12.75) with a range from 1 to 73. Race/ethnicity of the deceased was available in 121 of the cases. The counts for each identified race or ethnicity were as follows: Black (46.3%), White (39.0%), Hispanic (12.2%), and Native American (0.8%). Information on whether the deceased individual was armed or not was available in all but five cases. The armed/unarmed counts were fairly close with 56.8% of the suspects being considered armed.¹ Legal system follow-up information was available in 58 cases. Cases involving suicide or death of the officer before a final legal decision were excluded. Officers were criminally charged in 26 of the 58 cases.

After examining the circumstances of the reason for the interaction between the suspect/victim and the off-duty officer, four categories of contact were created: Intentional Homicide, Self-defense, Interventions for Others, and Other Interactions. This differs from White’s (2000) notion that there are generally two groups of off-duty incidents, incidents common to both on- and off-duty officers and incidents unique to off-duty officers. For example, in White’s assertion there is a merging of what is considered here to be self-defense and intentional homicide. Given the stark contrast in legal outcomes for the two categories in the current study, the data are analyzed separately. Other Interactions could fit into both of White’s groupings. Some examples of these are accidental discharges of weapons and mistaken identity, both of which could occur while on- or off-duty. The race, gender, and age of known victims, as well as the count of armed individuals for four distinct situations involving off-duty officer shootings are presented in Table 1.

Table 1.

Demographic Characteristics of Suspects/Victims Across Shooting Classifications.

	Officer intentional homicide (n = 22)	Officer targeted (self-defense) (n = 44)	Officer interventions for others (n = 41)	Officer other (n = 18)^a
Suspect/victim characteristics
Male	7 (31.8%)	44 (100%)	40 (97.6%)	15 (83.3%)
Female	15 (68.2%)	0 (0%)	1 (2.4%)	3 (16.7%)
Black	3 (13.6%)	28 (63.6%)	21 (51.2%)	5 (27.8%)
Hispanic	0 (0%)	5 (11.4%)	9 (22.0%)	2 (11.1%)
White	19 (86.4%)	11 (25.0%)	9 (22.0%)	10 (55.6%)
Native Amer.	0 (0%)	0 (0%)	0 (0%)	1 (5.6%)
Unknown	0 (0%)	0 (0%)	2 (4.9%)	0 (0%)
Armed	0 (0%)^c	28 (65.1%)^b	37 (94.8%)^b	4 (22.2%)
Mean age (SD)	35.73 (14.98)	29.40 (9.89)^b	32.98 (13.67)^b	33.67 (12.98)
Age range	1–57	15–53	15–68	16–73

Note.

Numbers do not total 123 victims because of two cases of overlap in self- and other defense.

One missing case.

Two missing cases.

Intentional Homicide by Off-Duty Officers

When examining the cases of all off duty shooting deaths, 18 officers were identified as having committed intentional homicide of another person. In these cases, the officer while off-duty committed a murder of another individual or individuals. Two incidents could not be identified as clearly intentional or not, both of which involved female officers and some form of combat with their domestic partner. Most of the remaining cases (totaling 22 victims) were instances of domestic violence involving the homicide of a partner, child, or ex-partner’s new relationship. However, two familial murders (dad, brother), one boss, and one friend of an ex domestic partner were included in the 18 officer homicides. Thirteen officers committed a homicide of a current or former domestic partner. Two officers killed a child of theirs, and one of these two also killed a grandchild. None of the victims of these homicides were considered armed at the time of the crime. Exactly one-half of the 18 officers committed suicide after their crime. All remaining officers were criminally charged and convicted.

Off-Duty Officers and Self-Defense

Of the remaining 101 off-duty officer shooting incidents, cases were examined for the potential that an officer was acting in self-defense. Cases of the defense of others is discussed below. Forty-four off-duty officer shooting incidents were classified as cases of possible self-defense. The majority of these cases (68.2%) were a direct result of criminal contact initiated on the off-duty officer (robbery, assault, home burglary). The other reasons for initial officer contact were scattered across a variety of reasons (road rage, domestic violence, disturbing the peace, theft, etc.).

All cases in this category involved male suspects. Racially, there was a disproportionate number of Blacks shot in these scenarios (see Table 1).

Coverage of the legal process beyond the incident is limited. Follow-up information was available in only 19 cases of the 43. It is interesting to note that six officers were charged with a crime in cases. The reason for contact in these six were as follows: auto theft, mutual combat with intoxication, traffic violations (2), domestic incident, simple battery. None of the cases in which an attempted robbery was the impetus for contact resulted in charges. Surprisingly, many suspects/victims in this category were not armed (approximately 35%) which could account for the decision to prosecute. An examination of follow-up legal charges and armed/unarmed status of the suspect/victim was significant, χ² (1, N = 55) = 29.22, p < .001, φ = .73.

The remaining 13 cases with follow-up information all resulted in justification from a prosecutor/district attorney, grand jury, or federal law enforcement agency, or the officer had passed away. Still, the majority of reports do not include follow up information for what happens to these officers at the next stage of the legal system.

Off-Duty Officers’ Interventions for Others

The classification of types of interactions between off-duty officers and the suspect/victim identified 41 cases of officers interceding on other’s behalf to stop a crime. Two of these cases also involved self-defense discussed above. These two officers were with another person at the time of the start of the interaction and subsequently a simultaneous target.

The reasoning for the initial encounter of the off-duty officer varied greatly. Robbery was the number one motive for the officer’s introjection into the situation at 24.3%. Second on this list was a shooting incident of some variety at 21.9%, and stabbing cases were third at 12.2%. Finally, theft accounted for three cases, or 7.3%, of the reasons for encounters between the officer and the suspect/victim. The vast majority of incidents involved armed individuals (94.9%). Like self-defense cases, much of the data on final disposition of the case is missing (22 missing). Of the 19 cases eligible for scrutiny, none of the officers were subsequently charged criminally.

Other Off-Duty Officer Interactions

A further 18 cases did not fit into the above three categories and are subsequently analyzed here. Five of the 18 cases involved accidental discharges of weapons resulting in the death of others. All five of those cases resulted in criminal prosecution of the off-duty officer. Three other cases were domestic in nature, three involved both parties in a mutual argument/confrontation, and three more were attempted traffic stops of suspects (after a potential identification of a suspect in another crime) while off-duty.

Interestingly, the other category of cases had the greatest percentage of legal outcome decisions. Data were available in 14 of the 18 cases. Only four of the 14 cases resulted in a justified/cleared or insufficient evidence outcome. The remaining 10 cases all proceeded with criminal charges against the officer. Potentially influencing these prosecutorial decisions was the fact that 12 of the victims were unarmed at the time.

Race and Off-Duty Officer Shootings

Because of the low number of Native Americans in the data, a Fisher’s Exact test examining the race/ethnicity of the suspect/victim across the four different categories² was conducted (see Table 1). This resulted in a significant outcome, p < .001, φ = .59. A visual inspection of the expected count found the largest discrepancies in the racial makeup of the suspects/victims was seen in Officer Initiated Homicide (mostly White) and the Officer Self-defense (mostly Black) categories. Because of the findings regarding race and the intentional homicides, the forthcoming statistical tests were run twice. The first analysis includes all data, with a repeat of the analysis excluding the intentional homicides.

As mentioned above, a dichotomous variable of White/Person(s) of Color (POC) suspect/victim was created. There was a significant finding with respect to the White/POC and armed/unarmed status of the deceased. The POC were more likely to be armed than their White counterparts, χ² (1, N = 116) = 16.87, p < .001, φ = .38. However, this effect was quite diminished when accounting for the intentional homicide victims, χ² (1, N = 95) = 4.79, p = .029, φ = .23.

The logistic regression³ test of armed or unarmed individuals was significant, Model χ² (5, N = 106) = 50.32, p < .001, Nagelkerke R² = 0.51. As can be seen in Table 2, three variables were significant in the model: was the officer being attacked, was the initial encounter a crime against person, and the variable POC/White. This result suggests (a) the officers were more likely to be attacked by armed individuals, (b) armed suspects were committing crimes against persons, more so than crimes against property, and (c) that there was a higher likelihood that POC suspects/victims were armed.

Table 2.

Logistic Regression on Whether the Suspect/Victim Was Unarmed (0) or Armed (1).

	B	Bias	SE	Sig	95% CI for B
	B	Bias	SE	Sig	Lower	Upper
Median household income^a	−1.580	−0.225	2.279	.421	−6.627	2.340
NCHS population	0.126	0.035	0.274	.558	−0.332	0.784
Initial encounter a crime against person	1.806	0.216	0.690	.003**	0.746	3.492
Was the deceased attacking the officer?	1.644	0.207	1.156	<.001***	0.459	3.488
POC or White	−1.371	−0.107	0.714	.015*	−2.947	−0.152
Constant	6.274	0.831	10.665	.491	−12.090	29.684

Note. N = 106. Initial encounter 0 = no 1 = yes. Deceased attack 0 = no 1 = yes. POC = 0; White = 1.

Log10 transformation of Median Household Income.

p < .05, **p < .01, ***p < .001.

The results of this logistic regression excluding Intentional Homicide cases are presented in Table 3. The removal of the homicide cases modified the findings in strength but not direction, χ² = 24.04 (5, N = 89), p < .001, Nagelkerke R² = 0.34. It did, however result in a change of significance for two of the three variables. Attack of an officer was no longer significant. Likewise, the dichotomous variable of POC or White deceased individual was no longer significant by the traditional cut-off (p = .055). Only the initial factor of a crime against person response remained significant in the model.

Table 3.

Logistic Regression on Whether the Suspect/Victim Was Unarmed (0) or Armed (1) After Removing Intentional Homicides.

	B	Bias	SE	Sig	95% CI for B
	B	Bias	SE	Sig	Lower	Upper
Median household income^a	−1.769	−0.403	2.460	.375	−7.587	2.585
NCHS population	0.207	0.062	0.358	.416	−0.305	1.139
Initial encounter a crime against person	1.858	0.287	0.959	<.001***	0.773	3.723
Was the deceased attacking the officer?	0.877	0.136	1.262	.200	−0.682	2.827
POC or White	−1.211	−0.109	0.784	.055	−3.052	0.142
Constant	7.696	1.684	11.583	.408	−13.343	34.419

Note. N = 89. Initial encounter 0 = no 1 = yes. Deceased attack 0 = no 1 = yes. POC = 0; White = 1.

Log10 transformation of Median Household Income.

***

p < .001.

Similarly, the model testing whether officers were attacked or not in a confrontation was significant, χ² (5, N = 106) = 48.84, p < .001, Nagelkerke R² = 0.51. The model (see Table 4) had three significant variables (armed status of the deceased, the initial encounter as a crime against person, and POC/White). The variable POC/White signified that the POC suspects were more likely to engage in an attack on the officer. Upon removal of intentional homicides (see Table 5), the model was reduced in strength, χ² (5, N = 89) = 18.76, p = .002, Nagelkerke R² = 0.30. Two of the previously significant variables dropped out of importance in the prediction. The officers being attacked or not was only predicted by whether the initial crime was against a person.

Table 4.

Logistic Regression on Whether the Suspect/Victim Was Attacking the Off-Duty Officer (1) or Not (0).

	B	Bias	SE	Sig	95% CI for B
	B	Bias	SE	Sig	Lower	Upper
Median household income^a	−1.871	−0.434	3.175	.486	−8.709	3.687
NCHS population	−0.125	−0.031	0.277	.588	−0.771	0.374
Initial encounter a crime against person	1.993	0.538	2.373	<.001***	0.739	4.657
Was the deceased armed?	1.677	0.216	0.732	.003**	0.603	3.520
POC or White	−1.131	−0.187	0.666	.041*	−2.872	−0.120
Constant	8.906	2.106	15.048	.486	−17.112	41.390

Note. N = 106. Initial encounter 0 = no 1 = yes. Deceased armed 0 = no 1 = yes. POC = 0; White = 1.

Log10 transformation of Median Household Income.

p < .05, ** p < .01, *** p < .001.

Table 5.

Logistic Regression on Whether the Suspect/Victim Was Attacking the Off-Duty Officer (1) or Not (0) After Removing Intentional Homicides.

	B	Bias	SE	Sig	95% CI for B
	B	Bias	SE	Sig	Lower	Upper
Median household income^a	−1.618	−0.525	4.367	.602	−10.251	5.235
NCHS population	−0.037	−0.035	0.329	.871	−0.713	0.504
Initial encounter a crime against person	1.856	1.195	4.238	.005**	0.394	20.612
Was the deceased armed?	0.964	0.137	0.972	.111	−0.310	2.760
POC or White	−0.832	−0.144	1.013	.182	−2.849	0.467
Constant	8.153	2.613	20.804	.570	−24.603	49.700

Note. N = 89. Initial encounter 0 = no 1 = yes. Deceased armed 0 = no 1 = yes. POC = 0; White = 1.

Log10 transformation of Median Household Income.

p < .01.

Finally, Table 6 presents the logistic regression variables of whether the initial encounter was the result of a crime against person or not, Model χ² (5, N = 106) = 39.37, p < .001, Nagelkerke R² = 0.42. Control variables of armed status and attacking/not attacking the officer were significant. The race of the suspect/victim was not a significant predictor of the variable. The removal of intentional homicide (see Table 7) did not change the direction of this model’s variables, only the strength of the association, χ² (5, N = 89) = 26.17, p < .001, Nagelkerke R² = 0.34.

Table 6.

Logistic Regression on Whether the Officer Was Reacting to a Crime Against Person (1) or Not (0).

	B	Bias	SE	Sig.	95% CI for B
	B	Bias	SE	Sig.	Lower	Upper
Median household income^a	1.563	0.293	1.952	.371	−1.863	6.121
NCHS population	−0.230	−0.018	0.264	.336	−0.794	0.260
Was the deceased armed?	1.813	0.183	0.916	.002**	0.775	3.394
Was the deceased attacking the officer?	1.815	0.544	2.710	.002**	0.641	3.985
POC or White	0.157	0.039	0.976	.818	−1.152	1.716
Constant	−9.644	−2.024	10.037	.228	−34.189	5.681

Note. N = 106. Deceased armed 0 = no 1 = yes. Deceased attack 0 = no 1 = yes. POC = 0; White = 1.

Log10 transformation of Median Household Income.

p < .01.

Table 7.

Logistic Regression on Whether the Officer Was Reacting to a Crime Against Person (1) or Not (0) After Removing Intentional Homicides.

	B	Bias	SE	Sig.	95% CI for B
	B	Bias	SE	Sig.	Lower	Upper
Median household income^a	2.026	0.293	1.991	.233	−1.388	6.388
NCHS population	−0.363	−0.038	0.277	.125	−1.006	0.119
Was the deceased armed?	1.863	0.425	1.883	<.001***	0.844	3.768
Was the deceased attacking the officer?	1.767	1.092	4.216	.009**	0.305	21.380
POC or White	0.182	0.005	0.827	.792	−1.362	1.910
Constant	−11.553	−2.732	10.925	.157	−40.006	3.958

Note. N = 89. Deceased armed 0 = no 1 = yes. Deceased attack 0 = no 1 = yes. POC = 0; White = 1.

Log10 transformation of Median Household Income.

p < .01, ***p < .001.

While limited in available data, the procedural outcomes of cases and race was tested. Off-duty officers were not more likely to be charged criminally for the death of either POC or White suspects/victims, χ² (1, N = 58) = 1.58, p = ns, φ = .17.

Discussion

The importance of understanding the causes and outcomes of police shootings is paramount in the current environment of police and community relationships. While much less common, off-duty police shootings are nevertheless still important moments to understand. Here, we were able to show that the most supported aspect of off-duty police interactions was the aid of another person. Being off-duty (or 10-7) does not always equate to actually being absolved of police responsibilities. While non-LEO citizens take their work home, the same is equally true of officers. Likewise, officers are not immune to the possibility of being a crime victim, nor a perpetrator. Indeed, these three situations account for most fatalities by the off-duty officer. The cumulative effect of self-defense and defense of others accounts for 82% of the fatalities in this study after the removal of intentional homicides.

Off-duty officers may choose to put themselves in harm’s way for the sake of defending others and stopping crime. This can be a perilous situation. Off-duty officers interject themselves into situations without the priming of a dispatcher to alert the officer as to what situation they are going into (Taylor, 2020). Taylor (2020) noted that officers rely on dispatcher’s information in deadly force situations. Instead, these officers are often caught off-guard by the situation. Charbonneau et al. (2017) found that half of the off-duty officers shot and killed by their LEO compatriots were in fact coming to the assistance of other civilians or even another officer. Two off-duty officers killed in Charbonneau et al.’s data were actually victims of another crime before the interaction with an on-duty officer. White (2000) noted just how confusing it may be for off-duty officers to determine the extent of their responsibilities. There is little consistency in the guidance that officers receive, if they receive any, from their departments on the administrative policy of off-duty officer obligations. The choice to intercede on behalf of another person was often justified by prosecutors or grand juries. As Gaston et al. (2021) have shown, violent crime is associated with the shooting deaths of male suspects irrespective of their race/ethnicity. It is not surprising that altruistic behaviors on the part of the off-duty officer would give grand juries and district attorneys pause in the pursuit of criminal charges.

Being a target of crime was the other major impetus for deadly force by officers. Historically, large numbers of officers killed were actually off-duty at the time of their death (Geller, 1982). Geller (1982) summarized two possible explanations: (a) administrative policy directing officers to take appropriate action when witnessing a crime and (b) the issue of officers, particularly Black officers, who were the victims of crime, such as robbery. It is unclear how many of the departments regulated off-duty officers to carry a weapon. There has been controversy surrounding the issue of mandatory firearm carrying as a departmental policy for decades (Fyfe, 1980). The issue of self-defense today and self-defense at the time of Fyfe’s work warrants further exploration. Fyfe (1980) found that over a five-year period more than half of all shootings were cause of off-duty shootings. A very small number (2.5%) were attributed to the defense of others. This is a notable shift in the likely causes of off-duty shootings.

Most situations in the Other Interactions category did not involve the officer introjecting themselves into domestic situations or arguments. This is encouraging considering the stereotype that these situations are more dangerous, emotional, and volatile than other situations. The low numbers of these interactions leading to off-duty shootings partially support the work of Pinchevsky and Nix (2018), who concluded that there is not an increased likelihood of fatal police shootings in domestic disturbance calls. This lack of off-duty misconduct is contrary to the work in New York City Police Department Studies of the 1970s (Fyfe, 1980) and more in line with the current research by Kane and White (2009). This shows a continued trend of better off-duty police behavior. However, there remains the recent findings that officers who engage in violent activities, do so more often when off-duty (Boateng et al., 2021). Predicting these behaviors goes well beyond duty status and is nuanced in several other areas. Another scenario folded into the Other Interactions construct were deaths at the hands of privately employed off-duty officers. The number of shooting incidents resulting from private security work was minimal. Deaths in these circumstances were exclusively the result of an armed individual who was menacing patrons or the business in question.

One case in the sub category of Other Interactions stands out because of the extensive news coverage associated with it. The death of Botham Shem Jean received a multitude of news outlet coverage. The off-duty officer, Amber Guyger, entered Jean’s apartment mistakenly believing it was her own before shooting Mr. Jean. Adding to the coverage was a racial component, as Mr. Jean was Black and Ms. Guyger is White. Guyger was subsequently charged, convicted, and sentenced to prison for the homicide. The State Court of Appeals upheld this conviction and sentence (Williams, 2022).

Martinot (2012) argued that the mainstream media demonize victims of police shootings and withhold personhood and humanity from nonwhite victims. In the late 1990s, the media’s coverage of police shootings invoked more legitimacy of the police’s actions than challenged the decision making of the officer and his or her department (Hirschfield & Simon, 2010). Durán (2016) found most local newspapers to be supportive of both the police and the district attorney (DA) when reporting on an individual shooting. This included the nearly unanimous decisions by DA’s not to charge officers in the shootings (on or off-duty).

News reporting on issues of race undergoes a change when a high-profile race-related shooting takes place. In Hirschfield and Simon’s (2010) study, the death of Amadou Diallo served as a historical point of comparison in the media’s reporting of race and police shootings. Race played a more central role in the newspaper’s reporting on police shootings after such an incident. Martinot’s (2012) assertions of an evil victim were not found in Guyger’s shooting. Instead, Jean was seen as a clear victim and Guyger the illegitimate perpetrator. Like the media coverage following the Diallo shooting, the racial component of the two was a prominent feature of the media coverage. The Jean case is contrary to assertions about the portrayal of police shootings a decade ago (Martinot, 2012). The coverage of Jean’s death could have continued a shift in media coverage on race and police shootings since the death of Michael Brown in 2014, similar to that reported by Hirschfield and Simon (2010). In the current study, the media reports of officer race were not greater in the cases of POC suspects/victims when compared to the overall percentage of such individuals.⁴

Finally, the cases of Intentional Homicides initiated by the officer warrant discussion. There is no evidence that the coverage of the officer involved domestic violence situations were different from media reports which would involve non-officers. Most cases fit into the spousal/consortial relationship category (Cohen et al., 1998). Only a single case involved an extrafamilial situation (the murder of a boss). There was no evidence to suggest that these few cases are indicative of any patterns among off-duty officers.

Limitations and Future Directions

This project was limited in the available data sources. Reliance on access to the cases by news media proved to be problematic in low profile cases/areas. The records of investigations into police shootings by these agencies far surpasses the limited information provided by any and all publicly available databases. Durán’s (2016) access to the complete files in the DA’s office took time to analyze but was limited to a single jurisdiction. Still, the data gathered from such a plethora of information can prove useful in the understanding of how police shootings justified or not. The greatest detailed information in this study was obtained via FOIA requests from state agencies. Future research on police shootings in general could benefit from an aggressive filing of these requests. Both prosecutorial and investigative decisions could be reviewed in such a manner from across the country.

A second limitation of this project is the sole reliance of fatal shootings. Numerous police shootings, both on- and off-duty, occur without the end result of a suspect’s death. It is possible that the data would be skewed in various directions should such data become available and analyzed alongside of the existing data on fatalities. Future research should update the work done in the 1970s on large department’s shooting data for off-duty officers.

Policy Implications

Previously Nix et al. (2017) recommended policy changes that implored “the U.S. government to move forward with its publication of a national police use-of-force database, including as much information about the officers involved as possible” (pp. 309–310). The Federal Bureau of Investigation (FBI, 2022b) moved forward with this program in January of 2019. However, this program’s participation is dependent on voluntary responses by law enforcement agencies. A quick examination will show any user of the platform that the data is severely lacking in many respects. Large states, such as California and Texas, report less than 5% of agencies participating and nearly 70% of sworn officers missing from the dataset. Other states, such as Tennessee, repeatedly show agencies are failing to respond on a regular basis, with some months showing “no response” from more than half of participating agencies. The data collected by the FBI needs a substantial enhancement to move beyond the superfluous nature currently reported. A good start would be a mandatory reporting by law enforcement agencies. The FBI’s projected ability to release information would accelerate rapidly under such a policy.

To a greater extent, police departments and police academies across the country need to examine their policy on training and communication of expectations to their personnel when off-duty. Some state statues codify a responsibility to disclose the duty status of the officer when a deadly weapon is used in an officer involved injury or death. Without a clear consensus of how officers understand their off-duty responsibilities, the risk of a misconstrued situation leading to a death only increases. The number of deaths resulting from private security work was minimal. However, there remains the recent findings that officers who engage in violent activities, do so more often when off-duty (Boateng et al., 2021).

Finally, there has been a lack of transparency in the investigation of off-duty police shootings by the agencies. Some states, like California, have created videos for dissemination to the public regarding the shootings, while others have never updated the public about any decision at any level (police, internal affairs, prosecutors/district attorneys). A comprehensive release of information about the case is a necessary component to ensure the impartiality of these investigations. Through FOIA requests, the author was able to take a close examination of the investigation material from some states, but the information gathered from the condensed releases of other states provided sufficient explanation to code data. Police departments and prosecutorial offices should be consistent with the release of material to the public to satisfy the trust placed in these investigations.

Footnotes

Acknowledgements

Thanks are extended to Frank V. Ferdik for thoughtful comments on an earlier draft of this paper and to Cameron S. Culhane for assistance with data collection and imputation.

Author Note

Portions of this research were presented at the March 2023 Annual Meeting of the American Psychology-Law Society, Philadelphia, PA.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

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

ORCID iD

Scott E. Culhane

Notes

Author Biography

Scott E. Culhane is a Professor in the Department of Criminal Justice at Austin Peay State University. He received his doctorate in legal psychology from the University of Texas at El Paso. His research focuses on three main areas: alibis, police shootings, and psychopathologies of serial homicide perpetrators.

References

Alpert

G. P.

(2016), Toward a national database of officer-involved shootings. Criminology & Public Policy, 15, 237–242. https://doi.org/10.1111/1745-9133.12178

Boateng

F. D.

Pryce

D. K.

Hsieh

M.-L.

(2021). The criminal police officer: Understanding factors that predict police crime in the United States. Crime & Delinquency, 69(9), 1700–1735. https://doi.org/10.1177/00111287211054732

Brunet

J. R.

(2008). Blurring the line between public and private sectors: The case of police officers’ off-duty employment. Public Personnel Management, 37(2), 161–174. https://doi.org/10.1177/009102600803700202

Cameron

(2014). The legal status of “off-duty” police. Blue Line Magazine, 26(9), 14–17.

Castle Rock v. Gonzales . (2005). 545 US 748.

Charbonneau

Spencer

Glaser

(2017). Understanding racial disparities in police use of lethal force: Lessons from fatal police-on-police shootings. Journal of Social Issues, 73(4), 744–767. https://doi.org/10.1111/josi.12246

Cohen

Llorente

Eisdorfer

(1998). Homicide-suicide in older persons. The American Journal of Psychiatry, 155, 390–396. https://doi.org/10.1176/ajp.155.3.390

Culhane

S. E.

Boman

J. H.

Schweitzer

(2016). Public perceptions of the justifiability of police shootings: The role of body cameras in a pre- and post-Ferguson experiment. Police Quarterly, 19(3), 251–274. https://doi.org/10.1177/1098611116651403

DeShaney v. Winnebago County Department of Social Services . (1989). 489 U.S. 189.

10.

Durán

R. J.

(2016). No justice, no peace: Examining controversial officer involved shootings. Du Bois Review: Social Science Research on Race, 13(1), 61–83. https://doi.org/10.1017/S1742058X16000059

11.

Durán

R. J.

Loza

(2017). Exploring the two trigger fingers thesis: Racial and ethnic differences in officer involved shootings. Contemporary Justice Review, 20(1), 71–94. https://doi.org/10.1080/10282580.2016.1262771

12.

Fagan

Campbell

A. D.

(2020). Race and reasonableness in police killings. Boston University Law Review, 100(3), 951–1016.

13.

FBI. (2022a). Mississippi burning (MIBURN) case. FBI records: The vault. https://vault.fbi.gov/Mississippi%20Burning%20%28MIBURN%29%20Case

14.

FBI. (2022b). National use-of-force data collection. https://www.fbi.gov/how-we-can-help-you/need-an-fbi-service-or-more-information/ucr/use-of-force (Fatal Encounters: https://fatalencounters.org/).

15.

Fatal Encounters. (2024). https://fatalencounters.org/

16.

Fridel

E. E.

Sheppard

K. G.

Zimmerman

G. M.

(2020). Integrating the literature on police use of deadly force and police lethal victimization: How does place impact fatal police–citizen encounters? Journal of Quantitative Criminology, 36(4), 957–992. https://doi.org/10.1007/s10940-019-09438-5

17.

Fyfe

J. J.

(1980). Always prepared: Police off-duty guns. The Annals of the American Academy of Political and Social Science, 452, 72–81.

18.

Gaston

Fernandes

A. D.

DeShay

R. A.

(2021). A macrolevel study of police killings at the intersection of race, ethnicity, and gender. Crime & Delinquency, 67(8), 1075–1102. https://doi.org/10.1177/0011128720977444

19.

Geller

W. A.

(1982). Deadly force—What we know. Journal of Police Science and Administration, 10(2), 151–180.

20.

Graham v. Connor . (1989). 490 U.S. 386.

21.

Gray

A. C.

Parker

K. F.

(2019). Race, structural predictors, and police shootings: Are there differences across official and “unofficial” accounts of lethal force? Crime & Delinquency, 65(1), 26–45. https://doi.org/10.1177/0011128718788044

22.

Hirschfield

P. J.

Simon

(2010). Legitimating police violence: Newspaper narratives of deadly force. Theoretical Criminology, 14(2), 155–182.

23.

Ingram

D. D.

Franco

S. J.

(2014). 2013 NCHS urban-rural classification scheme for counties. National Center for Health Statistics: Vital and Health Statistics, 2(166), 1–73. https://www.cdc.gov/nchs/data/series/sr_02/sr02_166.pdf

24.

International Association of Chiefs of Police (IACP). (2019). Law Enforcement Policy Center: Standards of conduct. Retrieved July 2019, from https://www.theiacp.org/sites/default/files/2020-06/Standards%20of%20Conduct%20June%202020.pdf

25.

Jennings

W. G.

Hollis

M. E.

Fernandez

A. J.

(2020). Deadly force and deadly outcome: Examining the officer, suspect, and situational characteristics of officer-involved shootings. Deviant Behavior, 41(8), 969–976. https://doi.org/10.1080/01639625.2019.1596532

26.

Kane

R. J.

White

M. D.

(2009). Bad cops. Criminology & Public Policy, 8(4), 737–769. https://doi.org/10.1111/j.1745-9133.2009.00591.x

27.

Lim

(2017). Police bias, use of deadly force, public outcry. Criminology & Public Policy, 16, 305–308. https://doi.org/10.1111/1745-9133.12293

28.

Mapping Police Violence. (2024). https://mappingpoliceviolence.org/

29.

Martinot

(2012). On the epidemic of police killings. Social Justice, 39(4), 52–75.

30.

Nix

Campbell

B. A.

Byers

E. H.

Alpert

G. P.

(2017). A bird’s eye view of civilians killed by police in 2015. Criminology & Public Policy, 16(1), 309–340. https://doi.org/10.1111/1745-9133.12269

31.

O’Neill

D. A.

Spence

W. R.

Lewinski

W. J.

Novak

E. J.

(2021). Training and safety: Potentially lethal blue-on-blue encounters. Police Practice & Research, 22(2), 1209–1228. https://doi.org/10.1080/15614263.2019.1617143

32.

O’Neill

Hartman

M. E.

O’Neill

D. A.

Lewinski

W. J.

(2018). Further analysis of the unintentional discharge of firearms in law enforcement. Applied Ergonomics, 68, 267–272.https://doi.org/10.1016/j.apergo.2017.12.004

33.

Oramas Mora

Terrill

Foster

(2022). A decade of police use of deadly force research (2011–2020). Homicide Studies, 27(1), 6–33. https://doi.org/10.1177/10887679221123591

34.

Phillips

S. W.

Kim

D.-Y.

(2021). Exploring officer-involved shootings with interaction effects: A deeper understanding of how race/ethnicity interacts with other factors in the use of deadly force. Criminal Justice & Behavior, 48(6), 755–775. https://doi.org/10.1177/0093854821997529

35.

Pica

Sheahan

C. L.

Pozzulo

Bennell

(2020). Guns, gloves, and tasers: Perceptions of police officers and their use of weapon as a function of race and gender. Journal of Police & Criminal Psychology, 35(3), 348–359. https://doi.org/10.1007/s11896-020-09365-3

36.

Pinchevsky

G. M.

Nix

(2018). Domestic disturbances and fatal police shootings: An analysis of the Washington Post’s data. Police Quarterly, 21(1), 53–76. https://doi.org/10.1177/1098611117735657

37.

Ross

C. T.

(2015). A multi-level Bayesian Analysis of racial bias in police shootings at the county-level in the United States, 2011–2014. PLoS One, 10(11), e0141854. https://doi.org/10.1371/journal.pone.0141854

38.

Ross

C. T.

Winterhalder

McElreath

(2021). Racial disparities in police use of deadly force against unarmed individuals persist after appropriately benchmarking shooting data on violent crime rates. Social Psychological and Personality Science, 12(3), 323–332. https://doi.org/10.1177/1948550620916071

39.

Scott

D. S.

Sadler

M. S.

Correll

(2017). A social scientific approach toward understanding racial disparities in police shooting: Data from the Department of Justice (1980–2000). Journal of Social Issues, 73(4), 701–722. https://doi.org/10.1111/josi.12243

40.

Shane

J. M.

Lawton

Swenson

(2017). The prevalence of fatal police shootings by U.S. police, 2015–2016: Patterns and answers from a new data set. Journal of Criminal Justice, 52, 101–111. https://doi.org/10.1016/j.jcrimjus.2017.05.001

41.

Sheppard

Tucker

(2022). Police shooting mortality: Investigating individual, incident, and contextual factors differentiating fatal and non-fatal police shootings. Journal of Ethnicity in Criminal Justice, 20(4), 319–344. https://doi.org/10.1080/15377938.2022.2122648

42.

Smith

M. R.

Tillyer

Engel

R. S.

(2022). Race and the use of force by police revisited: Post-Ferguson findings from a large county police agency. Police Quarterly, 26(4), 411–440. https://doi.org/10.1177/10986111221139442

43.

Stinson

P. M.

Sr. Liederbach

Freiburger

T. L.

(2012). Off-duty and under arrest: A study of crimes perpetuated by off-duty police. Criminal Justice Policy Review, 23(2), 139–163. https://doi.org/10.1177/0887403410390510

44.

Stoughton

S. W.

(2017). Moonlighting: The private employment of off-duty officers. University of Illinois Law Review, 2017, 1847–1900.

45.

Taylor

P. L.

(2020). Dispatch priming and the police decision to use deadly force. Police Quarterly, 23(3), 311–332. https://doi.org/10.1177/1098611119896653

46.

Tennessee v. Garner . (1985). 471 U.S. 1.

47.

Washington Post Police Shooting Database. (2024). https://www.washingtonpost.com/graphics/investigations/police-shootings-database/

48.

White

M. D.

(2000). Assessing the impact of administrative policy on use of deadly force by on- and off-duty police. Evaluation Review, 24, 295–318. https://doi.org/10.1177/0193841X0002400303

49.

Williams

(2022, March 30). Court upholds conviction, sentence of Amber Guyger, ex-Dallas officer who murdered Botham Jean. The Dallas Morning News. https://www.dallasnews.com/news/courts/2022/03/30/court-upholds-conviction-sentence-of-amber-guyger-ex-dallas-officer-who-murdered-botham-jean/

50.

Worrall

J. L.

Bishopp

S. A.

Zinser

S. C.

Wheeler

A. P.

Phillips

S. W.

(2018). Exploring bias in police shooting decisions with real shoot/don’t shoot cases. Crime & Delinquency, 64(9), 1171–1192. https://doi.org/10.1177/0011128718756038

51.

Zare

Meyerson

N. S.

Delgado

Spencer

Gaskin

D. J.

Thorpe

R. J.

Jr. (2022). Association between neighborhood and racial composition of victims on fatal police shooting and police violence: An integrated review (2000–2022). Social Sciences, 11(4), 153. https://doi.org/10.3390/socsci11040153