Abstract
Suicide by cop (SbC) occurs when an individual purposely engages in threatening behavior toward police officers in an attempt to be killed. Previous studies have found the prototypical SbC subject is male, mid-30s, with disrupted relationships, and mental health concerns, although these studies have almost exclusively relied on officer involved shootings or public information as sources of data. To address the dearth of knowledge for SbC cases involving no force or less lethal force, 419 SbC cases from the Los Angeles Police Department Mental Evaluation Unit were analyzed. Results revealed similar frequencies with regard to subject characteristics as in the previous literature; however, substantial differences were seen across incident and outcome characteristics, with a much lower rate of injury and death. Thirteen variables were associated with differing levels of force. The results of the present study paint a more positive picture of SbC outcomes for police and subjects alike.
According to the American Foundation for Suicide Prevention (2015), suicide accounts for more than 44,000 deaths per year, with an average of 121 suicides per day. In addition, for every completed suicide, there are approximately 25 attempts (American Foundation for Suicide Prevention, 2015). Suicidal behavior encompasses not only the unambiguous act of completed suicide but also a wide spectrum of attempts that range from highly lethal (i.e., the person only survived due to good fortune or a fluke) to minimally lethal (i.e., behavior unlikely to cause significant harm; Mann et al., 2005). Consequently, suicide, as we know it, evolves into a different phenomenon when a third variable, such as law enforcement, gets introduced.
Suicide by cop (SbC) is a method of suicide in which an individual purposely engages in threatening behavior toward police officers in an attempt to be killed (Mohandie, Meloy, & Collins, 2009). In such an instance, the person will come into contact with law enforcement, whether intentional or not, and behave or communicate in manner that suggests he or she has a desire for police officers to end his or her life. This can be seen in accounts of subjects pretending to brandish a weapon to police in order to provoke use of lethal force or disobeying commands and rushing police officers despite warnings that force will be used (Hutson et al., 1998; Lord, 2000; Mohandie et al., 2009). In these circumstances, suicidal intent is exhibited either through actions, verbal communications, or the ability of a subject to enlist a police officer to be instrumental in their death (Lord, 2000).
Research on Suicide by Cop
Early research on SbC relied heavily on analyzing SbC cases that were drawn from public resources, which included cases discussed in newspaper articles or on television, and those drawn from publicly available databases that track officer involved shooting (OIS). Those works that relied on newspaper and television resources were at the mercy of the reporting sources to accurately present information about such cases, and, therefore, information drawn from these studies is somewhat limited. An OIS can be defined as any incident in which an officer discharges his or her weapon (Hutson et al., 1998; Mohandie et al., 2009) and much of what is currently known about SbC incidents is based on information from OIS cases. Many of these OIS-based studies emerged in the late 1990s and primarily focused on the frequency with which SbC incidents occur. Quite obviously, not all OIS incidents are SbC cases. In fact, the general consensus of the research is that approximately 10% of OIS cases involve an SbC component (Kennedy, Homant, & Hupp, 1998; Parent & Verdun-Jones, 1998; Wilson, Davis, Bloom, Batten, & Kamara, 1998). At present, there are few studies that have looked at the frequency of SbC incidents that occur outside of an OIS; therefore, knowledge about the rates of SbC incidents that are successfully managed without the use of lethal force is limited.
Beyond looking at the frequency of occurrence, some studies have focused on gaining an understanding of the individuals who initiate SbC incidents. In 2000, Lord conducted a study that provided the first in-depth look at historical, personality, behavioral, and situational factors of SbC subjects. To identify SbC cases, Lord (2000) asked law enforcement personnel to select incidents in which individuals verbally or behaviorally exhibited their desire to be killed when confronted by police. Of the 64 cases analyzed, SbC victims tended to share some characteristics with individuals who committed suicide on their own. Specifically, a large percentage of the SbC victims had an identified mental illness, a history of drug and alcohol abuse, were experiencing stressful life events at the time of the incident, or had previously talked about wanting to commit suicide. Results also indicated that substance abuse (particularly of illicit drugs), previous suicide attempts, stressful life events, and threatening/intimidating behavior during the SbC incident were related, albeit weakly, to a fatal outcome (Lord, 2000).
As more became known about the characteristics of SbC subjects and incidents, Homant and Kennedy (2000b) argued that in order to help police identify these cases and respond more effectively, it was important to understand a subject’s motivation, thus they proposed a typology of SbC cases. To do so, they analyzed 143 SbC cases drawn from a variety of media and other public sources and, based on the characteristics and nature of the incident as well as the type of police response involved, they proposed three distinct types of SbC cases: direct confrontation, disturbed intervention, and criminal intervention. The direct confrontation cases comprised 31% of SbC incidents in their sample and entailed the subjects’ planning to attack police ahead of time with the explicit intent of being killed by law enforcement. The disturbed intervention cases comprised the majority (57%) of SbC incidents and entailed subjects’ acting in an irrational and emotionally disturbed manner, which eventually led to the SbC incident. The authors noted that the subjects in this type of case may be overtly suicidal or they may simply seize the opportunity of police arrival to be the means to their end. Finally, the criminal intervention cases made up 12% of SbC incidents and were instances that began with an “ordinary” crime, but police intervention was unwelcome and the subject exhibited preference of death (or chance of escape) over arrest and incarceration. While Homant and Kennedy’s (2000b) typology of SbC cases was a useful step forward, due to the unscientific nature of their data sources (i.e., public information, news media, etc.), replication of their findings is necessary to confirm the existence of these categories as well as the frequency with which they occur.
Some studies of SbC have sought to identify historical, demographic, incident, and behavioral characteristics that would significantly differentiate SbC cases from other OIS cases. In their study, Mohandie et al. (2009) reviewed over 707 OIS files from participating police and criminal justice agencies for incidents occurring between 1998 and 2006. Results of their review revealed 256 SbC cases and found these cases were more likely to result in death or injury of the subject than were other OIS cases. The prototypical SbC subject from Mohandie et al.’s (2009) study was male and in his mid-30s, had a history of suicidal ideation, had disrupted interpersonal relationships, and had prior or current mental health problems. Most of the subjects were armed and nearly half of those who were armed purposely discharged their weapon at police during the incident. The majority of these cases occurred spontaneously and subjects in these cases expressed suicidal intention verbally and typically threatened others both behaviorally and verbally. These characteristics were very much in line with findings from prior studies (Homant & Kennedy, 2000a; Homant, Kennedy, & Hupp, 2000; Lord, 2000, 2012, 2014; Lord & Sloop, 2010; Wilson et al., 1998).
Mohandie et al. (2009) also classified their sample of cases according to Homant and Kennedy’s (2000b) typology. However, they found notable differences in frequencies across the three types of SbC, with the Mohandie et al. (2009) sample having a much higher rate of criminal intervention (64%) cases and a much lower rate of direct confrontation (16%) and disturbed intervention (20%) cases. Based on their analyses, the authors concluded that SbC subjects were much more likely than regular OIS subjects to be deliberate, purposeful, and resolved in their actions to provoke lethal force.
In their study, Mohandie et al. (2009) point out a few methodological weaknesses inherent in only utilizing OIS databases when studying SbC cases. First, doing so employs a sampling bias by only including cases in which lethal force is deployed. Looking only through this lens creates the possibility that subjects involved in these situations may exhibit a higher degree of desperation and intentionality than suicidal subjects whose cases do not end in a use of lethal force. The authors postulated that perhaps SbC cases that do not evoke lethal force represent a different severity or range of psychopathology. In other words, are there important differences between SbC cases that end in use of lethal force and those that are managed with no force or less than lethal force? This point has also been raised by Lord (2014), who has conducted some of the only work done to shed light on SbC cases involving less lethal force.
Suicide by Cop in Cases With Nonlethal Use of Force
An interesting caveat found in the SbC literature is the lack of consistency in defining what constitutes an SbC case. Some studies have used definitions of SbC that require lethal force to have been deployed during an incident (Drylie & Violanti, 2008; Hutson et al., 1998), while others simply require a subject to engage in threatening behavior toward police in an attempt to be killed (Mohandie et al., 2009). Despite varying definitions, very few prior studies have reviewed incidents that did not involve lethal force but did involve subjects trying to have police kill them. Parent (1996, 1998), however, has discussed the idea of evolving tactical approaches within law enforcement and has argued that a combination of less lethal force (such as tasers or pepper spray) and better trained and informed interactions between officers and suicidal individuals may help resolve many SbC situations without resorting to lethal force. The notion of including SbC situations that involve no use of force or less lethal force demands a new understanding of SbC, one that steps away from using OIS-only samples, and moves toward a focus on the level of lethality deployed and the outcome therein.
Lord’s (2000) work differs from those preceding it in that she defined SbC as incidents in which individuals verbally or behaviorally exhibited their desire to be killed when confronted by police. Lord’s (2000) study provided insight into different possible outcomes of SbC because not all the incidents involved lethal force. In fact, 28% of the subjects were committed to a hospital, 25% were killed by police, 23% were arrested, 14% were injured by police, 8% committed suicide during the incident, and 2% had no action taken. Further supporting the notion that SbC cases may not always necessitate the use of lethal force, Homant and Kennedy (2000a) found that in SbC cases where less lethal force was used, 44% of cases resolved without the death of the subject. These outcomes certainly challenge the notion that SbC incidents are destined to result in death or injury of the subject and argue for the inclusion of all events in which there is indication of suicidal indent, regardless of the level of force used or the lethality of the outcome.
Given that Lord’s (2000) and Homant and Kennedy’s (2000a) studies are among the few to take this broader approach to analyzing SbC cases, further research along these lines is necessary for an increased understanding of the nature of SbC cases that fall outside the realm of OIS. This calls for a wider net to be cast on the SbC phenomenon as a whole in order to understand and include all incidents that might be considered SbC and to move away from the requirement of lethal force deployment or death of the subject. However, to address these gaps, studies must move beyond OIS databases as the primary source for information and this presents a challenge. OIS databases have been used because SbC cases are found within them and this information is typically public information and thereby accessible. The challenge in studying these cases is that access to an internal police database would be necessary to provide in-depth information about other relevant incidents. The obstacle of receiving internal data from police departments has been a great challenge; however, accessing these cases could serve to enhance our understanding of the nature of non-OIS SbC cases.
Present Study
Information drawn from OIS and high-profile cases has been helpful in creating a profile of the typical SbC subject, but the question remains as to whether looking only at these sources misses important information about SbC incidents that are resolved without the need for lethal force. In other words, is there a difference between the cases analyzed in prior research and those in which a subject is suicidal, plans to use the police to end their life, but the situation is managed by law enforcement without using force or with use of less lethal force? To explore this possibility, one must look at all possible SbC incidents, including those in which a shooting did not occur. Doing so may capture a different view of SbC subjects, and outcomes.
In addition, if the focus is taken off the survival of the subject and placed more on the outcome of the case, is a different view of SbC created? The reasoning for this approach lies in the usefulness of this information for police agencies. If these incidents rise to a level where use of lethal force occurs, the officer will be scrutinized, the police agency will come under fire, and in addition to the potential death of the subject, the officer is at risk to suffer negative emotional consequences as a result of the shooting, and possible taking of another’s life (Miller, 2006). For these reasons, it is useful to identify any associations between various aspects of cases where less lethal or no force is used and those cases where lethal force is used. This information is likely to be of use to police agencies to promote responses that are more successful in cases where less lethal force was used or—better yet—no force at all.
With these considerations in mind, the present study aimed to answer two primary research questions. First, when analyzing cases from a more inclusive, internal police database, what does the typical SbC subject look like? In particular, when compared with SbC cases drawn from an OIS database, such as those in the Mohandie et al.’s (2009) study, are the characteristics of these subjects and incidents similar? As well, does this more inclusive set of subjects fit within the previously established typology offered by Homant and Kennedy (2000b)? Looking at the characteristics of these cases can help determine whether taking this broader look at SbC cases offers additional insights beyond those drawn from OIS cases exclusively. As well, if the Homant and Kennedy (2000b) typologies hold up with the present sample, this can offer some reassurance that this broader look at the SbC phenomenon holds up even when including cases that involve no force or less lethal force. The second question to be answered was whether particular variables are associated with differing levels of force? Because preventative strategies are better formed by understanding what factors would encourage a subject’s survival, rather than their demise, it was of interest whether differing levels of force were associated with different case or subject characteristics.
Methods
Participants
Data for the present study were derived from a retrospective review of case reports that were categorized as SbC by the Los Angeles Police Department’s Mental Evaluation Unit (LAPD MEU). LAPD MEU officers are specially trained in crisis evaluation, negotiation, and mental illness and serve to consult with the patrol officers on calls involving mental health issues. As part of their work, the MEU maintains a database comprised of several hundred thousand reports that capture important information about the incidents to which the MEU officers respond. Information in this database includes subject demographic data, location of call and call type, and a narrative written by the responding officers. There are checkboxes that allow for classification and filtering of specific types of cases, among those classifications is an SbC designation. The reports used for this study were identified as SbC by responding officers.
In classifying cases, the MEU has a special designation process for deeming incidents as SbC that is based on the Police Officer Standards and Training’s (1999) definition. To be classified as SbC, subjects must verbally express that they wish to be killed by police or behaviorally assert themselves in an aggressive manner in a way that would encourage SbC. Many (if not most) of the MEU reports do not involve a shooting. A review of all incidents in the MEU database that occurred between January 2010 and December 2015 initially revealed 533 cases that were classified as SbC. These cases were preidentified by LAPD MEU personnel; therefore, the researchers did not make this designation. After reviewing all 533 cases, 114 were excluded from analysis due to the case later being declassified by the agency as not being SbC, there being no report available from which to code the variables, or the case only contained follow-up paperwork to a previous SbC incident. The final sample included 419 cases.
Measures
Permission was obtained from Mohandie et al. (2009) to use and adapt a six-page, 110-variable codebook that was used in their prior studies on SbC. The initial intent of Mohandie et al. differed slightly from the aims of this study, thus the codebook was adapted to fit with the data and reports to be coded in this study. This modified version of the codebook 1 was the basis for coding pertinent information from the MEU reports.
The adapted codebook included 86 variables that were divided into three main categories: incident, subject, and outcomes. Within the Incident category, the variables of interest included (a) overall characteristics of the incident (i.e., descriptive information, level of force used, and efforts police made) and (b) incident context (i.e., call type, severity of crime, type of intervention sought, setting, and location). Within the Subject category, the variables of interest included (a) demographics, (b) information about the communication (verbal and behavioral) of the subject (i.e., communication of suicidal intent, planning of SbC, and verbal/behavioral threats), (c) weapon use, (d) other SbC indicators (i.e., co-occurrence of crimes, whether or not the subject evidences running as escape behavior, noncompliance, resistance, aggression, inconsistent escape behavior, verbal will, a desire to be shot, destruction of property, or harm to others), (e) psychological factors (i.e., prior suicidal behaviors, descriptive information of past attempts, mental health diagnosis, presence of psychoticism, substance use, number of contacts with MEU, prior/current use of medications, and history of hospitalizations), and (f) historical factors (i.e., criminal history, history of violence, prior incarcerations, and current probation/parole). Within the Outcomes category, the variables of interest included (a) police tactics used and (b) the outcome of the incident. A final variable was created to capture the three levels of force of interest in this study (no force, less lethal, or lethal).
Procedure
The first author obtained LAPD-approved status as a volunteer and reported to LAPD headquarters in order to retrieve and code reports on police premises. Keeping all reports onsite at the LAPD headquarters served to protect the confidentiality of the information so as not to breach compliance with Health Insurance Portability and Accountability Act and other legal protections of private information. In an agreement with the Detective and Lieutenant of the LAPD MEU, the researcher reported to headquarters where a computer desk was secured and logged into by a supervisor. The supervisor opened the database and filtered the cases to only include SbC incidents dated from January 2010 to December 2015. From there, each case was reviewed and information from the reports was recorded on coding sheets containing the variables of interest for this study. The only documents to leave the premises were the deidentified coding sheets. All procedures used in this study were in compliance with legal and ethical requirements and both the LAPD and academic review boards that govern research protections for human participants approved the study.
Results
Suicide by Cop Case Characteristics
Subject characteristics
The mean age of subjects was 38.01 (standard deviation = 13.38) with a range of 14 to 76 years. The vast majority of subjects were males (83%), and there were a wide range of races/ethnicities represented. A majority of the sample had a confirmed or probable mental health diagnosis (67%) and of the confirmed group, schizophrenia was the most commonly diagnosed disorder (accounting for 19% of the total sample), followed by bipolar disorder and depression (16% and 14%, respectively). Several subjects had previously been hospitalized and were previously or currently prescribed medication for mental health purposes (40% and 38%, respectively; however, medication use was unknown in more than 38% of the sample for both). Table 1 presents a summary of the subject characteristics.
Comparison of Subject Characteristics.
Note. SD = standard deviation.
aVariables that had a large number of unknown cases (>20%).
Incident characteristics
The most common call types received by police were “mentally ill subject” (28%) and “suicidal subject” (23%). The majority of the contacts made with police were spontaneous (69%) and only 28% of subjects deliberately made contact. A large majority of the subjects made suicidal communications during the incidents (89%); most were considered aggressive during the incident (61%) but not as many were verbally or behaviorally threatening (38% and 43%, respectively). One quarter of subjects were armed during the incident, but most of those were armed with a knife as opposed to a firearm. For those subjects who were not armed but aggressed toward officers in a threatening manner, most rushed officers (7%) or had their hands in their pockets in a suspicious manner (4%). All variables listed earlier, with the exception of Call Type and Type of Weapon, were coded as being present or absent. Therefore, these categories are not mutually exclusive of one another. Table 2 presents a summary of incident characteristics.
Comparison of Incident Characteristics.
Note. n/a = not applicable.
aVariables that had a large number of unknown cases (>20%).
Additional incident characteristics that were not in the table include behavior that occurred prior to the actual incident: 38% of subjects verbally communicated their suicidal intentions prior to the incident (n = 160) with 53% of those communications specifying SbC (n = 85). Fifty-one percent of subjects verbally threatened others prior to the incident (n = 212). Fifty-three percent of subjects behaviorally threatened others prior to the incident (n = 221). Finally, in 43% of cases (n = 181), police called for another service (either an LAPD mental health team, crisis negotiations, or SWAT) and in 95% of the cases (n = 396), there were overt verbal efforts from officers.
Outcome characteristics
Results revealed that the majority of cases fell under the no force category (81%), with far fewer in the less lethal force (17%) and lethal force (2%) categories. Within the less lethal category, 30% involved a taser, 27% involved hands-on use of force, 21% involved a beanbag shotgun, 7% involved pepper spray, and 15% used a combination of the aforementioned methods. Five total subjects died as a result of the incident; however, one of these five killed himself/herself during the incident as opposed to being killed by police like the other four. A vast majority (82%) of the incidents were resolved by hospitalizing the subject. Table 3 presents a summary of the outcome characteristics.
Comparison of Outcome Characteristics.
Comparison to Prior Research
Because one of the primary aims of this study was to compare the results of this more inclusive and less lethal set of SbC cases to those derived from on OIS database, the present results are presented side by side with those of Mohandie et al. (2009) in Tables 1 to 3. This head-to-head comparison is meant to provide a contrast across cases that were coded in similar ways, albeit obtained from largely different data sets.
In Table 1, very similar numbers were seen in terms of the sample demographics, with the typical SbC subject being male, mid-to-late 30s, either Hispanic or Caucasian, and most likely diagnosed with a mental disorder. Greater differences, however, were seen between the present results and Mohandie et al. (2009) regarding incident characteristics. Table 2 shows a greater concentration of call types being categorized as “mentally ill subject” or “suicidal subject” in the present study, while the concentration in Mohandie et al. is less clear-cut and includes more “domestic violence,” “other,” and “observed event” calls. As would be expected, far higher percentages are seen in Mohandie et al. with regard to variables that are indicative of violence during the incident, such as verbal and behavioral threats, aggressive behavior, being armed (particularly with firearms), and being under the influence of alcohol.
The largest difference between the present study and Mohandie et al. (2009) was within the outcomes of the incidents (see Table 3). Not surprisingly, the comparison study resulted in 92% of the OIS cases involving the use of lethal force, whereas the present study only had 2% of cases resulting in lethal force. Moreover, the present study had far higher percentages of no force or less lethal force used; therefore, a large difference is seen in the rate of death and injury to the subjects. Specifically, the rate of death was 1% in the present study, compared with 51% in Mohandie et al. the rate of subject injury was 3% versus 40%, and the rate of injury to responding officers was less than 1% versus 16%, respectively. Finally, far more cases from the present sample resulted in hospitalization of the subject as opposed to arrest (82% vs. 7% in prior study). Overall, the comparison of the present results to those reported in the Mohandie et al.’s study reveals similarities in terms of subject characteristics, but notable differences across incident characteristics, and large differences in outcomes.
SbC Typology
With regard to the typologies proposed by Homant and Kennedy (2000b), the present study found that 56% of all cases were classified as disturbed intervention, 28% as direct confrontation, and 16% as criminal intervention. Table 4 presents a comparison of these results to those of Homant and Kennedy (2000b) and Mohandie et al. (2009). These results revealed strong similarities between the present results and those of Homant and Kennedy (2000b), but dissimilar results compared with Mohandie et al. (2009).
Comparison of Suicide by Cop Typologies Across Studies.
Differences in Use of Force
I Chi square analyses were conducted to determine whether different subject and incident characteristics are more strongly associated with different levels of force (no force, less lethal force, and lethal force). Prior to the analysis, the data were screened for missing values. For most variables, the data were complete; however, a few of the demographic and subject characteristics were missing a large number of values and, as such, were not included in these analyses. The only variable for which this is an exception was diagnosis, for which pairwise deletion of cases was used and only complete cases were analyzed.
No significant associations were for found for any of the demographic variables. The variables analyzed included gender, race/ethnicity, mental health diagnosis, the presence of psychosis, and homelessness. The results are presented in Table 5. For the relevant incident variables, the results revealed several significant associations across the differing levels of force. Both behavioral and verbal indicators of the subjects’ suicidal ideation as well as generally aggressive behavior were significant, while both kinds of contact with police (spontaneous and deliberate), SbC communication, and destruction of property were nonsignificant. Table 6 presents the results of the χ2 analyses alongside with the frequencies and percentage of cases falling at each of the three levels of force.
Crosstabulation of Levels of Force and Subject Variables.
aThese results should be interpreted with caution as there were a large number of cases with missing values.
Frequencies and Associations Across Varying Levels of Force for Incident Variables.
Note. Percentages are calculated from the number of subjects in each group (no force = 341, less lethal force = 71, and lethal force = 7), for all χ2 analyses df = 2. SbC = Suicide by cop.
To provide a visual representation of the differing levels of force seen in these variables, clustered bar graphs were created (see Figures 1–3). The percentages presented in the graphs are calculated from the number of subjects in each of the three level of force categories; therefore, each bar represents the percentage of cases from that level of force that are represented for that variable (i.e., spontaneous contact was seen in 69% of the 341 cases involving no use of force, 72% of the 71 cases involving less lethal force, and 29% of the 7 cases involving lethal force). These graphs illustrate a very high likelihood of lethal force being deployed with the presence of overt behavioral characteristics. Moreover, there appears to be a higher likelihood of no force being deployed when the subject is more verbal and communicative about their suicidal ideation. The most telling of the graphs is Figure 3 with the behavioral characteristics during the incident. Almost all of the variables listed on that graph were seen in 100% of the lethal force cases (except for inconsistent escape behavior) and most were also seen in 80% of less lethal force cases (except for inconsistent escape behavior and subject being armed). These results imply that there is a higher likelihood of force when behavioral indicators are present during the incident, although the findings for lethal force should be interpreted cautiously, as they are comprised of such low frequencies that the generalizability of the data may be limited.
Differences in use of force for SbC incident characteristics.
Differences in use of force for verbal characteristics during SbC incidents.
Differences in use of force for behavioral characteristics during SbC incidents.
Discussion
The vast majority of what is currently known about SbC cases has been derived from studies that utilize OIS databases and high-profile media cases as their sources, which means that virtually all of the knowledge about these incidents stems from cases where officers deployed lethal force. The present study aimed to expand upon the present state of the SbC literature by analyzing SbC cases from an internal police database where the police department has classified cases and where all cases that involve some degree of suicidal intent are included, regardless of whether or not lethal force was involved.
SbC Case Characteristics
The first aim of the present study was to determine, when using a wider range of cases, what the typical SbC subject and incident looked like and how similar those characteristics were to findings from the previous literature. It was expected that results would be similar to those from the previous literature, particularly from Mohandie et al. (2009) as this study included the largest sample of SbC cases to date and used similar procedures and coding strategies. The results revealed that the average SbC subject was male, with an average age of 38 years, a wide range of race/ethnicities represented, and a high probability of mental illness. The most common call types for this sample were “suicidal subject” and “mentally ill subject” and 89% of subjects verbally indicated their suicidal intent during the incidents. The outcomes of these cases show that most were resolved without any use of force (81%) and the vast majority of subjects were hospitalized (82%) rather than arrested.
In comparison to Mohandie et al. (2009), the present data show that while on the surface these cases look similar in terms of demographics, when looking closer at what happens during the incident and how those occurrences are resolved there are many differences. First, both samples involved more spontaneous than deliberate contact, but the present data show notably more calls involving suicidal or mentally ill subjects. Second, the present sample involved very few cases involving lethal force and a much lower likelihood of either the officers or subjects being injured or killed. In fact, the vast majority of subjects in the present sample were hospitalized, as opposed to being arrested. Thus, while it appears we are dealing with similar subjects, the actual incidents resolve in very different ways across these two samples.
While a definitive explanation of the source of these differences is beyond the scope of this one study, it is logical to conclude that the differences across the samples are a direct reflection of the differences in data sources. Quite obviously, OIS cases entail more lethality, as by definition they involve a shooting. Therefore, these cases carry a much higher likelihood of injury or death not only for subjects but also for officers (Mohandie et al., 2009). On the other hand, analysis of a broader range of cases reveals a more optimistic outlook for SbC cases, as the resulting rate of injuries and deaths drops quite substantially. When considering the source of the data for the present study, all cases were drawn from the LAPD MEU database. The MEU is a unit that is specifically designed to assist officers with cases wherein some degree of mental health issue is involved. When we look at SbC cases through this lens, we can conclude that having the MEU team support likely helped to resolve cases in a way that promotes the use of mental health resources as opposed to having only police tactics and varying levels of force with which to respond. While further research is necessary to conclude the MEU teams are the primary reason behind the greater successful resolutions in this sample, it is a logical inference that warrants further testing. Should these findings carry over to other departments and settings, this would provide strong support for departments who do not have them already, to implement mental health supports for officers, such as MEU teams and Crisis Intervention Team (CIT) training as these resources could be vitally important for officer and subject safety in cases involving SbC.
SbC Case Typology
The present results were classified according to the typology formulated by Homant and Kennedy (2000b). As the only prior study that attempted to use these typologies showed conflicting results, the present data provide a third look and indicated strong consistency with the frequencies of cases found by Homant and Kennedy (2000b). Therefore, it appears that, here also, data source likely matters when trying to categorize cases according to this typology. When looking solely through an OIS lens, most SbC cases involve a criminal element; however, when looking at a wider variety of cases, we see just over half of the cases being disturbed interventions, with fewer direct confrontations, and those cases involving criminal interventions to be the least frequent type of SbC case. On the whole, these three categories do seem to provide a useful way of categorizing incidents and given the high frequencies of disturbed intervention cases, the importance of mental health training and resources for officers is emphasized as a way to help increase the likelihood of better outcomes.
SbC Cases Differences by Level of Force Used
The second primary aim of this study was to determine whether any case characteristics were associated with differing levels of force. First, no associations were found across levels of force for any of the demographic variables. Given the intense focus on policing and race, it is important to explore whether certain superficial characteristics, like gender or race, may prompt a higher likelihood of lethal force, but in this study that did not appear to be the case. Second, when looking at incident characteristics, it is also reassuring to see that the variables associated with lethal force are the things that would evoke force in any police encounter. Specifically, when a subject is aggressive, armed, behaviorally threatening, or harming others, the chances of lethal force being used increased. Overall, these results provide encouragement that, even in these intense situations where a subject may be directly provoking officers in an SbC attempt, that use of lethal force appears to be used only in cases where a direct behavioral or weapon-based threat is present. Moreover, one does not see an escalation in force even in cases where a subject may directly communicate a desire to be killed by police or may be destructive to property but not a danger to others. Both spontaneous and direct contact cases were not associated with a higher likelihood of less force; thus, it appears that simply looking at the initial type of contact in a case does not relate to the potential outcome of that case. When considering these findings, it is important to keep in mind that no causal inferences can be made. Instead, the findings merely illustrate identified patterns and associations across the varying levels of force included in the study.
Implications
The results from the present analysis of cases drawn from an internal police database offers a different perspective of SbC compared with prior studies where cases were drawn solely from OIS databases or from public media sources. Past research has generally concluded that these cases are highly dangerous and typically involve the use of lethal force (Flynn & Homant, 2000). The broader perspective from the present analysis offers new insights. The similarities in demographic characteristics increase the level of confidence that the present study is talking about the same phenomenon that has been captured in the previous SbC literature; however, the differences are telling as well. What is arguably the most important take away from these findings is that the chances of these situations being resolved peacefully are probably greater than previously believed. Given that, of the 419 cases analyzed, only 3% of subjects were injured, 1% were killed, and one officer was injured; this represents a stark contrast to prior studies, like Mohandie et al. (2009), where the rates of subject death have been reported as high as 51%, with 40% of subjects and 16% of officers injured. In one of the few other studies where SbC cases involving less lethal force were included, Homant and Kennedy (2000a) also found lower rates of lethality. However, the rates from the current sample are still far lower than in this study as well. While the present results clearly need to be replicated, this first look from a wider lens provides more hope for successful resolution of SbC cases than seen in previous studies.
In considering these findings, it is important to note that the results are derived from one sample from one department that has a specialized team at the disposal of officers who need assistance when there is a possible mental health concern during a call. It is possible that the low level of lethality found in the present study could be particular to the LAPD MEU due to the extensive mental health training provided for their officers. Along these lines, it appears that Homant and Kennedy’s (2000b) typology may be a useful way to conceptualize and classify these cases. If the majority of SbC cases are disturbed interventions, as seen in the present sample and in Homant and Kennedy’s (2000b) original sample, then it could be beneficial to provide resources and training for officers to learn how to handle these cases in the most effective ways possible. At present, the most promising mode of accomplishing that is through the use of CIT (i.e., the MEU). Studies on CIT programs have found that they are having a positive impact and influence on how officers resolve calls involving persons with mental illness (Lord & Bjerregoord, 2014), and particularly, with improving officers’ attitudes toward and effectiveness in interactions with individuals with mental illness (Watson, Compton, & Draine, 2017). In addition, compared with their non-CIT-trained peers, CIT officers have a higher likelihood of redirecting a person with mental illness to a mental health intervention instead of a criminal justice one (Watson et al., 2010). While further research may help to clarify whether these findings are particular to this department, a red flag should be raised for departments who do not have such extensive training but could still be dealing with the same kinds of calls. For those departments that do not have these resources, a push to develop them is recommended.
Finally, in terms of differences seen across cases in levels of force used, it is encouraging to see that demographic and mental health factors were not associated with the use of lethal force; instead, we see higher levels of force occurring with the presence of aggressive and threatening behavior. These are the behaviors that would logically be expected to evoke a continuum of force, thus it appears that the police officers involved were in line with typical use of force protocol. However, in order to further study this phenomenon, it is important for police agencies across the country to adopt a formal method to identify and track SbC incidents. The cases in this study were only analyzed because the LAPD tracks them in a standardized fashion. The MEU database was created in order to track these incidents as well as other important variables so that they can be studied, used to implement changes, and used to develop and conduct trainings to help officers and support staff be as well equipped to handle these situations as possible. Other departments are encouraged to follow suit, as it is evident that much can be learned from studying what does and does not work in these types of incidents.
Limitation and Future Directions
Of course, as in all studies of SbC, the analysis is only as good as the information from the reports and the way in which it is coded. In retrospect, there are methodological weaknesses that should be considered. The codebook used by Mohandie et al. (2009) and adapted for the present study was useful, but adaptation was difficult in that the data sources for the studies were quite different. The codebook offered mostly categorical data, which made it impossible to conduct more advanced statistical analyses. As well, because each variable was coded independently of other variables, it was impossible to present or analyze data for combinations of variables that may have been of interest. Future studies looking into similar variables should keep these considerations in mind when creating and defining variables.
In terms of the data collection, there was only one coder for the data, thus there was no way to assess for interrater reliability. While the majority of the data was fairly objective, anytime there is but one coder for a large volume of data, the opportunity for error occurs and it would have been ideal to have a portion of the reports coded by a second rater to assure consistency. In addition, the database is a newer creation for the department, thus imperfections can be expected and for this study, 114 of the 533 cases that were initially analyzed had to be excluded. This large number of inaccurately coded or irrelevant reports leads to some degree of caution in terms of knowing whether all relevant reports were captured and all irrelevant reports were excluded. Finally, there were a large percentage of “unknown” responses to many of the demographic and mental health variables due to variability of available information and variability in report writing from officers. Although the variables are still important to consider, statistical analyses were limited by a large number of unknown responses.
There is still much to be discovered about SbC. Additional studies that utilize similar databases are crucial to expanding the knowledge in this area. While this study relied on the use of extensive data from a large number of cases, these cases were drawn from one department. Having similar tracking systems in other departments would provide the means for analysis of data from departments with a wider range size and geographic locations. Currently, aside from a few notable national databases (i.e., the Hostage Barricade Data System maintained by the Federal Bureau of Investigation; National Violent Death Reporting System maintained by the Center for Disease Control) that have been used in prior research (Lord, 2012, 2014; Lord & Sloop, 2010), there is little knowledge or awareness of this type of tracking taking place within individual departments across the country. Future studies of SbC cases that use internal reports that gather information about the full breadth of incidents may allow for additional work and insight about how these cases play out, particularly in situations where lethal force is not used and a violent death does not occur. Furthermore, such databases would be exceptionally useful if they tracked information about more nuanced issues, such as the opioid crisis and veterans’ issues, that may potentially come into play with SbC cases and which have not yet been explored in research on this topic.
Studies that compare SbC incidents from departments that do, and do not, have CIT training or specific MEUs would be beneficial to determine whether differences in outcomes are seen when this type of resource is in place. In addition, research like that of Best, Quigley, and Bailey (2004) and Lord and Sloop (2010) that aims to develop a system or tool that can be used for better screening and classification of SbC calls may help to better prepare responding officers to deal with these calls and to, ideally, increase the likelihood of a mental health intervention over a criminal justice one, when appropriate. The better trained and more well-equipped officers are to handle such calls, the higher the likelihood of the subject to not only survive, but also have their mental health concerns addressed in the appropriate setting. Finally, it is recommended that research involving a survey of officer perceptions of SbC encounters in conducted. Such as study would be useful to help determine whether officers believe these cases are possible to resolve without use of force and whether these perceptions about the likelihood of successful resolutions are in line with what is currently known about these cases.
Overall, the present findings show that when we take a broader look at SbC cases and go beyond just OIS cases, there are similarities across the subjects that are involved in these incidents, but differences in the outcomes. There are more involuntary hospitalizations and fewer fatalities and injuries to subjects and officers. Moreover, lethal force is seen in cases where such force would typically be called for due to the threatening behavior of the subjects. This new, broader, perspective appears to provide a more optimistic look at the SbC phenomenon and the ability of officers to manage these situations, particularly when they have the support of an MEU-type team or CIT training to aid in managing these intense and often unpredictable situations.
Footnotes
Authors’ Note
The results of this research were presented at the annual conference of the Society for Police and Criminal Psychology in September 2017 in San Diego, CA. This research would not have been possible without the support of the Los Angeles Police Department, and in particular, Kimberly Telesh, PhD; Charles Lennon, LCSW; Lieutenant Brian Bixler; and Scott Musgrove, PsyD.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.