Exploring Anti-LGBT Homicide by Mode of Victim Selection

Abstract

This study extends previous research by comparatively examining variations in situational circumstances of bias homicide targeting lesbian, gay, bisexual, and transgender (LGBT) victims. Rather than conceptualize categories of homicide by offender motive, this study disaggregates anti-LGBT homicide based on observable characteristics of victim selection by offenders. Anti-LGBT homicides are conceptualized as predatory and responsive offenses, which capture planned, unprovoked homicides and unplanned crimes in which victims play a role in the escalation of violence. Further situational distinctions are conceptualized within these categories of anti-LGBT homicide, and comparative analyses reveal important differences across these situational circumstances. Considering the large amount of variation found within anti-LGBT homicide, which often goes unrecognized by studies relying on official bias crimes data, these findings suggest that open-source crime data and nuanced examinations of anti-LGBT homicide may be useful tools for fashioning informed policy discussions of victim-based legal protections and law enforcement responses to bias crime.

Keywords

homicide bias crime hate crime LGBT victims masculinity

The United States is in the midst of a transitional period in which the social positioning of the lesbian, gay, bisexual, and transgender (LGBT) community continues to elevate. For example, the General Social Survey showed that while nearly 30% of Americans supported same-sex marriage in 2004, almost 40% did so in 2008 (Powell, Bolzendahl, Geist, & Steelman, 2010), and more recent polls show that 55% of Americans support same-sex marriage (Richey, 2014). Other indicators of change include the repeal of the “Don’t Ask, Don’t Tell” policy held by the United States military (Stolberg, 2010), and the increased acceptance of openly gay recruits into the military (Harris, 2010). The relatively recent passing of bias crime legislation, often referred to as the Matthew Shepard Act (2009), which added sexual orientation and gender identity victims to a growing list of other federally protected bias crime groups, is another indicator of change. The signing of this legislation followed decades of failed attempts by social movement organizations that have pushed for the inclusion of victims targeted because of their sexual orientation and gender identity (see Jenness, 1995; Jenness & Broad, 1997).

Yet, acceptance of the LGBT community has not been universal and violence against the LGBT community persists as a serious issue in the United States. Scholars have suggested that discriminatory violence against LGBT victims (and other forms of bias crime) can have uniquely harmful psychological effects on victims and the LGBT community more broadly (Dunbar, 2006; Garnets, Herek, & Levy, 1990; Herek & Berrill, 1992; Herek, Gillis, Cogan, & Glunt, 1997; Iganski, 2008; Lawrence, 1999; Rose & Mechanic, 2002). In addition to having psychological effects, bias-motivated attacks, such as those against LGBT victims, may be more likely to be followed by similar acts of discriminatory violence, thus resulting in further community disruptions and avoidable future acts of violence (Herek & Berrill, 1992; Iganski, 2008). The Anti-Violence Project recently found that there has been a trend of increasing homicides against the LGBT¹ community since 2007 (National Coalition of Anti-Violence Programs [NCAVP], 2012). Despite decreases in anti-LGBT violence generally, the NCAVP recorded 30 anti-LGBT murders in 2011, the highest homicide number ever reported (NCAVP, 2012).²

Thus far, our understanding of bias crime has been largely shaped by a general typology centered on differences in offender motives. In particular, criminologists have suggested that bias crimes are driven by thrill-seeking, retaliation, the need to defend territory, and the need to rid the world of certain social minority groups (Levin & McDevitt, 1993; McDevitt, Levin, & Bennett, 2002). This prominent typology has advanced our understanding of bias crime in the United States by reducing a variety of complex motivational circumstances into four categories. Nonetheless, there remain several issues with relying solely on this general approach to typifying bias crime. Two of those issues are discussed here and a more complete discussion follows below.

The first issue involves categorizing bias crime based on offender motives. Prior bias crime typologies, like crime typologies in general, have been used to organize ranges of criminal behaviors into sets of more meaningful categories (Dabney, 2004), and research on homicide motives has led the field in producing such typologies (e.g., Block & Block, 1992; Decker, 1993). Determining the actual motives of bias crime offenders, however, has been shown to be challenging even under ideal circumstances (see Boyd, Berk, & Hamner, 1996; Haider-Markel, 2002; Nolan & Akiyama, 1999). One way to expand our understanding of the different circumstances under which these crimes occur is to conduct focused research on specific subtypes of bias crime. Second, Levin and McDevitt’s (1993) general bias crime typology does not explicitly discern between crime types (e.g., violent and nonviolent) or victim groups (e.g., LGBT, racial/ethnic, religious, etc.), thus implicitly assuming homogeneity across forms of bias crime. Consequently, little is known about the specific circumstances under which specific types of bias crime, such as anti-LGBT homicide, are initiated and performed.

One study to date has examined anti-LGBT homicide as a unique form of bias crime based on offender mode of victim selection. Gruenewald (2012) compared characteristics of anti-LGBT homicides with a random sample of more than 600 homicides from the 2000 Uniform Crime Reports–Supplementary Homicide Report (SHR). This comparison sample was designed to represent a snapshot of the typical homicide for purposes of comparison. Although this sample could have included bias crime, it is statistically unlikely (Gruenewald, 2012). He found that anti-LGBT homicide offenders and situational characteristics differ in important ways from typical homicides. However, Gruenewald (2012) did not examine variation within anti-LGBT homicide or the similarities and differences in these homicide events based on how offenders selected their victims.

Therefore, the overall purpose of the current study is to advance knowledge about the comparative nature of anti-LGBT homicide across different modes of LGBT victim selection. Rather than focus primarily on bias offender motives, this study draws from Lawrence’s “discriminatory selection model” to categorize anti-LGBT homicide by offenders’ observable victim selection behaviors (Lawrence, 1999). While offenders may not always premeditatedly choose their victims, all homicides in the current study involve offenders who made deadly decisions to violently attack LGBT individuals in such a way to lead to the death of victims. In this study, anti-LGBT homicide refers to fatal acts of criminal violence in which victims were targeted in whole or in part because of their actual or perceived sexual orientation or gender identity. Sexual orientation refers to sexual preferences (e.g., bisexual, homosexual), while gender identity refers to how victims identify with a particular gender category (i.e., transgender). Anti-LGBT homicide data come from an open-source database known as the Extremist Crime Database (ECDB), which includes all known anti-LGBT homicides in the United States from 1990 to 2010 (see Gruenewald, 2012). Based on prior research and a systematic analysis of homicide “storylines” (see Agnew, 2006), anti-LGBT homicides are disaggregated into five categories defined by offender mode of victim selection and comparatively examined. Two questions will guide this research. First, how can anti-LGBT homicides be distinguished into subtypes that increase our understanding of these crimes? Second, how does the nature of anti-LGBT homicide events vary across the unique processes offenders use to select LGBT victims?

How scholars and law enforcement practitioners understand the nature of bias crime in the United States has been largely shaped by the four-pronged bias crime typology developed by Levin and McDevitt (1993; see also McDevitt et al., 2002). For this reason, we review the typology below. We also review two studies that demonstrate the general typology’s ability to capture nuances of bias crime offender motives. We then discuss the limited research on anti-LGBT homicide before addressing how the extant literature has neglected to adequately consider the different ways that anti-LGBT homicide offenders select their victims and the similarities and differences in modes of victim selection.

A Typology of Bias Crime

Levin and McDevitt (1993) have offered a general bias crime typology of offender motives based on interviews with bias crime victims, offenders, and police officials from the Community Disorders Unit of the Boston Police Department. Their original typology comprised three motivational categories that included thrill, defensive, and mission crimes. Later, after reanalyzing their data, McDevitt et al. (2002) produced a fourth category labeled retaliatory bias crime.³ Their typology has played a prominent role in how we understand bias crime and has been commonly integrated into manuals and other materials used to educate law enforcement and others (e.g., Department of Justice, 1998, n.d.). Yet, scholars have critiqued this bias crime typology on a number of counts. In one study, Phillips (2009) exposed possible limitations of their work by examining prosecuted bias crime from one New Jersey county. Applying the general typology to specific bias crimes, she found that more than one third of the analyzed cases could not be classified within the four designated motivational types, showing that the typology was not exhaustive. These bias crimes deemed unclassifiable by Phillips (2009) often included cases in which animus toward a social group was not the only motivating factor. Phillips (2009) also found that McDevitt et al.’s (2002) motive categories were not necessarily mutually exclusive and that there may be more room for interpretation than previously thought. Indeed, not being able to place offenders in discrete categories remains a problem. In another study, Fisher and Salfati (2009) empirically examined the McDevitt et al. (2002) typology by using a nonmetric multidimensional scaling technique known as Smallest Space Analysis (SSA) to test whether bias crime cases could be separated into designated categories (see also Guttman, 1968). This technique draws from Facet Theory and allows for the visualization of the similarities in homicide cases by identifying variables most likely to simultaneously occur. The characteristics closest to the center of the visual map produced by the analysis are more related to a specific form of criminal behavior and occur more frequently. They based their analysis on 91 bias homicides reported to the FBI’s Uniform Crime Report. Like Phillips (2009), they concluded that the motive-based categories were not mutually exclusive and that McDevitt et al. (2002) failed to clearly differentiate four distinct bias crime categories. In fact, the only clear division appeared between thrill-related and retaliatory-related crime variables, demonstrating how bias crime may be best differentiated by offenders who seek a sense of power and others who are committed to restoring lost honor. Consequently, Fisher and Salfati (2009) questioned the usefulness of creating typologies based on offender motivation alone.

Anti-LGBT Homicide

In one more recent study, Gruenewald (2012) integrates data from the open-source ECDB (Freilich, Chermak, Belli, Gruenewald, & Parkin, 2014) and the FBI’s Uniform Crime Reports-SHRs to comparatively examine anti-LGBT homicide with a random sample of homicides that more commonly occurred in the United States between 1990 and 2008. This study found that this particular type of bias homicide differed from more routine forms of homicide in regard to the weapons used, number of offenders, number of victims, and offenders’ demographics. Gruenewald (2012) concluded that future research on anti-LGBT homicide, as a unique form of bias crime, was needed to gain a more comprehensive understanding of this complex form of violence.

The only other empirical research known to the authors that specifically focuses on anti-LGBT homicide has been Stephen Tomsen’s (2002, 2006, 2009, Tomsen & Mason, 2001) study of antihomosexual homicide in New South Wales, Australia, which is based on interviews and a review of archival materials. Whereas Gruenewald (2012) failed to closely examine the situational variations in anti-LGBT homicide, Tomsen’s research recognized the heterogeneous nature of anti-LGBT fatal attacks by uncovering two general scenarios of antihomosexual lethal violence. The first scenario consisted of attacks between people, usually men, that occurred in a public space and were often “marked by a tone of outrage” (Tomsen, 2009, p. 66), whereas the second violent scenario was more confrontational in nature and typically occurred in private. Interestingly, Tomsen recognized that robbery often held an “incidental relation” to bias attacks, but occasionally, robbery was the principal motive of an antihomosexual killing (2009, p. 67).

Limitations of Prior Research

There are notable limitations of the few studies addressing the typification of bias crime and anti-LGBT homicide. In particular, with the exception of Gruenewald (2012), previous research on bias crime has relied primarily on official data, despite research showing that many bias crimes go unrecorded by police agencies due to discrepancies across states (Nolan & Akiyama, 1999; Perry, 2001), as well as differences across various practices between law enforcement agencies (Boyd et al., 1996; Cronin, McDevitt, Farrell, & Nolan, 2007; Haider-Markel, 2002; McDevitt et al., 2000; Nolan & Akiyama, 1999; Walker & Katz, 1995). Another limitation is that prior research on the categorization of bias crime (Fisher & Salfati, 2009; McDevitt et al., 2002; Phillips, 2009) has considered all bias crime without disaggregating by victim type. Relying on data that encompass all bias crime victim groups may lead to oversights regarding potential differences between unique target groups like LGBT victims. Most previous studies, with the exception of Fisher and Salfati (2009), have also lumped together all types of bias crime rather than focusing on a single type of violence. Doing so may erroneously assume homogeneity across disparate forms of lethal and nonlethal violence and may result in overgeneralizations regarding the relative nature of anti-LGBT violent offender behavior. Finally, bias crime has thus far been primarily categorized by differences in offender motives. Determining actual motives of bias crime offenders, however, remains challenging (see Boyd et al., 1996; Haider-Markel, 2002; Nolan & Akiyama, 1999). Scholars have yet to examine other schemes for categorizing bias crime based on the observable behaviors of offenders.

The Current Study

The current study extends prior research on bias crime in a number of ways. First, this study avoids relying on national crime data sources, which have been shown to be largely unrepresentative of the extent and nature of bias crime (Nolan & Akiyama, 1999; see also Sandholtz, Langton, & Planty, 2013). Instead, the current study relies on an event-based, open-source database known as the ECDB (see Freilich et al., 2014). By relying on concrete inclusion criteria based on commonly accepted indicators of bias that capture observable offender behavior, the ECDB seeks to capture a more representative set of anti-LGBT homicides occurring in the United States.

Second, the current study avoids overlooking differences across violent bias crime types targeting disparate victim groups by examining a single form of violence, criminal homicide,⁴ and a single protected group, LGBT victims. Drawing from criminological research on the disaggregation of homicides (Flewelling & Williams, 1999), this study avoids making the assumption that the nature of bias homicide is homogeneous for all crime types and victim groups. Instead, substantive differences in patterns of bias homicide are expected which are conditional to crime types and the victim groups that are targeted (see Stacey, 2011).

Third, this study extends prior research by placing focus on one particular dimension of crime, offender mode of victim selection, to classify anti-LGBT homicide events. Miethe, McCorkle, and Listwan (2005) have argued that victim selection is a major dimension of the criminal event. Moreover, one study on racial bias assaults in the United States suggested that “[b]igotry may serve as a factor in the selection of the particular victim rather than as the catalyst to the criminal act” (Messner, McHugh, & Felson, 2004, p. 608). In other words, if offender bias is an important factor in the selection of a victim, but is not a discernible motive for the crime, it is imperative to examine how anti-LGBT homicide may be disaggregated by offender selection processes, rather than motive. Drawing from Lawrence’s (1999) “discriminatory selection model,” the current study’s conceptualization of anti-LGBT bias crime is based on how, and not why, offenders discriminately select victims due to their membership in a particular social group or category. Lawrence (1999) argued that the discriminatory selection model is a superior way to classify bias crime compared with the “group animus” model, which suggests that offenders’ bigotry squarely drives bias-motivated violence. By relying on the discriminatory selection model, it is possible to avoid the impossible task of reading offenders’ minds and evaluating their feelings prior to their crimes. Instead, it is necessary to identify only observable indicators of biased victim selection by offenders, or indicators that demonstrate how offenders selected victims based on their perceived social minority status.

Just as it is not possible to gauge offenders’ level of animus toward victims, it is also not possible to discern in all cases if offenders necessarily intended to kill victims, as opposed to only injuring them. Intentions to kill, as opposed to intentions only to injure victims, remain largely unobservable. Therefore, the homicides included in the current study represent both premeditated homicides and unplanned, spontaneous violence that escalates into homicide. Furthermore, elements of premeditation and spontaneity are discussed below as key characteristics that vary across anti-LGBT homicide categories.

Method

The current study relies on anti-LGBT homicide⁵ data from the ECDB, which includes event-level data on several types of extremist crime occurring in the United States (Freilich et al., 2014).⁶ In this way, each anti-LGBT homicide is conceptualized as a multidimensional, dynamic event involving the interactions of victims and offenders within specified temporal and situational settings (see Sacco & Kennedy, 2002). The ECDB identifies anti-LGBT homicide from open-sources, including advocacy group reports, academic chronologies, and systematic media searches. The process of systematically identifying extremist homicides from open-sources has been detailed elsewhere (Freilich et al., 2014; Gruenewald, 2012)⁷ and will be briefly discussed here as well. The ECDB includes anti-LGBT homicides that are officially charged as bias crime as well as those that are not. Including all homicide events that were eligible to be classified as bias crime based on available bias indicators allows for a more representative selection of anti-LGBT homicide. Concrete inclusion criteria are relied upon to guide the process of identifying anti-LGBT homicide events from multiple sources (Gruenewald, 2012). The inclusion criteria consist of a list of observable bias indicators used to designate crimes as involving bias motivational circumstances. Bias indicators used by the ECDB were directly borrowed from materials used to guide police officers in the identification and investigation of bias crime (e.g., Department of Justice, 1998, n.d.), while others were modified to specifically address anti-LGBT violence. Anti-LGBT homicide bias indicators include (a) gender-based derogatory remarks and verbal harassment made by the offender toward the victim, (b) symbolic (sexualized) manipulation of the victim’s body as described in open-source materials, (c) symbolic location of the body (e.g., “a gay bar” or “gay cruising area”), (d) selection of a victim through a LGBT organization as described in open-source materials, (e) an official bias crime charge, (f) offender admission, and (g) prior anti-LGBT violence perpetrated by the offender against the victim or someone else in the hours, or even days, leading up to the homicide (i.e., “spree crimes”). Once identified as a potential bias homicide in open-sources, a single indicator was required to be present for the case to be included in the study. In other words, cases were included in the current study only if police, witnesses, advocacy groups, media, or others labeled them as bias crimes, and an observable behavioral indicator of discriminatory selection was evident in available materials. We chose to exclude homicides if the identity of an offender was unknown even if such an indicator of bias was present. Absent a known offender, it would not be possible to reasonably categorize a homicide based on observable evidence of offenders’ discriminatory selection decision-making processes. Because there is a large and unknown quantity of homicides targeting LGBT victims that involve unknown suspects, some may feel that the inclusion criteria are too conservative. Paradoxically, others may suggest that the stated inclusion criteria are too liberal because not all suspects were legally charged with committing a hate crime. We chose to balance these conflicting perspectives by including anti-LGBT homicides committed by offenders who were and were not charged with committing a “hate crime” while also excluding cases for which the discriminatory selection processes of known offenders could not be determined from available information. Once identified, included anti-LGBT homicides were open-source searched in nearly 30 web-based search engines by ECDB researchers and all relevant information was collected and organized by case (Freilich et al., 2014).

The total number of anti-LGBT homicides examined in the current study is 121.⁸ A number of offender-, victim-, and incident-level homicide characteristics are included in the current study to evaluate the comparative nature of anti-LGBT homicide. Relying on these variables discussed below, the current study comparatively examines the nature of all anti-LGBT homicides across various categories representing victim selection modes. These categories are discussed after detailing variable measurement schemes.

Anti-LGBT Homicide Offender and Victim Variables

In this study, only one offender for every homicide was considered.⁹ Offender demographic characteristics were measured with two binary-coded variables, race and age. Offender race (White) was measured as White (1) and non-White (0).¹⁰ In addition, offender age (juvenile) was measured as under the age of 18 (1) and 18 years of age and above (0) at the time of the crime. The sex of offenders was not included in analyses as all anti-LGBT offenders included in this study were male. Finally, the presence of evidence that offenders were under the influence of drugs or alcohol (alcohol/drug use) directly prior to the homicide was measured as “yes” (1) or “no” (0).¹¹

Like offenders, only one victim for every homicide was considered in the subsequent analyses. This should not influence victim-level findings as so few homicides involved multiple victim deaths (7%) and because there is no reason to expect that secondary victims should significantly vary on the variables of interest. The sex of victims (male) was measured as male (1) and female (0).¹² Victim race (White) was measured as White (1) and non-White (0). Victim age (juvenile) was measured as under the age of 18 (1) and 18 years of age and above (0) at the time of their death.

Anti-LGBT Homicide Incident Variables

Six binary-coded incident-level variables were included in this study to capture multiple dimensions of anti-LGBT homicide situations. First, where a homicide event occurred was of interest to the current study. In particular, whether a homicide occurred in a private residence (occurred in residence) was measured as yes (1) or no (0). Second, as Gruenewald (2012) found that anti-LGBT homicides were more likely to involve alternative weapons to firearms than a random sample of U.S. homicides, the use of alternative weapons (nonfirearm) was measured as yes (1) or no (0) to compare weapon use across victim selection categories. Third, Gruenewald (2012) also found that anti-LGBT homicides were not significantly more likely to involve victims and offenders unknown to one another. To examine how victim–offender relationships operated across victim selection categories, a victim–offender relationship variable (stranger) was included and measured as yes (1) or no (0). Fourth and fifth, this study was also concerned with who was present at the homicide event. Therefore, the presence of multiple offenders and multiple victims were measured (1 = yes, 0 = no). Sixth, the presence of bystanders (bystanders present) is potentially important in distinguishing between victim selection categories and was measured as yes (1) or no (0).

Two variables were also included to capture the dynamics of each violent transaction, which can reveal offender-based behavioral patterns across varying victim selection modes.¹³ First, to examine the role of verbal insults in the escalation of anti-LGBT homicides, the use of antigay slurs or other gender-based remarks just prior to the homicide was measured as yes (1) or no (0). Second, whether the homicide involved any sort of theft (profit-related circumstances) from the victim was measured as yes (1) or no (0).

An additional four variables were included in the analyses below to capture offenders’ behavior in the aftermaths of fatal anti-LGBT attacks. The first variable included to describe the aftermaths of these attacks was the manipulation of victims’ bodies by offenders, which was measured as yes (1) or no (0). Manipulation of victims may include sexualized or provocative posing postmortem, or the symbolic undressing or changing of victim attire. Second, offender mutilation of victims’ bodies (yes = 1, no = 0) was also included to capture whether or not offenders utilized weaponry to symbolically violate victims’ bodies after their death. Examples include severed body parts and symbols etched into the skin of victims. Finally, this study captured whether offenders revealed (offender revelation) homicide circumstances to others (yes = 1, no = 0) as well as whether offenders admitted that victims were targeted because of their sexual orientation or gender identity (selected admission of motive; yes = 1, no = 0).

Disaggregating Anti-LGBT Homicide

In this study, anti-LGBT homicide was disaggregated based on offender modes of victim selection. Subtypes of anti-LGBT homicide were established based on prior research and a systematic review of open-source materials for all 121 anti-LGBT homicides. We first drew from studies of homicide typologies (Fisher & Salfati, 2009) and anti-LGBT homicide (Tomsen, 2009). Fisher and Salfati (2009) claimed that

a model of hate homicide emerges that may be focused more on the role that the victim serves for the offender. That is, in some hate homicides the offender may have chosen victims who were easy targets, which allowed them to attain a sense of power/control or general excitement/thrill, and in other hate-motivated homicides a particular individual may have been targeted because offending against that individual allowed the offender to take back honor. (p. 130)

In addition, Tomsen (2009) suggested that “different forms of violence with homosexual victims, such as collective stranger violence or more private assaults occurring between male acquaintances and intimates, necessitate recognition of the distinct dynamics of these crimes” (p. 57). The current study extends these findings by focusing on two broadly defined situational circumstances of anti-LGBT bias homicide offending that distinguish between homicide events in which offenders appear to target LGBT victims to seek power and those in which offenders appear to desire to restore honor. We refer to these types of anti-LGBT homicide as predatory homicide and responsive homicide, respectively. Second, we systematically reviewed each anti-LGBT homicide case and produced accounts of offenders’ victim selection decision-making processes for each homicide. This involved constructing narratives or “storylines” of each fatal attack, including precursor attributes, violent transaction attributes, and homicide aftermath characteristics (Agnew, 2006; examples of homicide storylines can be found in the appendix). Dabney (2004) has suggested that scholars seeking to categorize crimes are best served to conceptualize crimes as “criminal events” (p. 6). Accordingly, the purpose of creating anti-LGBT homicide storylines in this study was to capture the processes of victim selection within the context of the criminal event.

Included in anti-LGBT homicide storylines, precursor attributes consisted of offenders’ behavior prior to violent interactions with victims. Commonly coded details included the use of alcohol by offenders and incriminating statements made by offenders prior to the attacks. Transaction attributes included verbal and physical interactions between victims, offenders, and bystanders, as well as other witnesses. The weapons used, who was involved, and the progressions of violence were described. Finally, the events following the attacks were captured. This often included details of how victims’ bodies were left following attacks, how victims were eventually found, and the behaviors of offenders following the anti-LGBT attacks.

In the majority of cases, a clear account of how offenders selected their victims, as well as how they carried out their crimes, was ascertainable from open-source materials. Some cases were eventually removed from the study because circumstances around victim selection could not be categorized due to insufficient available information from open-source materials. In other cases for which conflicting accounts of victim selection motives were provided, traditionally “trusted” sources of open-source information were relied upon (i.e., court documents, police records). The two categories of anti-LGBT homicide and associated subcategories are described in more detail below.

Predatory Homicide

This first broadly defined category of offender victim selection is predatory homicide, which consists of planned acts of violence against members of the LGBT community. The nature of planning involved prior to an attack has been shown to be an important distinguishing element of criminal event types (Miethe et al., 2005). Predatory anti-LGBT homicide offenders took time to orchestrate a plan for perpetrating the homicide prior to encountering victims, typically during the precursor phase of the criminal event. In cases of predatory anti-LGBT homicide, victims did not provoke offenders. In fact, victims often had little or no direct contact with the offenders prior to their initial interaction. In such cases, victims played no part in offenders’ decisions to target them for violence. From open-source data, two common profiles of predatory anti-LGBT homicide situations were identified, including “representative offenses” and “instrumental offenses.” The key distinction between these situational variants was whether or not offenders selected LGBT victims to murder as representatives of the LGBT community or, on the contrary, to rob victims. These two variants of anti-LGBT predatory violence are discussed below.

Representative offenses

Predatory representative offenders appear to select victims whose deaths will communicate symbolic messages regarding the social standing of the LGBT community. Selection of representative victims usually occurred in one of two scenarios. One scenario involved offenders who had intimate knowledge about victims’ activities and routines, resulting in the tracking down or stalking of their victims. Offenders traveled to victims’ residences or other locations where LGBT victims were known to frequent. In the second scenario of predatory representative attacks, offenders lured victims to remote locations to ensure privacy to commit the homicide. For example, victims often arranged meetings through Internet chat rooms and other electronic LGBT relationship services. Many of these homicides involved elements that suggest offenders were “thrill seeking.” For example, some attacks began with groups of young men who sought out vulnerable LGBT victims for no apparent reason, suggesting that it was simply for the thrill of the hunt and the attack (see Levin & McDevitt, 1993). Regardless of the particular situational variant, all predatory representative homicide offenders remained unprovoked. It is likely that these offenders selected victims to express animosity toward particular LGBT victims and to send symbolic messages of the dangers associated with being a member of the LGBT community (see Perry, 2001). Considering the nature of the data, it is not possible to prove that each representative offender was attempting to send a symbolic message; however, available evidence indicating that offenders sought out these victims without having first been confronted suggests that offenders desired to express their own animosity toward members of the LGBT community (see Perry, 2001). Moreover, it is probable the offenders also sought to share this message with a broader community, including LGBT individuals and the general population, to make a larger social statement.

Instrumental offenses

Instrumental homicides were the second subcategory of predatory anti-LGBT homicide identified in this study. In a predatory instrumental attack, offenders selected victims based on their sexual orientation or gender identity primarily to rob them. Attacks were considered instrumental offenses because anti-LGBT homicide victims were selected as means to another end (i.e., robbery; see Block & Block, 1992; Meithe & Drass, 1999). The defining characteristic of anti-LGBT instrumental offenses was that crimes were primarily profit oriented. While police may be hesitant to investigate and classify instrumental bias offenses, which involve complex, mixed-motive circumstances, these events were considered bias crimes in the current study because LGBT victims of robbery homicides were discriminately selected based on sexual orientation or gender identity.

Again, LGBT victims were not necessarily targeted because of deep-seated “hatred” toward gays or lesbians. Motives for victim selection were often much more mundane (see also Tomsen & Mason, 2001). For instance, LGBT victims were often targeted because offenders viewed them as “easy prey.” In some cases, gay men were viewed as less likely to physically fight back during a robbery. Stereotypes of gay men as wealthy may have also influenced offenders’ decisions to target LGBT victims. These predatory killings took on different forms that varied by the relational distance between offenders and victims. Some homicides, for example, occurred during initial encounters between victims and offenders at a predetermined meeting time and place (usually under the guise of a sexual encounter). Other instrumental homicide offenders established short-term relationships with victims to earn their trust before robbing and murdering them. Although extremely rare, some offenders committed these sorts of violent crimes serially over long periods of time and across different parts of the country.

Responsive Homicide

The second overarching category of anti-LGBT homicide victim selection was referred to as responsive homicide. Responsive homicide refers to expressive violent acts with “little rational planning” (see also Block & Block, 1992, p. 65). In contrast to predatory anti-LGBT homicide offenders, responsive offenders typically did not take care to plan the offense prior to the homicide transaction. Anti-LGBT responsive homicide was also different from the predatory form of homicide as offenders selected victims in response to real or perceived affronts. Some scholars in the past have suggested that some forms of bias violence (i.e., racial assaults) are unique from conventional violence as the former usually comes in the form of unprovoked “bullying,” while the latter tends to be dispute related (Messner et al., 2004, p. 609). In contrast, the current study found that approximately half of LGBT homicide victims were targeted because of perceived offensive behavior by the victim. In this way, the victim played an integral role in these dispute-related transactions. The affront was usually perceived by the offender as a personal attack related to their sexual orientation or gender identity or, less commonly, as a general affront to traditional norms of sexual behavior. Discussed below are three situational variants in which anti-LGBT responsive homicide occurred.

Gay bash offenses

The first and most common type of responsive homicide scenario was the “gay bash” offense. In all cases of responsive gay bash homicide, offenders selected victims who offenders believed to have insulted or disrespected them in some way. Victims and offenders often exchanged insults during verbal confrontations. In some cases, offenders may be disrespected or threatened simply by the physical presence of a member of the LGBT community.¹⁴ Offenders, therefore, took it upon themselves to “punish” selected victims for their perceived wrongdoing (Perry, 2001). This is not to say that victims usually initiated the deadly transaction. More often, it was simply a reciprocal act on the part of victims, such as returning verbal insults to the offender. Again, gay bash homicide was unique from predatory representative homicide as gay bash attacks were not premeditated and they involved some form of victim provocation.

Undesired romantic or sexual advance offenses

The second most common situational variant of responsive anti-LGBT homicide was “undesired advance” homicide. These attacks occurred when offenders selected victims in response to a real or perceived sexual (or romantic) advance made by the victim toward the offender. Advances often occurred in intimate settings such as victims’ residences. In many instances, both victim and offender were under the influence of drugs or alcohol prior to the attack. Under these circumstances, victims and offenders often found themselves alone under the pretenses of consuming drugs or alcohol in private settings. At some point during the interaction, victims either erroneously concluded that offenders were interested in a physical relationship or, as in some cases, offenders changed their mind about participating in a physical relationship. Fatal attacks came either directly following perceived personal affronts or several days later when offenders returned to confront victims.¹⁵

Mistaken identity offenses

Much less common, the third type of anti-LGBT responsive homicide, “mistaken identity,” occurred following cases of offenders mistaking the sex of victims. In most situations of mistaken identity, a male offender participated in a sexual encounter with another anatomically male victim who was perceived to be female. It is likely that victims of responsive mistaken identity homicides were male-to-female transgendered prostitutes or other transwomen. The realization of victims’ anatomical sex occurred prior to, during, and after sexual encounters. Offenders who mistook the sex of victims held them responsible for withholding information regarding his or her sex. In these cases, offenders felt duped, which then turned quickly into humiliation and rage.

A Description of Anti-LGBT Homicide

Anti-LGBT homicides are distributed across victim selection categories relatively proportionately with the exception of mistaken identity attacks (4.1%). As shown in Table 1, the most common type of anti-LGBT homicide was the predatory representative homicide. All offenders included in the current study were male and nearly 70% of offenders were White. The mean age of offenders was approximately 25 years and ranged from 14 to 60 years old. The vast majority (95%) of victims’ sexes were coded as male, though some may have identified as female or transgender in regard to their gender identity. Of the victims with data available on race, approximately 57% of victims were White. Of the non-White victims for whom race was determinable, approximately 19% were Black and 17% were Hispanic. The remaining victims were coded as “Other race.”

Table 1:

Offender, Victim, and Situational Characteristics of All Anti-LGBT Homicides (N = 121)

	Percentage
Anti-LGBT homicide subcategories
Predatory–representative	28.1
Predatory–instrumental	22.3
Responsive–gay bash	23.1
Responsive–undesired advance	22.3
Responsive–mistaken identity	4.1
Offender characteristics
Offender White	69.7
Offender juvenile	7.8
Offender drug/alcohol use	14.9
Victim characteristics
Victim male	95.0
Victim White	56.5
Victim juvenile	9.3
Situational characteristics
Occurred in residence	41.2
Bystanders present	16.0
Nonfirearm	72.7
Stranger	30.7
Multiple offenders	42.0
Multiple victim deaths	6.6
Attack characteristics
Gender-based remarks	26.4
Profit-related circumstances	32.2
Aftermath characteristics
Offender manipulation of body	6.6
Offender mutilation	8.3
Offender revelation	15.7
Offender admission of motive	50.4

Note. LGBT = lesbian, gay, bisexual, and transgender.

The most common location for an anti-LGBT homicide was inside a residence (41.2%). As shown in Table 1, most offenders relied on weapons other than firearms to kill their victims, and usually committed homicide alone against single victims when bystanders were not present. Interestingly, most offenders knew their victims, even if only after a brief relationship, prior to the homicide. Fatal attacks involved evidence of use of gendered slurs about 26% of the time and profit motives were apparent in approximately 32% of the cases for which profit was not the primary motive (instrumental attacks). After the attacks, very few offenders manipulated or mutilated victims’ bodies, though it did occur. Finally, there was evidence that offenders revealed their crimes to others in nearly 16% of the anti-LGBT homicide cases, while more than 50% of offenders eventually admitted that they targeted victims because of their sexual orientation or gender identity.

Results

A bivariate comparative analysis was used below to more closely examine offender-, victim-, and incident-level variables across subcategories of predatory and responsive anti-LGBT homicide. Each homicide characteristic shown below in Table 2 was binary-coded and treated as an independent variable for purposes of comparison. For each independent variable, every homicide subcategory presented in Table 2 is compared with every other homicide subcategory. This resulted in a series of 2 × 2 comparisons for each of the homicide characteristics of interest. While chi-square tests for significance were used to comparatively examine anti-LGBT homicide subcategories across the different modes of victim selection, Fishers Exact Tests were used to make comparisons when then number of homicides in comparative categories was small. In Table 2, superscript letters corresponding to each of the homicide subcategories are used to denote statistically significant differences across the subcategories of anti-LGBT homicide. In regard to race, representative homicide offenders were proportionately more likely to be White, though differences across categories were not statistically significant. The anti-LGBT homicide category proportionately most likely to be committed by a juvenile was undesired advance offenses. In addition to being committed disproportionately by juveniles, undesired advance homicides were significantly more likely to involve offenders who were under the influence of drugs or alcohol (or both) directly prior to the fatal attacks.

Table 2:

Offender, Victim, and Situational Characteristics of Anti-LGBT Homicides by Mode of Victim Selection (N = 121)

	Predatory		Responsive
	Representative (n = 34)	Instrumental (Robbery) (n = 27)	Gay Bash (n = 28)	Undesired Advance (n = 27)
Offender characteristics
Offender White	80.6	69.6	61.9	73.7
Offender juvenile	0.00	3.8^u	8.3	22.2ⁱ
Offender drug/alcohol use	5.7^U	3.8^U	15.2^u	34.5^g,R,I
Victim characteristics
Victim male	91.4	100.0	87.9	100.0
Victim White	75.0^G	70.0^G	20.8^R,I	52.6
Victim juvenile	5.9	0.00	12.5^I	10.3^I
Situational characteristics
Occurred in residence	35.3^g,u	46.2^G	12.5^r,I,U	62.1^r,G
Bystanders present	2.9^U,G	0.00^U	41.9^R,U	17.2^G,r
Nonfirearm	77.1	69.2	75.8	72.4
Stranger	28.1^g,U	36.0^U	64.0^r,I,U	3.6^R,I,G
Multiple offenders	26.5^I,g	61.5^R,U	50.0^r	32.1^I
Multiple victim deaths	8.6	0.00	6.1	6.9
Attack characteristics
Gender-based remarks	22.9^G	7.7^G	57.6^R,I,U	10.3^G
Profit-related circumstances	17.1	100.0	9.1	20.7
Aftermath characteristics
Offender manipulation of body	8.6	7.7	15.2	3.4
Offender mutilation	8.6	11.5	9.1	6.9
Offender revelation	17.1	15.4	6.1	24.1
Offender admission of motive	60.0^G	65.4^G	12.1^U,R,I	62.1^G

Note. U = undesired advance; I = instrumental; G = gay bash; R = representative. Capital letters indicate that differences are significant at alpha level p ≤ .01, whereas lowercase letters indicate differences are significant at alpha level p ≤ .05. Responsive cases of mistaken identity are not included in the table due to the limited number of known cases (n = 5). LGBT = lesbian, gay, bisexual, and transgender.

Table 2 also presents findings for victims of anti-LGBT homicide. Like offenders, the vast majority of all victims were male and this did not significantly vary across victim selection categories. Predatory anti-LGBT homicides were generally more likely than responsive homicides to involve White victims. Gay bash victims were proportionately the least likely to be White, significantly less than all other offender categories, with the exception of undesired advance homicide offenses. Victims of predatory anti-LGBT homicides were also proportionately less likely to be juveniles. In fact, all predatory instrumental attacks were perpetrated against adults (and most often by adults).

The location of responsive anti-LGBT homicides varied significantly. Cases involving undesired advances were significantly more likely to occur inside the home and outside of the purview of others compared with all other victim selection categories. Gay bash homicide, on the contrary, occurred outside the home and usually in public locations (e.g., bars). This supports international research that has found that anti-LGBT homicides often occurred at locations that were favored by gay male clientele (Tomsen & Mason, 2001, p. 266). Further evidence that gay bash homicides were more public in nature was the finding that a significant proportion of cases involved one or more bystanders. In contrast, predatory forms of anti-LGBT violence rarely involved bystanders. These premeditated attacks, instead, were orchestrated to occur in secret and outside the purview of others.

As for the weapons used to commit anti-LGBT homicide, most attacks were committed with weapons other than a firearm regardless of the mode of victim selection. Weapons included knives, blunt objects, and other bodily weapons, and more than 30% of all cases involved the use of more than one type of weapon.¹⁶

Of all anti-LGBT homicides, approximately 30% of cases involved a victim unknown to the offender (see Table 1). While there were no significant differences across victim–offender relationships for predatory homicides, responsive anti-LGBT homicides varied significantly. Responsive gay bash homicides involved significantly more victims unknown to offenders. On the contrary, very few victims of undesired advances were strangers to their offenders. This is not surprising as a romantic or sexual advance on a stranger is much less likely than that of an acquaintance.

Approximately 42% of all anti-LGBT homicide events involved multiple offenders, as shown in Table 1. This supports prior research that has found group-based bias crime offending to be relatively common (Herek et al., 1997; National Gay and Lesbian Task Force, 1995). In this study, the homicide type least likely to involve more than one offender was the predatory representative offense. This is likely due to the solo nature of “pick-up crimes” or crimes committed under the guise of an anticipated sexual rendezvous. On the contrary, predatory instrumental offenses and responsive gay bash offenses were significantly more likely to be perpetrated by multiple offenders when compared with predatory representative offenses. Group offending increases the likelihood that predatory instrumental attacks against LGBT victims will result in an effective robbery. Offending with others also allows for the diffusion of blame across offenders (Levin & McDevitt, 1993). Moreover, Short (1997) has argued that the presence of others may fuel the escalation of violence, especially when alcohol is involved, as offenders may be more willing to demonstrate masculinity and bravado through brutal violent acts.

Relatively few (approximately 7%) anti-LGBT homicide events involved multiple victim deaths. Those cases that did involve multiple deaths usually involved offenders who targeted victims that were intimately known to one another. Predatory instrumental attacks were the only form of anti-LGBT homicide that involved no multiple victim attacks. One plausible explanation is the increased risk associated with targeting more than one robbery victim at a time.

The current study found that approximately a quarter of all anti-LGBT homicides involved the use of gender-based derogatory remarks (e.g., faggot, queer) toward the victim by the offender prior to and during fatal attacks. Responsive gay bash homicides were significantly more likely to include the use of inflammatory gender-based insults. On the contrary, predatory instrumental homicides were the least likely to include antigay slurs, significantly different from gay bash homicide.

It is not surprising that predatory instrumental homicides were significantly more likely to involve profit motives, as profit-seeking is the defining feature of this mode of victim selection. More interesting, though, is that less than 10% of responsive gay bash homicides included any sort of profit-oriented behavior. This supports the assertion that gay bash attacks primarily constitute responses or punishments for behaviors that violated norms of traditional sexual identity and behaviors.

The aftermaths of responsive homicides significantly varied in regard to offender admission of an anti-LGBT motive. Specifically, offenders committing attacks against victims making alleged sexual or romantic advances were significantly more likely to admit that they targeted victims because of their sexual orientation or gender identity compared with responsive gay bash offenders. This may be because offenders felt justified in their responses to unwanted advances. Of all anti-LGBT homicide offenders, 50.4% admitted that they targeted victims because of their sexual orientation or gender identity (see Table 1). Offenders who committed responsive gay bash homicides were significantly less likely to make such admissions and less likely to state that victims were selected because of their sexual orientation or gender identity.

Discussion

Flewelling and Williams (1999) suggested that “a key issue encountered by homicide researchers is how to best categorize homicides into significant and relatively homogeneous types, under the assumption that different types of homicide may have different patterns and causes” (p. 96). One of the primary purposes of this study was to advance knowledge about the comparative nature of anti-LGBT homicide across different modes of LGBT victim selection. We drew from Lawrence’s (1999) discriminatory selection model to disaggregate homicides based on how they behaviorally went about selecting particular victims. Drawing from prior research and homicide storylines, two broadly defined categories of anti-LGBT homicide, namely, predatory homicide offenses and responsive homicide offenses, were identified. The development of these categories and related subcategories extended prior literature by focusing on a single form of bias and a single crime type. In this way, the authors avoided making the questionable assumption that all violent crimes and all bias crimes follow the same patterns.

Homicide types were initially differentiated by whether or not attacks were planned and unprovoked or, on the contrary, occurred in response to a personal affront (neither planned nor unprovoked). One could argue that predatory homicides are committed primarily to exert power over victims. In addition to the thrill that comes with stalking and hunting victims, offenders may be seduced by the power that comes through the ultimate domination of the planned fatal attack. Moreover, responsive attacks may be committed as acts of “righteous slaughter” that serve to restore what is considered good and right and the honor lost through victims’ affronts against the offender (see Katz, 1988). It is through responsive acts of violence that offenders regain honor by “saving face” and demonstrating their masculinity (Bufkin, 1999; Perry, 2001). International research on anti-LGBT killings in New South Wales, Australia, by Tomsen and Mason (2001) provides support for the generalizability of the current study’s findings. Through an analysis of homicide scenarios, their research demonstrated two general scenarios of anti-LGBT murder, including, first, a “planned attack in a public place on a victim who is homosexual or presumed to be, and who is usually a complete stranger” and second, “a personal dispute between two men over sexual activity or an alleged sexual advance [that] led to fatal violence between parties who were friends or acquaintances” (pp. 265-266). Thus, anti-LGBT homicide events may be premeditated, unprovoked acts of violence (predatory) or these fatal attacks may be dispute-related (responsive). Tomsen and Mason’s (2001) research on scenarios of anti-LGBT violence also demonstrates that the motivational circumstances and modes of victim selection do not always support the typical profile of a bias crime, or crimes involving attacks by strangers in public places against symbolic and random victims (see also Tomsen, 2002).

A key purpose of this study was then to conduct quantitative comparisons across anti-LGBT homicide victim selection subcategories. The first type of predatory homicide was the representative offense. Predatory representative offenses align closest to the public’s shared understanding of the typical bias crime, as they consist of expressive forms of unprovoked violence. Representative attacks do not occur by happenstance or arise abruptly. Instead, offenders commit these attacks cautiously and most commonly in the absence of others. Offenders are more likely driven by bigotry and significantly less likely to be under the influence of drugs or alcohol and are less distracted by other motives, such as profit. Whether done for the thrill of the hunt or as part of a larger mission to eliminate the symbolic threat posed by the LGBT community (see Levin & McDevitt, 1993), these predatory offenses send a clear anti-LGBT lifestyle message.

The second type of predatory anti-LGBT homicide was the instrumental offense. Predatory instrumental homicides, which involved robbery of LGBT homicide victims as a primary motive, were unique from predatory representative homicides in regard to the number of suspects involved, as instrumental anti-LGBT homicides were significantly more likely to involve multiple offenders. The difference in offenders’ victim selection processes, which are not necessarily captured in these findings, should not be overlooked. Offenders who plan from the outset to rob LGBT victims likely represent a different type of predatory killer. As instrumental offenders are driven primarily by profit, the selection of victims arguably has less to do with animus toward members of the LGBT community and more with the strategic victimization of “easy prey.” For this reason, critics may question the inclusion of instrumental homicides in a study of anti-LGBT violence. Determining the relative roles of bias and profit-related circumstances in a single homicide event undoubtedly presents challenges. This study, however, suggests that the discriminatory selection of LGBT victims in instrumental offenses is just as harmful for victims and victim communities regardless of the added complexity of robbery motives. Moreover, findings from this study demonstrated that instrumental anti-LGBT homicides were in many ways similar to the ideologically motivated representative anti-LGBT homicide. Therefore, ignoring this subtype of homicide would sacrifice a more complete understanding of an important element of the anti-LGBT violence story.

The second category of anti-LGBT homicide was responsive anti-LGBT homicide, which differed from predatory homicide in important ways. Responsive gay bash homicides were the most common type of responsive homicide and tended to involve victims and offenders who were unknown to one another prior to the homicide. In many cases, offenders became acquainted with their victims only minutes prior to the homicide. Gay bash homicides occurred most often in public locations (often outdoors) and were significantly more likely than other types of anti-LGBT homicides to occur in the presence of other bystanders who were not directly involved in the attack. Indeed, it is likely that the presence of others escalated the seriousness of violent transactions as offenders attempted to save “face” and respond to perceived affronts by victims (Luckenbill, 1977). This study found that a key indicator of gay bash homicide was the excessive taunting and gender-based, verbal degradation by offenders toward victims (see also Comstock, 1991). This finding is consistent with prior theoretical accounts of anti-LGBT violence, which have suggested that offenders are able to regain a lost sense of honor and to demonstrate masculinity through the use of this language (Bufkin, 1999; Perry, 2001; Tomsen & Mason, 2001; Van Der Meer, 2003). Importantly, however, gay bash homicides were not routine acts of fatal violence in which offenders simply used common gay slurs. Instead, for a homicide to be categorized as a gay bash case, there must have been evidence that an offender’s use of gender-based derogatory remarks were clear indicators for why victims were selected.

The second most common type of responsive anti-LGBT homicide involved responses to undesired romantic or sexual advances. This type of anti-LGBT homicide was significantly more likely to be perpetrated by young males who were under the influence of drugs or alcohol, or both, in the “safety” of private residences. One explanation for this extreme reaction to undesired advances by young males was their need to demonstrate masculinity once it had been threatened (Bufkin, 1999; Perry, 2001). Indeed, for some young males, violence is the only available response to a real or perceived threat to one’s masculinity (see Messerschmidt, 2012). Also important, the victims who were allegedly responsible for the unwanted advances were not significantly more likely to be under the age of 18. Many sexual advance cases involved adult victims and juvenile offenders. In many of these cases, sexual advances stemmed from situations involving the provision of drugs and alcohol by adult victims to juvenile offenders within the privacy of victims’ homes. Youthful offenders often claimed “self-defense” following sexual advances and suggested that victims became violent following the rejection of the undesired advance or feared that victims would eventually turn violent. Interestingly, many offenders openly admitted to murdering victims as a necessary response to such an advance, or as a result of “gay panic” by offenders (see also Dressler, 1995; Lunny, 2003; Mison, 1992; Tomsen, 2006, 2009).

The last subcategory of responsive homicide was mistaken identity homicide. In these cases, offenders mistook their victims for the opposite sex. Mistaken identity cases were by far the least common type of anti-LGBT homicide captured by the ECDB and were not included in the main comparative analyses presented above due to their small number. Despite their relative rarity, one finding is worth mentioning. Mistaken identity cases almost always involved transgendered victims. The number of mistaken identity cases in the United States may be larger than official reports suggest, because such homicides are unlikely to be reported or investigated as possible anti-LGBT attacks (see Witten & Eyler, 1999). In addition, the number of mistaken identity cases in the current study may be very small because offenders were often not identified for potential cases of mistaken identity homicide. Because of the uncertainty surrounding the circumstances of victim selection, these events did not meet the inclusion criteria and were not included in the current study. In many cases, for instance, a transgendered prostitute would be found murdered in a public location, thus leading advocacy groups to speculate that an offender who felt duped by the victim perpetrated the homicide. Speculation was not enough, however, to be included in the current study. With the inclusion of antitransgender crimes in federal bias crime laws following the passing of the 2009 Matthew Shepard Act, it is possible that the number of identifiable mistaken identity cases will increase over time.

There are several implications for research and policymakers stemming from the findings of this study. While conventional wisdom maintains that bias crime tends to be predatory acts of violence, this study also revealed the responsive nature of anti-LGBT homicide events. Therefore, it is important to recognize that bias-motivated violence does not necessarily symbolically target LGBT victims at random, but can also be largely fueled by misunderstandings, disputes, and escalated confrontations between victims and offenders in which offenders feel threatened by the victim’s sexual orientation or gender identity. Recognizing differences across homicide events within the context of how victims are selected can assist in the prevention of lethal anti-LGBT victimization. Police and policymakers should consider the varying modes of victim selection and roles of victims in violence when making decisions regarding bias crime legislation and the enforcement of bias crime laws for LGBT victims. Particular strategies for preventing future acts of anti-LGBT violence may depend on how victims are likely to be selected. To prevent future responsive anti-LGBT homicide victimization, such as sexual advance offenses, educational or public awareness campaigns could be utilized to inform the public about the risks of going home with more distant acquaintances and strangers, especially when drugs and alcohol are likely to be consumed. However, the content of messages regarding predatory forms of anti-LGBT violence should focus more on the risks of participation in anonymous meetings and Internet-based relationships.

Second, law enforcement bias crime training should also focus on research findings regarding the differences across modes of anti-LGBT victim selection categories. Evidence of prior confrontations between victims and offenders, or mixed-motive circumstances, should not necessarily rule out the occurrence of an anti-LGBT homicide. More research should also be done to advance a comprehensive understanding of bias crime victim selection strategies and these findings should be integrated into bias crime educational materials.

Third, researchers should seek to utilize alternative data sources to study rare forms of violence, such as anti-LGBT homicide. Until national crime databases include the information required to empirically study these forms of violence, researchers must rely on alternative open-source information and innovative research designs to advance knowledge about anti-LGBT homicide and other serious forms of bias-motivated violence.

There are also a couple notable limitations to the current research. First, a number of anti-LGBT homicide events that occurred during the time frame of the study remain unidentified. It is possible that authorities and victim’s families wish not to “out” victims or draw attention to victims’ LGBT statuses following their deaths in many cases. Therefore, this study focused only on observable anti-LGBT homicide events in the United States. How unidentified cases may systematically vary from those included in the study remains unknown. Second, this study focused only on fatal anti-LGBT attacks. While it is imperative to disaggregate anti-LGBT homicide separately, it is also necessary for future research to compare categories of lethal anti-LGBT violence with nonlethal anti-LGBT violence. Although this study was not able to conduct careful observations of offenders’ intentions to kill their victims, as opposed to simply injuring them, it is likely that many responsive homicides were devoid of intent to kill. Thus, it is possible that future research may find several similarities between responsive homicides and nonfatal acts of anti-LGBT violence.

Conclusion

In this study, we narrowed anti-LGBT homicide into two broadly defined categories and related subcategories: predatory offenses (including representative and instrumental subcategories) and responsive offenses (including gay bash, undesired romantic advance, and mistaken identity subcategories). A number of significant differences were identified at the incident, suspect, and victim levels. Future research should attempt to replicate this study’s findings on anti-LGBT bias violence and other types of bias crime. While there is admittedly no “gold standard” for typifying homicides (Flewelling & Williams, 1999, p. 99), the hope is that the utility of approaching bias homicide disaggregation based on victim selection of particular target groups will be further tested and this approach will be expanded. Doing so will allow law enforcement to better identify and prosecute specific forms of bias crime while also informing policymakers on the complex nature of bias crime incidents, offenders, and victims.

Footnotes

Appendix

Examples of Anti-LGBT Homicide Event Precursor, Transaction, and Aftermath Attributes

Precursor Attributes	Transaction Attributes	Aftermath Attributes
Offenders referred to victim(s) as girls when they left “gay bar” and victims reciprocated verbally	Offenders made contact with the victim by asking him where the gay bar was and then beat him to death with blunt objects	Offenders left victim’s body in remote outdoor location
Offenders lured victim away from “gay bar” under guise of going to a party then took him to a remote location where he was robbed	Offenders beat victim before dragging him to their garage and strangling him to death	Offenders drove body to remote area and buried his body in a shallow grave
Offenders binged on alcohol before “cruising for homosexuals” and identifying victim leaving gay bar	Offender stabbed victim repeatedly (20 or more times)	Offenders placed victim’s body in car then dumped body on the side of the road to be found later by police
Victim met offender at undisclosed location prior to returning to victim’s home together where the offender realized that victim was a biological male	Offenders stabbed victim 35 times with a knife and beat him with blunt objects, including a bat and sticks	Offenders set fire to the car carrying the deceased victims and fled
Offender and victim met online and arranged a meeting at a hotel	Offenders tortured victim with a metal rod until he died	Offenders threw victim’s body off of a canyon
Offender picked up victim who was hitchhiking	Victim was pummeled, stomped, and kicked unconscious during a party with many onlookers	Offender dismembered victim’s body
Offender robbed victim at a “homosexual meeting spot,” forced the victim into his vehicle, and drove him to a remote area	Offender cut the throat of a bound victim	Offenders dumped victim’s body under a highway overpass
Offender met victim at a karaoke bar before going on a walk together in which the victim made a sexual advance toward the offender	Offender beat victim and put him in a headlock until his body went limp	Victim’s body was left in the snow to be found days later

This research was supported by the U.S. Department of Homeland Security through a pre-doctoral fellowship with the National Consortium for the Study of Terrorism and Responses to Terrorism, Grant N00140510629.

However, any opinions, findings, and conclusions or recommendations in this article are those of the authors and do not necessarily reflect views of the U.S. Department of Homeland Security.

Notes

Jeff Gruenewald is an assistant professor in the Department of Sociology & Criminal Justice at the University of Arkansas. His research interests include homicide, bias crime, terrorism/homeland security, and media representations of violence.

Kristin Kelley is a graduate student in the Department of Sociology at Indiana University. Her research interests include gender, social psychology, and anti-LGBT homicide.

References

Agnew

(2006). Storylines as a neglected cause of crime. Journal of Research in Crime & Delinquency, 43, 119-147.

Belli

Freilich

J. D.

(2009). Situational crime prevention and non-violent terrorism: A “soft” approach against ideologically motivated tax-refusal. Crime Prevention Studies, 25, 173-206.

Block

C. R.

Block

R. L.

(1992). Homicide syndromes and vulnerability: Violence in Chicago community areas over 25 years. Studies on Crime and Crime Prevention, 1, 61-87.

Boyd

E. A.

Berk

R. A.

Hamner

K. M.

(1996). “Motivated by hatred or prejudice”: Categorization of hate-motivated crimes in two police divisions. Law & Society Review, 30, 819-850.

Bufkin

J. L.

(1999). Bias crime as gendered behavior. Social Justice, 26, 155-176.

Chermak

S. M.

Freilich

J. D.

Parkin

W. S.

Lynch

J. P.

(2012). American terrorism and extremist data sources and selectivity bias: An investigation focusing on homicide events committed by far-right extremists. Journal of Quantitative Criminology, 28, 191-218.

Chermak

S. M.

Freilich

J. D.

Suttmoeller

(2013). The organizational dynamics of far-right hate groups in the United States: Comparing violent to non-violent organizations. Studies in Conflict & Terrorism, 36, 193-218.

Comstock

G. D.

(1991). Violence against lesbians and gay men. New York, NY: Columbia.

Cronin

S. W.

McDevitt

Farrell

Nolan

J. J.

III . (2007). Bias-crime reporting: Organizational responses to ambiguity, uncertainty, and infrequency in eight police departments. American Behavioral Scientist, 51, 213-231.

10.

Dabney

D. A.

(2004). Crime types: A text/reader. Belmont, CA: Wadworth.

11.

Decker

S. H.

(1993). Exploring victim–offender relationships in homicide: The role of individual and event characteristics. Justice Quarterly, 10, 585-612.

12.

Dressler

(1995). When “heterosexual” men kill “homosexual” men: Reflections on provocation law, sexual advances, and the “reasonable man” standard. The Journal of Criminal Law and Criminology, 85, 726-763.

13.

Dunbar

(2006). Race, gender, and sexual orientation in hate crime victimization: Identity politics or identity risk? Violence and Victims, 21, 323-337.

14.

Federal Bureau of Investigation. (2011). Uniform crime reports: Hate crime statistics 2011. Retrieved from http://www.fbi.gov/about-us/cjis/ucr/hate-crime/2011/tables/table-4

15.

Fisher

Salfati

C. G.

(2009). Behavior or motivation: Typologies of hate-motivated offenders. In Perry

Blazak

(Eds.), Hate crimes: Hate crime offenders (Vol. 4, pp. 103-135). Westport, CT: Praeger.

16.

Flewelling

R. L.

Williams

K. R.

(1999). Categorizing homicides. In Smith

M. D.

Zahn

M. A.

(Eds.), Homicide, a sourcebook of social research (pp. 96-106). Thousand Oaks, CA: SAGE.

17.

Freilich

J.D.

Chermak

S.M.

Caspi

(2009). Critical events in the life trajectories of domestic extremist White supremacist groups: A case study analysis of four violent organizations. Criminology & Public Policy, 8(3), 497-530.

18.

Freilich

J. D.

Chermak

S. M.

Belli

Gruenewald

Parkin

W. S.

(2014). Introducing the United States Extremist Crime Database (ECDB). Terrorism and Political Violence, 26, 372-384. doi:10.1080/09546553.2012.713229

19.

Garnets

Herek

G. M.

Levy

(1990). Violence and victimization of lesbians and gay men: Mental health consequences. Journal of Interpersonal Violence, 5, 366-383.

20.

Gruenewald

(2011). A comparative examination of far-right extremist homicide events. Homicide Studies, 15, 177-203.

21.

Gruenewald

(2012). Are anti-LGBT homicides unique? Journal of Interpersonal Violence, 27, 3601-3623.

22.

Gruenewald

(2013). Studying bias-motivated homicide: An examination of lethal violence against homeless persons using open-source data. International Journal of Criminology & Sociology, 2, 538-549.

23.

Gruenewald

Chermak

Freilich

J. D.

(2013a). Distinguishing “loner” attacks from other domestic extremist violence. Criminology & Public Policy, 12, 1-27.

24.

Gruenewald

Chermak

Freilich

J. D.

(2013b). Lone wolves and far-right terrorism in the United States. studies in conflict and terrorism, 36, 1005-1024.

25.

Gruenewald

Pridemore

W. A.

(2012). A comparison of ideologically-motivated homicides from the new Extremist Crime Database and homicides from the Supplementary Homicides Reports using multiple imputation by chained equations to handle missing values. Journal of Quantitative Criminology, 28, 141-162.

26.

Guttman

(1968). A general nonmetric technique for finding the smallest coordinate space for a configuration of points. Psychometrika, 33, 469-506.

27.

Haider-Markel

(2002). Regulating hate: State and local influences on hate crime law enforcement. State Politics and Policy Quarterly, 2, 126-160.

28.

Harris

(2010, October 19). US military accepts openly gay recruits. The Guardian. Retrieved from http://www.guardian.co.uk/world/2010/oct/20/us-military-openly-gay-recruits

29.

Herek

G. M.

Berrill

(1992). Hate crimes: Confronting violence against lesbians and gay men. Newbury Park, CA: SAGE.

30.

Herek

G. M.

Gillis

Cogan

Glunt

(1997). Hate crime victimization among lesbian, gay, and bisexual adults. Journal of Interpersonal Violence, 12, 195-215.

31.

Iganski

(2008). Hate crime and the city. Bristol, UK: Policy Press.

32.

Jenness

(1995). Social movement growth, domain expansion, and framing processes: The gay/lesbian movement and violence against gays and lesbians as a social problem. Social Problems, 42, 145-170.

33.

Jenness

Broad

(1997). Hate crimes: New social movements and politics of violence. New Brunswick, NJ: Transaction Publishers.

34.

Katz

(1988). Seductions of crime. New York, NY: Basic Books.

35.

Lawrence

F. M.

(1999). Punishing hate: Bias crimes under American law. Cambridge: MA: Harvard University Press.

36.

Levin

McDevitt

(1993). Hate crimes: The rising tide of bigotry and bloodshed. New York, NY: Plenum.

37.

Luckenbill

(1977). Criminal homicide as a situated transaction. Social Problems, 25, 176-186.

38.

Lunny

(2003). Provocation and “homosexual” advance: Masculinized subjects as threat, masculinized subjects under threat. Social & Legal Studies, 12, 311-333.

39.

McDevitt

Balboni

J. M.

Bennett

Weiss

J. C.

Orchowsky

Walbolt

(2000). Improving the quality and accuracy of bias crime statistics nationally: An assessment of the first ten years of bias crime data collection (Executive Summary). Boston, MA: Center for Criminal Justice Policy Research, Northeastern University.

40.

McDevitt

Levin

Bennett

(2002). Hate crime offenders: An expanded typology. Journal of Social Issues, 58, 303-317.

41.

Messerschmidt

J. W.

(2012). Gender, heterosexuality, and youth violence: The struggle for recognition. Lanham, MD: Rowman & Littlefield.

42.

Messner

S. F.

McHugh

Felson

R. B.

(2004). Distinctive characteristics of assaults motivated by bias. Criminology, 42, 585-618.

43.

Miethe

T. D.

Drass

K. A.

(1999). Exploring the social context of instrumental and expressive homicide: An application of qualitative comparative analysis. Journal of Quantitative Criminology, 15, 1-21.

44.

Miethe

T. D.

McCorkle

R. C.

Listwan

S. J.

(2005). Crime profiles: The anatomy of dangerous persons, places, and situations (3rd ed.). New York, NY: Oxford University Press.

45.

Mison

(1992). Homophobia as manslaughter: The homosexual advance as insufficient provocation. California Law Review, 80, 133-178.

46.

National Coalition of Anti-Violence Programs. (2012). Hate violence against lesbian, gay, bisexual, transgender, queer, and HIV-affected communities. Retrieved from http://www.avp.org/storage/documents/Reports/2012_NCAVP_2011_HV_Report.pdf

47.

National Gay and Lesbian Task Force. (1995). Anti-gay violence, victimization and defamation in 1994. Washington, DC: Author.

48.

Nolan

J. J.

Akiyama

(1999). An analysis of factors that affect law enforcement participation in hate crime reporting. Journal of Contemporary Criminal Justice, 15, 111-127.

49.

Perry

(2001). In the name of hate: Understanding hate crimes. New York, NY: Routledge.

50.

Phillips

(2009). The prosecution of hate crimes: The limitations of the hate crime typology. Journal of Interpersonal Violence, 24, 883-905.

51.

Powell

Bolzendahl

Geist

Steelman

(2010). Counted out: Same-sex relations and Americans’ definitions of family. New York, NY: Russell Sage Foundation.

52.

Richey

(2014, March 27). Poll finds broad, rapid shift among Americans toward gay marriage. The Christian Science Monitor. Retrieved from http://www.csmonitor.com/USA/Politics/2014/0327/Poll-finds-broad-rapid-shift-among-Americans-toward-gay-marriage

53.

Rose

S. M.

Mechanic

M. B.

(2002). Psychological distress crime features, and help-seeking behaviors related to homophobic bias incidents. American Behavioral Scientist, 46, 14-26.

54.

Sacco

V. F.

Kennedy

(2002). The criminal event: Perspectives in space and time (2nd ed.). Belmont, CA: Wadsworth.

55.

Sandholtz

Langton

Planty

(2013). Hate crime victimization, 2003-2011. Bureau of Justice Statistics Special Report, U.S. Department of Justice. Retrieved from http://www.bjs.gov/content/pub/pdf/hcv0311.pdf

56.

Short

J. F.

(1997). Poverty, ethnicity, and violent crime. Boulder, CO: Westview.

57.

Stacey

(2011). Distinctive characteristics of sexual orientation bias crimes. Journal of Interpersonal Violence, 26, 3013-3032.

58.

Stolberg

S. G.

(2010, December 22). Obama signs away “don’t ask, don’t tell.” The New York Times. Retrieved from http://www.nytimes.com/2010/12/23/us/politics/23military.html

59.

Tomsen

(2002). Hatred, murder and male honour: Anti-homosexual homicides in New South Wales, 1980-2000. Canberra: Australian Institute of Criminology.

60.

Tomsen

(2006). Homophobic violence, cultural essentialism and shifting sexual identities. Social & Legal Studies, 15, 389-407.

61.

Tomsen

(2009). Violence, prejudice and sexuality. New York, NY: Routledge.

62.

Tomsen

Mason

(2001). Engendering homophobia: Violence, sexuality and gender conformity. Journal of Sociology, 37, 257-273.

63.

U.S. Department of Justice. (1998). National hate crimes training curricula. Author. Retrieved from https://www.ncjrs.gov/pdffiles1/nij/186784.pdf

64.

U.S. Department of Justice. (n. d.). Hate crime training: Core curriculum for patrol officers, detectives & command officers. Author. Retrieved from http://www.justice.gov/archive/crs/pubs/hct.pdf

65.

Van Der Meer

(2003). Gay bashing: A rite of passage? Culture, Health, & Sexuality, 5, 153-165.

66.

Walker

Katz

C. M.

(1995). Less than meets the eye: Police department bias crime units. American Journal of Police, 14(1), 29-48.

67.

Witten

T. M.

Eyler

A. E.

(1999). Hate crimes and violence against the transgendered. Peace Review, 11, 461-468.