As this special issue illustrates, mass shooting definitional criteria remain a source of controversy and disagreement among academics and practitioners (see also: Duwe, 2020; Freilich et al., 2020; Huff-Corzine & Corzine, 2020; Silva & Greene-Colozzi, 2019). Published works often provide several pages in the methodology dedicated to defining the phenomenon of interest, and because of disagreement among scholars, many authors are also compelled to defend their rationale for the definition in anticipation of a dissenting reviewer. This rigorous attention to definition has been valuable for the field: encouraging careful consideration of methodology and database compilation for open-source scholars and setting a precedence for establishing clear-cut definitions in published articles. To this end, progress has been made toward making distinctions between different definitions and types of mass shootings. It is important to briefly illustrate these current advancements.

Drawing from the Federal Bureau of Investigation (FBI) definition of “mass murder,” the Congressional Research Service (CRS) offers one of the broadest and most widely accepted mass shooting definitions: “a multiple homicide incident in which four or more victims are murdered with firearms—not including the offender(s)—within one event, and in one or more locations in close geographical proximity” (Krouse & Richardson, 2015, p. 10). Importantly, they developed and compared three mass shooting types: familicide, felony, and public. Familicide mass shootings refer to incidents involving the offender’s immediate or extended family, most of whom are killed in private residences or sparsely populated locations. Felony mass shootings refer to incidents involving underlying criminal activity (e.g., robbery/burglary, drug violence, gang violence). Public mass shootings refer to incidents occurring in public or populated locations involving victims who are often killed indiscriminately.

Studies and datasets use variations of this mass shooting definition to examine all mass shootings (e.g., Duwe, 2020; Gun Violence Archive; Mass Shooting Tracker) and to compare the different types of mass shootings (e.g., Silva, 2022a; USA Today). Most scholars agree that there are important differences between public and private/residential mass shootings (Duwe, 2004, 2020; Schildkraut & Turanovic, 2022), and that private/residential mass shootings constitute unique forms of violence with critically different response and prevention strategies. For example, the perpetrators who commit these private/residential attacks tend to be psychologically and behaviorally different from public mass shooting offenders (Krouse & Richardson, 2015; Silva, 2022a). Additionally, the seemingly random victims involved in public mass shootings means these incidents often receive more public attention, media coverage, and political discourse (Duwe, 2020; Silva & Capellan, 2019). While we acknowledge that familicide and felony shooting incidents are equally tragic to public incidents, the current commentary is specially focused on public mass shootings.

Public mass shooting scholars generally agree on three definitional elements: (1) the incident must occur in public or populated locations; (2) it must involve at least some random and/or symbolic victims; and (3) it should occur over a short period of time (usually within 24 hours; see Peterson & Densley, 2019; Schildkraut & Elsass, 2016). This definition largely aligns with the FBI’s (2021) definition of active shooter incidents, which refers to individuals “actively engaged in killing or attempting to kill people in a populated area [using firearms]”. The FBI definition includes cases that occurred in public, and it excludes cases where the shooter’s actions were motivated by a separate criminal act, like gang or drug violence, as well as contained residential or domestic disputes (FBI, 2021). Thus, despite different names, examinations of public mass shootings and active shooter incidents are largely assessing the same phenomenon. However, there is one key distinction that separates the two: victim/casualty count.

In line with the CRS definition, much of the current public mass shooting research only includes attacks that involved four or more victim deaths. This minimum victim criterion is practiced by some of the leading scholars in the mass shooting field (Duwe, 2020; Fox & Levin, 2022), and incorporated into many of the most-cited publicly available mass shooting databases (e.g., The Violence Project, The Washington Post). However, we believe that a consequence of this definition is the exclusion of relevant cases characterized by mass shooting intent. In other words, they often exclude incidents that did not progress into a “completed” mass shooting (i.e., involving four or more fatalities) due to pre-incident exposure or situational and environmental aspects of the attack. While the FBI offers a broader victim count, many intended public mass shooting incidents are also excluded from the FBI’s active shooter definition and reports.

To this end, we propose a unifying public mass shooting definition that expands the phenomenon to include predatory, intentional public shooting incidents regardless of casualty count. Below, we outline the value, rationale, and strategies for expanding the casualty count to include fewer than four shooting victims and argue for the importance of a no-minimum casualty count definition (including zero victims) of public mass shootings that emphasizes mass intent rather than the completion of the shooting.

Proposed Public Mass Shooting Definition

Drawing from previous research (Freilich et al., 2020; Newman et al., 2004; Schildkraut & Elsass, 2016; Silva, 2021, 2022c; Silva & Greene-Colozzi, 2022a), we propose the following definition:

A public mass shooting is a gun violence incident (or intended incident), perpetrated by (or intended to be perpetrated by) one or two offenders,¹ targeting one or more public or populated locations, within a 24-hour period. At least some of the victims (or intended victims) must be chosen at random or for their symbolic value. If the perpetrator(s) does not kill four or more victims, they must demonstrate behavioral evidence suggesting mass victim intent. A public mass shooting excludes any incident involving familicide or other criminal activity.²

Like the CRS breakdown of all mass shootings into three victim type- and location-based categories, we propose a breakdown of public mass shootings into four casualty-count and target-reached outcome categories: completed, attempted, failed, and foiled. Each outcome category is defined below.

Aligning with the most commonly accepted definitions, a completed mass shooting is an incident involving four or more fatalities (excluding the perpetrator). We strongly suggest researchers use the term “completed” mass shooting instead of the phrasing used in previous public mass shooting research measuring “success”. This alternative phrasing avoids glorifying offenders as being “successful”.

An attempted mass shooting is an incident involving fewer than four deaths, but at least one gunshot casualty (fatality or injury). These shootings are sometimes referred to as active shooter incidents, although examinations of active shootings also include completed and failed outcomes (FBI, 2021; O’Neill et al., 2016). We suggest researchers use the term “attempted” mass shooting instead of active shooting or active shooter incident to avoid confusion. In other words, the perpetrator attempted to kill four or more individuals (and was able to incur at least one gunshot casualty) but did not reach the completed fatality threshold. For example, in 2019, a 19-year-old opened fire at a synagogue after bringing a semi-automatic rifle and a tactical vest containing five magazines with ten rounds each (Paul & Mettler, 2019). The shooting incurred one fatality and three injuries before his rifle malfunctioned. Given the amount of ammunition he brought to the scene, his predatory crime scene behavior (roaming and searching out new victims), his attempt to livestream the shooting, as well as his pre-incident actions (posting a manifesto online prior to the shooting), we would classify this case as an attempted mass shooting.

A failed mass shooting is an incident involving zero victim casualties, but the perpetrator arrived at their intended target with their firearm, and they tried to carry out their attack on that day. This type of mass shooting progresses beyond the preparation stage (meaning it was not pre-operationally foiled) and the shooter successfully arrived at and opened fire (or tried to open fire) on their intended target. However, during the event, they did not incur any gunshot casualties. For example, in 2019, a 22-year-old man opened fire fear near a courthouse before making his way to the glass door of the building and opening fire inside (Hutchinson, 2019). The perpetrator was armed with a semi-automatic rifle, he had 150 rounds of ammunition, and he was wearing body armor. However, there were officers on the scene who immediately returned fire and killed the shooter. This case would qualify as a failed mass shooting since the perpetrator had arrived at the scene with ample ammunition to harm many victims and opened fire, but external circumstances caused the shooting to fail without any victim casualties.

A foiled mass shooting is an incident involving zero victim casualties that did not progress past the planning and preparation stage into actualized initiation. While the incident was set into motion, it was stopped before the offender could arrive at or enter their target location with their firearm. In other words, the offender’s plan did not fail during the incident, as they were thwarted before they could try and shoot victims at their intended target. For example, in 2019, a 13-year-old boy leaked information to fellow students suggesting he was planning a mass shooting targeting his middle school (Ellis et al., 2019). During a search of the boy’s home, law enforcement seized an AR-15 rifle, 100 rounds of ammunition, a hand-drawn map of the school, and a list of students and staff members. However, this incident was thwarted before the student could finalize his plans and arrive at the target location, and it would thus be considered a foiled mass shooting.

Ultimately, this proposed definition incorporates the predatory active shooter component of the FBI’s definition, allowing for attempted public mass shootings. It also aligns with the common academic definitions of public mass shootings and active shootings by excluding certain types of mass gun violence (felony and family). Further, it expands upon the FBI’s active shooter concept by including intended public mass shootings that were foiled before initiation. Understood together, this proposed definition allows for a more robust sample/population size, while still providing a targeted and unambiguous assessment of a specific gun violence phenomenon. We acknowledge the inherent challenge of measuring intent and provide guidelines and suggestions for measuring the level of mass intent associated with the shooting. By refocusing on the perpetrator’s behavioral indicators of intent, rather than the completed-ness of the shooting, we feel our definition has the potential to expand understanding of public mass shootings through the inclusion of diverse incidents that are linked by a common theme of intent, rather than by outcome.

Importance of No-Minimum Casualty Count

A no-minimum casualty count definition of public mass shootings has value for both scholarly advancements and policy implications. It improves overall understanding of the phenomenon by including all mass shooting outcomes and therefore offers crucial insights for harm mitigation, intervention, and prevention. Additionally, this definition provides an opportunity for individual analysis of each outcome type, as well as comparative analyses among different outcome types, which can help explain why some mass shooters complete the attack while others are thwarted. Below, we highlight some empirical benefits of our proposed definition that we consider to be particularly critical.

Mass shooting studies are often limited by a small sample/population size. Public mass shootings remain one of the rarest forms of gun violence (Duwe, 2020). This broader definition will allow for more robust and advanced analyses of the phenomenon despite the low incidence baseline. The FBI’s (2021) active shooter report has already demonstrated how lower minimum victim criterion can advance sample size. For instance, The Violence Project Database (Peterson & Densley, 2019) lists 94 incidents from 2000 to 2020, but the FBI active shooter reports document 373. The major difference between the incidents in The Violence Project Database and those in the FBI report is whether the shooting was completed, which we argue is situationally dependent. Furthermore, a no-minimum victim definition can be easily filtered by casualties should researchers want to examine only completed cases or only those cases with fatalities and injuries. Researchers and data-users have the option of modifying the data to accommodate a minimum casualty criterion, which is impossible to do with datasets that only collect cases meeting a certain fatality threshold.

Relatedly, the proposed definition allows researchers to perform important comparisons of attempted versus completed public mass shootings that are not possible with a single incident outcome of fatalities. Some studies use expanded datasets that include attempted public mass shootings to determine common mass shooter characteristics and factors influencing rates of victim casualties (Blair et al., 2021 Blau et al., 2016), but these continuous measures of casualties are still limited because they do not determine what makes attempted shootings unique from completed shootings. These comparisons have enormous potential to provide policy-relevant insights for harm mitigation during a public mass shooting event. Capturing these attempted public mass shootings also fills a gap in previous completed public mass shooting research by including incidents involving fewer than four deaths, which nevertheless involve more than four victims shot.³ For example, the 2014 Fort Hood shooter “only” incurred three deaths, and as such, this incident would not fit the four-death cutoff used in many completed public mass shooting studies. However, during this incident, an additional 12 victims were also injured via gunshot. We argue that the purposeful public shooting of 15 people should be considered a mass shooting regardless of the number of fatalities. To this end, our suggested inclusion of attempted mass shootings captures incidents where four or more people were shot, but systematic and situational factors such as police and EMT responses, victim or witness defense, and to some degree, luck, prevented the deaths of four or more victims.

While some scholarly databases have endeavored to include and measure attempted mass shootings, many expanded datasets often overlook failed mass shooting outcomes involving no fatalities or injuries. Failed incidents represent the ideal outcome of a mass shooting that reaches the intended target (i.e., no gunshot victims). Examinations and comparisons of failed mass shootings also provide valuable implications for harm mitigation and situational crime prevention during incidents that reach the intended target (Freilich et al., 2020). Our initial research finds that situational crime prevention techniques, such as target hardening and access control, may help prevent bloodshed when individuals without legitimate access plan and initiate a public mass shooting (Author, 2022; Author, 2022a, 2022b). Additionally, internal measures, including lockdowns, place managers, and armed security, can also help prevent victimization in open access or semi-protected locations—even if the perpetrator successfully enters the building (Author, 2022; Author, 2022a, 2022b). Nonetheless, there is still much to be learned about failed mass shootings, and it is surprising how little attention has been directed at this outcome considering failed mass shootings should be the primary goal when prediction and prevention strategies are inadequate.

Finally, foiled incidents present the ideal outcome of a public mass shooting because there are no gunshot casualties, and there are no individuals psychologically victimized or traumatized by a failed mass shooting experience. We suggest that foiled shootings – not solely completed or attempted shootings – should be a primary focus for researchers aiming to prevent public mass shootings. These types of incidents have been largely overlooked and expanding the public mass shooting definition to include foiled cases can offer crucial insight for threat assessment techniques and public safety professionals. While examinations of foiled outcomes may not be appropriate in all public mass shooting studies, like those focused on situational crime prevention, that does not mean they should be excluded from studies entirely.

Ultimately, we believe that the systematic exclusion of attempted, failed, and foiled cases may result in biased conclusions pertaining to perpetrator psychosocial and emotional characteristics, planning and preparatory behaviors, and important pre-operational actions, such as leakage and gun obtainment. Nevertheless, there are challenges in defining and determining the intent to commit a public mass shooting when the outcome is ambiguous (i.e., zero casualties).

Considerations for Case Criteria and Database Compilation

Despite its value, the proposed definition does require special consideration when deciding what makes a public shooting a mass attack, if not the number of fatalities. Public mass shooters seek out random and targeted victims (Fox & Levin, 2022), but should their attacks result in a non-completed fatality outcome, it becomes difficult to assess the intent of the attack. Suicide is a common comorbidity with public mass shooting attacks (Lankford, 2015; Newman et al., 2004; Peterson & Densley, 2019), and in the absence of expressed intent, suicide notes, or legacy tokens, the motive and intention of a suicidal attack can be frustratingly abstruse to law enforcement and researchers. Nonetheless, we believe the benefits outweigh the costs, and we propose considerations when curating data collection, as well as strategies for addressing these potential limitations.

First, scholars should consider when to begin their examinations of public mass shootings. Many scholars use the 1966 University of Texas shooting as a starting point because it introduced the second wave of mass shootings (i.e., public mass shootings) into the public consciousness (Duwe, 2004; Peterson & Densley, 2019; Schildkraut & Elsass, 2016). However, given the reliance on open-source media coverage for building datasets, time period effects (i.e., older incidents receiving less coverage) are likely to influence these examinations (Silva & Greene-Colozzi, 2019). A consequence of these longer time periods is a decreasing likelihood of finding and coding cases from earlier years using open sources (Chermak & Gruenewald, 2006; Chermak et al., 2012). And, when broadening the definition to include attempted, failed, and foiled attacks, the risk of selection bias becomes even more pronounced. Thus, depending on the outcomes included in the research, scholars may choose to begin in 1999: the year of the Columbine shooting, which instigated a media frenzy and introduced the concept of a “mass shooter” into the modern American cultural lexicon. Regardless of what direction a researcher chooses to take with timeline range, clarity and transparency are key (for further insight, see Greene-Colozzi et al., 2021; Silva & Greene-Colozzi, 2019).

Second, including foiled plots requires scholars to determine threat credibility and assess whether plots were serious versus fantasies, police overreactions, or media hyperbole (Larkin, 2009). According to O’Toole (2000), some threats should be classified as more serious than others, yet studies are often vague in their definitions of threatened and thwarted violence. For example, Langman and Straub’s (2019) study includes any averted school violence attack that is “planned… and prevented before any loss of life occurred”, but they provide a limited indication of what this “planning” entails. This ambiguity surrounding what constitutes a planned attack may unintentionally populate the database with “false positives”, or cases identified as a threat that were not likely to occur. However, studies of unsuccessful terrorist plots provide some guidance to address this issue, with scholars suggesting incidents must include a target or target type explicitly identified, and at least one overt step (e.g., collecting weapons, reconnoitering the target, etc.) made toward carrying out the plot (Dahl, 2011; Klein et al., 2017). In the mass shooting arena, Sarteschi (2016) offers similar guidelines for determining “high” level mass homicide threat credibility, defining high credibility “actionable” cases as those where the threatener possessed the means to carry out the threat (i.e., bombs, guns, etc.), or had plans or materials (i.e., manifesto, etc.) for securing the necessary means to carry out the attack (p. 89). While all threats should be taken seriously, the general goal for empirical data collection should be to avoid including false positives.

Finally, given the ambiguity that can occur with a lower or zero casualty definition, scholars should aim to narrowly focus on public mass shootings perpetrated to kill mass random (but intentional) victims. For example, consider the comparison of two hypothetical incidents occurring in a crowded mall. In the first incident, the perpetrator had a previous altercation with an acquaintance and upon seeing that acquaintance in the mall, opens fire with a single handgun in response to that specific altercation. However, due to poor aim and/or high emotion, this shooter kills several innocent bystanders. In the second incident, the perpetrator is an employee of the mall who was recently terminated and has returned with a gun and several hundred rounds of ammunition. This perpetrator targets former co-workers while also aiming and firing at random shoppers in the mall and manages to wound one victim before mall security intervenes. While both these incidents are seemingly public mass shootings, the perpetrator in the second case displayed a motivation to commit continuous random mass violence. The perpetrator in the first case unintentionally committed random mass violence by hitting collateral victims in pursuit of a single, specific target. We argue for exclusion of the first case because it lacked mass intent, and inclusion of the second case (despite the low casualty count) because the perpetrator displayed predatory behavior suggestive of mass intent.

An Empirical Strategy for Measuring Mass Intent

Currently, little public mass shooting research directly addresses the issue of mass intent outside of victim casualties. We propose a novel coding strategy that addresses the challenge of determining mass intent with varied incident outcomes. Specifically, we encourage readers to use observable behavioral indicators before the public mass shooting and during the incident to assign an intent rating score to each individual incident. Prior research has documented a wide variety of discrete public mass shooter behaviors and characteristics relevant to planning, preparation, and perpetration (Clemmow et al., 2022; Cowan & Cole, 2022; Fox & Levin, 1994; Newman et al., 2004; O’Toole, 2000), including bringing spare guns and ammunition, leakage, wearing a costume to invoke fear or mimic violent role models, aiming and firing at victims and witnesses, having a targeted or symbolic grievance against the location, stockpiling guns and ammunition, or training to shoot. Importantly, these indicators are readily observable to guardians and handlers (Felson, 1995) surrounding the perpetrator and to witnesses and survivors at the crime scene. By incorporating these objective indicators of intent, we become less reliant on the completed outcome of the shooting as confirmation of mass intent, which, as discussed above, is likely to be dependent on factors outside of the perpetrator’s control.

We suggest coding thirteen behavioral indicators of mass intent during incidents that reach the intended target: public location, populated location, perpetrator aiming at victims, perpetrator aiming at others, verbal intent or leakage, written intent or leakage, hunting behavior, costuming, displaying a known grievance against the location, having a hit list of targeted victims, obtaining a gun for the shooting,⁴ training with firearms before the shooting, and bringing spare ammunition and/or firearms to the crime scene. Each of these indicators is reviewed briefly in Table 1, except for public and populated locations since these are inherent in the definition of a public mass shooting.