Understanding More Than 50 Years of Mass Public Shootings as a Function of Routine Activities

Abstract

Mass public shootings have drawn considerable attention from the public, policymakers, and researchers, yet despite what is known about these events, assessments to date have failed to consider their timing as a function of the locations where they occur. Using data on 401 U.S. mass public shootings occurring between 1966 and 2020, we examine these events’ temporal patterns. The findings suggest that the occurrence of mass public shootings may not be as random as once assumed but instead mirror the routine activities of the perpetrators, their victims, and the shootings’ locations. Considerations for prevention and response policies also are offered.

Keywords

mass shootings school shootings routine activities temporal patterns place

Mass shootings continue to be a cause for concern among the general public and within the scholarly community. To date, researchers have examined diverse issues related to the phenomenon, including, but not limited to, differences among perpetrators (Arluke et al., 2018; Lankford, 2015, 2016), the impact of such events on both individuals (Lowe & Galea, 2017; Schildkraut et al., 2021) and policy (Luca et al., 2020), potential causes (Kwon & Cabrera, 2019; Schildkraut & Muschert, 2013; Skeem & Mulvey, 2020), and even the role of media coverage in contagion (Meindl & Ivy, 2017; Towers et al., 2015). Less is known about when mass shootings occur beyond examination of patterns of prevalence across decades (Duwe, 2020; Lin et al., 2018; Schildkraut, 2021). In other words, despite what we know about mass shootings, what we do not know is when specifically (e.g., day, time) they are most likely to occur and how such patterns relate to the locations where they take place. This information, however, may have important implications for policy efforts as “knowing where crimes (of any type) occur is simply not enough; to prevent criminal activity, we must know where and when crime occurs for more effective prevention” (Hewitt et al., 2020, p. 2; emphasis added).

Researchers long have contended that crime more generally does not occur randomly across space or time (Brantingham & Brantingham, 1981; Cohen & Felson, 1979; Felson & Poulsen, 2003; Hawley, 1950; Newton & Felson, 2015). Instead, it occurs during the routine activities of people’s everyday lives when a motivated offender and suitable target converge in the absence of capable guardianship (Cohen & Felson, 1979; Felson & Cohen, 1980). Understanding this convergence requires consideration of both spatial and temporal patterns of routine, as well as illegal, activities (Felson & Cohen, 1980). Research to date, however, often focuses on the spatial distribution of crime while failing to also assess the temporal dimension (Cohen & Felson, 1979; Felson & Poulsen, 2003; Newton & Felson, 2015; Ratcliffe, 2002).

Some scholars (e.g., Elsass et al., 2016; Schildkraut et al., 2019) have argued that despite being episodic violent crimes, the events themselves—regardless of the characteristics of the perpetrators and victims—can be understood as a function of routine activity theory based on where and when they occur (see also Newton & Felson, 2015; Rossmo, 1995). To explore this argument, we analyzed the timing (day of week and hourly intervals) of 55 years of mass public shootings with particular focus paid to the locations at which these events occur. This line of inquiry is relevant not only to better understanding the phenomenon of mass shootings, as there has yet to be a comprehensive study assessing the relationship between timing and place, but also for applications of routine activity theory itself.

Temporal Variations of Crime

Routine activity theory (Cohen & Felson, 1979) has its origins in Hawley’s (1950) human ecology perspective, which asserts that assessments of community activities must account for both space and time. Central to the Hawley’s (1950) perspective are three key temporal components: (1) rhythm, which refers to the regular frequency at which events occur; (2) tempo, or the number of events that occur in each unit of time; and (3) timing, referring to the coordination between seemingly interdependent yet different activities (p. 289). Cohen and Felson (1979) expanded on this perspective in applying it to the rising crime rates of the 1960s and 1970s, contending that in the context of crime, the spatial and temporal structure of both motivated offenders and suitable targets’ routine activities must be accounted for, particularly when considering how their pathways ultimately cross (see also Brantingham & Brantingham, 1993, 2013; Felson & Cohen, 1980). More specifically, the timing of routine activities, which include work, education, leisure, and other activities, both at and outside of the home, is an important element to consider in explaining crime rates because they may influence the rate at which direct-contact predatory crimes can occur (Cohen & Felson, 1979; Felson & Cohen, 1980).

Newton and Felson (2015) argue that “disaggregation of crime by place and by time, for example hour of day, day of week, month of year, season, or school day versus non-school day is extremely relevant for theory” (p. 1). Certainly, crime rates are variable within each of these intervals of time (Cohen & Felson, 1979). The question then becomes what is the best unit to study temporal crime patterns. Some have argued that large temporal units (e.g., assessing patterns monthly, quarterly, or annually) can lead to the loss of important variation that is needed to study the dynamic nature of crime (Felson & Poulsen, 2003; Ratcliffe, 2002). Even when considering a week, which encompasses 168 seemingly unique hours, crime rates can shift rapidly (Newton & Felson, 2015). It is for this reason that Felson and colleagues suggest that temporal analyses of dynamic crime patterns should focus on hourly intervals, which captures more variation than almost any other temporal unit (Felson & Eckert, 2019; Felson & Poulsen, 2003; Newton & Felson, 2015).

Despite the importance of understanding temporal variation in crime patterns, there have been few in-depth studies, particularly using the hourly interval on its own or in conjunction with the location where incidents occur. Generally, findings suggest that temporal patterns vary not only by hour but also by the type of crime (Andresen & Malleson, 2015) and other extrinsic factors. Snyder and Sickmund (2006), for example, found that rates of violent crimes fluctuate not only by time but also by age of the victim. Attacks against juveniles vary throughout the day but are highest between 3:00 p.m. and 4:00 p.m. when youth typically are leaving school for the day. Conversely, violent victimization perpetrated against adults typically peaks between 9 p.m. and midnight (Snyder & Sickmund, 2006). In a separate study, Snyder (2000) found that sexual offenses against children also peaked between 3:00 p.m. and 4:00 p.m., as well as during intervals usually dedicated to breakfast (8:00 a.m.–9:00 a.m.), lunch (12:00 p.m.–1:00 p.m.), and dinner (6:00 p.m.–7:00 p.m.). Sexual offenses targeting adults were more common in late evening hours, with the peak incidence rate happening between midnight and 2:00 a.m. (Snyder, 2000). Similarly, using a broader sample of more than 2,250 sexual offenses over a 4-year period in a Canadian city, Hewitt et al. (2020) found that these crimes were more likely to occur between 11 a.m. and 4 p.m., although the authors did not explicitly state, but rather implied, that the victims of such incidents likely were children (thereby mirroring Snyder’s (2000) findings).

The temporal patterns of crime also may be relative to the location at which they occur. Cusimano et al. (2010), for instance, found that assaults against adults occurring during evening hours (8:00 p.m.–11:59 p.m.) in Toronto were more likely to occur in areas where there were greater concentrations of social housing communities and homeless shelters. Conversely, those assaults that took place in the early morning hours (midnight to 3:59 a.m.) occurred downtown, where there is a higher concentration of nightlife establishments. Moreover, the time interval with the highest rate of injuries (2:00 a.m.–2:59 a.m.) corresponds to the hour immediately after the close of bars and clubs for the night (Cusimano et al., 2010). In another study analyzing more than 6 million cybercrime attacks targeting one university’s computer network over a 3-year period, Maimon et al. (2013) found that such attacks were more likely to occur during the institution’s official business hours (9:00 a.m.–4:59 p.m.) than outside them. Further, assessing 9 years of data on robberies occurring at the Prudential Center arena in Newark, New Jersey, Kurland (2019) found that increases in such crimes occurred in each hour before, during, and after NHL hockey games, concerts, and Disney-sponsored events when there was increased traffic around the venue.

Ruderman and Cohn (2021) examined patterns of multiple-victim shootings using a broader definition than the present study.¹ Although their primary focus was to examine the relationship between the occurrence of multiple-victim shootings and temperature, the authors found that rates of events were higher on certain holidays (Fourth of July, Christmas, and New Year’s), as well as on weekend days as compared to weekdays (Ruderman & Cohn, 2021). While the authors concluded that there was temporal variation in rates of multiple-victim shootings, they did not assess these patterns in hourly intervals, which may provide important context for further understanding the occurrence of such events. Similarly, they failed to disaggregate the events across location types to assess what differences, if any, were present in temporal patterns based on where they occurred.

Importantly, one of the challenges facing this line of inquiry is the absence of precise timing information encoded within the crime data (Ratcliffe, 2002). In order to analyze temporal patterns of crime, particularly in hourly intervals, the timing that these events occur must be known, which often is not the case (Felson & Poulsen, 2003). As Ratcliffe (2002) points out, time data for crimes are not likely to be captured unless the victim is present, the crime is witnessed by or directly reported to law enforcement, or it is detected by some form of technology (e.g., CCTV, cell phone recordings) that includes time stamping. Given, however, that in the context of mass public shootings, all three conditions are usually met (victims present, immediate reports to law enforcement, and time-stamped records), these events lend themselves to a focused temporal analysis of their occurrence.

The Current Study

The current study seeks to explore the temporal patterns of mass public shootings to assess with more precision than previous research when and where these events take place. We must first acknowledge that how mass shootings are defined—including the language used to describe such events (e.g., mass shooting, mass public shooting, mass murder, multiple victim shooting)—is a point of considerable contention within the scholarly community (for a thorough review, see Schildkraut & Elsass, 2016). As a result, there is no generally agreed-upon definition nor universal database to catalog such events (Bjelopera et al., 2013; Freilich et al., 2020).

At the center of this debate is how to conceptualize the “mass” in mass shootings (see Freilich et al., 2020). Early definitions included cases where four or more individuals were killed, and, depending on the source, this number either included (Krouse & Richardson, 2015) or excluded (Bjelopera et al., 2013) the perpetrator(s). Consequently, these studies focused on the most extreme events (Bjelopera et al., 2013) while simultaneously excluding other mass casualty incidents. For example, these studies both omitted the 1998 shooting at Thurston High School in Springfield, Oregon, where there were 27 total casualties—but only two of them fatalities. More recent cases, such as the 2019 STEM School shooting in Highlands Ranch, Colorado (1 killed, 8 injured), similarly would be excluded from research utilizing this threshold.

Several prominent publicly available databases of mass shootings, such as those compiled by The Washington Post (Berkowitz & Alacantra, 2021; see also Duwe, 2020) and the Associated Press/USA Today/Northeastern University (Fox, 2021), rely on the Congressional Research Service’s (Bjelopera et al., 2013) four-fatality threshold. The Mother Jones database (Follman et al., 2021), which also originally used this parameter, revised its definition in 2013 to include cases with three or more fatalities. Several sources include more narrow parameters, excluding shootings that occur in the commission of a crime, are related to gangs or drugs, or where the offenders target family and friends. Still others, such as the Gun Violence Archive (n.d.), include entirely different numerical parameters (four or more individuals shot, excluding the perpetrator) with no exclusionary criteria. Consequently, the number of events captured in these sources varies considerably, from just a few cases (Berkowitz & Alacantra, 2021; Follman et al., 2021) to two dozen (Fox, 2021) to more than 400 (Gun Violence Archive, n.d.) per year, on average.

For the present study, we employ Schildkraut and Elsass’s (2016) definition, which reflects a broadening of Newman et al. (2004) generally accepted definition of a school shooting (e.g., Muschert, 2007; Rocque, 2012). Noting similarities between rampage shootings both in and out of schools, Harris and Harris (2012) recommended this definitional expansion as part of a call to expand research on mass violence events. Specifically, Schildkraut and Elsass (2016) define mass (public) shootings as:

[A]n incident of targeted violence carried out by one or more shooters at one or more public or populated locations. Multiple victims (both injuries and fatalities) are associated with the attack, and both the victims and the location(s) are chosen either at random or for their symbolic value. The event occurs within a single 24-hour period, though most attacks typically last only a few minutes. The motivation of the shooting must not correlate with gang violence or targeted militant or terroristic activity. (p. 28)

The authors later clarified that the latter part of their definition, “targeted militant or terroristic activity,” excluded state-sponsored group killings (e.g., ISIS, al Qaeda) but not ideologically motivated lone-wolf actors who pledge allegiance to such groups (Schildkraut, 2021). Their definition (or variations of it) has been used by other scholars, including Freilich et al. (2020) and Greene-Colozzi and Silva (2020), and overcomes a number of challenges of these other databases in that it is narrow enough to exclude shootings that lack similarity to events like Columbine while being broad enough to encompass those that do, irrespective of an arbitrary number of fatalities (see also Schildkraut et al., 2018).

Using this definition, we built a comprehensive database designed to capture all mass public shootings occurring in the U.S. between 1966 and 2020. The start of the period represents the beginning of what Duwe (2004, 2020) refers to as the second wave of mass murder, which has been characterized by mass public shootings rather than by familicides that were more common in the 1920s and 1930s (the first wave). Consistent with previous research (e.g., Greene-Colozzi & Silva, 2020; Lankford, 2015; Schildkraut et al., 2018), we first culled existing databases of events, including the Federal Bureau of Investigation’s [FBI] active shooter reports (Blair & Schweit, 2014; FBI, 2016, 2018, 2019, 2020, 2021a, 2021b) and the aforementioned media and open-source outlets, for all cases that matched Schildkraut and Elsass’s definition. We then conducted additional exhaustive web-based keyword searches (e.g., mass shooter, active shooter, school shooter, rampage shooter) across multiple search engines (Google, Google Scholar, Yahoo, Lexis-Nexis, and ProQuest). The results were screened to identify cases that matched the definitional criteria, and those that could be cross-referenced across at least three independent sources were added to the database. The final database included 401 mass public shootings occurring between 1966 and 2020.

Using these same sources, several different variables were constructed for each case to assess when and where each shooting occurred. We identified the location of the shooting first relative to the perpetrator’s relationship to the place (e.g., schools for current or former students, workplaces for current or former employees). This relationship is important consider as it likely enhances the perpetrator’s familiarity with the place in respect to access and obstacles (e.g., security).² For shootings that were random, meaning that there was no clear connection between the perpetrator and the location, we coded the type based on the primary function of the space (e.g., place of worship, dining/nightlife establishment, shopping/entertainment). For the timing of the shootings, the day of the week was determined by searching the full date (month, day, year) in Google. The hourly interval was recorded based on the reported starting time of the shooting (as stated in media stories or after-action reports). As with the case inclusion, the reported start times of the events were cross-referenced across at least three sources prior to being recorded. Across all 401 cases, there were no missing data on any of these three variables.

Table 1 presents the descriptive statistics for the sample of cases included in the present study. More than half of the incidents occurred at workplaces (29.2%) and schools (25.2%). Mass shootings were more than five times more likely to happen on weekdays (Monday through Friday) than weekends (Saturday and Sunday), with Thursdays and Fridays having the greatest number of events. These events also were most likely to occur in morning (42.6%) and afternoon (32.7%) hours. They were least likely to happen in the early morning hours (12:00 a.m.–5:59 a.m.). The full distribution with hourly intervals is presented in the Appendix and these data are used in all analyses.

Table 1.

Descriptive Statistics for Key Variables (N = 401).

	n	%
Location type
School	101	25.2
Workplace	117	29.2
Place of Worship	16	4.0
Restaurant/Nightlife	29	7.2
Shopping/Entertainment	23	5.7
Government/Military	19	4.7
Multiple (Spree)	37	9.2
Other	59	14.7
Day of Week
Sunday	35	8.7
Monday	66	16.5
Tuesday	59	14.7
Wednesday	65	16.2
Thursday	73	18.2
Friday	72	18.0
Saturday	31	7.7
Time of Daya
Early Morning (12 am–5:59 am)	44	11.0
Morning (6 am–11:59 am)	171	42.6
Afternoon (12 pm–5:59 pm)	131	32.7
Evening (6 pm–11:59 pm)	55	13.7

Note. Frequency percentages may not total to 100.0% for each measure due to rounding error.

Due to space constraints, the time data are presented here in 6-hour intervals. For the full distribution by hour, please refer to the appendix.

Patterns of Timing of Mass Shootings

We first assessed the relationship between day of week and hour of the shooting. Figure 1 presents an abbreviated distribution with 6-hour intervals rather than hourly intervals for ease of understanding. The results of a Chi-square test indicate that there is a statistically significant relationship between the hourly interval in which the shootings occur and the day of week they happen (χ² = 181.88, Monte Carlo p < .01).³ Mass shootings occurred most frequently in the early morning hours on Saturdays and Sundays. More specifically, on Saturday mornings, the greatest number of incidents occurred between midnight and 1:59 a.m., while shootings were more likely to occur between 2:00 a.m. and 2:59 a.m. on Sundays. Those shootings that occur most frequently in the morning hours took place on Mondays and Thursdays. More specifically, the 7:00 a.m. to 7:59 a.m. hour on Mondays had more shootings occur than any other interval in the dataset (n = 11); Thursday morning shootings were more frequent between 10:00 a.m. to 10:59 a.m. Afternoon shootings were most likely to happen on Wednesdays (particularly between 1:00 p.m. and 1:59 p.m.) and Fridays (highest frequency between 3:00 p.m. and 3:59 p.m.). Evening shootings were most likely to take place on Wednesdays, though incidents were nearly evenly distributed across the hourly intervals. Shootings that occurred on Tuesdays were distributed somewhat evenly throughout the day, though less frequent during the evening hours. Notably, of the 168 unique hourly intervals in a week, no shootings occurred within 35 (20.8%) of them.

Figure 1.

Distribution of U.S. mass public shootings by day and time, 1966 to 2020.

We next considered the potential relationship between when (based on the day of the week) and where (location type) mass shootings occur. The results of a Chi-square test indicate that the occurrence of these events on specific days of the week does differ significantly based on the type of location (χ² = 103.93, Monte Carlo p < .001). To identify potential patterns, we plotted the distribution of mass shootings by day of week disaggregated by location type, which is presented in Figure 2. There are several interesting findings. Workplace shootings had a nearly even distribution throughout the weekdays (though highest on Wednesdays), with few events taking place on weekends. Similarly, school shootings also were most likely to occur on weekdays (slightly more pronounced on Mondays, Thursdays, and Fridays), though events did infrequently occur on weekends at extracurricular activities, such as sporting events or school dances. Conversely, shootings at places of worship were most likely to happen on Sundays, but they also occurred relatively frequently on Thursday and Saturdays. Mass shootings occurring at restaurant or nightlife establishments (e.g., bars, nightclubs) more commonly happened on Sundays, Wednesdays, and Saturdays, while those occurring at shopping and entertainment-centered locations (e.g., movie theaters, bowling alleys) were nearly evenly distributed through the full week, though highest on Fridays. Events occurring at government buildings were most likely to occur on Mondays and Thursdays, while those that spanned multiple locations (spree) took place with identical frequency on Tuesdays, Thursdays, Fridays, Saturdays, and Sundays. Shootings that spanned other locations—including, but not limited to, airports, medical facilities, gyms, supermarkets—were nearly identically distributed across all 7 days of the week.

Figure 2.

Distribution of U.S. mass public shootings by day and location type, 1966 to 2020.

Figure 3 shows how mass shootings are distributed across time of day in hourly intervals when disaggregated by location type. As with day of week, the results of a Chi-square test reveal a statistically significant relationship between the hourly interval and location type (χ² = 310.38, Monte Carlo p < .001). Shootings occurring at schools and workplaces were most likely to happen in the mornings and afternoons when such locations typically are open for normal operations. School shootings in particular occurred most frequently between 7:00 a.m. and 8:59 a.m. (arrival), 10:00 a.m. and 11:59 a.m. (lunch), and 2:00 p.m. and 3:59 p.m. (dismissal). Incidents taking place at these times accounted for more than 68% of all school shootings (25.7%, 19.8%, and 22.8%, respectively). Nearly 70% of workplace shootings occur between 6:00 a.m. and 9:59 a.m. (37.3%) and 2:00 p.m. and 5:59 p.m. (31.7%), when employees often are arriving or leaving. More specifically, the 8:00 a.m. to 8:59 a.m. hour reflected the highest incidence of events (14.5%), followed by the 3:00 p.m. to 3:59 p.m. interval (12.0%).

Figure 3.

Distribution of U.S. mass public shootings by time and location type, 1966 to 2020.

Other locations follow similar patterns. Mass shootings at places of worship are most likely to occur in the morning, with half of the incidents occurring between 10:00 a.m. and 11:59 a.m. Shootings at restaurants and nightlife establishments were more likely to happen in the late evening and early morning hours, with the greatest frequency of events occurring during the 10:00 to 10:59 p.m. (20.7%) and 2:00 to 2:59 a.m. (17.2%) intervals. Shopping/entertainment establishments were most likely to experience mass shootings in the afternoon and evening hours, with a combined 35% occurring at the 3:00 p.m. to 3:59 p.m. and 6:00 p.m. to 6:59 p.m. intervals. Government and military installations were most likely to experience mass shootings in the morning (8:00 a.m.–10:59 a.m. = 26.3%) and evening (6:00 p.m.–7:59 p.m. = 26.3%) hours. Like workplaces, these time intervals may reflect employees arriving, leaving, or even a shift change, though the hourly intervals with the highest percentage of incidents (7:00 p.m.–7:59 p.m. = 21.1%; 10:00 a.m.–10:59 a.m. = 15.8%) occur slightly later in the day compared to more traditional employment locations. Spree shootings were most likely to occur in the morning (9:00 a.m.–10:59 a.m. = 18.9%) and afternoon (4:00 p.m.–5:59 p.m. = 16.2%) hours. Shootings at other locations were distributed throughout the day, likely due to the variability in types of places included in the category. Notably, nearly 60% of these incidents occurred between 9:00 a.m. and 5:59 p.m., the same timeframe in which many people typically engage in other routine activities like going to work or school. Trips to salons, gyms, or doctors’ offices also often are incorporated into individuals’ daily schedules, which may explain this finding.

A set of supplementary analyses considered the relationship between day and time (hourly interval) by location type.⁴ In first assessing the results of the Chi-square analyses for each location type, significant relationships were found only for schools (χ² = 173.76, Monte Carlo p < .01) and workplaces (χ² = 196.25, Monte Carlo p < .05), suggesting that there were more nuanced patterns of when the shootings occurred at each location. For the remaining location types, however, there was no statistically significant relationship found between the day and hour of occurrence.

As noted above, for schools, shootings were most likely to occur on Fridays, Thursdays, and Mondays. The supplementary analyses revealed specific intervals on these respective days when such incidents were most likely to happen. On Mondays, for instance, shootings were more likely to occur between 7:00 a.m. and 7:59 a.m. (28.6% of incidents happening on that day), followed by 10:00 a.m. to 10:59 a.m. and 2:00 p.m. to 2:59 p.m. (19.0% each). Conversely, on Fridays, shootings more frequently occurred later in the day, with most incidents occurring between 2:00 to 2:59 p.m. (21.7%) as well as the 10:00 a.m. to 10:59 a.m. and 3:00 p.m. to 3:59 p.m. intervals (17.4% each). On Thursdays, shootings were most likely to happen between the 8:00 a.m. and 8:59 a.m. interval (18.2%), followed by an even distribution (13.6% each) between the hours of 7:00 a.m. and 7:59 a.m., 11:00 a.m. to 11:59 a.m., and 3:00 p.m. to 3:59 p.m. As with the distribution in Figure 3, these intervals correspond to times associated with arrival, lunch, and dismissal.

Although workplace shootings were nearly evenly distributed across weekdays, there were specific times during the period where such incidents were more likely to occur. The highest frequency of incidents on each workday occurred during the morning hours, though at different intervals. Shooting on Tuesdays, Wednesdays, and Fridays were most likely to occur between 8:00 a.m. and 8:59 a.m. (representing 20.0%, 13.3%, and 20.0% of the day’s incidents, respectively), while those on Mondays and Thursdays most frequently happened during the 9:00 a.m. to 9:59 a.m. interval (20.0% and 16.7% of daily shootings, respectively). Shootings on Wednesdays also occurred with identical frequency to the morning hour between 10:00 a.m. and 10:59 a.m. as well as 3:00 p.m. to 3:59 p.m. (with each interval representing 13.3% of the day’s incidents).

Discussion

While there is growing attention among researchers to the issue of mass shootings, one neglected consideration is the temporal variation of these events. To fill this gap, we analyzed the timing (day of week and hourly intervals) of more than 50 years of mass public shootings with particular focus paid to the locations at which these events occur. At the same time, we consider the subsequent findings through the lens of routine activity theory not only to assess arguments previously raised by scholars (Elsass et al., 2016; Schildkraut et al., 2019), but to also guide corresponding policy recommendations. In line with previous research (e.g., Brantingham & Brantingham, 1981, 2013; Cohen & Felson, 1979; Felson & Poulsen, 2003; Hawley, 1950; Newton & Felson, 2015; Rossmo, 2000), our findings suggest that crime—in this case, mass public shootings—does not occur randomly but rather during the daily routine activities of the perpetrators and their victims and in line with the rhythms of the locations where they happen.

In first considering where such events occur, the majority (55%) of mass shootings in the dataset occurred at the perpetrators and victims’ workplaces or schools. These location types are where individuals typically spend the largest consecutive segment of their day. The Bureau of Labor Statistics (2020) estimates that employed individuals spend approximately 7.6 hours per day at their workplaces, and students attend classes at their schools for an average of 5.4 hours per day. Outside of sleeping, these activities consume the largest time share of a person’s day (Bureau of Labor Statistics, 2020).⁵ Even those locations beyond workplaces and schools, including, but not limited to, places of worship, shopping centers, entertainment venues, restaurants, and nightlife, still are places that are frequented within the context of people’s daily activities (e.g., leisure). Importantly, location selection by perpetrators who commit mass shootings may be a function not only of their and their intended targets’ routine activities, but also their familiarity with and relative ease of access to these spaces (see also Brantingham & Brantingham, 1993).

In addition to perpetrators selecting locations that are largely part of their and their intended targets’ routine activities, the timing of mass shootings also mirrors the rhythms associated with these places. Workplace, school, and government/military shootings, for example, are more likely to happen on weekdays rather than weekends, which is when most activities related to these spaces occur. Conversely, mass shootings at places of worship are more likely to occur on the weekends when services are being held. Attacks at shopping outlets, entertainment venues (e.g., movie theaters, bowling alleys), restaurants, and nightlife also are more likely on weekends (or the Friday before the weekend) when individuals may have more time to invest in recreational or leisure activities outside of school and work.

Freilich et al. (2020) have argued that “certain public settings – or parts of ones – at certain times are more at risk for a public mass violence incident” (p. 276), and our findings confirm this. The time of day, conceptualized in hourly intervals (see Felson & Poulsen, 2003; Newton & Felson, 2015), at which mass shootings take place also similarly largely reflects the temporal patterns of routine activities relative to the locations where they occur, particularly for workplaces and schools. Workplace shootings are more likely to occur at times when employees are arriving for or leaving work. Relatedly, most school shootings occur at times that correspond with arrival, lunch, and dismissal, while those attacks that occur at places of worship correspond to times when services are being held. What is particularly noteworthy about these specific time intervals is that these also are the periods within a given day that people congregate together in these locations—and certain spaces within them. As such, the timing of these shootings may be less about chance and more about a rational choice made by offenders based on a calculation about when they will be able to inflict the most harm.

In what is perhaps one of the more extreme examples, the perpetrators of the 1999 shooting at Columbine High School in Jefferson County, Colorado, recognized the patterning of their school day and factored this into their planning (Schildkraut & Muschert, 2019). Intent on causing the most harm possible, they surveilled the cafeteria—the place students were most likely to assemble in large quantity during a school day—prior to the shooting to find the exact minute that the space would have the most people in it. Upon determining that nearly 500 people would be in the cafeteria at precisely 11:17 am, they used this as the set point for the timers on the two 20-pound propane tank bombs (which failed to detonate) they placed in the space on the day of the attack. Applying this line of thinking to mass shootings beyond schools and workplaces also likely explains the timing of incidents in more open spaces, such as shopping areas and dining and nightlife establishments—the intervals at which shootings have occurred at the greatest frequencies in these locations also reflect the times patronage is at its highest. With more potential targets in these spaces, it increases the possible casualty counts (both fatalities and injuries) of the shootings.

It must be noted, however, that these findings should not be used to attempt to predict when mass shootings will occur. Given the small base rate of events, averaging approximately 20 per year (Schildkraut, 2021), there is not sufficient power at this time to try and forecast future events (Borum et al., 1999; Rocque, 2012). In other words, these findings cannot be used to identify when or where future events will occur. Still, they provide important context about the rhythm of mass shootings, which adds to the growing research on these events. Specifically, the present study overcomes the challenges of studying temporal patterns of crime raised by Ratcliffe (2002) because the timing of each event in the population was known, and it expands on the findings of Ruderman and Cohn (2021) by accounting for both hourly intervals in the temporal patterns and consideration of the type of location where these events occur.

Policy Implications

In the aftermath of mass shootings, there is considerable focus from policymakers and the public alike on the offenders who committed these acts (Ruderman & Cohn, 2021). Efforts regularly are made to profile such individuals to predict who may carry out similar attacks in the future (Shapiro, 2018). From a routine activity perspective, however, such efforts fail given that there are countless possible motivations (Cohen & Felson, 1979; Hollis et al., 2013). Consequently, “there is no one-size-fits-all profile of who carries out a mass shooting in the United States” (Victor, 2018, para. 1), and policy responses that are narrowly focused on the perpetrators are likely to fail.

Rather than considering why crimes like mass shootings are committed, it is important to shift the focus to how they occur to guide prevention efforts (Felson & Clarke, 1998; Freilich et al., 2020; Mandala & Freilich, 2018). By situating mass public shootings in the context of routine activities, our findings highlight the importance of considering when the opportunity for these events to occur is increased. On certain days and at certain times, motivated offenders may be more likely to converge not only with suitable targets, but greater numbers of them compared to other days and times. To disrupt this convergence and reduce opportunity, response efforts should focus on increasing the presence of capable guardianship in these public spaces.

Situational crime prevention (SCP), which also draws upon routine activity theory (see Clarke, 1995), is one promising approach that can be used to guide efforts toward reducing the opportunity for mass shootings to occur (see also Freilich et al., 2020). SCP involves analyzing specific types of crime and the situational factors that lead to their commission, and then utilizing techniques designed to change those factors and subsequently reduce the opportunity for the crime to occur (Freilich & Newman, 2017). Techniques of SCP⁶ may be categorized into strategies that are designed to prevent offenders from committing crimes regardless of motivation or remove the opportunity altogether (referred to as “hard” interventions), and those that reduce situational factors that may increase an individual’s motivation to commit crime (called “soft” interventions; Cornish & Clarke, 2003). Hard intervention techniques emphasize increasing the effort for the offender, such as through target hardening, access control, or location entry/exit screening, and increasing the risks associated with the crime, such as through the extension of guardianship or increased natural and formal surveillance (Cornish & Clarke, 2003; Freilich & Newman, 2017). Soft interventions can include techniques designed to reduce the rewards associated with committing the crime, eliminating potential provocations, and removing excuses that can be used by the offender to justify their actions (Cornish & Clarke, 2003; Freilich & Newman, 2017). Importantly, as opportunities for crime vary across different situations, those techniques that will be the most appropriate to prevent the incident or mitigate its harms also will be context specific (Clarke, 1995).

In the context of mass shootings, certain SCP techniques may be more useful than others at reducing the opportunity for such events to occur. Increasing the presence of law enforcement (including school resource officers) or security personnel, for example, at the times that mass shootings are more likely to occur can serve dual purposes. First, increasing the presence of this capable guardianship—and, by extension, the supervision that often is lower at these critical times—can potentially thwart an attack if the offender perceives the risks to be increased. Second, were a shooting to still occur at such times, particularly as target density is higher due to people congregating together, having these individuals present can expedite response times, which can mitigate the potential harms caused (see also Freilich et al., 2020). This technique not only would work in locations like workplaces and schools but also more open spaces, such as malls, concerts, and other venues where there is added room to disperse crowds to avoid tighter clustering of potential targets.

Importantly, like other types of crime (e.g., Hewitt et al., 2020), our findings indicate that mass shootings are not uniformly distributed across place and time. While it is unlikely that public places will change anything specific to the location, particularly for a statistically rare event like mass shootings, better understanding of the timing of such incidents can allow for a focus on the assignment of resources (see also Felson & Poulsen, 2003). Locations such as workplaces, places of worship, malls, bars, nightclubs, and even supermarkets often hire private security or off-duty officers to keep employees and visitors safe. As these locations typically do not have unlimited financial resources, identifying those times where the place is the most vulnerable can serve to maximize the return on investment. Similarly, while schools may employ armed guards throughout the day, they may reassign personnel to specific locations within the building at times where there are more potential targets in open spaces. While, for example, multiple security personnel may be needed to conduct entry searches at arrival, fewer students will be entering the school later in the day; consequently, several of these guards could be reassigned to the cafeteria during lunch periods and supported by the presence of administrators.

Locations may supplement the presence of security through the use of target hardening approaches, which Clarke (1995) notes is “the most obvious way of reducing criminal opportunities” (p. 110), particularly in “soft targets” and “civilian-centric” places (Hesterman, 2015, p. 1) like those included in the present study. For schools and workplaces, this often takes the form of door locks, which has been noted to be the most effective life-saving device in a mass shooting situation (Martaindale et al., 2017; Sandy Hook Advisory Commission, 2015). While the presence of door locks does not prevent such events from occurring, these devices do help to minimize injury and the loss of life by creating distance between the perpetrators and their intended targets. Notably, in just three active shooter events has anyone been killed behind a locked door and in none of those instances was it because the door lock failed (Schildkraut & Muschert, 2019). Not only do door locks create physical distance, but they also add time on the clock for potential targets, which serves to reduce the rewards for the offender. With most mass shootings over in 5 minutes or less (Blair & Schweit, 2014; Blair et al., 2013), perpetrators do not have time to defeat the locks and instead focus on those targets who are more accessible. As such, place managers, such as school administrators and workplace supervisors, should focus on not only ensuring that classrooms and offices can be secured (and that occupants of these spaces have the necessary keys to be able to do so) but also that locations where targets converge in larger groups can similarly be secured. At Columbine, for example, although the perpetrators had an unprecedented 50 minutes in which they controlled the school, coupled with four firearms and nearly 100 improvised explosive devices, they never attempted to breach a locked door; instead, the majority of the fatalities and injuries that occurred that day happened in the school’s library—a space where dozens of students and teachers were congregated but one that could not be locked (Schildkraut & Muschert, 2019). The presence of door locks can be supplemented by lockdown drills and associated procedures to ensure that building occupants know how to correctly (and quickly) respond if a situation arises (Schildkraut & Nickerson, 2020).

Other SCP strategies also are regularly used at locations like schools and workplaces, such as controlled entry (e.g., single point entries, electronic access control, visitor management systems) and entry screening procedures (e.g., metal detectors; see, generally, Jonson, 2017; Schildkraut & Muschert, 2019). It bears noting, however, that the use of such techniques does not always prevent mass shootings from occurring. Sandy Hook Elementary School utilized entry control techniques on the day of the attack, but the perpetrator bypassed them by shooting through the glass entryway and accessing the building (Sandy Hook Advisory Commission, 2015). Similarly, Red Lake High School had entry screening in place on the day of their attack; the perpetrator shot and killed the security guard working the metal detector and proceeded into the school (Connolly & Harris, 2005). Having a layered approach (e.g., door locks, additional security personnel) in such situations is crucial to mitigating harm and providing redundancy in case one mechanism fails.

Additionally, the choice of techniques employed also must account for more nuanced situational factors. Controlled entry options, for instance, may be successful at keeping out visitors to a location but would do little to stop attackers who can access the building in a manner that legitimately bypasses the technique, such as an employee who swipes into their workplace with their credentials or a student who comes in during the arrival period prior to the school narrowing to single point entry for the day. These options also may be less successful in more open and less regulated spaces like malls, restaurants, and nightclubs that have multiple entrances with no restrictions on access. As such, locations should evaluate what specific techniques may be most appropriate for their respective environmental and opportunity structure challenges (see also Freilich et al., 2020). Further, as offenders adapt to their changing surroundings (Duntley & Buss, 2011), continual assessment of SCP techniques employed through this dynamic process will be needed.

Finally, policy initiatives must be grounded in research, and our findings highlight an important consideration in this context. Whereas Ruderman and Cohn (2021) found that mass shootings occurred three times as frequently on weekends, our findings showed the opposite—that such incidents were more likely to happen on weekdays. The research on mass shootings, as noted, suffers from a lack of agreement on how to define such events (see also Bjelopera et al., 2013; Freilich et al., 2020), and the differences in definitions and the corresponding databases used in the two studies almost certainly account for their disparate findings. At the same time, accounting for the circumstances surrounding mass shootings—or failing to do so, as in the Ruderman and Cohn (2021) study—can have important implications for how best to understand the opportunity for such events to occur in the context of routine activity theory. For example, in focusing on any instance of gun violence where four or more people, excluding the perpetrator, were shot (injured or killed), Ruderman and Cohn (2021) suggest that the greater occurrence of such incidents on weekends is due to the variability in routine activities. In the present study, however, we focus on a more nuanced segment of premeditated mass gun violence incidents that omits more common—and often more spontaneous and random—types of incidents such as felony and gang shootings. By accounting for this context, our findings suggest that instead it may be the greater predictability of routine activities such as employment and education, which make up the largest share of a person’s time, that facilitates the occurrence mass public shootings. As researchers continue to work to identify and assess prevention and harm mitigation strategies, accounting for this context will be critical.

Limitations

The present study, however, is not without its limitations. First, although the database used is one of the most comprehensive (Schildkraut, 2021), it is possible that not every single mass shooting that meets the definitional criteria has been captured due to a reliance on media accounts to identify such events. Second, this study focuses specifically on mass public shootings. It is possible that other types of multiple or mass victim shooting events, including those committed by gang members, that occur during felony or terrorist activities, or that happen in private spaces, may exhibit different temporal patterns from those uncovered here. Future research may wish to explore such considerations. Similarly, there may be differences in the temporal patterns, both broadly and relative to location type, based upon situational (e.g., lone vs. multiple offenders) and offender-specific (e.g., cited mental illness, variation in motivation) characteristics. The low base rate of incidents and high rate of missing data, however, prevent us from being able to assess such considerations in the present study; as more information and cases become available, future studies may analyze such possibilities. Finally, the analyses do not consider any additional factors beyond the time and place dimensions of these events, such as the presence of security, which may be a factor that differentiates not only outcomes (e.g., response time or time to incident conclusion; casualty rates) but also whether these events occur in the first place (e.g., failed vs. completed incidents). By better understanding the various factors that converge to create the opportunity for such tragedies to occur, we can continue to work toward the opportunity to prevent them.

Footnotes

Appendix

Distribution of U.S. Mass Public Shootings by Day and Time, 1966 to 2020.

	Day of Week
	Sunday	Monday	Tuesday	Wednesday	Thursday	Friday	Saturday	Total
Time of Day
00:00–00:59	2	2	1	2	1	3	3	14 (3.5)
01:00–01:59	1	1	1	0	0	2	3	8 (2.0)
02:00–02:59	5	0	1	1	1	0	2	10 (2.5)
03:00–03:59	1	0	2	0	0	0	0	3 (0.7)
04:00–04:59	1	0	1	1	0	1	2	6 (1.5)
05:00–05:59	0	0	1	0	0	1	1	3 (0.7)
06:00–06:59	0	3	2	1	2	3	0	11 (2.7)
07:00–07:59	0	11	3	5	6	3	3	31 (7.7)
08:00–08:59	2	7	7	7	6	5	1	35 (8.7)
09:00–09:59	0	6	6	1	6	5	3	27 (6.7)
10:00–10:59	3	7	3	6	9	8	2	38 (9.5)
11:00–11:59	4	5	5	1	7	5	2	29 (7.2)
12:00–12:59	1	2	5	6	2	1	1	18 (4.5)
13:00–13:59	2	1	2	8	5	3	0	21 (5.2)
14:00–14:59	0	6	5	6	8	6	0	31 (7.7)
15:00–15:59	1	5	4	5	8	9	3	35 (8.7)
16:00–16:59	0	2	3	1	3	4	0	13 (3.2)
17:00–17:59	3	1	3	1	0	4	1	13 (3.2)
18:00–18:59	0	1	1	1	3	2	2	10 (2.5)
19:00–19:59	2	3	1	2	4	1	1	14 (3.5)
20:00–20:59	0	1	0	3	1	0	0	5 (1.2)
21:00–21:59	2	0	0	2	0	4	0	8 (2.0)
22:00–22:59	3	1	2	3	1	1	0	11 (2.7)
23:00–23:59	2	1	0	2	0	1	1	7 (1.7)
Total	35 (8.7)	66 (16.5)	59 (14.7)	65 (16.2)	73 (18.2)	72 (18.0)	31 (7.7)	401 (100.0)

Note. Row and column totals presented as frequencies with percentage of total cases in parentheses.

Declaration of Conflicting Interests

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

Funding

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

ORCID iD

Jaclyn Schildkraut

Notes

Author Biographies

Jaclyn Schildkraut is an Associate Professor of Criminal Justice at the State University of New York (SUNY) at Oswego. Her research interests include mass/school shootings, homicide trends, mediatization effects, moral panics, and crime theories. She is the co-author of Mass Shootings: Media, Myths, and Realities and Columbine, 20 Years Later and Beyond: Lessons from Tragedy, and has published in journals such as Journal of School Violence, Homicide Studies, American Journal of Criminal Justice, Journal of Qualitative Criminal Justice & Criminology, Crime, Law and Social Change, and Journal of Homeland Security and Emergency Management.

H. Jaymi Elsass is Lecturer in the School of Criminal Justice at Texas State University. Her primary research interests include episodic violent crime, moral panics, fear of crime, and juvenile delinquency. She is the co-author of Mass Shootings: Media, Myths, and Realities and has published in American Journal of Criminal Justice, Crime, Law and Social Change, Security Journal, and Criminology, Criminal Justice, Law & Society, Journal of Crime and Justice, and Sociological Inquiry, as well as several edited volumes.

Mark C. Stafford is a Professor at Texas State University’s School of Criminal Justice and Criminology. He also has been a Professor at Washington State University and the University of Texas at Austin, where he was a founder of the Center for Criminology and Criminal Justice Research. Professor Stafford has been a postdoctoral fellow at the Center for Advanced Study in the Behavioral Sciences at Stanford University and at the University of Colorado. He was the IEAT-FORD Chair of Criminality, Violence, and Public Policy in the Institute of Interdisciplinary Advanced Studies at the Federal University of Minas Gerais, Belo Horizonte, Brazil. Professor Stafford has published extensively on deterrence, victimization and fear of crime, and the causes of crime and juvenile delinquency. He is a co-author of the recently published book, Sex Crimes and Sex Offenders (Routledge).

Jacob Sheingold is a second year law student at Albany Law School and a graduate of the criminal justice program at the State University of New York at Oswego. His research interests include law, psychology, and crime trends.

References

Andresen

M. A.

Malleson

(2015). Intra-week spatial-temporal patterns of crime. Crime Science, 4, 12. https://doi.org/10.1186/s40163-015-0024-7

Arluke

Lankford

Madfis

(2018). Harming animals and massacring humans: Characteristics of public mass and active shooters who abused animals. Behavioral Sciences & the Law, 36(6), 739–751. https://doi.org/10.1002/bsl.2385

Berkowitz

Alacantra

(2021, May 21). The terrible numbers that grow with each mass shooting. The Washington Post. https://www.washingtonpost.com/graphics/2018/national/mass-shootings-in-america/

Bjelopera

J. P.

Bagalman

Caldwell

S. W.

Finklea

K. M.

McCallion

(2013). Public mass shootings in the United States: Selected implications for federal public health and safety policy. Congressional Research Service. https://sgp.fas.org/crs/misc/R43004.pdf

Blair

J. P.

Nichols

Burns

Curnett

(2013). Active shooter events and response. CRC Press.

Blair

J. P.

Schweit

K. W.

(2014). A study of active shooter incidents, 2000-2013. U.S. Department of Justice, Federal Bureau of Investigation. https://www.fbi.gov/file-repository/active-shooter-study-2000-2013-1.pdf/view

Borum

Fein

Vossekuil

Berglund

(1999). Threat assessment: defining an approach for evaluating risk of targeted violence. Behavioral Sciences & the Law, 17(3), 323–337. https://doi.org/10.1002/(SICI)1099-0798(199907/09)17:3<323::aid-bsl349>3.0.CO;2-G

Brantingham

P. L.

Brantingham

P. J.

(1981). Notes on the geometry of crime. In Brantingham

P. J.

Brantingham

P. L.

(Eds.), Environmental criminology (pp. 27–54). Waveland Press.

Brantingham

P. L.

Brantingham

P. J.

(1993). Nodes, paths and edges: Considerations on the complexity of crime and the physical environment. Journal of Environmental Psychology, 13(1), 3–28. https://doi.org/10.1016/s0272-4944(05)80212-9

10.

Brantingham

P. L.

Brantingham

P. J.

(2013). The theory of target search. In Cullen

Wilcox

(Eds.), The Oxford handbook of criminological theory (pp. 535–553). Oxford University Press.

11.

Bureau of Labor Statistics. (2020). American time use survey – 2019 results (Report No. USDL-20-1275). U.S. Department of Labor. https://www.bls.gov/news.release/pdf/atus.pdf

12.

Clarke

R. V.

(1995). Situational crime prevention. Crime and Justice, 19, 91–150. https://doi.org/10.1086/449230

13.

Cohen

L. E.

Felson

(1979). Social change and crime rate trends: A routine activity approach. American Sociological Review, 44(4), 588–609. https://doi.org/10.2307/2094589

14.

Connolly

Harris

J. F.

(2005, March 23). Rampage in Minn. mirrors other cases. The Washington Post. https://www.washingtonpost.com/archive/politics/2005/03/23/rampage-in-minn-mirrors-other-cases/a032d505-367e-4a52-a4b7-3b09d3c7cd6e/

15.

Cornish

D. B.

Clarke

R. V.

(2003). Opportunities, precipitators, and criminal decisions: A reply to Wortley’s critique of situational crime prevention. Crime Prevention Studies, 16, 41–96.

16.

Cusimano

Marshall

Rinner

Jiang

Chipman

(2010). Patterns of urban violent injury: A spatio-temporal analysis. PLoS One, 5(1), e8669. https://doi.org/10.1371/journal.pone.0008669

17.

Duntley

J. D.

Buss

D. M.

(2011). Homicide adaptations. Aggression and Violent Behavior, 16(5), 399–410. https://doi.org/10.1016/j.avb.2011.04.016

18.

Duwe

(2004). The patterns and prevalence of mass murder in twentieth-century America. Justice Quarterly, 21(4), 729–761. https://doi.org/10.1080/07418820400095971

19.

Duwe

(2020). Patterns and prevalence of lethal mass violence. Criminology & Public Policy, 19(1), 17–35. https://doi.org/10.1111/1745-9133.12478

20.

Elsass

H. J.

Schildkraut

Stafford

M. C.

(2016). Studying school shootings: Challenges and considerations for research. American Journal of Criminal Justice, 41(3), 444–464. https://doi.org/10.1007/s12103-015-9311-9

21.

Federal Bureau of Investigation. (2016). Active shooter incidents in the United States in 2014 and 2015. U.S. Department of Justice. https://www.fbi.gov/file-repository/activeshooterincidentsus_2014-2015.pdf/view

22.

Federal Bureau of Investigation. (2018). Active shooter incidents in the United States in 2016 and 2017. U.S. Department of Justice. https://www.fbi.gov/file-repository/active-shooter-incidents-us-2016-2017.pdf/view

23.

Federal Bureau of Investigation. (2019). Active shooter incidents in the United States in 2018. Advanced Law Enforcement rapid Response Training (ALERRT). Center at Texas State University and the Federal Bureau of Investigation, U.S. Department of Justice. https://www.fbi.gov/file-repository/active-shooter-incidents-in-the-us-2018-041019.pdf/view

24.

Federal Bureau of Investigation. (2020). Active shooter incidents in the United States in 2019. Advanced Law Enforcement rapid Response Training (ALERRT). Center at Texas State University and the Federal Bureau of Investigation, U.S. Department of Justice. https://www.fbi.gov/file-repository/active-shooter-incidents-in-the-us-2019-042820.pdf/view

25.

Federal Bureau of Investigation. (2021a). Active shooter incidents in the United States in 2020. Advanced Law Enforcement rapid Response Training (ALERRT). Center at Texas State University and the Federal Bureau of Investigation, U.S. Department of Justice. https://www.fbi.gov/file-repository/active-shooter-incidents-in-the-us-2020-070121.pdf/view

26.

Federal Bureau of Investigation. (2021b). Active shooter incidents: 20-year review, 2000-2019. U.S. Department of Justice. https://www.fbi.gov/file-repository/active-shooter-incidents-20-year-review-2000-2019-060121.pdf/view

27.

Felson

Clarke

R. V.

(1998). Opportunity makes the thief: Practical theory for crime prevention (Police Research Series No. 98). Home Office. https://popcenter.asu.edu/sites/default/files/opportunity_makes_the_thief.pdf

28.

Felson

Cohen

L. E.

(1980). Human ecology and crime: A routine activity approach. Human Ecology, 8(4), 389–406. https://www-jstor-org-443.web.bisu.edu.cn/stable/4602572

29.

Felson

Eckert

M. A.

(2019). Crime and everyday life (6th ed.). SAGE.

30.

Felson

Poulsen

(2003). Simple indicators of crime by time of day. International Journal of Forecasting, 19(4), 595–601. https://doi.org/10.1016/s0169-2070(03)00093-1

31.

Follman

Aronsen

Pan

(2021). US mass shootings, 1982-2021: Data from Mother Jones’ Investigation. https://www.motherjones.com/politics/2012/12/mass-shootings-mother-jones-full-data/

32.

Fox

J. A.

(2021, March 26). You’re right to be confused about the number of mass shootings. USA Today. https://www.usatoday.com/story/opinion/2021/03/26/there-has-been-significant-inconsistency-tallying-mass-shootings-column/6987322002/

33.

Freilich

J. D.

Chermak

S. M.

Klein

B. R.

(2020). Investigating the applicability of situational crime prevention to the public mass violence context. Criminology & Public Policy, 19(1), 271–293. https://doi.org/10.1111/1745-9133.12480

34.

Freilich

J. D.

Newman

G. R.

(2017). Situational crime prevention. Oxford Research Encyclopedia of Criminology and Criminal Justice. https://doi.org/10.1093/acrefore/9780190264079.013.3

35.

Greene-Colozzi

E. A.

Silva

J. R.

(2020). Contextualizing firearms in mass shooting incidents: A study of guns, regulations, and outcomes. Justice Quarterly. Advance online publication. https://doi.org/10.1080/07418825.2020.1818805

36.

Gun Violence Archive. (n.d.). General methodology. https://www.gunviolencearchive.org/methodology

37.

Harris

J. M.

Harris

R. B.

(2012). Rampage violence requires a new type of research. American Journal of Public Health, 102(6), 1054–1057. https://doi.org/10.2105/AJPH.2011.300545

38.

Hawley

(1950). Human ecology: A theory of community structure. The Ronald Press Company.

39.

Hesterman

(2015). Soft target hardening: Protecting people from attack. CRC Press.

40.

Hewitt

A. N.

Andresen

M. A.

Beauregard

Brantingham

P. L.

(2020). Dangerous times? A routine activities examination of the temporal patterns of sexual offenses over time. Justice Quarterly. Advance online publication. https://doi.org/10.1080/07418825.2020.1856399

41.

Hollis

M. E.

Felson

Welsh

B. C.

(2013). The capable guardian in routine activities theory: A theoretical and conceptual reappraisal. Crime Prevention and Community Safety, 15(1), 65–79. https://doi.org/10.1057/cpcs.2012.14

42.

Jonson

C. L.

(2017). Preventing school shootings: The effectiveness of safety measures. Victims & Offenders, 12(6), 956–973. https://doi.org/10.1080/15564886.2017.1307293

43.

Krouse

W. J.

Richardson

D. J.

(2015). Mass murder with firearms: Incidents and victims, 1999-2013. Congressional Research Service. https://sgp.fas.org/crs/misc/R44126.pdf

44.

Kurland

(2019). Arena-based events and crime: An analysis of hourly robbery data. Applied Economics, 51(36), 3947–3957. https://doi.org/10.1080/00036846.2019.1587590

45.

Kwon

Cabrera

J. F.

(2019). Socioeconomic factors and mass shootings in the United States. Critical Public Health, 29(2), 138–145. https://doi.org/10.1080/09581596.2017.1383599

46.

Lankford

(2015). Mass shooters in the USA, 1966–2010: Differences between attackers who live and die. Justice Quarterly, 32(2), 360–379. https://doi.org/10.1080/07418825.2013.806675

47.

Lankford

(2016). Are America’s public mass shooters unique? A comparative analysis of offenders in the United States and other countries. International Journal of Comparative and Applied Criminal Justice, 40(2), 171–183. https://doi.org/10.1080/01924036.2015.1105144

48.

Lin

P. I.

Fei

Barzman

Hossain

(2018). What have we learned from the time trend of mass shootings in the U.S.? PLoS One, 13(10), e0204722. https://doi.org/10.1371/journal.pone.0204722

49.

Lowe

S. R.

Galea

(2017). The mental health consequences of mass shootings. Trauma Violence & Abuse, 18(1), 62–82. https://doi.org/10.1177/1524838015591572

50.

Luca

Malhotra

Poliquin

(2020). The impact of mass shootings on gun policy. Journal of Public Economics, 181, 104083. https://doi.org/10.1016/j.jpubeco.2019.104083

51.

Maimon

Kamerdze

Cukier

Sobesto

(2013). Daily trends and origin of computer-focused crimes against a large university computer network: An application of the routine-activities and lifestyle perspective. The British Journal of Criminology, 53(2), 319–343. https://www-jstor-org-443.web.bisu.edu.cn/stable/23640017

52.

Mandala

Freilich

J. D.

(2018). Disrupting terrorist assassinations through situational crime prevention. Crime and Delinquency, 64(12), 1515–1537. https://doi.org/10.1177/0011128717718488

53.

Martaindale

M. H.

Sandel

W. L.

Blair

J. P.

(2017). Active-shooter events in the workplace: Findings and policy implications. Journal of Business Continuity & Emergency Planning, 11(1), 6–20.

54.

Meindl

J. N.

Ivy

J. W.

(2017). Mass shootings: The role of the media in promoting generalized imitation. American Journal of Public Health, 107(3), 368–370. https://doi.org/10.2105/AJPH.2016.303611

55.

Mooney

C. Z.

(1997). Monte Carlo simulation (Vol. 116). SAGE Publications.

56.

Muschert

G. W.

(2007). Research in school shootings. Sociology Compass, 1(1), 60–80. https://doi.org/10.1111/j.1751-9020.2007.00008.x

57.

Newman

K. S.

Fox

Harding

Mehta

Roth

(2004). Rampage: The social roots of school shootings. Basic Books.

58.

Newton

Felson

(2015). Editorial: Crime patterns in time and space: The dynamics of crime opportunities in urban areas. Crime Science, 4, 11. https://doi.org/10.1186/s40163-015-0025-6

59.

Ratcliffe

J. H.

(2002). Aoristic signatures and the spatio-temporal analysis of high volume crime patterns. Journal of Quantitative Criminology, 18(1), 23–43. https://www-jstor-org-443.web.bisu.edu.cn/stable/23366730

60.

Rocque

(2012). Exploring school rampage shootings: Research, theory, and policy. The Social Science Journal, 49(3), 304–313. https://doi.org/10.1016/j.soscij.2011.11.001

61.

Rossmo

D. K.

(1995). Place, space, and police investigations: Hunting serial violent criminals. In Weisburd

Eck

J. E.

(Eds.), Crime and place (pp. 217–235). Criminal Justice Press.

62.

Rossmo

D. K.

(2000). Geographic profiling. CRC Press.

63.

Ruderman

Cohn

E. G.

(2021). Predictive extrinsic factors in multiple victim shootings. The Journal of Primary Prevention, 42(1), 59–75. https://doi.org/10.1007/s10935-020-00602-3

64.

Sandy Hook Advisory Commission. (2015). Final report of the Sandy Hook Advisory Committee. http://www.governor.ct.gov/malloy/lib/malloy/SHAC_Doc_2015.02.13_draft_version_of_final_report.pdf

65.

Schildkraut

(2021). Can mass shootings be stopped? To address the problem, we must better understand the phenomenon. Rockefeller Institute of Government. https://rockinst.org/issue-area/2021-can-mass-shootings-be-stopped/

66.

Schildkraut

Elsass

H. J.

(2016). Mass shootings: Media, myths, and realities. Praeger.

67.

Schildkraut

Elsass

H. J.

Meredith

(2018). Mass shootings and the media: Why all events are not created equal. Journal of Crime and Justice, 41(3), 223–243. https://doi.org/10.1080/0735648x.2017.1284689

68.

Schildkraut

Muschert

G. W.

(2013). Violent media, guns, and mental illness: The three ring circus of causal factors for school massacres, as related in media discourse. Fast Capitalism, 10(1), 159–173. https://fastcapitalism.uta.edu/10_1/schildkraut10_1.html

69.

Schildkraut

Muschert

G. W.

(2019). Columbine, 20 years later and beyond: Lessons from tragedy. Praeger.

70.

Schildkraut

Naman

B. M.

Stafford

M. C.

(2019). Advancing responses to mass shootings using a routine activity approach. Crime Prevention and Community Safety, 21(4), 346–361. https://doi.org/10.1057/s41300-019-00077-3

71.

Schildkraut

Nickerson

A. B.

(2020). Ready to respond: Effects of lockdown drills and training on school emergency preparedness. Victims & Offenders, 15(5), 619–638. https://doi.org/10.1080/15564886.2020.1749199

72.

Schildkraut

Sokolowski

E. S.

Nicoletti

(2021). The survivor network: The role of shared experiences in mass shootings recovery. Victims & Offenders, 16(1), 20–49. https://doi.org/10.1080/15564886.2020.1764426

73.

Shapiro

(2018, February 22). Dissecting the distinctive profile of school shooters: ‘There always is a trail of what they are about to do.’ ABC News. https://abcnews.go.com/US/dissecting-distinctive-profile-school-shooters-trail/story?id=53197511

74.

Skeem

Mulvey

(2020). What role does serious mental illness play in mass shootings, and how should we address it? Criminology & Public Policy, 19(1), 85–108. https://doi.org/10.1111/1745-9133.12473

75.

Snyder

H. N.

(2000). Sexual assaults of young children as reported to law enforcement: Victim, incident, and offender characteristics. U.S. Department of Justice. https://bjs.ojp.gov/content/pub/pdf/saycrle.pdf

76.

Snyder

H. N.

Sickmund

(2006). Juvenile offenders and victims: 2006 national report. Office of Juvenile Justice and Delinquency Prevention. https://www.ojjdp.gov/ojstatbb/nr2006/downloads/NR2006.pdf

77.

Towers

Gomez-Lievano

Khan

Mubayi

Castillo-Chavez

(2015). Contagion in mass killings and school shootings. PLoS One, 10(7), e0117259. https://doi.org/10.1371/journal.pone.0117259

78.

Victor

(2018, February 17). Mass shooters are all different. Except for one thing: Most are men. The New York Times. https://www.nytimes.com/2018/02/17/us/mass-murderers.html