At this point in the course, students have the basic tools they need to critique a journal article. Please select one of the three articles I posted for this week's module on Canvas.Read through the a

Received 19 March 2016 Received in revised form 19 September 2016 Accepted 20 January 2017 Available online 27 February 2017 Keywords:

Mass school shootings School violence School size Student–teacher ratios School characteristics a b s t r a c t Background: School shootings have approached epidemic levels in recent years. While men- tal illness is undoubtedly involved in nearly all cases of mass school violence, we sought to determine how environmental context may exacerbate preexisting personal factors.

The present study investigated the associations between mass school shootings, school enrollment size, student–teacher ratios, and student transitions.

Method: Our sample consisted of twenty-two mass school shooting cases between January 1995 and June 2014. Information about school shootings was gathered using preexisting school shooting databases and news media reports. Using state and national databases, data regarding school size and student–teacher ratios of incident schools were collected.

Information about schools where shooters previously attended, as well as state average school statistics, were also obtained.

Findings: Schools where mass shootings occurred had significantly higher enrollments than their state average counterparts. Additionally, students who committed a mass school shooting were significantly more likely to have previously attended a school with a smaller student body and/or a lower than state average student–teacher ratio.

Conclusion: Our findings are consistent with previous literature indicating that smaller schools are less likely to experience acts of mass violence. Additionally, our results suggest that transitioning from a smaller, more supportive school to a larger, more anonymous school may exacerbate preexisting mental health issues among potential school shooters.

The results of this study have significant implications for educational policy reform.

© 2017 Western Social Science Association. Published by Elsevier Inc. All rights reserved. 1. Introduction Mass school shootings have become increasingly preva- lent over the last twenty years. In fact, Agnich (2015) reports that the number of mass school shootings nearly doubled in the thirty years spanning from 1981 to 2010.

Despite thoughtful and strategic attempts to implement ∗ Corresponding author at: Psychological Science, Vassar College, Box 53, 124 Raymond Ave., Poughkeepsie, NY 12604, USA.

E-mail addresses: [email protected] , [email protected] (A.A.

Baird). policies preventing school violence, reports of school shootings, stabbings, and beatings have continued. This suggests that there may be parts of the larger context of mass school-based violence that are not being consid- ered. While mass school violence is a rare occurrence and accounts for less than one percent of all annual homicides of youth ages five to eighteen (Borum, Cornell, Modzeleski, & Jimerson, 2010 ; Modzeleski et al., 2008 ) its ramifications are so severe that it necessitates further investigation.

Research on mass school violence began in earnest in the 1990s, the decade in which the tragic shootings at Columbine High School, Thurston High School, and West- side Middle School took place. Between these three mass http://dx.doi.org/10.1016/j.soscij.2017.01.0090362-3319/© 2017 Western Social Science Association. Published by Elsevier Inc. All rights reserved. 262 A.A. Baird et al. / The Social Science Journal 54 (2017) 261–270 shootings, 24 students were killed and nearly 80 students were injured. While school violence dates back much fur- ther than the 1990s and encompasses a more diverse range of interpersonal harm, mass school shootings continue to mark one of the rarest, albeit most horrific, subsets of vio- lent school incidents.

Despite recent research efforts, mass school shootings remain one of the least understood types of school violence.

Previous work has indicated that mass school violence is markedly different from other types of school violence that involve interpersonal disputes or are related to criminal behaviors such as drug dealing and gang violence (Agnich, 2015; Casella, 2001 ; Flores de Apodaca, Brighton, Perkins, Jackson, & Steege, 2012 ; Kimmel, 2008 ; Verlinden, Hersen, & Thomas, 2000 ). While school violence occurs more fre- quently in low-income, urban schools that are populated mostly by students of color, research has shown that the vast majority of these incidents are rooted in interper- sonal disputes and usually only involve a small number of students (Kimmel & Mahler, 2003; Muschert, 2007; Rocque, 2012 ). On the other hand, mass school violence is more frequently characterized by a white, middle-class male(s) entering a rural/suburban school with the inten- tion of harming a large number of individuals who hold symbolic value, and/or are usually more socially distant from the shooter (Kimmel & Mahler, 2003; Newman, 2004; Rocque, 2012; Thompson & Kyle, 2005 ). Unpredictable and fatally destructive events such as these help to explain why so many students feel unsafe at school on a daily basis ( Kingery, Coggeshall, & Alford, 1998 ).

Following the horrific shootings of the 1990s (e.g., Columbine, Jonesboro), a number of task forces were created through the Secret Service, the Department of Edu- cation, and other governmental agencies, with the purpose of researching and preventing school violence. Previous sci- entific investigations of mass school violence have focused primarily on describing the characteristics of individual perpetrators (Fein et al., 2002 ; Leary, Kowalski, Smith, & Phillips, 2003 ; O’Toole, 1999 ). These studies have sought to construct a specific profile of individual(s) at greatest risk for committing an act of mass violence at their school.

In spite of failing to deliver a single predictive profile of a school shooter, this line of inquiry has been quite success- ful in identifying a number of consistent qualities among most school shooters. For example, nearly all perpetrators are male and have a history of mental illness and/or famil- ial instability (Farrington, 2007; Flores de Apodaca et al., 2012 ; Harding, Fox, & Mehta, 2002 ; Thompson & Kyle, 2005 ). Additionally, the vast majority of school shooters report feelings of exclusion, social isolation, rejection, and even abuse at the hands of peers (Böckler, Seeger, Sitzer, & Heitmeyer, 2011 ; Farrington, 2007; Harding et al., 2002; Leary et al., 2003; Rocque, 2012 ). Peer-to-peer conflicts are particularly impactful for middle and high school students, as adolescents begin to shift their focus from parent sup- port to peer networks during this period of development.

As a result, adolescents consistently show a heightened sensitivity to peer evaluations, and feel instances of social exclusion more deeply (Somerville, 2013 ). Thus, although parental relationships remain important throughout devel- opment, peer interactions may be the most important predictors of long-term social and emotional stability dur- ing the teen years (Brown, 2004; Levin & Madfis, 2009 ).

Although the aforementioned retrospective studies have generated reasonable and consistent explanations for why students commit acts of mass violence within their schools, they have not been particularly useful in predict- ing future school shootings. Psychopathology and access to firearms have been the focus of much study and contro- versy in the discussion of school violence. Though these factors are undoubtedly involved in nearly all cases of school shootings (Fein et al., 2002; Langman, 2011 ), they similarly lack considerable predictive value. For example, while factors such as mental illness, access to firearms, prior victimization, and rejection are consistently cited as perpetrator characteristics, many school-aged children possess some, or even all, of these risk factors and yet only an infinitesimal percentage of these students commit acts of mass violence. In other words, due to the complexity of the personal, environmental, and situational differences that accompany acts of mass violence, at present no com- bination of risk factors can definitively predict whether a violent incident will actually occur. By definition, any type of risk factor analysis is merely an evaluation of the like- lihood that a specific phenomenon will occur. One might think of a risk factor analysis in terms of a balloon. Each risk and protective factor affects an individual differently.

No single risk factor can fill the balloon entirely with air, and other protective factors might deflate the balloon slightly; however, a combination of factors can fill the balloon to its maximum capacity until it eventually pops.

While the discovery of a reliably predictive formula for school shootings is highly unlikely, expansion of poten- tial risk factor analyses would provide greater predictive value and could potentially aid in the creation of environ- ments that are less conducive to school violence. Though numerous researchers have studied individual factors asso- ciated with school violence, few have examined the effects of the environmental contexts in which school shootings take place. The sparse literature that exists has indicated that school violence is significantly more likely to occur at schools with higher enrollment and larger student–teacher ratios (DeVoe et al., 2003; Flores de Apodaca et al., 2012; Kaiser, 2005 ). Larger schools with less faculty support may precipitate violent behaviors due to a number of character- istics associated with these school environments; however, issues of anonymity and support may be the factors of greatest import.

By virtue of one’s placement within a large group set- ting, students attending schools with high enrollments may be more likely to experience feelings of anonymity. The idea that physical and emotional distance between indi- viduals significantly affects behavior has been explored frequently in the literature. Many researchers have found that altruistic behaviors decrease as social proximity decreases, as exemplified by the fact that individuals are more likely to help close friends or family members than they are strangers (Fry, 2008; Rachlin & Jones, 2008 ). Pre- vious work also indicates that empathy, which may be inversely correlated with social distance, is a significant inhibitor of violent aggression (Björkqvist, Österman, & Kaukianinen, 2000 ; Jolliffe & Farrington, 2004; Miller & A.A. Baird et al. / The Social Science Journal 54 (2017) 261–270 263 Eisenberg, 1988; Unger, 1992 ). Furthermore, it has been reported that empathy may play an important role in reducing the likelihood of violence amongst mentally ill populations (Harris & Picchioni, 2013 ). Many researchers have indicated that individuals are more likely to engage in antisocial behavior when they believe their identity is anonymous and/or that they will not be held accountable ( Nogami & Takai, 2008; Rogers & Ketchen, 1979; Wright, 2013 ). This research underscores the diffusion of respon- sibility that occurs when the distance between potential victims and perpetrators is vast. With increased physical and emotional distance, potential perpetrators can more easily dehumanize those around them, and are thus less likely to evaluate the consequences of their actions in terms of real human life. Large schools, and the increased anonymity that comes with them, are also associated with a perceived lack of responsibility (Goldkind & Farmer, 2013 ), which could not only limit a shooter’s perception of their responsibility to care for their school community, but may also inhibit other students or teachers from taking action if a shooter indicates a threat of violence prior to an incident occurring.

Large schools are also associated with increased alienation and isolation, which can trigger feelings of loneliness—an emotion to which adolescents are particu- larly vulnerable (Böckler et al., 2011; Flores de Apodaca et al., 2012; Page, 1990 ). In the case of school violence, feelings of alienation and isolation coupled with a sense of anonymity may lead at-risk individuals to believe that the only route to relief from these feelings is a public display of violence. There is considerable evidence that media cover- age of school shootings has had a profound influence on the way in which some at-risk youth manifest their pathology.

The relentless media coverage of mass school shootings has not only given instant national recognition to students who commit these acts, but has also offered a blueprint of sorts for at-risk individuals who crave recognition in the face of social exclusion or isolation (Zimring & Hawkins, 1997; Lawrence & Mueller, 2003 ). Thompson and Kyle (2005) explain how peer interac- tions foster moral development through the observation and replication of modeled behaviors. Without substantive opportunities for social learning through a peer-mediated system of rewards and consequences, adolescents will likely experience delayed moral development which, in turn, facilitates a vicious cycle of continued social exclu- sion and increased anti-social behavior. Combined with the fact that larger schools are prone to decreased stu- dent satisfaction scores and lower participation rates, a student’s continued lack of social integration can lead to the development of the types of depressive and/or aggressive tendencies frequently cited by those who are personally familiar with the perpetrators of mass violence (Pittman & Haughwout, 1987 ).

In cohesive schools with high levels of teacher support, subtle indications of potentially violent behavior including bullying, victimization, breach of school conduct, weapon- carrying, and expressions of violent intentions (Flores de Apodaca et al., 2012 ), may be addressed quickly and successfully by teachers, administrators, peers, or other members of the school community. In large schools, how- ever, individuals with social or behavioral difficulties may be overlooked and, as a result, may “fall through the cracks”.

Failing to address potential sources of student distress not only places the vulnerable student at risk for engaging in antisocial behaviors, but also further contributes to stu- dents’ perceptions of anonymity and isolation within the school environment. Levin and Madfis (2009) discuss predispositions to engaging in mass school violence through a model of chronic and acute strain, whereby chronic strains lay the groundwork for violent or aggressive behavior, while acute strains serve as the immediate precipitants of a destructive act. For example, an individual who struggles with psy- chopathology, generally experiences a lack of integration into the larger peer network, and receives little support from the school community (chronic stressors) may be driven over the edge by single-event “catalysts” such as receiving a bad grade on an exam, not securing a spot on a sports team, or breaking up with a significant other (acute stressors). On a broader scale, decreased support for indi- viduals in response to both chronic and acute strains could potentially create a tipping point for those on the edge of psychological decompensation. Again, though the com- bination of risk factors is often unique to the individual, any combination of chronic and acute stressors may pro- vide a dangerous cocktail for engaging in mass violence. In the context of school environments specifically, anonymity and lack of support could thus prove to be a lethal com- bination of chronic stressors, particularly for those already predisposed to other forms of chronic strain at home, or for those who experience an acute stress from an emotionally tumultuous life event.

One particularly stressful life event for the adolescent student is transitioning to a new school. Adjusting to any type of unfamiliar environment can be difficult for young people, but transitioning to a larger, less supportive envi- ronment may prove overwhelming, especially for students already struggling to find their place in a community of peers. It is conceivable that for adolescents who are already predisposed to feeling unimportant and alienated by virtue of their changing bodies, attitudes, and perhaps develop- ing psychopathology, confronting a large and impersonal environment for the first time may provoke toxic levels of hostility and anger. This idea, that the combination of biological predisposition and environmental factors wreak more havoc than either alone, is well established in the clinical psychology literature (see Rosenthal, 1970 for a review). The diathesis–stress model of mental illness posits that psychopathology emerges in people whose vulnera- bility to a disorder is exacerbated by environmental stress. Gentile and Bushman (2012) describe this idea in reference to aggression, explaining how biological males with a pre- disposition for aggressive behavior are even more likely to exhibit aggression when they have a history of victimiza- tion or low parental guidance in their lives.

In a relatively small classroom environment where there are fewer students demanding teachers’ social and emotional resources, students’ preexisting or developing vulnerabilities may be minimized by a supportive edu- cational setting that allows at-risk individuals to practice their social and emotional behavior and receive rapid and 264 A.A. Baird et al. / The Social Science Journal 54 (2017) 261–270 pertinent feedback, which in turn helps foster a health- ier physical and behavioral self. Students accustomed to a highly supportive system may feel jarred by its absence in a new school setting, particularly during periods of stress.

In other words, it is possible that transitioning to a larger school with less faculty support not only serves as a poten- tial stressor for students, but also removes the protective factors associated with once attending a small, supportive school.

Thus, we propose that large, impersonal school settings may create a unique and dangerously toxic environment for the individual at risk, and that transitions to these larger, less supportive schools from smaller, more personal schools may significantly contribute to violent outbursts.

Through their writings, videos, and eventually violent actions, school shooters are insisting that people pay atten- tion to them. The young people who commit acts of large-scale school violence are all collectors of injustices, and these injustices tend to revolve around the perception of being ignored or put down by teachers and peers in their school and by feelings of anonymity and lack of support.

The present study seeks to further explore the asso- ciations between school size, student–teacher ratios, and mass school shootings based on case studies that occurred within the last twenty years. We hypothesized that schools with larger than average enrollments and/or higher than average student–teacher ratios would be more prone to acts of mass violence. We also predicted that school shoot- ers would be significantly more likely to have transitioned from a smaller than average school to a larger than average school and/or from a school with a lower than aver- age student–teacher ratio to a school with a higher than average student–teacher ratio (i.e., high support to low support). In addition to analyzing enrollment size and student–teacher ratios of the schools where shootings occurred, this study is the first of its kind to incorpo- rate information regarding perpetrators’ previous school characteristics, highlighting the potential impact of transi- tioning from a smaller, more supportive environment to a larger, less supportive environment. 2. Methods The focus of the present study was on well-documented cases of mass school shootings in the United States from January 1995 to June 2014. We began in 1995 because this was the year in which shootings began to be covered more frequently and extensively in the media. We drew and cross-referenced cases from widely-reported national news media, preexisting school shooting databases, and previously published works detailing a thorough his- tory of school shootings (Newman & Fox, 2009 ; see Sommer, Leuschner, & Scheithauer, 2014 for review; also see Appendix I for a complete list of utilized databases).

Because our research questions focused only on inci- dents in which adolescents committed acts of mass aggression against individuals within their school commu- nity, we developed specific criteria for selecting incidents in order to examine comparable cases. As previously men- tioned, school violence is a complex phenomenon that encompasses a wide variety of incidents, each of which holds vastly different motivations and implications. In order to more accurately identify the underpinnings of mass school shootings, it was necessary to limit the sample pool to a set of distinctly similar events. Though incidents of mass school violence involving knives, bombs, or other weaponry have occurred within the last 20 years, only cases which involved the use of guns were included in our sam- ple, as research has indicated considerable distinctions in the social and psychological underpinnings of mass attacks based on weaponry choice (Agnich, 2015 ).

The present study focuses only on shootings that occurred on middle or high school campuses, in episodes that injured or killed two or more students, teachers, or staff members. Importantly, we excluded cases in which victims were personally targeted due to interpersonal dis- putes, domestic violence, drug trafficking, or gang activity.

In other words, we only studied cases in which the shooter appeared to choose his victims at random. Employing a similar case selection rationale to Levin and Madfis (2009) , shootings that occurred off of school grounds, and shoot- ings perpetrated by individuals not currently enrolled at the targeted school, were also excluded from analysis.

These criteria yielded a total of 22 incidents that were used in our analyses; however, two of these incidents involved multiple shooters for a total of 24 perpetrators. Thus, while analyses regarding school characteristics drew from a sam- ple of 22 cases, descriptive statistics regarding perpetrator characteristics drew from a sample of 24 cases. In both multi-shooter cases, the perpetrators attended the same school prior to matriculating at the incident school, which allowed a single count of these schools to be used in our analyses (see Table 1 for a more comprehensive overview of shooting incidents). 2.1. Data acquisition Once the sample criteria were determined, data regard- ing the number of students enrolled at the schools where the incidents took place, as well as student teacher ratios of these schools, were located and recorded using public records of school characteristics. Data were also collected regarding the average school enrollment and student–teacher ratio for the state in which the incident occurred.

We also collected data about the size and student–teacher ratio of the school attended by the perpetrator prior to their matriculation at the school where the shooting took place. Information for this variable was collected by finding the perpetrator’s home address and cross-referencing it with district boundaries and feeder patterns (see Appendix II for a complete list of resources used in regard to school characteristics).

During data analysis, certain variables were re-labeled in order to assist during the reporting process. For the pur- pose of our study, “incident school” is synonymous with the school in which a shooting took place and “previous school” is defined by the school that a shooter attended prior to enrolling at the incident school. “School level” represents the distinction between high schools, middle schools, and elementary schools. “School category”, on the other hand, is used to describe the comparison between incident school A.A. Baird et al. / The Social Science Journal 54 (2017) 261–270 265 Table 1 Incident characteristics. Location Date Urbanicity of school locationAge of shooter(s)Race(s) of shootersTotal harm (deaths)Victim type Shooter suicide Lynnville, Tennessee November 15, 1995 Distant rural 17 White 3 (2) Both Moses Lake, Washington February 2, 1996 Remote town 14 White 4 (3) Both Bethel, Alaska February 19, 1997 Fringe rural 16 White/Alaska native 4 (2) Both West Paducah, Kentucky December 1, 1997 Fringe rural 14 White 8 (3) Students Jonesboro, Arkansas March 24, 1998 Fringe rural 11, 13 White, white 15 (5) Both Edinboro, Pennsylvania April 24, 1998 Fringe rural 14 White 4 (1) Both Springfield, Oregon May 21, 1998 Small city 15 White 27 (5) Students Richmond, Virginia June 15, 1998 Large suburb 14 Black 2 (0) Adults Littleton, Colorado April 20, 1999 Large suburb 17, 18 White, white 37 (14) Both X, X Conyers, Georgia May 20, 1999 Large suburb 15 White 6 (0) Students Fort Gibson, Oklahoma December 6, 1999 Distant town 13 White 4 (0) Students Santee, California March 5, 2001 Large suburb 15 White 15 (2) Students Red Lake, Minnesota March 21, 2005 Remote rural 16 Native American 8 (8) Both X El Cajon, California March 22, 2005 Large suburb 18 White 5 (0) Both Jacksboro, Tennessee November 8, 2005 Fringe rural 14 White 3 (1) Adults Reno, Nevada March 14, 2006 Midsize city 14 Latino 2 (0) Students Cleveland, Ohio October 10, 2007 Large city 14 White 4 (0) Both X Omaha, Nebraska January 5, 2011 Large city 17 Unknown (non-white) 2 (1) Adults X Taft, California January 10, 2013 Distant town 16 White 3 (0) Students Sparks, Nevada October 21, 2013 Small city 12 Latino 3 (1) Both X Roswell, New Mexico January 14, 2014 Fringe rural 12 White 2 (0) Students Troutdale, Oregon June 10, 2014 Large suburb 15 White 2 (1) Both X Note: Shooter names were intentionally excluded so as to avoid increasing media coverage and incentivizing violence. characteristics and state average characteristics, and serves as a pseudo experimental control.

Despite limitations in terms of causal inference, the case study method is the only appropriate method for study- ing a rare phenomenon like mass school-based violence for which experimental methods are not feasible (Leary et al., 2003 ). 3. Results 3.1. Incident characteristics As previously mentioned, our selection criteria yielded a total of 22 cases. Out of these 22 incidents, six occurred at middle schools, 14 at high schools, and two at com- bined middle and high schools (i.e., grades 5/6–12). Using urbanicity measures provided by the National Center for Education Statistics, we determined that most schools were situated in rural or suburban locales (77.27%), and only two cases occurred in large cities (9.09%). Deaths ranged from 0 to 14 people per shooting (M = 2.23, SD = 3.35), injuries ranged from 0 to 23 per shooting (M = 5.18, SD = 6.38), and the total number of victims ranged from 2 to 37 per shoot- ing (M = 7.41, SD = 8.91). Three cases involved the death or injury of adults only, eight of students only, and 11 of both students and adults.

All 24 shooters in our sample were male and ranged in age from 11 to 18 years old (M = 14.75, SD = 1.87). Seven perpetrators took their own lives during the school shoot- ings in question. Using data compiled from news media reports and previously published papers, we determined that the shooters in our sample were predominantly white (79.17%). Given that we did not interview the shooters, their relatives, or their friends, we were unable to mea- sure more personal characteristics and histories such as personality traits, socioeconomic status, and mental health (please see Table 1 for incident and shooter characteristics). 3.2. School characteristics Complete descriptives of school characteristics within our sample can be found in Table 2. Enrollment data were normalized using logarithmic transformation, while arcsine transformation was used to normalize student–teacher ratio data. Our analyses revealed that when compared with the average size school in their respective states, schools where shootings occurred had significantly higher enrollments, t(21) = 2.68, p = .02. Shoot- ers’ previous school enrollments, however, were found to be comparable to their state average counterparts. Using a repeated measures ANOVA we discovered a significant interaction between the size of the school (relative to state average) and school level, F(1,42) = 10.67, p < .01. In practi- cal terms, school shooters experienced a 95% increase in enrollment size during their transition from one school level to the next, while their state average counterparts only experienced a 25% increase in enrollment size (see Fig. 1).

Examination of student–teacher ratio data showed that schools where shootings occurred had student–teacher ratios comparable to those of their state average coun- terparts. Interestingly, however, schools attended prior to the school where the shootings occurred had significantly smaller student–teacher ratios than their state average counterparts, t(21) = −3.68, p < .01 (see Fig. 2). These find- ings support the notion that school shooters may have experienced more stressful school transitions, in that they moved from schools with significantly higher potential fac- ulty support to schools with significantly larger student bodies and fewer teachers per student. 266 A.A. Baird et al. / The Social Science Journal 54 (2017) 261–270 Table 2 School characteristics. Location Date Enrol SAEnrol SFRatio SASFRatio PrevEnrol PrevSAEnrol PrevSFRatio PrevSASFRatio Lynnville, TN Nov. 15, 1995 710 853 13:1 16:1 710 853 13:1 16:1 Moses Lake, WA Feb. 2, 1996 734 446 20:1 20:1 429 446 17:1 20:1 Bethel, AK Feb. 19, 1997 507 461 17:1 11:1 327 505 15:1 23:1 West Paducah, KY Dec. 1, 1997 599 637 19:1 19:1 485 550 20:1 20:1 Jonesboro, AR Mar. 24, 1998 395 480 13:1 19:1 590 364 13:1 20:1 Edinboro, PA Apr. 24, 1998 702 543 14:1 14:1 342 543 14:1 14:1 Springfield, OR May 21, 1998 1,500 722 25:1 21:1 537 521 19:1 18:1 Richmond, VA June 15, 1998 986 1,181 18:1 15:1 533 728 10:1 12:1 Littleton, CO Apr. 20, 1999 1,677 629 20:1 11:1 519 524 18:1 21:1 Conyers, GA May 20, 1999 1,567 1,205 16:1 10:1 877 830 16:1 17:1 Fort Gibson, OK Dec. 6, 1999 424 371 17:1 20:1 424 330 16:1 20:1 Santee, CA Mar. 5, 2001 1,669 999 25:1 23:1 615 752 16:1 21:1 Red Lake, MN Mar. 21, 2005 320 469 11:1 11:1 185 552 17:1 21:1 El Cajon, CA Mar. 22, 2005 2,867 999 25:1 23:1 881 862 19:1 21:1 Jacksboro, TN Nov. 8, 2005 1,439 868 18:1 15:1 516 600 15:1 15:1 Reno, NV Mar. 14, 2006 966 673 20:1 20:1 584 673 20:1 20:1 Cleveland, OH Oct. 10, 2007 270 733 9:1 17:1 192 466 19:1 17:1 Omaha, NE Jan. 5, 2011 1,557 691 10:1 13:1 884 264 16:1 13:1 Taft, CA Jan. 10, 2013 941 999 13:1 23:1 712 615 17:1 22:1 Sparks, NV October 21, 2013 710 673 19:1 20:1 493 673 16:1 20:1 Roswell, NM Jan. 14, 2014 654 381 17:1 15:1 449 381 15:1 15:1 Troutdale, OR June 10, 2014 2,806 592 25:1 23:1 1,025 426 16:1 20:1 M = 1,090.9 M = 709.3 M = 17.5 M = 17.2 M = 559.5 M = 566.3 M = 16.2 M = 18.5 Enrol = school enrollment.

SA = state average statistic in school of comparable level (i.e., elementary, middle, or high school).

SFRatio = student–faculty ratio.

Prev = school that shooter previously attended. Fig. 1. Enrollment by school category and school level; * = p < .05. 4. Discussion Our analyses affirmed previous reports that over the past 20 years, mass school shootings were more likely to occur at larger than average schools. Importantly, our anal- yses also revealed that mass school shootings occurred more frequently when students transitioned to a school larger than the one that they previously attended, and when students transitioned from a school with a lower than state-average student–teacher ratio to one with a higher student–teacher ratio. Our findings regarding school size are consistent with existing literature indicating that larger schools have higher rates of crime and violence (Arum & LaFree, 2008; DeVoe et al., 2003; Flores de Apodaca et al., 2012; Kaiser, 2005 ) and that student–teacher ratios can contribute to at-risk students’ proclivity towards violence ( Flores de Apodaca et al., 2012; Gottfredson & DiPietro, 2011 ). Moreover, our findings indicate that although a student’s current school environments may influence the likelihood of engaging in mass violence, the stress of a transition from a small, supportive school to a larger, inad- equately staffed school may be an important predictor of mass school shootings. A.A. Baird et al. / The Social Science Journal 54 (2017) 261–270 267 Fig. 2. Student–teacher ratio by school category and school level, * = p < .01. Interestingly, many researchers have highlighted the prevalence of mass school shootings in tight-knit, subur- ban communities and the relative infrequence of these events in large metropolitan areas (Levin & Madfis, 2009; Newman, 2004; Rocque, 2012 ). Placing these findings within the context of our own research, it is possible that students attending urban high schools are more accus- tomed to larger group settings. Whether walking to the grocery store or moving through the subway, those who grow up in urban environments are generally more accus- tomed to large numbers of people and have a great deal of experience in crowded, impersonal settings. In the model we are proposing, it is the transition from an intimate school setting (where each student receives a degree of personal attention from teachers and peers) to a large- scale high school (with its potential for relative anonymity) that is much more overwhelming than transitioning from one impersonal setting to another. Additionally, it could be posited that those accustomed to the diversity of a city might be more apt to accept individuals who do not fit the “mainstream” model of student, while smaller, more homogenous communities might be more prone to exclusionary practices which further isolate already ill- supported, vulnerable students. 4.1. Implications Our research indicates that creating smaller schools with lower student–teacher ratios could significantly decrease the frequency with which mass school shootings occur. Previous research has revealed that, aside from being associated with fewer incidents of school violence, smaller schools generally house students who demonstrate higher levels of academic achievement (Cotton, 1996 ; Darling- Hammond, Ancess, & Ort, 2002 , Kahne, Sporte, de la Torre, & Easton, 2008 ), higher reading comprehension (Lee & Smith, 1997 ), and even lower dropout rates (Kahne et al., 2008; Pittman & Haughwout, 1987; Werblow & Duesbery, 2009 ).

What’s more, a number of researchers have determined that smaller schools are actually cost effective, with smaller schools costing significantly less per graduate than their larger school counterparts (Arum & LaFree, 2008; Kaiser, 2005 ; Stiefel, Latrola, Fruchter, & Berne, 1998 ). Our own research indicated that, in our sample, both enrollment size and student–teacher ratio were positively correlated with per capita income for the city/town in which the school was situated. 1This finding implies that wealthier areas have larger schools and proportionally fewer teach- ers within these schools. Given the vast research indicating the benefits of smaller schools and increased faculty sup- port, it would be wise for districts, particularly those with the luxury of financial options, to break down their system of larger, factory-like schools into smaller, more intimate settings where students can thrive and remain supported during school transitions.

Of course, it will take much more than simply changing the size and student–teacher ratio of a school to implement real, effective change. School climate also plays a consider- able role in the development of a healthy atmosphere that fosters peaceful communication between school commu- nity members (Osher et al., 2004; Ozer, 2006; Verlinden et al., 2000 ). Schools that foster a collaborative (as opposed to competitive) learning environment, offer supportive ser- vices, and facilitate a sense of mutual respect between students and faculty members have been shown to expe- rience few, if any, bouts of violence on school grounds ( Böckler et al., 2011; Borum et al., 2010; Golkind & Farmer, 2013 ; Kneese, Fullwood, Schroth, Pankake, 2003 ; Pittman & Haughwout, 1987 ).

In addition to facilitating in-school reforms, commu- nities would be wise to consider creating or further developing community-wide programs to assist in men- tal health services, peer mentoring, academic tutoring, and other activities and programs geared towards cre- ating happier, healthier, and more socially integrated students. Arum and LaFree (2008) discuss the intersections of school and community programming, explaining how 1Incident school enrollment and town/city per capita income, r(20) = .43, p < .05; incident school student–teacher ratio and town/city per capita income, r(20) = .47, p < .03. 268 A.A. Baird et al. / The Social Science Journal 54 (2017) 261–270 individuals from districts with small schools and access to welfare services are significantly less likely to experi- ence adult incarceration. Clearly, the associations between physical and emotional well-being and school services are very complex. Focusing on an interdisciplinary approach to violence prevention that includes recognition of the psychological, situational, and larger environmental com- ponents of school violence is undoubtedly more effective than approaching the issue from any single perspective. 4.2. Study limitations The present study was conducted using a case study method, which has considerable drawbacks due to lack of experimental conditions. With this said, the methodol- ogy employed in our study proved the only appropriate strategy by which to compare and analyze the research questions at hand. We incorporated state average enroll- ments and student–teacher ratios in an effort to create a pseudo-control group and reference point against which to contrast the size and support networks of schools where mass shootings occurred.

The overwhelming majority of people who study mass school violence do so with the intention of learning how to prevent future incidents. Our findings contribute impor- tant pieces to the puzzle of predicting when and where mass school shootings will occur. With this in mind, it is important to reiterate that our findings offer their greatest predictive utility when considered among the myriad of additional personal, environmental, and situational factors known to contribute to school shootings. 4.3. Future directions It is important to note that school level, enrollment, and student–teacher ratio were the only types of school characteristics collected and reported in the present study.

School type (public, private, charter, alternative, etc.) and evaluations of school climate were not included, but could prove to be important variables in the assessment of risk factors. Future researchers would be wise to engage in a more nuanced examination of school characteristics. This type of analysis may provide greater insight into the inter- action of environmental and personal characteristics, and may further explain why students vary in their responses to transitions from one school to the next. Students tran- sitioning from alternative schools (which are often smaller than average) to more traditional schools may be a partic- ularly interesting subset of students to understand more clearly as these are the students who are most frequently predisposed to dysfunctional behaviors and who transition to and from schools of varying enrollment sizes and support levels.

Preliminary analyses indicated a possible association between school violence and a number of factors includ- ing seasonality, day of the week, geographic region, and per student expenditure. Due to a limited sample size and given that these factors are beyond the scope of the present dis- cussion, we kept analyses of these variables to a minimum.

However, given the potential for a more comprehensive risk factor analysis of school shootings, it would be prudent for future research to investigate the effects of these factors in a more integrative approach. This would entail a thor- ough analysis of personal, situational, and environmental variables, each of which represents a number of facets that are critical for a comprehensive predictive model of mass school violence.

Additionally, though we intentionally limited our sam- ple to a specific subset of school violence so as to ensure that the analyses were performed using comparable cases, future researchers might consider designing a comparative study taking into account disparities in various compo- nents of school violence. For example, researchers could compare school violence on the basis of school level (mid- dle vs. high school vs. postsecondary), weapon(s) involved, number of deaths, or presence of interpersonal dispute, to name a few potential points of reference. Comparative studies would allow for further differentiation between subsets of school violence, and could potentially aid in the targeted prevention of each type of case.

Lastly, and perhaps most importantly, our findings sug- gest that investigations into supportive school services (i.e., counseling) and programs that foster collaboration between community members and a deeper sense of school pride (i.e., athletics and other extracurriculars) may provide an avenue through which to better understand pro- tective mechanisms against school violence. Determining how these programs optimally function to assist students in integrating themselves into a broader community may be an important endeavor for future research. 5. Conclusion Our findings regarding school size, support, and tran- sitions offer an important contribution by highlighting contextual variables that may be used in predictive models of mass school shootings. Our data confirm previous inves- tigations that demonstrated a relationship between large schools and mass school violence. The present study adds to our understanding of context by showing that students who commit mass shootings in these larger schools are sig- nificantly more likely to have transitioned from relatively smaller, more supportive schools. This adds important nuance to the idea that larger schools are more prone to mass violence by demonstrating that it may not be size alone, but rather the stress associated with losing the interpersonal connections and social support provided in smaller school settings that best predicts mass school shootings.

Acknowledgements We are grateful to Shari H. Silver and Christopher G.

Baird for their assistance with preparing this manuscript.

This work was supported in part by a grant (awarded to AAB) from the Tatlock Endowed Fund for Strategic Fac- ulty Support, and fellowship funding (awarded to AAB and EVR) from Vassar College’s Undergraduate Research Sum- mer Institute. A.A. Baird et al. / The Social Science Journal 54 (2017) 261–270 269 Appendix I. KPTV-KPDX Broadcasting Corporation (2014). “Guns on campus”, Fox 12: Oregon, http://www.kptv.com/link/ 554163/guns-on-campus-campus-shootings.

Stop School Shootings (2013). “Map of school shoot- ings”, http://www.StopTheShootings.org.

Newman, K.S. (2004). Rampage: The social roots of school shootings. New York, NY: Basic Books.

Pearson Education (2014). Timeline of worldwide school and mass shootings. Infoplease, http://www.

infoplease.com/ipa/A0777958.html.

Sommer, F., Leuschner, V., & Scheithauer, H. (2014).

Dynamics in the development of school shootings—A systematic review. International Journal of Developmental Science, 8, 3–24.

Vossekuil, B., Reddy, M., Fein, R., Borum, R., & Modze- leski, W. (2000). Safe school initiative: An interim report on the prevention of targeted violence in schools. Washington, DC: U.S. Secret Service, National Threat Assessment Center. Appendix II. Common Core of Data (CCD), “State Nonfiscal Survey of Public Elementary/Secondary Education”, 1998–99 v.1b.

Great Schools, “School and District Boundaries Map”, http://www.greatschools.org/school-district-boundaries- map/.

Institute of Education Sciences: National Center for Education Statistics, “Search for schools and colleges”, http://nces.ed.gov/globallocator/.

Public Schools K-12, “Public School Statis- tics/Demographics”, http://www.publicschoolsk12.com.

US News and World Reports, “Best High Schools”, http://www.usnews.com/education/best-high-schools.

References Agnich, L. (2015). A comparative analysis of attempted and completed school-based mass murder attacks. American Journal of Criminal Jus- tice, 40(1), 1–22. Arum, R., & LaFree, G. (2008). Educational attainment, teacher–student ratios, and the risk of adult incarceration among U.S. birth cohorts since 1910. Sociology of Education, 81(4), 397–421. Björkqvist, K., Österman, K., & Kaukianinen, A. (2000). Social intelli- gence − empathy = aggression? Aggression and Violent Behavior, 5(2), 191–200. Böckler, N., Seeger, T., Sitzer, P., & Heitmeyer, W. (2011). School shootings:

International research, case studies, and concepts for prevention. New York, NY: Springer Science and Business Media. Borum, R., Cornell, D., Modzeleski, W., & Jimerson, S. (2010). What can be done about school shootings? A review of the evidence. Educational Researcher, 39, 27–37. Brown, B. B. (2004). Adolescents’ relationships with peers. In R. M. Lerner, & L. Steinberg (Eds.), Handbook of adolescent psychology. Hoboken, NJ:

John Wiley & Sons Inc. Casella, R. (2001). Being down: Challenging violence in urban schools. New York, NY: Teachers College Press. Cotton, K. (1996). Affective and social benefits of small-scale schooling. In ERIC Digest. Charleston, WV: ERIC Clearinghouse on Rural Education and Small Schools. Darling-Hammond, L., Ancess, J., & Ort, S. W. (2002). Reinventing high school: Outcomes of the coalition campus school project. American Educational Research Journal, 39(2), 639–673. DeVoe, J. F., Peter, K., Kaufman, P., Ruddy, S. A., Miller, A. K., Planty, M., et al. (2003). Indicators of school crime and safety: 2003. Washington, DC: U.S. Departments of Education and Justice. Farrington, D. P. (2007). Childhood risk factors and risk-focused preven- tion. In M. Maguire, R. Morgan, & R. Reiner (Eds.), The Oxford handbook of criminology. Oxford: Oxford University Press. Fein, R. A., Vossekuil, B., Pollack, W. S., Borum, R., Modzeleski, W., & Reddy, M. (2002). Threat assessment in schools: A guide to managing threatening situations and to creating safe school climates. Washington, D.C: United States Secret Service and United States Department of Education. Flores de Apodaca, R., Brighton, L. M., Perkins, A. N., Jackson, K. N., & Steege, J. R. (2012). Characteristics of schools in which fatal shootings occur.

Psychological Reports, 110(2), 363–377. Fry, D. P. (2008). Aggression and altruism. In Encyclopedia of violence, peace, and conflict (2nd ed., pp. 14–28). Elsevier. Gentile, D. A., & Bushman, B. J. (2012). Reassessing media violence effects using a risk and resilience approach to understanding aggression. Psy- chology of Popular Media Culture, 1(3), 138–151. Goldkind, L., & Farmer, G. L. (2013). The enduring influence of school size and school climate on parents’ engagement in the school community.

School Community Journal, 23(1), 223–244. Gottfredson, D. C., & DiPietro, S. M. (2011). School size, social capital, and student victimization. Sociology of Education, 84(1), 69–89. Harding, D. J., Fox, C., & Mehta, J. (2002). Studying rare events through qual- itative case studies: Lessons from a study of rampage school shootings.

Sociological Methods and Research, 31(2), 174–217. Harris, S. T., & Picchioni, M. M. (2013). A review of the role of empathy in violence risk and mental disorders. Aggression and Violent Behavior, 18(2), 335–342. Jolliffe, D., & Farrington, D. P. (2004). Empathy and offending: A system- atic review and meta-analysis. Aggression and Violent Behavior, 9(5), 441–476. Kahne, J. E., Sporte, S. E., de la Torre, M., & Easton, J. Q. (2008). Small high schools on a larger scale: The impact of school conversions in Chicago.

Educational Evaluation and Policy Analysis, 30(3), 281–315. Kaiser, D. A. (2005). School shootings, high school size, and neurobiological considerations. Journal of Neurotherapy, 9(3), 101–115. Kimmel, M. (2008). Profiling school shooters and shooters’ schools: The cultural context of aggrieved entitlement and restorative masculinity.

In B. Agger, & D. Luke (Eds.), There is a gunman on campus: Tragedy and terror at Virginia Tech. Lanham, MD: Rowman and Littlefield. Kimmel, M. S., & Mahler, M. (2003). Adolescent masculinity, homopho- bia, and violence: Random school shootings, 1982–2001. American Behavioral Scientist, 46(10), 1439–1458. Kingery, P. M., Coggeshall, M. B., & Alford, A. A. (1998). Violence at school:

Recent evidence from four national surveys. Psychology in the Schools, 35(3), 247–258. Kneese, C., Fullwood, H., Schroth, G., & Pankake, A. (2003). Decreasing school violence: A research synthesis. In M. S. Fishbaugh, G. Schroth, & T. R. Berkeley (Eds.), Ensuring safe school environments: Exploring issues, seeking solutions. Mahwah, NJ: Lawrence Erlbaum Associates,Inc. Langman, P. (2011). Thirty-five rampage school shooters: Trends, patterns, and typology. In N. Böckler, T. Seeger, P. Sitzer, & W. Heitmeyer (Eds.), School shootings: International research, case studies, and concepts for prevention. New York, NY: Springer Science and Business Media. Lawrence, R., & Mueller, D. (2003). School shootings and the main-bites- dog criterion of newsworthiness. Youth Violence and Juvenile Justice, 1(4), 330–345. Leary, M. R., Kowalski, R. M., Smith, L., & Phillips, S. (2003). Teasing, rejec- tion: and violence: Case studies of the school shootings. Aggressive Behavior, 29, 202–214. Lee, V., & Smith, J. (1997). High school size: Which works best and for whom? Educational Evaluation and Policy Analysis, 19, 205–227. Levin, J., & Madfis, E. (2009). Mass murder at school and cumulative strain.

American Behavioral Scientist, 52(9), 1227–1245. Modzeleski, W., Feucht, T., Rand, M., Hall, J., Simon, T., & Butler, L. (2008). School-associated student homicides—United States 1992–2006.

Morbidity and Mortality Weekly Report, 57(2), 33–36. Muschert, G. W. (2007). Research in school shootings. Sociology Compass, 1(1), 60–80. Newman, K. S. (2004). Rampage: The social roots of school shootings. New York, NY: Basic Books. Newman, K., & Fox, C. (2009). Repeat tragedy: Rampage shootings in American high school and college settings, 2002–2008. American Behavioral Scientist, 52(9), 1286–1308. Nogami, T., & Takai, J. (2008). Effects of anonymity on antisocial behavior committed by individuals. Psychological Reports, 102(1), 119–130. O’Toole, M. E. (1999). The school shooter: A threat assessment perspective.

Quantico, VA: Critical Incident Response Group, National Center for the Analysis of Violence Crime, Federal Bureau of Investigation. Osher, D., VanAcker, R., Morrison, G. M., Gable, R., Dwyer, K., & Quinn, M. (2004). Ecological perspectives and effective practices for combat- 270 A.A. Baird et al. / The Social Science Journal 54 (2017) 261–270 ing school aggression and violence. Journal of School Violence, 3(2/3), 13–37. Ozer, E. J. (2006). Contextual effect in school-based violence prevention programs: A conceptual framework and empirical review. The Journal of Primary Prevention, 27(3), 315–340. Page, R. M. (1990). High school size as a factor in adolescent loneliness.

The High School Journal, 73(3), 150–153. Pittman, R. B., & Haughwout, P. (1987). Influence of high school size on dropout rate. Educational Evaluation and Policy Analysis, 9(4), 337–343. Rachlin, H. A., & Jones, B. (2008). Altruism among relatives and non- relatives. Behavioural Processes, 79(2), 120–123. Rocque, M. (2012). Exploring school rampage shootings: Research, theory and policy. The Social Science Journal, 49, 304–313. Rogers, R. W., & Ketchen, C. M. (1979). Effects of anonymity and arousal on aggression. The Journal of Psychology: Interdisciplinary and Applied, 102(1), 13–19. Rosenthal, D. (1970). Genetic theory and abnormal behavior. New York, NY:

McGraw-Hill. Somerville, L. H. (2013). The teenage brain: Sensitivity to social evaluation.

Current Directions in Psychological Science, 22(2), 121–127. Sommer, F., Leuschner, V., & Scheithauer, H. (2014). Dynamics in the development of school shootings—A systematic review. International Journal of Developmental Science, 8, 3–24. Stiefel, L., Latrola, P., Fruchter, N. A., & Berne, R. (1998). The effects of size of student body on school costs and performance in NYC high schools. New York, NY: Institute for Education and Social Policy. Thompson, S., & Kyle, K. (2005). Understanding mass school shootings:

Links between personhood and power in the competitive school envi- ronment. The Journal of Primary Prevention, 26(5), 419–438. Unger, P. (1992). Causing and preventing serious harm. Philosophical Stud- ies: An International Journal for Philosophy in the Analytic Tradition, 65, 227–255. Verlinden, S., Hersen, M., & Thomas, J. (2000). Risk factors in school shoot- ings. Clinical Psychology Review, 20(1), 3–56. Werblow, J., & Duesbery, L. (2009). The impact of high school size on math achievement and dropout rate. The High School Journal, 92(3), 14–23. Wright, M. F. (2013). The relationship between young adults’ beliefs about anonymity and subsequent cyber aggression. Cyberpsychology, Behav- ior, and Social Networking, 16(12), 858–862. Zimring, F., & Hawkins, G. (1997). Crime is not the problem: Lethal violence in America. New York, NY: Oxford University Press.