University of Southern California Dissertations and Theses

Unraveling the threads of tragedy: a public health exploration of adverse childhood experiences as precursors to targeted school violence and protective factors for mitigation strategies

(USC Thesis Other)

Unraveling the threads of tragedy: a public health exploration of adverse childhood experiences as precursors to targeted school violence and protective factors for mitigation strategies

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content UNRAVELING THE THREADS OF TRAGEDY:
A PUBLIC HEALTH EXPLORATION OF ADVERSE CHILDHOOD EXPERIENCES
AS PRECURSORS TO TARGETED SCHOOL VIOLENCE
AND PROTECTIVE FACTORS FOR MITIGATION STRATEGIES
by
Alejandro Vargas, Jr.
A Dissertation Presented to the
FACULTY OF THE USC SOL PRICE SCHOOL OF PUBLIC POLICY
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF POLICY, PLANNING AND DEVELOPMENT
May 2024
©Copyright 2024 Alejandro Vargas, Jr.

ii
Dedication
In profound gratitude to the esteemed USC Sol Price School of Public Policy, I extend
my heartfelt appreciation to the remarkable individuals who have shaped my academic journey.
To the students who trod this path before me, your resilience and achievements have
paved the way, and I am profoundly humbled to follow in your footsteps.
To the dedicated professors, staff, and illustrious alumni of the Sol Price School, your
collective legacy has been my guiding light, inspiring me to reach higher and strive for
excellence.
A special acknowledgment to my dissertation chair, Dr. William D. Leach, PhD, and my
esteemed committee members, Dr. Astrid Heger, MD, and Dr. Bruce Hoffman, PhD, your
wisdom, unwavering support, and patience have been instrumental in the realization of this
dissertation. You have not only challenged my academic limits but have also been my steadfast
cheerleaders, instilling in me faith in my abilities. With immense gratitude, I submit this work as
a testament to the transformative influence of the USC Sol Price School of Public Policy.
To the valiant men and women of law enforcement whose unwavering commitment to
duty shines brightest in the most challenging of times, you stand as guardians in a world fraught
with uncertainties, embodying a resolve that remains steadfast regardless of the circumstances.
Your profession, a unique blend of courage, compassion, and unyielding tenacity, often unfolds
far from the public eye, where the true depth of your sacrifices and the weight of your
responsibilities are known only to your brethren in blue.
Each day, you don a badge not as a mere symbol but as a testament to a promise to
protect, to serve, and to uphold the fabric of our society, even when faced with adversity that
would deter others.

iii
In the quiet moments beyond the chaos and the clamor, where reflection meets reality,
your sense of duty resonates with a profound understanding that what you do is not just a job, but
a calling. This calling, often laced with unsung heroism and silent battles, is where true grit and
humanity intertwine, crafting a fraternity bound by an unspoken oath.
To you, the keepers of peace, navigators of turmoil, and upholders of justice, your selfgratification, deeply personal and profoundly noble, is a beacon of hope and safety in our
community. Your dedication is not just your strength but our nation’s as well, a relentless force
that endures, inspires, and safeguards the ideals we cherish.
To my dearest family and friends, the pillars of strength and the bedrock of my journey,
your unwavering love, steadfast support, and genuine desire to witness my success have been the
driving forces behind every triumph. In moments of triumph and struggle alike, your
encouragement has been my anchor, grounding me when the seas were rough and lifting me to
new heights when the skies were clear.
This achievement is not just mine but a reflection of the collective love and
encouragement that you have generously bestowed upon me. Your belief in my potential has
been the catalyst for my perseverance, and I am profoundly grateful for the countless sacrifices,
late-night conversations, and shared joys that have marked this path.
To my family and friends, your immeasurable impact on my life and accomplishments is
beyond words, and I submit this dissertation with heartfelt gratitude, knowing that every step was
taken with your love propelling me forward.
For Ave Maria.

iv
Acknowledgements
Committee in Charge of Candidacy
Dr. William D. Leach, PhD, is the committee chairperson and a faculty member at the Sol
Price School of Public Policy, University of Southern California.
Dr. Astrid Heppenstall Heger, MD, is the second committee member and a faculty
member of Clinical Pediatrics at the Keck School of Medicine, University of Southern
California, and the founder and executive director of the Violence Intervention Program.
Dr. Bruce Hoffman, D.Phil., is the third committee member and a faculty member at the
Edmund A. Walsh School of Foreign Service, Georgetown University.
Committee Members in More Detail
Dr. William D. Leach, PhD, is a professor in the online master’s program in public
administration at the Price School, and his research on collaborative policymaking and
implementation has been published in top public administration, public policy, and political
science journals. He has directed over $1.0 million of research sponsored by the National
Science Foundation and private foundations, and he has provided scientific and policy advice to
state and federal agencies, including the U.S. Government Accountability Office (GAO) and the
National Research Council.
Dr. Leach graduated magna cum laude from UC Berkeley, earned a master’s degree in
natural resource management from the University of Michigan, and obtained a PhD in
environmental policy from the UC Davis Graduate Group in Ecology.
Dr. Astrid Heppenstall Heger, MD, is a professor of clinical pediatrics at the USC Keck
School of Medicine and the founder and executive director of the Violence Intervention Program
(VIP) at Los Angeles County-USC Medical Center in East Los Angeles. Dr. Heger has a

v
bachelor’s degree and a master’s degree from the University of Southern California. She founded
the Center for the Vulnerable Child (CVC) in 1984 for the evaluation of child abuse. The CVC
was the world’s first medical-based child advocacy center, evaluating over 10,000 child abuse
and child sexual assault victims every year. This program has been replicated in hundreds of
programs around the world. Today, renamed the Violence Intervention Program (VIP), Heger
established the first-of-its-kind, “one-stop-shop” community Family Advocacy Center, providing
medical, mental health, protective, legal, and social services to victims of family violence and
sexual assault throughout Los Angeles County. The VIP also serves over 4,000 victims of elder
and dependent adult abuse. Dr. Heger serves as a consultant to the Los Angeles County Coroner
in cases involving a child’s death or sexual assault at all ages.
Dr. Bruce Hoffman, D.Phil., is a highly respected expert in the field of counterterrorism
and insurgency. He is currently the Shelby Cullom and Kathryn W. Davis senior fellow for
counterterrorism and homeland security at the Council on Foreign Relations. He also serves as a
tenured professor at Georgetown University’s Edmund A. Walsh School of Foreign Service.
With over four decades of experience studying terrorism and insurgency, Dr. Hoffman is
also a visiting professor of terrorism studies at the University of St. Andrews in Scotland. He
previously held the corporate chair in counterterrorism and counterinsurgency at the RAND
Corporation and was director of RAND’s Washington, D.C., office and vice president for
external affairs.
Dr. Hoffman has served as a commissioner on the Independent Commission to Review
the FBI’s Post-9/11 Response to Terrorism and Radicalization (9/11 Review Commission), where
he was a lead author of the commission’s final report.

vi
He has also advised on counterinsurgency to the Strategy, Plans, and Analysis Office at
Multi-National Forces-Iraq Headquarters in Baghdad, Iraq, and on counterterrorism to the Office
of National Security Affairs, Coalition Provisional Authority, in Baghdad in 2004.
In recognition of his contributions to the field, Dr. Hoffman has received numerous
awards and accolades. In 1994, he was awarded the U.S. Intelligence Community Seal
Medallion, the highest level of commendation given to a nongovernment employee, by the
director of central intelligence.
He is also the editor-in-chief of Studies in Conflict and Terrorism, the leading scholarly
journal in the field.
Dr. Hoffman is the author of several acclaimed books, including the revised and updated
edition of Inside Terrorism, which has been in print for two decades. He has also written The
Evolution of the Global Terrorist Threat: From 9/11 to Osama bin Laden’s Death and
Anonymous Soldiers: The Struggle for Israel, 1917–1947, which was awarded the Washington
Institute for Near East Studies’ gold medal for the best book published in 2015 on Middle
Eastern politics, history, and society.

vii
Table of Contents
Dedication....................................................................................................................................... ii
Acknowledgments.......................................................................................................................... iv
List of Tables....................................................................................................................................x
List of Figures................................................................................................................................ xi
Abstract......................................................................................................................................... xii
Vita Auctoris ................................................................................................................................ xiv
Chapter 1: Introduction....................................................................................................................1
Research Problem ................................................................................................................1
Study Purpose ......................................................................................................................3
Research Questions..............................................................................................................4
Key Research .......................................................................................................................6
Summary..............................................................................................................................9
Chapter 2: Literature Review: Public Health Theory ....................................................................10
Introduction........................................................................................................................10
Primordial Prevention ........................................................................................................12
Primary Prevention ............................................................................................................15
Identifying Risk Factors.........................................................................................16
Risk Factors at the Community Level—Primary Prevention ................................17
Protective Factors at the Community Level...........................................................18
Secondary Prevention (Intervention).................................................................................22
Necessary Inputs....................................................................................................23
The Roles of Different Agencies and Actors .........................................................28
Barriers to Intervention..........................................................................................34
Tertiary Prevention (Treatment).........................................................................................34
Summary............................................................................................................................35
Chapter 3: Literature Review: School Shootings...........................................................................37
Introduction........................................................................................................................37
Theory................................................................................................................................37
Background........................................................................................................................41
School Shooters .................................................................................................................45
Methods..............................................................................................................................49
Recommendations..............................................................................................................53
Comparisons with Other Groups .......................................................................................58
Summary............................................................................................................................59

viii
Chapter 4: Methods........................................................................................................................60
Introduction........................................................................................................................60
Research Design.................................................................................................................60
Data Sources ......................................................................................................................67
Instrumentation ..................................................................................................................68
Procedure ...........................................................................................................................71
Data Processing and Analysis............................................................................................71
Summary............................................................................................................................75
Chapter 5: Results: JRIC Comparisons and Time Analyses..........................................................76
Introduction........................................................................................................................76
School Perpetrators: Descriptive Statistics........................................................................76
Other Perpetrators: Descriptive Statistics..........................................................................86
Bivariate Analyses .............................................................................................................99
Analysis of Trends over Time..........................................................................................118
Summary..........................................................................................................................121
Chapter 6: Results: ACEs in JRIC...............................................................................................123
Introduction......................................................................................................................123
Law Enforcement Contacts..............................................................................................123
Violence to Self................................................................................................................128
Violence to Others............................................................................................................133
Violence from Others.......................................................................................................142
Vagrancy ..........................................................................................................................148
Conclusion .......................................................................................................................155
Chapter 7: Results: ACEs in TASSS............................................................................................157
Introduction......................................................................................................................157
Descriptive Statistics........................................................................................................157
Fisher’s Exact Tests .........................................................................................................160
Summary..........................................................................................................................200
Chapter 8: Results: Difference in Proportions Tests....................................................................202
Chapter 9: Results: Comparing TASSS with the General Population .........................................204
Chapter 10: Discussion and Conclusion ......................................................................................208
Introduction......................................................................................................................208
Discussion........................................................................................................................208
JRIC: School Versus Other Perpetrators..........................................................................208
JRIC: Adverse Childhood Experiences............................................................................210
TASSS: Adverse Childhood Experiences........................................................................211

ix
Difference in Proportions Tests........................................................................................213
TASSS Versus the General Population ............................................................................213
Hypothesis Tests ..............................................................................................................214
Previous Literature and Theory .......................................................................................215
Prevention-Oriented Approach ............................................................................215
Public Health Model ............................................................................................216
Other Theory........................................................................................................222
Leakage ................................................................................................................224
Limitations...................................................................................................................................225
Future Research ...........................................................................................................................226
Recommendations........................................................................................................................228
Conclusions..................................................................................................................................231
References....................................................................................................................................232

x
List of Tables
Chapter 5: Results: JRIC Comparisons and Time Analyses
Table 5.1 Summary of Main Factor: School Violence Perpetrators ........................................77
Table 5.2 Summary of Other Perpetrators by Dataset .............................................................87
Table 5.3 Summary of Main Factor: Other Perpetrators..........................................................88
Table 5.4 Cross-tabulation of Law Enforcement Contacts with Perpetrator Group ..............102
Table 5.5 Cross-tabulation of Emotional Distress with Perpetrator Group ...........................104
Table 5.6 Cross-tabulation of Suspicious Travel with Perpetrator Group .............................105
Table 5.7 Cross-tabulation of Violence to Others with Perpetrator Group............................106
Table 5.8 Cross-tabulation of Drug Abuse with Perpetrator Group.......................................107
Table 5.9 Cross-tabulation of Radicalized by Associate with Perpetrator Group..................109
Table 5.10 Cross-tabulation of Personal Grievances with Perpetrator Group .......................110
Table 5.11 Cross-tabulation of Extremist Media Consumption with Perpetrator Group.......111
Table 5.12 Cross-tabulation of Social Media Platform Used with Perpetrator Group ..........113
Table 5.13 Cross-tabulation of Level of Education with Perpetrator Group .........................116
Table 5.14 Regression Analysis of Incidents on Adjusted Year for School Violence Perps..119
Table 5.15 Regression Analysis of Incidents on Adjusted Year for Other Perpetrators.........120
Chapter 6: Results: ACEs in JRIC
Table 6.1 Cross-tabulations with Law Enforcement Contacts: Fisher’s Exact Tests............125
Table 6.2 Cross-tabulations with Violence to Self: Fisher’s Exact Tests..............................129
Table 6.3 Cross-tabulations with Violence to Others: Fisher’s Exact Tests..........................135
Table 6.4 Cross-tabulations with Violence from Others: Fisher’s Exact Tests.....................143
Table 6.5 Cross-tabulations with Vagrancy: Fisher’s Exact Tests.........................................150
Chapter 7: Results: ACEs in TASSS
Table 7.1 Frequencies of ACE Measures...............................................................................158
Table 7.2 Frequencies of ACE External Factors....................................................................159
Table 7.3 Frequencies of Environmental Sustainable Design Measures...............................160
Table 7.4 Cross-tabulation of Domestic Violence with K-12 Dropout..................................161
Table 7.5 Cross-tabulation of Domestic Violence with Street Gang Member.......................162
Table 7.6 Cross-tabulation of Domestic Violence with Criminal Record..............................163
Table 7.7 Cross-tabulation of Domestic Violence with Struck a Student..............................164
Table 7.8 Cross-tabulation of Workplace Shooting with Struck an Administrator................165
Table 7.9 Cross-tabulation of Workplace Shooting with Struck Someone Else ....................166
Table 7.10 Cross-tabulation of Psychological Issues with Street Gang Member..................167
Table 7.11 Cross-tabulation of Psychological Issues with Victim Gang Affiliated ...............168
Table 7.12 Cross-tabulation of Psychological Issues with Struck a Teacher.........................169
Table 7.13 Cross-tabulation of Psychological Issues with Struck an Administrator.............170
Table 7.14 Cross-tabulation of Divorced/Separated with K-12 Failure ................................171
Table 7.15 Cross-tabulation of Social Stratum with K-12 Failure.........................................172
Table 7.16 Cross-tabulation of Social Stratum with Street Gang Member............................173

xi
Table 7.17 Cross-tabulation of Social Stratum with Victim Gang Affiliated ........................174
Table 7.18 Cross-tabulation of Significant Family Problems with Criminal Record ............175
Table 7.19 Cross-tabulation of Significant Family Problems with Struck Someone Else.....176
Table 7.20 Cross-tabulation of Recent Death with K-12 Failure...........................................177
Table 7.21 Cross-tabulation of Recent Death with Struck a Student.....................................178
Table 7.22 Cross-tabulation of Recent Death with Struck an Administrator.........................179
Table 7.23 Cross-tabulation of Loss of Social Standing with Struck a Student ....................180
Table 7.24 Cross-tabulation of Loss of Social Standing with Struck an Administrator ........181
Table 7.25 Cross-tabulation of Peer Aggression with K-12 Suspension ...............................182
Table 7.26 Cross-tabulation of Peer Aggression with K-12 Failure ......................................183
Table 7.27 Cross-tabulation of Peer Aggression with Street Gang Member.........................184
Table 7.28 Cross-tabulation of Peer Aggression with Victim Gang Affiliated......................185
Table 7.29 Cross-tabulation of Peer Aggression with Struck a Teacher................................186
Table 7.30 Cross-tabulation of Peer Aggression with Struck a Student ................................187
Table 7.31 Cross-tabulation of Peer Aggression with Struck an Administrator ....................188
Table 7.32 Cross-tabulation of Metal Detector with K-12 Expulsion ...................................190
Table 7.33 Cross-tabulation of Metal Detector with K-12 Dropout ......................................191
Table 7.34 Cross-tabulation of Metal Detector with Struck a Student ..................................192
Table 7.35 Cross-tabulation of Metal Detector with Struck an Administrator ......................193
Table 7.36 Cross-tabulation of School Guard/Resource Officer with Criminal Record .......194
Table 7.37 Cross-tabulation of School Guard/Resource Officer with Struck an Admin. ......195
Table 7.38 Cross-tabulation of School Police Officer with K-12 Suspension.......................196
Table 7.39 Cross-tabulation of School Police Officer with Street Gang Member.................197
Table 7.40 Cross-tabulation of School Police Officer with Criminal Record .......................198
Table 7.41 Cross-tabulation of School Police Officer with Struck a Teacher .......................199
Table 7.42 Cross-tabulation of School Police Officer with Struck an Administrator............200
List of Figures
Chapter 5: Results: JRIC Comparisons and Time Analyses
Figure 5.1 Relationship between the Number of Incidents of School Violence and Year.....119
Figure 5.2 Relationship between the Number of Incidents of Other Violence and Year.......121

xii
Abstract
The research problems focused on in this study consisted of the following questions: What
factors or characteristics delineate perpetrators of school violence from perpetrators of other
forms of violence? What are the associations between adverse childhood experiences (ACEs),
external factors, and Environmental Sustainable Design measures? How do the JRIC and TASSS
samples differ? How does the TASSS sample compare with the general population?
A total of 15 research questions were developed in order to examine these research
problems. This study’s purpose was to reduce the incidence of school violence and, specifically,
the incidence of school shootings, particularly in the United States, while taking a public health
approach to the problem. This study examined perpetrators of school violence as an independent
group, as well as the similarities and differences between perpetrators of school violence and
perpetrators of other forms of violence in order to determine what factors might delineate
perpetrators of school violence both independently and in comparison with other perpetrators of
violence.
Further analyses examined associations between ACEs and related measures in both
datasets, as well as analyses comparing the JRIC and TASSS samples, and the TASSS sample
with the general population. A quantitative approach was taken in this study, incorporating both
descriptive and correlational methods, with perpetrators of school violence as well as other
perpetrators of violence examined on a descriptive level, and with these two groups compared
using inferential statistical tests primarily consisting of Fisher’s exact tests, along with
independent-samples t-tests and regression analyses. Associations between ACEs and related
measures were examined using Fisher’s exact tests, while comparisons between samples took the
form of difference in proportions tests.

xiii
The results of the analyses conducted for this study found significant differences when
comparing perpetrators of school violence with perpetrators of other forms of violence with
respect to age, level of education, contact with law enforcement, emotional distress, suspicious
travel, violence to others, drug abuse, having been radicalized by an associate, having had a
personal grievance, extremist media consumption, social media platforms used, and social media
activities engaged in. Significant associations between ACEs and related measures were found in
both samples, as well as significant differences between the JRIC and TASSS samples, as well as
between the TASSS sample and the general population.
These results allowed for a better understanding of perpetrators of school violence. These
results also allowed for a series of recommendations to be provided that, if implemented, should
substantially reduce the incidence of school violence, and specifically, school shootings in the
United States.

xiv
Vita Auctoris
Alejandro (Alex) Vargas, a distinguished individual with a commendable 28-year tenure,
has retired as a lieutenant from the Los Angeles Police Department. Mr. Vargas’s extensive
professional background encompasses multifaceted roles, such as patrol duties, uniformed crime
suppression, and investigations spanning a broad spectrum, including domestic violence, sexual
assault, homicide, mental health crises, stalking, workplace violence, school place violence, and
domestic and international terrorism.
Notably, he culminated his career as the deputy director of the Los Angeles Joint
Regional Intelligence Center (JRIC), underscoring his significant contributions to intelligence
operations.
Mr. Vargas holds a Bachelor of Science degree from California State University, Long
Beach, attesting to his foundational academic achievements. Furthermore, he earned a Masters in
Security Studies from the esteemed Naval Postgraduate School in Monterey, California,
augmenting his academic repertoire with specialized knowledge of security studies.

1
Chapter 1: Introduction
In this chapter, an overview of this study is presented. This includes the presentation of
this study’s research problem that was examined along with the research questions included
within this study. The overall purpose of this study and what it intended to contribute to this body
of literature and practice are also presented with a brief discussion of this study’s methods.
This chapter also includes a discussion of key literature in this area of research and the
important gap that remains in this literature, which highlights the need for further study to help
fill this gap.
Chapter 2 and Chapter 3 both consist of independent literature reviews, the first focusing
on public health theory, and the second focusing on school shooters themselves. Chapter 4
discusses this study’s methods in detail, with Chapters 5 through 9 presenting and discussing the
results of the analyses conducted for this study. Finally, Chapter 10 includes a discussion of how
the results of these analyses speak to this study’s research questions, and discusses which null
hypotheses were rejected, and which failed to be rejected on the basis of the results of the
analyses conducted. These results are also discussed alongside previous literature and theory, as
well as the implications of the results obtained in this study, limitations of this study, possibilities
for future research, and recommendations and conclusions.
Research Problems
This study’s first research problem is the following: What factors or characteristics
delineate perpetrators of school violence from perpetrators of other forms of violence? This
research problem is restated in greater specificity as a total of 15 research questions, along with
the same number of null and alternative hypotheses, so that this research problem could be

2
examined in finer detail. These research questions are presented later in this chapter as well as in
Chapter 4, with these null and alternative hypotheses presented in Chapter 4.
This study’s second research problem asks: What are the associations between Adverse
Childhood Experiences (ACEs), External Factors, and Environmental Sustainable Design
measures? Both the JRIC and The American School Shooting Study (TASSS) datasets were
examined in order to explore this research problem.
The third and fourth research problems included in this study consisted of the following:
How do the JRIC and TASSS samples differ? How does the TASSS sample compare with the
general population? The additional associated research questions are presented later in this
chapter, with the associated null and alternative hypotheses presented in Chapter 4.
Theoretically and conceptually, this study was built upon the foundation of public health
and its approach to better understanding the factors that precipitate school shootings. With this
study taking a prevention-based approach, its analysis was focused upon the exploration of
factors associated with perpetrators of school violence. This took the form of descriptive analysis
in which the available data on perpetrators of school violence were presented simply, with the
same being done with respect to perpetrators of other forms of violence.
In addition, bivariate analyses were also conducted in order to statistically examine the
extent to which perpetrators of school violence may be similar to, or different from, other
perpetrators. Further analyses were also conducted in order to determine whether significant
trends over time were present with respect to the number of cases of violence found.
While all of the above analyses were conducted on the JRIC dataset, additional analyses
examining ACEs and their correlates were conducted on both the JRIC and TASSS datasets.

3
Additional analyses consisted of difference in proportions tests comparing the JRIC and TASSS
samples and the TASSS sample with the general population.
By describing features of perpetrators of school violence and examining how they relate
to other perpetrators, it was an aim of this study to determine factors or characteristics that can
describe the school shooter, or more generally, the perpetrator of school violence.
While previous literature has already clearly shown that no detailed and specific “profile”
of the school shooter can be determined, perpetrators of school violence tend to share certain
commonalities. It was an aim of the study to further explore what these factors or commonalities
might be in the interest of more effectively being able to screen students and assist those who
may eventually conduct an act of school violence so that their path in this direction can be
redirected before a tragedy occurs.
Study Purpose
The purpose of the study, broadly, was to help reduce the incidence of school violence,
and specifically, school shootings, particularly in the United States. More specifically, this
study’s purpose was to examine perpetrators of school violence as an independent group and the
similarities and differences between perpetrators of school violence and perpetrators of other
forms of violence in order to determine what factors may delineate perpetrators of school
violence independently and in comparison with other perpetrators of violence.
Further areas encapsulated within this study’s purpose consisted of an examination of the
correlates of ACEs, an exploration of how the JRIC and TASSS samples differ, and how the
TASSS sample differs from the general population. The goal in conducting these analyses was to
provide a set of factors that could be applied to screen students for those who may be more likely

4
to perpetrate an act of school violence in the future, so that these individuals can receive the
assistance they need so that this future risk is minimized.
In this way, this study aimed to not only contribute to this body of literature, helping to
fill important gaps in our understanding of school violence, but also to provide empirically
supported evidence that can be applied to create more effective screens for those who have a
tendency towards violence, with the aim of reducing the incidence of school violence in
America.
Research Questions
As discussed earlier in this chapter, this study’s research problem incorporated the
following questions: What factors or characteristics delineate perpetrators of school violence
from perpetrators of other forms of violence? What are the associations between adverse
childhood experiences (ACEs), external factors, and environmental sustainable design measures?
How do the JRIC and TASSS samples differ? How does the TASSS sample compare with the
general population?
In total, 11 research questions were developed from this initial research problem focusing
on specific data included in the dataset analyzed in this study. These 11 research questions are as
follows:
• Research Question 1: Do incidents of school violence and incidents of other forms of
violence differ on the basis of year?
• Research Question 2: Do perpetrators of school violence and perpetrators of other forms
of violence differ with regard to demographics?
• Research Question 3: Do perpetrators of school violence and perpetrators of other forms
of violence differ with regard to weapons, the military, and law enforcement contact?

5
• Research Question 4: Do perpetrators of school violence and perpetrators of other forms
of violence differ with regard to online behavior and social media use?
• Research Question 5: Do perpetrators of school violence and perpetrators of other forms
of violence differ with regard to travel?
• Research Question 6: Do perpetrators of school violence and perpetrators of other forms
of violence differ with regard to violence?
• Research Question 7: Do perpetrators of school violence and perpetrators of other forms
of violence differ with regard to abuse?
• Research Question 8: Do perpetrators of school violence and perpetrators of other forms
of violence differ with regard to radicalization?
• Research Question 9: Do perpetrators of school violence and perpetrators of other forms
of violence differ with regard to grievances?
• Research Question 10: Do perpetrators of school violence and perpetrators of other forms
of violence differ with regard to extremist media consumption?
• Research Question 11: Do perpetrators of school violence and perpetrators of other forms
of violence differ with regard to other psychological and sociological measures?
Then, research questions 12 and 13 were developed from this second research problem,
What are the associations between Adverse Childhood Experiences (ACEs), External Factors,
and Environmental Sustainable Design measures?
• Research Question 12: Are Adverse Childhood Experiences (ACEs) associated with
External Factors?
• Research Question 13: Are Adverse Childhood Experiences (ACEs) associated with
Environmental Sustainable Design measures?

6
The final two research problems, How do the JRIC and TASSS samples differ? and How
does the TASSS sample compare with the general population?, are associated with the following
two research questions, respectively:
• Research Question 14: Do the JRIC and TASSS samples differ?
• Research Question 15: Does the TASSS sample differ from the general population?
All research questions were examined, and the associated null hypotheses were tested
statistically using a quantitative methodology that incorporated descriptive and correlational
components. While the descriptive analysis simply presented the data available separately for
perpetrators of school violence and for other perpetrators of violence, the correlational
component of these analyses were examined, primarily by using Fisher’s exact test, as to whether
there were any significant differences between these two groups of individuals with regard to the
data included in the available dataset. These tests enabled the determination of whether
perpetrators of school violence significantly differ from perpetrators of other forms of violence
with regard to the items included in this study’s 11 research questions.
The following research questions examining ACEs and their correlates were also
examined using Fisher’s exact test. The research questions asking how the JRIC and TASSS
samples differ and how the TASSS sample differs from the general population with both
examined using difference in proportions tests.
Key Research
While substantial research has been conducted in this area in the past, an opportunity
exists to fill important gaps that remain in this area of study. With regard to theory, psychological
theories have focused on cognitive, emotional developmental, and clinical factors in the study of
school shooters. This includes social learning theory, the frustration-aggression hypothesis, and

7
theories of aggression (Grøndahl & Bjørkly, 2016). Currently, no single model exists that
accurately predicts who will commit a school shooting (Grøndahl & Bjørkly, 2016), and no
specific profile has been developed for school shooters (Langman, 2013b).
While researchers have not been able to create a distinct profile of school shooters,
commonalities and factors do exist. Research has found that mass shooters tend to exhibit severe
childhood trauma, being in a crisis, mental health disorders, and leaking plans in advance
(Peterson, 2021). Risk factors among perpetrators of school violence include multiple motives
being present, including those of grievances with classmates; firearms acquired from the home;
psychological, behavioral, or developmental symptoms; an interest in violence; social stressors;
negative home life factors; being a victim of bullying; a history of school disciplinary actions;
prior contact with law enforcement; and other concerning behaviors (Alathari et al., 2019).
School violence is a growing problem in the United States, with mass shootings
becoming more frequent and with a greater number of deaths incurred over time (Peterson,
2021). Of the 167 mass shootings in the past 50 years in the United States, 20% occurred in the
past five years. The average death toll from mass shootings has increased from eight per year in
the 1970s to 51 per year currently, while 7.6% of mass shootings occurred in kindergarten
through 12th grade (elementary, middle, and high schools) and 5.3% took place in colleges or
universities (Peterson, 2021).
Studies have emphasized the importance of taking a prevention-oriented approach in
reducing targeted school violence in part as law enforcement is rarely able to effectively respond
until after such incidents are over, and with these incidents continuing to occur despite the
implementation of various physical security measures (Alathari et al., 2019).

8
Comprehensive targeted prevention programs have been recommended in schools to
reduce the incidence of these acts of violence. These programs aim to identify potential future
perpetrators of violence, determine their level of risk, and implement intervention strategies to
minimize that risk (Alathari et al., 2019). Alathari et al. (2021) also provide a set of
recommendations focusing specifically on targeted school violence, with these including a focus
on intervention by the school and other students, along with the community, parents, and
families.
Many recommendations have been provided by researchers in this area with regard to
how school shootings may be prevented, and these include improved mental health services in
schools, threat assessment, as well as prevention programming (Lenhardt et al., 2010; Lenhardt
et al., 2018). A public health approach to target school violence aims to reduce violence through
the nurturing of healthy individuals while also maintaining their civil and privacy rights and
other civil liberties (Brown, 2022).
The phases, or tiers, of a public health model consist of the following: primordial (or
upstream) prevention, primary prevention, secondary prevention (or intervention) and tertiary
prevention (or treatment). The Centers for Disease Control and Prevention (CDC) lists risk
factors at the individual, family, social, and community levels (CDC, 2020), while factors that
influence the risk of targeted violence can be delineated as either “push factors” or “pull factors”
(Weine et al., 2017). The former factors drive people away from normal everyday life, while the
latter draw people toward violent behavior. With regard to school shootings, the main pull factor
might be the alleviation of a grievance.
While a great deal of effort has been made in order to differentiate school shooters from
other perpetrators of violence, in reality, similarities still exist (Ferguson et al., 2011). Primarily,

9
they differ with regard to the location of their attacks (Lankford, 2013). However, they also differ
as suicide terrorists are mentally healthy and motivated by ideology, while school shooters tend
to be mentally troubled and motivated by personal problems (Carey, 2007). Beyond this
distinction, school shooters share similarities with suicide terrorists, suggesting that they may
have similar psychologies and behaviors (Lankford, 2013).
All of the above research points to an important gap in the literature that this study hopes
to fill. Specifically, how are perpetrators of school violence similar to and different from other
perpetrators of violence and the general population? After examining the correlates of ACEs,
how can these results be applied in order to better screen potential perpetrators of school
violence so that they can be provided with the assistance they need to avoid a violent act from
ever occurring.
Summary
This chapter outlined the background of the important problem examined in this study,
the purpose of this study, and what this study aimed to contribute to the body of extant literature
and in terms of more practical contributions to the field.
The research questions posed in this study were also presented here with an overview of
this study’s methods and some key research studies that this current study drew upon. Finally, the
need for further research in this area, which this study aimed to provide, was discussed. The next
two chapters of literature reviews will discuss public health theory in relation to the important
problem of school violence, and the second literature review chapter specifically examines
literature relevant to school shootings and their perpetrators.

10
Chapter 2: Literature Review—Public Health Theory
Introduction
Mass shootings in the United States have become more frequent in recent decades, with
more deaths occurring at most of these incidents (Peterson, 2021). A total of 20% of all 167 mass
shootings in the past 50 years in the United States have taken place in the past five years, with
more than half of these 167 mass shootings since 2000 and one-third since 2010. While the
average death toll from mass shootings was eight per year in the 1970s, it is currently 51 per year
(Peterson, 2021).
Of the 167 mass shootings in the last 50 years examined by Peterson (2021), 7.6%
happened in kindergarten through 12th grade (elementary, middle, and high schools), and 5.3%
happened in colleges or universities (Peterson, 2021). While researchers have been at a loss at
trying to create a distinct profile of school shooters, commonalities and factors exist. Mass
shooters as a group have suffered severe childhood trauma, been in a crisis, exhibited mental
health disorders (especially psychosis), and leaked plans in advance. Conceptual links have also
been found between suicide and homicide (Peterson, 2021).
Studies have also emphasized the efficacy and the importance of taking a preventionoriented approach to reducing targeted school violence (Alathari et al., 2019). This is particularly
important due to the fact that law enforcement rarely and effectively responds until after such
incidents are over, even as these incidents are continuing to occur despite the implementation of
various physical security measures in schools.
Comprehensive targeted prevention programs have been recommended in schools to
prevent these acts of violence by identifying potential future perpetrators of violence,
determining their level of risk, and implementing intervention strategies to minimize that risk

11
(Alathari et al., 2019). A low threshold for intervention has also been suggested so that schools
can identify at-risk students before their behavior reaches the point at which it may negatively
impact other students’safety (Alathari et al., 2019).
Risk factors, or commonalities, among perpetrators of school violence have multiple
motives: grievances with classmates; firearms acquired from the home; psychological,
behavioral, or developmental symptoms; an interest in violence; social stressors; negative home
life factors; a victim of bullying; a history of school disciplinary actions; prior contact with law
enforcement; and other concerning behaviors (Alathari et al., 2019). These same factors were
generally echoed by Alathari et al. (2023) when studying perpetrators of mass violence more
generally.
In this more recent paper, she suggests the following in order to reduce the incidences of
violence: community bystander reporting; concern over individuals exhibiting an unusual interest
in violent topics; workplace violence prevention plans in businesses, with an intervention for
those who present a risk of violence; a careful consideration of strategies for resolving
interpersonal grievances; immediate intervention in cases where an individual is sharing final
communications or committing final acts; communities helping individuals who are managing
stressful life circumstances, mental health problems, or a personal crisis; and noting that many
who commit mass shootings are legally prohibited from carrying firearms (Alathari et al., 2023).
Alathari et al. (2021) also provides a set of refocused recommendations pertaining
specifically to targeted school violence, with these generally consisting of those
recommendations given for general mass violence. However, she also focuses here on
intervention on the part of the school and other students, as well as the community, parents, and

12
families. She also highlights the importance of school resource officers in preventing targeted
school violence (Alathari et al., 2021).
A public health approach to targeted school violence seeks to prevent elevated threats of
domestic violence (National Security Council, 2021) through the nurturing of healthy individuals
in a community while maintaining the civil rights, privacy rights, and civil liberties of its
members (Brown, 2022). This approach focuses on the health and well-being of the communities
in which potential perpetrators of school shootings operate, as well as the perpetrators
themselves. As such, this is a departure from the go-to approach based on surveillance and law
enforcement and extends the onus of responsibility from the shooter and law enforcement out
toward the community, social change groups, and advocates on both sides of the political
spectrum (Alathari et al., 2021). The focus is also on understanding the policies and
circumstances that lead to an attack, rather than focusing primarily on the perpetrators
themselves.
This public health approach focusing on prevention has strong potential, and its focus
aimed at mitigating the elevated threat of school violence pays close attention to all phases of a
potential attack. A public health approach would also reflect the government’s “multi-faceted”
response to domestic terrorism at the federal level (National Security Council, 2021). The phases
or tiers of a public health model are primordial (or upstream) prevention, primary prevention,
secondary prevention (or intervention), and tertiary prevention (or treatment), and these tiers are
discussed in further detail in the following sections.
Primordial Prevention
The primordial, or upstream, prevention tier of a public health model is concerned with
the baseline policies and conditioning of community well-being, which influence the health of

13
everyone in the community. The means to intervene in a potential school shooting are available
at this stage through the creation of a healthy, functioning community.
The primordial prevention tier can be thought of as promoting the protective factors
known to reduce the risk of targeted violence, such as school shootings. These include policies
for reducing drug and alcohol use within the community, reducing adverse childhood
experiences (ACEs), and implementing policies addressing mental health issues and dealing with
bullying.
According to the USSS, over one-third of school shooters have a history of substance use
or abuse (Alathari et al., 2021). A society or community with a low prevalence of substance or
alcohol abuse is therefore less likely to witness targeted violence. Substance Abuse and Mental
Health Services Administartion (SAMHSA) states that policies for mitigating drug and alcohol
use include laws against hate crime and policies on alcohol sales. SAMHSA also posits that
faith-based resources and after-school activities are important community-based protective
measures against drug and alcohol use, with after-school activities being one of the support
service functions provided by the Center for Prevention Programs and Partnerships’ (CP3) Local
Prevention Frameworks (Brown, 2022).
Adverse childhood experiences (ACEs) are a known driver of substance and alcohol
abuse. Campbell et al. (2016) found that behavioral symptoms typical of an unhealthy
community and certain medical complaints are associated with ACEs. The study identified links
between specific forms of abuse or negative experiences and certain negative outcomes.
Behavioral symptoms linked to ACEs include heavy drinking, smoking, and risky behavior. In
particular, strong links were established between verbal abuse and heavy drinking, and between
sexual abuse/verbal abuse and smoking/risky behavior. Medical conditions linked to ACEs

14
include diabetes, heart problems, strokes, and depression. Specifically, sexual and verbal abuse
categories were found to be associated with diabetes and depression. A public health approach to
school shootings at the primordial level should therefore include policies designed to reduce the
prevalence of ACEs, and so, in turn, reduce the prevalence of troubling community behavior and
medical complaints known to lead to targeted violence.
Mental illness is another primordial factor that can be managed with the goal of reducing
the risk of targeted violence, including that of school shootings. Over two-thirds of school
shooters show evidence of a mental health condition in the time leading up to their attack
(Alathari et al., 2021). These conditions may be caused by a combination of genetic
predispositions and severe life stressors. The USSS categorizes these conditions under
psychological, behavioral, and neurological, and states that for most plotters, their mental health
condition is psychological rather than behavioral or neurological. This means that symptoms
such as depression and anxiety, or more severe conditions such as personality disorders, are
likely to manifest. Where stressors have contributed to an individual’s mental health condition,
these most commonly have to do with family, social issues, or academic problems. These
stressors are much more common among perpetrators of school shootings than criminal stressors
or issues of poor physical health.
Bullying is another issue that can be addressed at the primordial level. Nearly 50% of
school shooters report having been bullied by their classmates (Alathari et al., 2021). This is an
issue that spans numerous health issues within society, and as such would form part of the
primordial tier in a public health model to school shootings. However, of particular importance is
the fact that school officials often know about the bullying a perpetrator of a school shooting
experiences, but they are unable or unwilling to respond in an effective way (Alathari et al.,

15
2021). This knowledge feeds into the need for communication within the community as
discussed under the primary and secondary prevention tiers.
Primary Prevention
While the primordial tier concerns itself with the general health of a community, the
primary prevention tier relates to policies and strategies designed to reduce the prevalence of a
specific health issue. In the case of school shootings, the primary prevention tier is concerned
with addressing the risk factors that lead to the spread of psychological issues that result in
school shootings taking place. A public health approach proposes that school shootings be treated
like any other public health concern. Through this lens, the issue can be viewed as the
transmission of a disease and, as such, can be thought of as an epidemiological problem
(Bentson, 2006).
Data show that the number of school shootings has increased over the past 20 years from
5 to 13 per year in 1999–2004 to 8 to 41 per year in 2016–2021, with a low of 8 being due to the
COVID-19 pandemic (del Carmen et al., 2022). Of the 342 cases since 1999, 176 have been
cases of targeted attacks. Far and away, the most common location for school shootings to take
place on the basis of campus or school type is high schools (62%). A little over half of the cases
relate to perpetrators enrolled in the schools as students. Research suggests that elevated activity
around home-grown violence, such as school shootings, will likely persist (Joint Regional
Intelligence Center, 2022), though the nature and number of violent attacks will likely fluctuate
as a function of prevailing sociopolitical events. Actions are likely to be large-scale episodes
rather than an epidemic of smaller scale events. There is, therefore, a clear and apparent need to
address school violence through policies that address the issues specifically.

16
Creating a society resilient to the threat of school shootings is difficult, with efforts to
reduce DVE events, in fact, paradoxically increasing the drivers that lead to their occurrence
(Joint Regional Intelligence Center, 2022). For example, moves to intervene in violent extremism
can cause members of the community to feel singled out and, as a result, drive them toward
wanting to commit a violent act. Attempts to even discuss the need to change laws on gun
ownership can lead concerned individuals to commit violence in opposition. However, the
government knows that it needs ways to deal with the threat of individuals responding to
legislation or action on violence with violence and being determined to do so (National Security
Council, 2021).
Identifying Risk Factors
The CDC lists violent behavioral risk factors at the individual, family, social, and
community levels (CDC, 2020). A risk factor can be defined as a characteristic affecting the
probability of an event taking place (Weine et al., 2017). At the individual level, risk factors
influencing the probability of incidents involving violence include experiences of violent
victimization, attention deficit hyperactivity disorder (ADHD), signs of aggressive behavior
during early development, a history of drug and alcohol abuse, poor behavioral control, cognitive
and IQ deficits, and the demonstration of anti-social beliefs. At the family level, the CDC lists a
number of risk factors relating to parenting, including overly strict or particularly lax parenting,
disinterested parents, low parent income and education, parental criminality, and parental
substance use.
Factors influencing the risk of targeted violence can also be distinguished between “push
factors” and “pull factors” (Weine et al., 2017). The former are factors that drive people away
from normal everyday life, such as oppression or family difficulties. The latter are factors that

17
draw people toward violent behavior. In the case of school shootings, the main pull factor would
be the promise of alleviating a perceived grievance. In understanding the factors that influence
the risk of a school shooting taking place, it is perhaps useful to categorize factors both by level
(individual, family, community) and according to the “push-pull” dyad.
Another important distinction in risk factor assessment is between static and dynamic
factors (Weine et al., 2017). Traditional risk assessment models have focused on static factors,
such as a person’s record with law enforcement. More modern risk assessment tools seek to
include factors that can change over time, such as personal attitudes and beliefs, social
associations, and current behavior. These dynamic factors are better identifiers of potential
risks—and can be influenced by outsiders—and can therefore be used to start a person’s course
of treatment. In addition to understanding the risk factors, it is important to also understand the
protective factors that reduce the likelihood of a person committing targeted violence, despite
being exposed to the risk factors. Such factors at the individual level include being married,
maintaining stable employment, and having no previous history of violence (Weine et al., 2017).
Risk Factors at the Community Level—Primary Prevention
At the community level, risk factors for violent extremism include poor economic
opportunities, minimal social interaction, population transiency, family disruption, and
disorganized neighborhoods. Economic position is often cited as a factor among perpetrators of
violent behavior. The “aggrieved entitlement” of perpetrators can frequently be linked back to
their economic standing, experience of poverty, or lack of economic opportunity (CDC, 2020).
This entitlement dictates their belief in a right to something they do not have and, as a result,
they resort to violence. At the level of the primary prevention tier, efforts to mitigate against the

18
risk of targeted violence, such as school shootings, can therefore be combined with general
efforts to improve the economic conditions within a community.
The role of social interactions within a community in attenuating the risk of school
shootings is two-fold. First, it provides an open environment where members feel connected,
reducing cases of severe exclusion—a later stage identified in many school shooting cases.
Second, the USSS (Alathari et al., 2021) and the Federal Bureau of Investigation (FBI) both
emphasize the importance of healthy social interactions for the creation of a community
environment in which bystanders can speak to someone about concerns they may have regarding
a particular individual. Southers (2013) refers to this as a “Mosaic of Engagement,” where
community policing of potential violent extremism, such as school shooting, is built into the
interactions within a community.
The government’s strategy against domestic terrorism also emphasizes the importance of
“renewing partnerships” across all levels and sectors of society (National Security Council,
2021). The roles of social interactions within a community are therefore relevant at both the
primary and secondary prevention tiers of a public health model as applied to school shootings.
Protective Factors at the Community Level
There are tangible steps that can be taken by local or government agencies to increase
communities’ capacities to prevent people from becoming at-risk individuals. For example,
investment in community mental health services and taking the necessary steps in order to
improve collaboration between judicial courts and mental health services are both ways that state
agencies can increase the likelihood of providing treatment to individuals before they become at
risk (Amman et al., 2017). Besides forming an important part of a primary prevention tier, such
developments could also provide benefits as part of a secondary prevention or intervention tier.

19
The FBI’s Behavioral Assessment Unit (BAU) emphasizes the role of language in
reducing the risk of school shootings (Amman et al., 2017). The BAU asserts that, in order to
diminish the appeal of performing a school shooting, the use of terms that instill a sense of power
(e.g., “lone shooter”), or those hat romanticize school shooters (e.g., “lone wolf”) should be
avoided. Instead, language should be used that conveys a sense of criminality and wrong-doing
(e.g., “offender” or “assailant”). In this vein, the BAU also advises against the publishing of
shooters’ names, photos, or backgrounds on media outlets in order to avoid any sense of legacy
or pride.
The government’s strategy on domestic terrorism makes mention of online material as an
influence on at-risk individuals (National Security Council, 2021). Internet communication
platforms have been combined with the long-standing history of domestic terrorism within the
U.S. in many cases of targeted violence. The primary prevention tier should, therefore, consider
the impact of online material and internet communities on at-risk individuals. The CP3 discussed
earlier is also tasked with setting up digital forums with a focus on highlighting how technology
can be a protective factor rather than a risk factor and to build resilience against harmful material
in the online arena (Brown, 2022).
A cornerstone of national security is “engagement through partnership” and such
partnerships constitute a component of primary prevention of terrorist threats (Weine et al.,
2017). In efforts to prevent school shootings, the right partnerships also need to be made. The
White House’s Strategic Implementation Plan on countering violent extremism emphasizes the
importance of community policing (Weine et al., 2017). This entails the partnering of affected
communities with law enforcement as a means for joint problem-solving and for the building of
trust and cooperation. It should be noted, however, that most police departments in the U.S.

20
would say that community policing is part of their operating methods and always have been.
Local law enforcement can also explore “tailored engagements” that involve the police
proactively engaging with a community on a specific emerging trend, such as school shootings.
These practices have worked well in engaging Muslim communities against home-grown
terrorist threats, and they have also been used successfully by the Los Angeles Police
Department (LAPD) in countering violent extremism. Research has shown that community
policing improves citizens’satisfaction and trust in the police. However, more research is needed
to determine whether it leads to a measurable reduction in violent crime.
Liaison activities are another example of how law enforcement can preempt violent crime
(Weine et al., 2017). The LAPD has been performing liaison activities with the Muslim
community through annual forums, with their goal being to improve communication between the
community and the police and to help the police better understand the community they are
dealing with. Similar liaison activities with schools and universities could help law enforcement
understand the communities in which school shootings take place. The LAPD reports that these
efforts can be thought of as “building a framework to defeat a framework”— a philosophy that
could help break the cycle of targeted school violence. Many of the practices used in these
liaison activities between the LAPD and the Muslim community could also be used in preventing
targeted school violence, including one-on-one relationships between police and community
leaders (e.g. teachers), facilitating youth involvement, and teaching community members about
the role of law enforcement.
Protecting against violent behavior, such as school shootings, at the community level,
also includes policies at the school level (CDC, 2020). This includes schooling with intense
supervision, clear rules, firm discipline, and high teacher/parent engagement. Furthermore,

21
school leadership has a responsibility to convey information on possible threats to threat
managers, even at the risk of doing a disservice to the individuals or school community
concerned (Amman et al., 2017). The CP3 has also begun to expand its program for helping
schools to prevent targeted violence. Over the course of 2021, the CP3 expanded its
“Invent2Prevent” Prevention Innovation Labs from 25 colleges to 40 colleges and 10 high
schools (Brown, 2022). The program invites students to devise novel ways of preventing targeted
school violence through competitions.
The Department of Homeland Security (DHS) issued a grant program in 2021 for
community projects aiming to mitigate targeted violence (Department of Homeland Security,
2021). The grant program is intended principally for projects that improve communities’
awareness of what a targeted violence threat looks like, and its awareness of both protective
factors against and risk factors for targeted violence. In addition, projects can also seek to
improve community engagement between as many sets of actors as possible for the creation of
sustainable partnerships and increased communication.
At the community level, one possible means of promoting the social interactions needed
to reduce violent extremism is through the DHS Center for Prevention Programs and
Partnerships (CP3). This center seeks to prevent acts of targeted violence, such as school
shootings, through Local Prevention Frameworks established by Field Operations Teams. On the
ground, these teams are comprised of Regional Prevention Coordinator (RPC) placements. In
2020 and 2021, the center awarded $30 million in grants for projects designed to prevent
domestic violent extremism (DVE) across 21 states (Brown, 2022).
The main function of these frameworks is to build trusted partnerships between all
community stakeholders (Brown, 2022). Such partnerships may help to create communities with

22
a lower susceptibility to targeted violence. Local Prevention Frameworks are also tasked with the
provision of support services, which can reduce the number of at-risk individuals. These services
include career centers, after-school groups, mentoring, and counseling. RPCs curate prevention
programs designed to reduce the risk factors and promote protective factors associated with
targeted violence.
The setting up of a Local Prevention Framework is a two-stage process. The first stage is
concerned with identifying all stakeholders throughout the community, including public
employees (government officials, medical personnel, mental health and social service
professionals, school staff, emergency personnel, and law enforcement) and members of the civic
community (members of faith organizations and respected community members).
The second stage concerns drawing up a list of all available resources. This should
include all programs on prevention and relevant social service programs. One of the CP3’s key
resources is its Community Awareness Briefing (CAB). This is a two-hour presentation that can
be used by community leaders to teach about efforts to prevent targeted violence and
radicalization (Brown, 2022).
Secondary Prevention (Intervention)
The secondary prevention tier of a public health model seeks to prevent an injury or
disease from occurring following exposure to risk factors. It occurs at the preclinical stage of an
illness. In the case of school shootings, the secondary prevention or intervention tier would
become relevant once a person has been identified as having been exposed to the risk factors
pertaining to the mental conditioning associated with that of a school shooting perpetrator.
A secondary prevention tier would require certain inputs, as well as certain actions on the
part of different segments of the community. The inputs would be strategy, relevant data,

23
interpretation of relevant data, and understanding of possible cases of leakage. Relevant
segments of the community include schools, law enforcement, mental health services, and the
judicial system.
Necessary Inputs
When considering school shooters from an epidemiological perspective, it is important to
decide whether the goal is one of elimination or eradication (Bentson, 2006). Elimination refers
to reducing the number of cases within a certain geographic area to zero, while eradication refers
to a permanent reduction to zero on a global scale. In the case of school shooters, elimination
would refer to reducing the cases to zero within a state or a community, while eradication would
refer to its permanent removal from U.S. culture.
In immunizing against a physical disease, intervention agencies can use either a mass or
ring vaccination strategy. A mass vaccination strategy involves immunizing an entire population
with the goal of protecting everyone who may come into contact with it. A ring vaccination
strategy refers to identifying cases of the disease and immunizing only those within a fixed
geographic radius. The former is more comprehensive, but the latter enables resources to be
concentrated more specifically and perhaps more effectively. Bentson (2006) states that
understanding how the ideology of terrorism spreads through a population could help individuals
decide how best to contain it. A better understanding how the ideas spread would enable better
containment strategies to be derived based on models from the epidemiology of physical
diseases. To date, such work in the extant literature is lacking.
Disease management also requires the discerning use of available data. The primary
prevention tier of a public health approach to school shootings needs to use the available data on
previous shooting events in studying the phenomenon. This step must be taken if a society is to

24
be developed where school shootings are no longer a threat. Through its Policy and Research
Team, the CP3 works to ensure that the Center’s approach to violence prevention is based on
empirical evidence and on the sharing of information between different agencies (Brown, 2022).
Data on a person’s mental health are critical in assessing the risk of violence, particularly
in cases where leakage does not manifest (Amman et al., 2017). The body of research into
targeted violence points increasingly toward a set of behaviors in individuals who are in the
process of planning an attack. Knowing what these behaviors are and how to respond to them
would be key components of a secondary prevention tier action plan, particularly in cases where
leakage does not occur. Information sharing on potential risks is also an important part of
preventing domestic targeted violence (National Security Council, 2021).
It is important to note that attempts to identify at-risk individuals through the study of
known risk factors alone have not proven successful (Weine et al., 2017). There are currently no
consistently reliable means of intervening in an at-risk individual’s behavior based simply on the
study of available risk factors. The literature shows that intervention methods need to look at risk
indicators rather than risk factors, with risk indicators consisting of the behaviors or
characteristics typical of a person conspiring to commit a subversive act. Previous studies have
found the following behaviors to relate to a person about to perform an act of targeted violence
(Weine et al., 2017):
1. Awareness within the public of a person’s grievance
2. Awareness within the public of a person’s ideology
3. Leakage
4. Stockpiling weapons

25
The “Providing Alternatives to Hinder Extremism” (PATHE) initiative established in L.A.
is a possible means of profiling behaviors that may enable an act such as a school shooting to be
prevented. PATHE is a process for profiling the behaviors typical of someone about to commit an
act of targeted violence (Hussin & Gunaratna, 2019). It has been used by the LAPD as a means
of assessing risk, and can be used by different disciplines (e.g., law enforcement, mental health,
and legal) as a means of identifying behaviors.
RTI International (2017) explains how current risk assessment models view risk as
“contextual, dynamic and continuous.” In assessing the risk of an individual enacting
ideologically motivated violence, difficulty in assessing the risk is compounded by the myriad of
factors involved. However, there are some stand-out trends in cases of active shooters. For
example, the frequency of active shooters by day in 2021 was noticeably higher on certain days,
and at certain times of day (Federal Bureau of Investigation, 2022).
Another factor that can be drawn on in identifying potential attackers is their history with
law enforcement. The JRIC (2022) states that contact with law enforcement is among the top 10
risk factors with regard to perpetrators of domestic violent extremism, with around 30% of
perpetrators having a history of law enforcement contact. Nearly one-third of school shooters
were found to have had some interaction with law enforcement prior to the attack (Alathari et al.,
2021).
Identified behaviors can be combined with a person’s individual risk factors to formulate
an assessment of risk. The CDC discusses factors relating to an individual’s risk of committing a
violent act at the individual, family, social, and community levels (CDC, 2020). These frequently
reflect the protective factors that are to be promoted at the primary prevention tier. Though
behaviors and risk factors are frequently indicative of a person’s proximity to committing a

26
violent extremist act, such as a school shooting, it is critical that these judgments are made
accurately. Such judgments are frequently drawn from an unstructured, clinical approach and are
deemed unreliable and difficult to replicate in different settings. To address this, structured
judgment protocols (SJPs), such as the Violent Extremist Risk Assessment (VERA), have been
established (Pressman, 2009). However, such tools are at the conceptual research stage, and they
are based on little, if any, empirical data.
The FBI BAU describes an important feature of school shootings that can be used in the
secondary prevention or intervention tier of a public health model. The BAU explains that school
shootings can be either impulsive or premeditated, and that the majority of school shootings are
very much premeditated (Federal Bureau of Investigation, 2022). This is evidenced by the fact
that witnesses of school shootings often report the total lack of emotion in the perpetrators;
emotions that can be present at these events are simply those resulting from the commission of
an impulsive action.
The evidence also shows the often lengthy and detailed planning that goes into these
attacks. With that in mind, the likelihood of extensive planning behind plots devised by at-risk
individuals creates an opportunity for secondary prevention, once the at-risk individuals have
been identified. The BAU also emphasizes the importance of intervening as soon as a threat is
detected. This allows for a measured and effective response so that the individual is appropriately
treated before they can commit a violent act.
In secondary intervention strategies, it is also important to consider the concept of
“aggrieved entitlement” discussed under the primary prevention tier (del Carmen et al., 2022).
The source of this entitlement in perpetrators of violence, in general terms, is typically an
economic issue. In perpetrators of school shootings, the source is more typically that of a

27
relational issue, frequently taking the form of a grievance between a male perpetrator and a
female (del Carmen et al., 2022). Efforts to intercept possible school shootings should therefore
involve asking questions to the right people in the school community about an at-risk
individual’s previous relational experiences.
Leakage is an important concept in establishing the secondary prevention tier when
applying the public health model to school shootings. Leakage is a potential perpetrator’s
communication to a third party about their intent to commit a violent act (Weine et al., 2017). It
is the most common means for identifying a potential perpetrator of a school shooting (Federal
Bureau of Investigation, 2022), and it can be as simple as an inappropriate or violent statement.
The JRIC (2022) reports that it is the most common risk factor present in violent
perpetrators after social media use. It is important as well not to ignore threatening language
from potential school shooters who are female. Despite the large majority of school shooters
being male, the FBI’s BAU expresses the need to treat threatening language from females that
could be construed as leakage as importantly as that from males (Federal Bureau of
Investigation, 2022).
However, leakage on its own is not enough to initiate secondary prevention. It is only
useful if those who receive it (i.e., bystanders) are able and willing to convey it to the right
people in a timely manner. This echoes the DHS slogan of “If you see something, say
something,” which instructs witnesses of worrying behavior patterns to report it to local law
enforcement (Department of Homeland Security, 2022). The campaign also informs the public of
the type of behavior they should be worried about. The thought or emotion expressed through
leakage needs to be both understood by the third party and reported in the right way to the right
people. Notably, two-thirds (66%) of mass attack perpetrators exhibited behaviors that elicited

28
concern in other people (Brown, 2022). In over 50% of cases, these behaviors caused observers
to fear for their safety. One of the CP3’s key resources is its Community Awareness Briefing
(Brown, 2022). This aims to improve awareness around targeted violence as well as terrorism.
Understanding by the third party requires familiarity with the motivations of extremist
violence, such as that resulting from white supremacism, extreme religious views, or worrisome
political ideologies (Joint Regional Intelligence Center, 2022). Among the different ideologies
that can lead to targeted domestic violence, white supremacism is one of the most dominant
(National Security Council, 2021).
The Roles of Different Agencies and Actors
Previous shooting incidents have also provided insights into the changes in the
involvement of public agencies that are needed as part of a secondary prevention or intervention
tier. For example, the shooting at Virginia Polytechnic in 2007 led the state to improve
procedures for performing emergency evaluations and to change the criteria needed to commit an
at-risk individual (Amman et al., 2017). Such changes to government agency mandates may be
needed as part of a public health model to school shootings within this secondary tier.
Social services are perhaps the government agency sector’s heaviest burden with respect
to managing the threat of an at-risk individual (Amman et al., 2017). This impacts its ability to
perform effectively and safely. The responsibilities of social services within an effective
secondary prevention tier need to be shared with other government agencies if social service
functions are to be deployed effectively.
Healthcare providers are equally overburdened. This burden can be reduced and their role
enhanced through awareness training around how to assess the threat posed by an individual.
They should also be open to sharing the risk of a threat with the right people where privacy laws

29
permit. In California, mental health services providers have the legal right to detain any
individual deemed to be a risk to themselves or others for up to 72 hours for a detailed
assessment (LAPD, 2019). Effective involvement of mental health services into a secondary
prevention tier would ensure that the providers of these services know when and how to exercise
these rights and how to liaise with law enforcement effectively. Efforts to provide mental
healthcare services with more robust capacities have been made as far back as the 1990s.
The role of law enforcement needs to be considered carefully. Though at-risk individuals
may be in the process of committing a criminal act, it is difficult for already stretched law
enforcement resources to be allocated to a potential threat that may or may not result in an actual
violent event (Amman et al., 2017). The extent to which local law enforcement can preempt
attacks based on awareness education should perhaps be tailored to each specific local force as a
function of available resources. The CP3 provides a Law Enforcement Briefing (LAB) that
serves to train LE personnel on a public health approach to recognizing the risk indicators of a
person at risk of committing an act of targeted violence.
In California, law enforcement should be aware of the fact they have the powers to seize
any firearm in the possession of someone that has been deemed a risk by a mental health services
provider (LAPD, 2019). Law enforcement agents are encouraged to perform a search of the
Department of Justice’s Automated Firearms System to determine whether an at-risk individual
is in possession of firearms.
As of 2016, law enforcers in California also have the power to conduct a “welfare check”
of any person deemed as being a risk to themselves or others. A “Gun Violence Restraining
Order” is another new power endowed to law enforcement officers, which allows them to restrict
an at-risk individual’s ability to purchase firearms or ammunition (LAPD, 2019).

30
In Los Angeles, the Police Department has put in place a special Mental Evaluation Unit
(LAPD & Los Angeles County Department of Mental Health, 2020). The purpose of this unit is
to respond to mental health crises and to prevent the incarceration of mentally ill individuals. The
unit collects and shares data on potential threats and integrates mental health professionals into
law enforcement so that responses from diverse sets of actors can be deployed (i.e., codeployment). Such initiatives could be established more widely in the U.S. in response to the
threat of school shootings and would constitute a component of the secondary prevention
(intervention) tier of a public health model.
In 1992, the implementation of a System-wide Mental Assessment Response Team
(SMART) program was begun in parts of California to address needs of the mentally ill
population (LAPD, 2019). These SMART teams consisted of a mental health practitioner and a
member of law enforcement personnel. By 2016, SMART operations had become a 24/7
operation in parts of the state. One of the key goals of SMART teams was to prevent
incarceration of the mentally ill and to transfer them from law enforcement over to mental health
services providers so that law enforcement can return to normal duties.
The LAPD also has the Mental Health Crisis Response Program (MHCRP). This
program brings together various stakeholders within the mental health community and ensures
that the Chief of Detectives liaises with the LA County Department of Mental Health (LAPD,
2019). Through this program, the Chief of Detectives is also expected to analyze state and
federal legislation regarding people with mental illness. Importantly, the program also requires
that the Chief of Detectives maintains records on all mental health crisis responses within the
department.

31
In California, as stated, law enforcement has the right to seize firearms from a person
deemed as being a danger to themselves or others due to a mental health condition (LAPD,
2019). Furthermore, vendors of firearms should be made aware of laws that put them at risk of
heavy fines or imprisonment for supplying firearms to anyone who has “communicated a serious
threat of violence” to a psychotherapist. Even though at-risk individuals may not be undergoing
psychotherapy, awareness among firearm vendors of the risk they face in selling firearms to atrisk individuals could help to prevent at-risk individuals from obtaining the firearms used in
school shootings.
Of greater use is the role of the judicial system (Amman et al., 2017). Previous
experiences have shown that the discretion and wisdom of prosecutors can be a highly effective
tool in managing the threat posed by an at-risk individual. Likewise, lawmakers and the courts
can enhance their roles by applying their powers prudently. Stakeholders within the legal system
can apply the needed shrewd use of powers by first establishing baseline familiarity with the
dynamics between mental health issues and violence. In people of concern, parole agencies are,
or should be, directly responsible for observing behavior and responding. Any violations of
parole may signify issues, though parole agencies may require training on risk factors if they are
to interpret specific behaviors and violations.
Members of the public also need to be made aware of their responsibility in effectively
creating a secondary prevention tier. In particular, employers are in a unique position for
observing at-risk individuals on account of how much time they spend with an at-risk individual.
However, an appropriate response to a threat of violence on the part of an employer is a difficult
thing to gauge. Often, an at-risk individual performs their violent act long after they have left a
given place of employment. In such cases, previous employers need to be prepared for fully open

32
dialogue with law enforcement and other agencies. The CP3 Community Resilience Exercise
(CREX) is a half-day workshop run with members of the community to stimulate thoughts
around how they would act when faced with a potential act of targeted violence or to an act that
is underway (Brown, 2022). Bystander training is also in the process of being developed for a
more thorough approach to training members of the public in how to respond to the unfolding of
targeted violence.
The CP3’s Local Prevention Frameworks mentioned under the primary prevention tier
(CP3, 2022) can also play a role in the secondary prevention tier through the building of trusted
partnerships between all community stakeholders (CP3, 2022). These frameworks perform
various roles that can form part of a secondary prevention tier. In addition to educating the public
of the risk factors associated with school shooters, a key function in these frameworks is to
provide help to individuals before they resort to violence. These frameworks also include
members of law enforcement, meaning that at-risk individuals can be intercepted using
investigations and arrests if deemed necessary (Brown, 2022).
These frameworks also bring together different agencies and services so that members of
the public with concerns have people they can report to in confidence. One important feature of
these frameworks is their uniqueness to each community. Each framework can be designed to
accommodate differences in resources, infrastructure, population, laws, and political climate
within each community (CP3, 2022). These CP3 frameworks are established by RPC field agents
whose role is to educate and assist the whole of society. This is achieved through the building of
partnerships between different stakeholder groups and building frameworks within communities.
The agents also increase awareness in communities of targeted violence threats, and what
members of the community can do to prevent them—whether in the community or online.

33
The DHS grant program in 2021 for community projects focuses on the mitigation of
targeted violence listed under primary prevention, which is also being opened to projects whose
goals seek to enable members of the local community to act in the face of a potential threat,
particularly in enabling them to engage online. These projects should result in community
members having multiple ways to divert at-risk individuals to appropriate intervention services.
Through whichever means an at-risk individual is identified, the next step in a secondary
prevention tier is for intervention to take place before the violence does. A principal component
to threat management is de-escalating the potential threat by convincing the at-risk individual
that violence does not solve the grievance they have (Joint Regional Intelligence Center, 2022).
Furthermore, it is important that threat management teams are cognizant of all agencies and
services available to them in responding to an identified at-risk individual. The goal of this
engagement is to divert at-risk individuals toward options for treatment before a school shooting
takes place.
The CP3 Local Prevention Frameworks can also play a role in threat management. An
important secondary function of these frameworks is building teams consisting of educators,
psychologists, faith leaders, medical personnel, law enforcement, and social workers in making
accurate assessments of possible at-risk individuals (CP3, 2022). The goal of these teams is to
provide an early intervention into cases so that school shootings do not take place.
The JRIC (2022) also outlines the importance of evolving intervention models in the face
of evolving threats of domestic violence extremism, which includes school shootings. Such
developments in DVE events include the adoption of new technologies and the use of tactics
from foreign terrorists. The best defense against an evolving DVE threat is the bringing together

34
of typically disparate sets of community actors, including law enforcement, mental health, and
even counter-terrorism. The entire spectrum of stakeholders needs to be informed and trained.
Barriers to Intervention
Identifying an at-risk individual is only a first step in neutralizing the threat that they
pose. Familiar barriers can prevent an intervention before they choose to act. Some of these are
more difficult to overcome than others. A major barrier difficult to overcome may be the
resistance of parents to any third-party intervention, particularly if the individual concerned is yet
to perform a criminal act. Another barrier can be the lack of information or weak information
sharing.
Tertiary Prevention (Treatment)
The tertiary prevention or treatment tier of a public health model seeks to treat a public
health issue that has already taken hold in an individual. The goal is to attenuate the
complications of the health issue or to reduce the chances of a relapse into illness. In the case of
school shootings, this tier of the public health model is aimed at handling perpetrators of
violence so their behavior does not escalate into something as serious as a school shooting or
helping them not to re-offend if they have already performed a violent act (Brown, 2022).
Once the threat has been de-escalated at the secondary prevention tier, treatment of the
individual should be performed at the tertiary prevention tier. This seeks to prevent the individual
from planning such activity again. The Local Prevention Frameworks designed to play a role in
the primary and secondary prevention tiers of a public health model are also designed to play a
role at the tertiary prevention or treatment tier. These frameworks are designed to include
personnel from corrections, probation, and parole to implement specialized correctional

35
programming to individuals who have been deemed at-risk or who have been a perpetrator in a
targeted school shooting (Brown, 2022).
The VERA protocol mentioned previously details two important psychological processes
in the intervention process. The first of these is “de-radicalization,” which is the process of
rejecting or modifying a set of beliefs or ideology (Pressman, 2009). The second is
disengagement, which constitutes a person leaving a group that may have played a role in their
radicalization. It often precedes de-radicalization. The tertiary prevention tier of a public health
model should ensure that both of these processes take place, and that their effects are effective in
the long term.
The study of dynamic factors can also play a role in the tertiary prevention or treatment
tier of a public health approach to school shootings. Dynamic factors change over time and can
be used by outsiders such as health workers. By observing a person’s dynamic risk factors,
treatment plans can be designed based on these factors, while also seeking to influence these
factors through the process of treatment itself (Weine et al., 2017).
The JRIC (2022) states that perpetrators of DVEs, such as school shooters, are likely to
have an elaborate history of violence toward others. With this in mind, the tertiary prevention tier
of a public health approach to school shootings may need to treat those with a record of violence
and not just treating those who have performed a school shooting.
Summary
Numerous factors and commonalities have been found in previous literature pertaining to
perpetrators of mass shootings, as well as targeted school violence, while the development of a
reliable profile has evaded researchers. However, enough data are present such that targeted
prevention plans can be developed in order to reduce the incidence of mass and school violence.

36
Such a response is necessary as once law enforcement is mobilized to respond to such an
incident of violence, it is almost always too late.
This chapter outlined these various risk factors, as well as recommendations for violence
reduction based on past empirical research. Also, a public health approach focusing on
prevention was discussed in detail, with the phases or tiers elaborated on: primordial prevention
(or upstream), primary prevention, secondary prevention (or intervention) and tertiary prevention
(or treatment).
The following chapter discusses the results of the literature review conducted by focusing
more specifically on school shootings and school shooters.

37
Chapter 3: Literature Review: School Shootings
Introduction
Targeted school violence has posed and continues to pose a serious threat to students in
the United States (Lenhardt et al., 2010; Lenhardt et al., 2018). Over time, researchers have
developed a series of models for assessing the risk of individuals and understanding the warning
signs that they present. While many researchers have attempted to build a profile of the school
shooter, the reality is that school shooters come from a wide variety of backgrounds and life
experiences and exhibit characteristics that do not necessary conform to any particular profile.
Theory
Generally, research regarding school shootings has suffered from a lack of welldeveloped theory. Often, research has been descriptive and pragmatic rather than theory-driven.
Psychological theories aimed at understanding the cognitive, emotional developmental, and
clinical factors are important in the analysis of school shooters (Grøndahl & Bjørkly, 2016).
Social learning theory, the frustration-aggression hypothesis, and theories of aggression, such as
the General Aggression Model are also all relevant (Grøndahl & Bjørkly, 2016). Currently, “no
single explanatory model is developed to inform risk assessment of who will cross the line and
actually start to kill their schoolmates and teachers” (Grøndahl & Bjørkly, 2016, p. 9). However,
many researchers have attempted to create models to better understand what might lead a
perpetrator to commit such atrocities.
Levin and Madfis (2009) created a sequential and additive/cumulative strain model with
the goal of allowing for a complex explanation of a complex phenomenon. The Levin and
Madfis model has five steps: “prolonged stress, strain without social control, acute stress,
planning, and finally the atrocity” (Grøndahl & Bjørkly, 2016, p. 2). School shooters often begin

38
feeling prolonged stress or chronic strain through bullying and failures in interpersonal
relationships (Bonanno & Levenson, 2014). Perpetrators move into the second stage due to a
lack of social connections and meaningful relationships that might help them cope with stress
(Bonanno & Levenson, 2014). The third stage, acute stress, occurs when a potential attacker
experiences a loss perceived as catastrophic (Bonanno & Levenson, 2014). From this stage, the
attacker moves to the planning phase that occurs at least two days before the attack (Bonanno &
Levenson, 2014). In the final stage, an attack is actually carried out, which requires access to and
proficiency in firearms (Bonanno & Levenson, 2014). It is notable that many people experience
the first three stages, and some even plan an attack, but very rarely is an attack actually carried
out (Bonanno & Levenson, 2014).
Meloy and O’Toole (2011) created a similar model that categorizes warning behaviors
that may indicate that a person is preparing to commit a violent act. Clues about a student’s
intention to commit a violent act are called leakage. The warnings of an attack are: 1) pathway
warning behavior—researching, planning, and preparation; 2) fixation warning behavior—a
pathological fixation on a person or cause; 3) identification warning behavior—the psychological
desire to be identified with weapons, military, or law enforcement paraphernalia; 4) novel
aggression warning behavior—committing novel acts of aggression; 5) energy burst warning
behavior—increasing the frequency or variety of any activities related to the target; 6) leakage
warning behavior—communicating with a third party about the intent to do harm; 7) last resort
warning behavior—directly communicating a threat; and 8) directly communicated threat
warning behavior—displaying increasing desperation or distress in word or deed (Meloy et al.,
2014; Abel et al., 2022). Meloy and colleagues (2014) researched the validity of these warning

39
behaviors and their predictive power. In almost all cases, both for attackers and students of
concern, leakage was present (Meloy et al., 2014).
Most people learn of a shooter’s intention to attack shortly before the event, giving them
little time to respond. School climate, belief in the threat, and assessment of the likeliness and
immediacy of attack all affect a bystander’s decision to come forward (Abel et al., 2022). It is
important to note that all warning behaviors are dynamic, not static (Meloy et al., 2014). The
factors that result in a person committing a school shooting are always a web of causes (Böckler
et al., 2014). The FBI categorizes threats of an attack at three different levels: 1) indirect, vague
and lacking realism; 2) includes details, but no preparatory steps have been taken; 3) direct,
specific, and plausible, with concrete steps being taken (Wetterneck et al., 2005).
Researchers have also attempted to categorize the shooters themselves in order to better
understand their motivations and goals. Ioannou and colleagues (2015) identified three categories
of school shooters: the disturbed school shooter, the rejected school shooter, and the criminal
school shooter. The disturbed school shooter shows signs of emotional and/or mental health
problems, indicating “that school shooting emerges out of the person’s self-destructive emotions”
(Ioannou et al., p. 13). Their analysis found that perpetrators who live with their grandparents
and those with a family history of suicide scored higher in this theme. The rejected school
shooter is one who feels rejected as the result of a breakup, abuse, or a suspension or expulsion.
They have also attempted suicide in the past. The analysis found that perpetrators who were
adopted or lived with foster parents, those who had a history of family criminality, and those
with parental drug and/or alcohol abuse, scored higher in this theme (Ioannou et al., 2015). The
criminal school shooter relates to a perpetrator with general criminal behavior.

40
Similarly, Langman (2009) describes three types of school shooters: psychopathic,
psychotic, and traumatized. Langman finds, through a simple comparison of facts and details
about cases, that psychopathic shooters tend to have been around legal gun usage, while
traumatized shooters tend to use guns illegally (Langman, 2009). Traumatized shooters come
from broken homes, have at least one parent with substance abuse problems, and have at least
one parent with a criminal history (Langman, 2009). They also have family role models who
“engaged in criminal behavior involving the misuse of firearms” (Langman, 2009, p. 84). They
also tend to have a peer encouraging them to commit the attack and who sometimes even
accompanies them in the attack (Langman, 2009). Traumatized shooters also tend to have
endured some sort of physical or sexual abuse and to come from low-income families (Langman,
2016).
Psychotic shooters have symptoms of schizophrenia or schizotypal personality disorder,
with paranoid delusions, delusions of grandeur, and auditory hallucinations (Langman, 2009).
They are misfits in their families (Langman, 2009). Psychopathic shooters show narcissism, a
lack of empathy, a lack of conscience, and sadistic behavior (Langman, 2009). They come from
families with a legal history of firearm usage, and they are fascinated with firearms (Langman,
2009). They are also sadistic, deriving pleasure from having power over others and inflicting
pain on them (Langman, 2009). They also recruit peers to accompany them in the attacks
(Langman, 2009). Psychopathic shooters are often influenced by previous shooters, with many
drawing inspiration specifically from the Columbine massacre (Langman, 2016a, 2016b). Millon
and Davis (1998) identify subtypes of psychopathy that differentiate between perpetrators.
Abrasive psychopaths are contentious and quarrelsome, claiming the moral high ground for their
actions (Langman, 2013a). Explosive psychopaths are very sensitive to feelings of betrayal and

41
personal failure (Langman, 2013a). Some psychopaths are also adept at impression management,
hiding their true violent intentions (Langman, 2013a).
Farr (2018), on the other hand, categorizes school shooters in accordance with their
personal troubles: psychiatric disorder, family turbulence, and situational volatility. These
groupings have some overlap and are not as systematic as the groupings used by Ioannou et al.
(2015). Farr (2018) argues that school shooters fail to meet the standards of adolescent insider
masculinity (be cool, prove heterosexuality, repudiate femininity, and be tough), resulting in
acting out through means of violence. Volatile shooters were the most difficult to understand as
their classmates had either mixed or positive feelings toward them, and they had fewer problems
with their perceived masculinity (Farr, 2018).
Similarly, Sommer et al. (2014) showed that social dynamics play a significant role in
school shootings. However, the perpetrator’s perception of the social dynamics was much more
relevant than the actual social dynamic, along with their coping and emotional regulation
strategies (Grøndahl & Bjørkly, 2016).
Researchers have also sought to categorize types of mass violence in schools. Langman
and Staub (2019) identified three types of school shooting incidents: 1) large-scale, random
attacks planned in advance, 2) small-scale targeted attacks, and 3) unplanned attacks, where the
perpetrator brought a gun without the intention of using it. Most of the cases explored in this
study fall in the first category.
Background
Contrary to popular belief, school shootings are not a new phenomenon. In 1966, a 25-
year-old murdered 16 individuals and injured more than 30 at the University of Texas (Ferguson
et al., 2011). “The 1990s saw an unusual string of” mass shootings at academic institutions

42
(Ferguson et al., 2011, p. 142). From the late 1960s until the 1990s, the U.S. experienced a wave
of violent crime, particularly including youth violence, that likely contributed to the wave of
mass shootings in the 1990s (Ferguson et al., 2011). Guidance from the APA and FBI to assess
threats has often been vague and unreliable due to the difficulty in building a profile for suspects
of mass shootings (Ferguson et al., 2011).
The news in the late 1990s and early 2000s tended to focus on external factors, especially
video games, as the culprit for school shootings rather than internal factors such as “depression
and psychopathic traits of the perpetrators themselves” (Ferguson et al., 2011, p. 145). However,
research examining the association between video games and school shootings has been
inconsistent, with video games ultimately having been shown to “play no causal role in violent
behavior” (Ferguson et al., 2011, p. 150). While the accusations about the impact of video games
on school shootings can be considered a moral panic (Ferguson et al., 2011), other scholars argue
that violent media can contribute to attacks. For instance, Stephen King’s book Rage was quoted
by two different attackers as part of their inspiration (Ioannou et al., 2015).
Generally, people assume that school shootings have similar causes because they are
similar events (Grøndahl & Bjørkly, 2016). Many studies emphasize the fact that perpetrators are
lonely, alienated, and victims of bullying, but these “characteristics can be found in many
students who never do show any signs of violence” (Grøndahl & Bjørkly, 2016, p. 2). School
shootings typically occur in suburbs, which led Kiilakoski and Okasnen to posit that smaller
communities may have “a more normative approach to how one ‘should’ be than in larger cities”
(Grøndahl & Bjørkly, 2016, p. 2).

43
According to Vossekuil et al. (2002, p. 18), “there is no accurate or useful ‘profile’ of
students who engaged in targeted school violence” as there are many students who would fit the
supposed profiles that exist yet have never committed such atrocities.
Overwhelmingly, school shooters tend to be white, male, and heterosexual (Farr, 2018).
However, this is not universally true, and there is no one profile for all school shooters
(Langman, 2013b). Men tend to feel much more comfortable expressing anger outwardly and
using physical mechanisms to cope with anger, while boys also feel more hostility and have less
confidence in adults at school than their female peers (Farr, 2018). Masculine norms generally
encourage boys to be aggressive, act tough, and not show emotions. About half of school
shooters who felt they were bullied were described by classmates as being bullies themselves.
Some perpetrators have experienced sexualized physical bullying at the hands of their peers.
Rage and resentment toward women, including ex-girlfriends and unrequited crushes, were a
common theme among school shooters. Almost all shooters have told others about their exploits,
whether as an attempt to show off their masculinity or as a threat (Farr, 2018).
Almost all school shooters are adolescents, so understanding neurobiology is helpful in
understanding this phenomenon (Rothe, 2022). Two major changes include an increase in
seeking high-intensity experiences and a natural growing interest in peers and potential romantic
partners. The parts of the brain that are still developing and the social brain overlap significantly.
The opinions of peers also become more important as the desire for relationships increases.
Consequently, rejection and isolation feel much more serious. Bullying and cyberbullying are
prevalent in adolescence, with 71% of attackers experiencing bullying prior to their attack. The
attacks are almost never impulsive, and usually result from a personal loss that creates a
perceived failure and loss of status (Rothe, 2022).

44
School shootings tend to occur in affluent areas where violence is rare (Landau, 2012).
Generally, perpetrators come from middle- and upper-class families because school tends to hold
more meaning for them (Landau, 2012). In recent years, young people have become more
comfortable reporting threats to adults in order to prevent attacks (Landau, 2012). School
shooters do not tend to have a history of drug abuse, prior violence, or criminal behavior
(Keatley et al., 2020). About half of school shootings take less than 15 minutes, and the majority
of school shootings are not stopped by law enforcement (Bonanno & Levenson, 2014).
From 1760 to 2010, there have been over 310 “documented shootings on school
property” in the U.S. (Duplechain & Morris, 2014, p. 145). From 2010 to 2014, there have been
at least 80 school shootings (Duplechain & Morris, 2014). According to Crawford (2002), in
over 80% of cases, at least one person knew about the shooter’s plan, and two or more people
knew in almost 60% of cases.
The CDC lists three categories of risk factors from school shooters: 1) personal risk
factors like learning disorders, a history of aggressive behavior, high emotional stress, and
exposure to violence; 2) familial risk factors like harsh, lax, or inconsistent discipline, low
parental involvement, and poor family functioning; and 3) community and societal risk factors
like low economic opportunity and low community participation (Duplechain & Morris, 2014).
Snell and Volockh (2005) listed the risk factors for school violence: poverty, illegitimacy,
domestic violence and abuse, society-wide violence, drug culture, immigration, population
mobility, discrimination, violent cultural imagery, materialism, and competitiveness and high
expectations. Snell and Volockh (2005) noted that while the top three disciplinary problems in
schools in the 1950s were talking out of turn, chewing gum, and students making too much
noise, the top issues faced today are drug abuse, alcohol abuse, and pregnancy.

45
Shootings involving four or more victims have substantially increased since 2017
(Flannery et al., 2021). Current strategies used by schools to reduce the risk of attacks include
having armed school resource officers, practicing active shooter drills, and preparing to run, hide,
or fight. However, students remain disproportionately fearful when it comes to the risk of an
active shooter. Numbers of deaths and injuries are trending downward, with occasional spikes
from high-profile shooting incidents. Schools are safer than they have been, and may be the
safest place for a child to be. About “4.6 million homes in the U.S. have loaded, unlocked guns
that youth can access” (Flannery et al., 2021, p. 239).
In 93% of cases, a peer receives prior knowledge of attacks (leakage) and the perpetrator
displays behavior indicating their plan to harm others (Wetterneck et al., 2005). In 88% of cases,
an adult reported concern about the perpetrator prior to the attack (Wetterneck et al., 2005). Life
turning points represent a significant moment that leads to accelerated involvement in deviance
(Abel et al., 2022).
School Shooters
Substantial literature has been written describing potential factors that relate to school
shooters. Some of this work simply describes how school shooters are similar to or different
from the general population, while other work examines factors that may precipitate school
shootings or may cause students to be more likely to become school shooters. School shooters
are “often motivated by personal problems that correspond with Durkheim’s (1897) theories of
suicide and common risk factors for suicide” (Lankford, 2013, p. 2). These include social
marginalization, problems with family, work, or school, and other crises (Lankford, 2013).
Durkheim’s (1897) theory of suicide states that there are four types of suicide: egoistic, caused
by social marginalization and isolation; anomic, caused by a lack of purpose; fatalistic, caused by

46
a desire to escape pain; and altruistic, caused by a feeling that collective needs far outweigh
individual self-worth. A total of 78% of perpetrators either attempted suicide or expressed
suicidal ideation prior to the incident (Ioannou et al., 2015). School shooters tend to suffer from
depression (71%) and have a psychiatric history (57%) (Ioannou et al., 2015). A majority of
perpetrators also experience bullying and marginalization. Additionally, 31% of perpetrators
displayed past violent behavior and 27% had been arrested previously, according to Vossekuil
and colleagues (2002). According to Farr’s (2018) sample of 29 shooters in the U.S. between
1995 and 2015, 10 shooters showed signs of psychosis or psychopathy, with eight having
experienced hallucinations or delusions.
School shootings are rarely just a part of general criminal activity as only about 7.5% of
cases seem to fit this criterion (Ioannou et al., 2015). School shootings almost always involve
significant planning and premeditation (Ioannou et al., 2015). Often, perpetrators show a
fascination with weapons and portrayals of violence. About a third of offenders had evidence of
violent writing in the past. About 60% of school shooters lived with their grandparents or had a
family history of suicide, showing that family background has a relationship with offending.
Feelings of rejection, whether real or perceived, have been shown to be a potential cause of
school shootings as perpetrators seek to send a message to those who hurt them (Ioannou et al.,
2015). Dutton and colleagues (2013) find evidence that suggests that many school shooters have
paranoid personalities and persistently feel “wronged.” Paranoid individuals have “closed
information processing systems,” which means they ignore any knowledge that might correct
their thinking, so all things serve to confirm their beliefs and make them more extreme (Dutton
et al., 2013, p. 551). These individuals often believe that they are inferior, and that others can
perceive their inferiority. They also shift all of the blame for their situation onto others, and they

47
view all hurt from others as intentional (Dutton et al., 2013). Generally, school shooters lack
skills at solving social problems, resulting in their resorting to violence (Wike & Fraser, 2009).
One key factor that is often overlooked when studying school shooters is the existential
concerns that they engage with prior to their attacks (Pfeifer & Ganzevoort, 2016). Perpetrators
of school shootings “seem to consider their deed meaningful” (Pfeifer & Ganzevoort, 2016, p.
133). They tend to believe that they are superior, and they face an existential crisis when others
question their superiority. Often, they use their violent actions to demonstrate their perceived
superiority to others (Pfeifer & Ganzevoort, 2016).
While no specific traits indicate that someone will commit mass violence, perpetrators of
school shootings tend to have poor coping skills and anger management problems (Wetterneck et
al., 2005). Among school shooters, “[m]ass-rampage offenders are significantly different” from
targeted attackers or unplanned attackers (Abel et al., 2022, p. 788). Perpetrators are often
members of deviant groups, and they tend to have negative role models (Lenhardt et al., 2010).
67% of perpetrators were involved in peer groups that focused on violence (Lenhardt et al.,
2010). Motives behind school shootings are varied but include grievances with family,
classmates, or staff, a relationship breakup, disciplinary action, suicidal ideation, a desire to kill,
or a desire to achieve fame and notoriety (Dowdell et al., 2022). Many shooters had a history of
disciplinary action at school and poor grades, although this is not always the case (Dowdell et al.,
2022). Most importantly, Langman (2009) notes that school shooters cannot be completely
understood because they are not normal people—they are children with serious psychological
problems.
School shooters choose the school as their place of attack because it is where they
experience the most suffering (Farr, 2018). Some shooters also blame classmates and teachers for

48
discrediting them, challenging them, treating them unfairly, or failing to help them (Farr, 2018).
However, according to Langman (2013a), not all school shooters are students. Some are former
students, employees, or people with no relation to the school. Langman’s study of college
campus shooters found a large diversity in race, challenging the notion that all attackers are
white males. College-aged shooters were also much more likely to fall into multiple categorizes,
whereas all secondary school shooters could be classified into one type easily (Langman, 2013a).
All targeted shooters were above 28, and almost all random shooters were under 27. Almost all
random attackers had experienced military failures such as a dishonorable discharge or a
rejection from the military. A total of 63% of shooters experienced academic failure on some
level. All of the targeted shooters were facing financial crisis. Often, shooters tend to have failed
in areas where their family members have succeeded (Langman, 2013a). While it is assumed that
all school shooters are white males, there have been female perpetrators, and one-third of the
North American shooters studied by Langman in 2013 were racial or ethnic minorities
(Langman, 2013b). Langman (2013b) also notes that older perpetrators are associated with
attacks having a greater number of victims.
A study of social media usage in school shooters found that 80% used Facebook, 40%
used Instagram, 28% used Snapchat, 24% used YouTube, and 16% used Twitter (Dowdell et al.,
2022). A total of 72% had at least one adverse childhood experience, and 60% had been bullied.
Shooters with more adverse childhood experiences were more likely to have had some
psychiatric treatment. Additionally, 76% of shooters had social media posts with disturbing
content, with 44% posting a photo of a gun and 40% posting threats on more than one platform.
Shooters who posted a photo of a gun were more likely to die at the scene of their attack
(Dowdell et al., 2022).

49
Methods
Much research on school shooters has been conducted by reviewing primary and
secondary sources from cases and drawing conclusions. Langman (2013b) studied 35 rampage
shooters in schools to categorize them and expand his typology. Psychotic and psychopathic
shooters killed twice as many people as traumatized shooters (Langman, 2013b). Over time,
shooters have become more suicidal, with a much higher percentage of shooters being suicidal
after 1998. Psychotic shooters and older shooters are most likely to be suicidal. A total of 54% of
shooters had targeted non-random victims including bullies, women, teachers, or family
members. Traumatized shooters most frequently had non-random victims, while psychotic
shooters killed the greatest number of people and viewed their attacks as serving a greater
purpose. Psychopathic shooters were the least harassed and had the most atypical attacks,
including killing from a distance (Langman, 2013b). Similarly, Langman (2009) researched 10
rampage shooters through case studies with an emphasis on what was known about them prior to
the attacks. Later, Langman (2013a) studied school shootings on college campuses, analyzing 16
different attacks from a psychological perspective. More recently, Farr (2018) analyzed 29
school shootings in the U.S. between 1995 and 2015.
Similar to previous case studies, Pfeifer and Ganzevoort (2016) studied the statements
that seven school shooters made prior to their attacks in order to assess the role of existentialism
in their actions. The five main existential problems were found to be death, isolation, identity,
freedom, and meaning. The statements from each shooter were coded and then analyzed based
on the existential material. Each shooter was found to be concerned about death and identity,
with almost all shooters being very concerned about death and identity. Five of the seven
shooters were concerned about isolation (Pfeifer & Ganzevoort, 2016).

50
Meloy et al. (2014) used case studies to derive eight warning behaviors for attacks and
studied the difference in warning behaviors between those who carried out attacks and those who
exhibited warning signs but never intended to attack. They studied nine German school shooters
and 30 “students of concern” who displayed warning signs but never attacked (Meloy et al.,
2014, p. 203). The study found that attackers showed significantly more fixation warning
behavior, identification warning behavior, novel aggression warning behavior, and last resort
warning behavior (Meloy et al., 2014). In almost all cases, both for attackers and students of
concern, leakage was present (Meloy et al., 2014). This indicates that “most cases in which
leakage is present . . . do not result in targeted violence” (Meloy et al., 2014, p. 208).
A final study using qualitative methods that will be discussed here consists of that
conducted by Keatley et al. (2020), who developed a separate timeline for 16 school shootings
using Crime Script Analysis. The 16 cases were studied and coded into scripts. The five scenes in
the scripts were: 1) influences, the factors leading to the crime; 2) operational, the preparatory
actions taken by the attacker; 3) reconnaissance, the planning of the day and gathering
information about location; 4) activity, the attack; and 5) withdrawal, the exit from the scene
(Keatley et al., 2020).
Other researchers approached their studies with quantitative methods. Grøndahl and
Bjørkly (2016) performed a thorough review of all available quantitative literature and
previously utilized methodologies. They found that empirical research is scarce, research quality
is poor, and the use of theories is largely absent in previous research on school shootings
(Grøndahl & Bjørkly, 2016). However, some quantitative research has been helpful and provides
interesting results. Lankford performed an analysis of 81 suicide attacks in the U.S. occurring
between 1990 and 2010, including 12 terrorist strikes, 18 rampage shootings, 16 school

51
shootings, and 35 workplace shootings (Lankford, 2013). The analysis focused on the primary
attacker for each incident. Data were collected by analyzing previous scholarship and collecting
transcripts and news stories from the event. The data were analyzed using chi-square and
ANOVAs to test for significant differences between suicide terrorists, rampage shooters,
workplace shooters, and school shooters (Lankford, 2013).
Similarly, Ioannou et al. (2015) analyzed 40 school shooting cases from 1966 to 2012 in
seven different countries with the goal of exploring characteristics of perpetrators. Data were
sourced from governments, news articles, books, and journal articles, with all details confirmed
by at least two sources. The data were analyzed using SSA-I, Smallest Space Analysis. The
analysis aimed to “explore the co-occurrences of offender characteristics,” with the goal of
identifying the similarities and differences between perpetrators (Ioannou et al., 2015, p. 11).
Each case was then classified as a disturbed, rejected, or criminal school shooter. Of the 40
incidents, 24 (60%) were classified as disturbed, nine (22.5%) were classified as rejected, three
(7.5%) were classified as criminal, and four (10%) were unclassified. Disturbed was found to be
the most common type of school shooter, followed by rejected, with criminal being the least
common (Ioannou et al., 2015).
Wetterneck et al. (2005) used quantitative methods to determine how individuals assess
the risk of school violence based on different background characteristics and threat levels. Two
groups, one of parents and one of young adults (18-25), were asked to complete a questionnaire
relating to their background. Each participant was then shown one of four vignettes with either a
high or low background risk and a realistic or unrealistic threat. The participants then filled out a
questionnaire about the perceptions of the potentially violent student in the vignette. The results
were fit with a two-factor measurement model and assessed for fit with five metrics. Results

52
found that parents reported a greater risk for violence in the student described in the vignette than
young adults. Females were more likely to feel to need to intervene than males. Both young
adults and parents were able to assess when there was a higher threat and a higher need to
intervene (Wetterneck et al., 2005).
Using basic quantitative methods, Lenhardt et al. (2010) studied the impact of
environmental factors like school climate and culture, peer and social influences, and whether
the perpetrator disclosed his intentions. A total of 15 school shooting events were analyzed for
this study, looking at primary sources as well as news sources. Two independent evaluators
coded the materials for quantitative analysis. In 53% of cases, the schools had too few counselors
for the number of students in the school, while 73% of cases had ineffective school response
systems (Lenhardt et al., 2010). In a similar study using basic quantitative methods, Abel et al.
(2022) used case studies of 20 attacks from 1990 to 2016 to determine whether there is a
difference between fatal and non-fatal school shooters and whether there is a difference between
low self-control and high self-control shooters. The mean ages were similar for all categories,
with high-risk offenders found to be twice as lethal as low-risk offenders. Most online warning
signs are communicated via text, instant messaging (IM), email, Instagram, Facebook, or Twitter
(now X). However, most warning signs occurred offline in conversations with peers. Most peers
tended to assume that the perpetrator was joking and did not report the warning signs (Abel et
al., 2022).
Some researchers focused specifically on a single potential warning sign or risk factor for
perpetrators. Dowdell and colleagues (2022) researched risk factors for school shooters and their
social media usage. They studied 25 cases of males aged 12-26 who attacked a middle school,
high school, or college which they currently or previously attended. A methodical process was

53
used to record variables, relying on multiple sources for each recorded variable. A chi-square
analysis was used to test for significant differences in social media use, adverse childhood
experiences, and shooter outcome by the characteristics examined (Dowdell et al., 2022). Hall et
al. (2019) researched whether shooters were on psychotropic medications and how that may have
impacted their attacks. They studied 49 incidents and the mental health treatment of each
perpetrator. A total of 43% of shooters had received some mental health treatment before their
attacks, while 47% of the perpetrators had been prescribed a psychotropic medication at some
point prior to the attack. The majority of perpetrators were not found to be on a psychotropic
medication, and the researchers suggested that the public should not believe that these
medications cause people to act violently (Hall et al., 2019).
Recommendations
Various recommendations have been provided by previous researchers in this area as to
how school shootings may be prevented. This includes that of improved mental health services in
schools, threat assessment, as well as prevention programming (Lenhardt et al., 2010; Lenhardt
et al., 2018). One recommendation is that law enforcement “pay closer attention to individuals
who are struggling with significant personal problems” (Lankford, 2013, p. 13). While some
students may believe attacks are sometimes justified, “only an infinitesimal minority actively
consider engaging in violence” (Lankford, 2013, p. 13). Noticing someone who is struggling at
school or work, having family problems, dealing with an unexpected crisis, or who is socially
isolated may be the best way to prevent violence. Assessing posts online about bullying, low
social status, or loneliness may also be important indicators (Lankford, 2013). Abel et al. (2022,
p. 807) note that the presence of warning signs on obscure online sites indicate a “need for
increased guardianship by parental figures.”

54
Most warning signs are evident shortly before the attack, so schools must develop
policies that take advantage of the small window of opportunity that is available to them (Abel et
al., 2022). Schools should take the time to address adolescent masculinity issues in their
curriculum (Farr, 2018). Discussion-based forums about gender performance with both mixedsex and same-sex groups should be regularized to assist young men in talking about masculine
norms and their challenges (Farr, 2018).
Schools should educate student bodies about school shootings so that students can look
for the warning signs in others, and adults should be encouraged to come forward with concerns
while maintaining protections for boys who are identified as at risk (Farr, 2018). Schools should
become safe communities where diversity is respected, bullying is not tolerated, and students are
encouraged to share any concerns they have (Rothe, 2022). Threat assessments should be done in
a way that protects student privacy and considers any threats made, the student’s family
background, and any major losses the student has experienced (Rothe, 2022).
While characteristics like anger, consumption of violent media, and rejection are present
in school shooters, they should be applied with caution as they are difficult to quantify and most
often do not lead to individuals committing mass atrocities (Ioannou et al., 2015). Smaller school
sizes could contribute to students feeling less alienated and isolated and thus help prevent future
attacks (Lenhardt et al., 2010). Adding social and emotional skills to the curriculum, building
prevention programs, partnering with the community, and building a better communication
infrastructure could all be effective means for preventing mass violence (Lenhardt et al., 2010).
Pfeifer & Ganzevoort (2016) argue that schools must educate students about how to express their
feelings in a healthy way. Also, religious education could be a way to combat the existential
crises faced by students (Pfeifer & Ganzevoort, 2016). Religious education should focus on

55
“religious, ethical, and existential questions” and allow for diversity among students (Pfeifer &
Ganzevoort, 2016, p. 135). Widham et al. (2005) argue that religious factors can protect students
from getting involved in at-risk behavior.
One major contributing factor to school shootings is the ease of access to firearms by
obtaining illegal firearms or using a legal firearm that was stored negligently (Flannery et al.,
2021). One potential solution is smart guns, which include a feature that prevents unauthorized
firearm access. However, these are not sold by any major manufacturers. Some legislative
recommendations include raising the minimum age for purchasing a long gun from 18 to 21 and
banning semiautomatic firearms (Flannery et al., 2021). Raising firearm age limits is not enough,
though. Governments should consider laws that prevent juvenile gun access in homes. Other
measures used internationally include a 28-day waiting period to purchase a firearm and the
acquisition of a license before purchase. Some argue that arming teachers would be a helpful
protective solution, but this measure would likely create more risks than benefits (Flannery et al.,
2021).
The expansion of mental health services alone will not prevent school shootings;
“increased coordination and collaboration” is more essential (Flannery et al., 2021, p. 244).
Collaborative threat assessment teams are more effective than standard school mental health
care. Efforts to “improve student social and emotional learning” are important and may indirectly
contribute to reduced gun violence (Flannery et al., 2021, p. 247). Positive behavioral
interventions and support by multidisciplinary teams have been shown to be the most effective
way to address concerns with students. Multidisciplinary threat assessment teams with mental
health professionals, law enforcement, and community members should be used to address and
prevent troubling behavior (Flannery et al., 2021). In addition to emphasizing the need for

56
interdisciplinary threat assessment teams, Bonanno and Levenson (2014) add that these teams
must have relationships both inside and outside of the school for effective responses.
Silk and Mire (2017) support the creation of a Behavioral Intervention Team to deal with
all student behavioral issues that meets regularly and consists of a diverse team representing
faculty, law enforcement, counseling, disability support, and student conduct officers. To ensure
safety, the goal of the team should be proactive response rather than reactive (Silk & Mire,
2017). The team should lead with empathy by fostering engagement, observation, and immersion
(Silk & Mire, 2017).
For college students, it can be important to include high school personnel in the threat
assessment process (Langman, 2013a). For example, Seung Hui Cho, who waged a violent attack
on Virginia Tech’s campus, had hinted about his desire to repeat the Columbine massacre in high
school, but the college officials were unaware of these threats (Langman, 2013a). Also, the
Threat Assessment Team should be able to deal with potential violence from anyone in a campus
community. While rare, staff or faculty can be attackers (Langman, 2013a). Langman (2013a, p.
13) also recommends that a school should set up an online alert system “to detect threats posted
online that identify a school by name.”
High-profile security responses like metal detectors and surveillance cameras are not
effective measures for combating violence (Lenhardt et al., 2010). Rather than practicing drills
that increase fear and anxiety, schools should implement “unobtrusive security measures . . .
through environmental design” to deter attacks (Flannery et al., 2021, p. 248). An ongoing risk
assessment model is essential to assess a risk as it evolves (Lenhardt et al., 2010). Sufficient
staffing and resources are necessary to effectively assess risks and mitigate threats (Lenhardt et
al., 2010).

57
On the other hand, some school security measures can help prevent school shootings
before they occur. Recommendations can be described as consisting of four categories: 1)
limiting access to the school, 2) prohibiting weapons on campus, 3) increasing surveillance, and
4) increasing measures for reacting in a crisis (e.g., telephones in classrooms, duress alarms, and
student drills) (Bonanno & Levenson, 2014). However, it is important to balance security
measures with the potential for negative effects on the school environment. One essential step in
the mitigation of threats is making sure that each student has a positive connection with “at least
one adult within the school with whom he or she can discuss concerns” (Bonanno & Levenson,
2014, p. 7). Similarly, Redlener (2006) argues that important steps include: limiting points of
entry, installing wireless panic alarms in schools and strategically placed telephones, building
relationships between school officials and local law-enforcement, teaching students and teachers
to be situationally aware, and encouraging parents to advocate for improved school security.
Additionally, schools should note that anytime a student talks about killing, this is a warning sign
and should be taken seriously (Duplechain & Morris, 2014).
Wetterneck and colleagues (2005) argue that students need a safe way to be able to report
threats to adults in a fair and responsible way. Dowdell et al. (2022) recommend that all health
care providers who work with adolescents provide a safe space for the patient to disclose their
thoughts and concerns. It is important for health-care workers to use a trauma-informed response
when working with adolescents who have been exposed to trauma or violence in order to build a
safe, stable, and nurturing relationship with the adolescent (Dowdell et al., 2022). A support
system for students who are experiencing bullying is also necessary. Students should also be
given a way to report threatening or disturbing posts from peers on social media to authorities
(Dowdell et al., 2022). Also, “[p]revention programs may benefit from tailoring their messages

58
based on gender,” as females are more likely to report a perceived threat than males (Wetterneck
et al., 2005, p. 163).
Comparisons with Other Groups
Much effort has been made to differentiate school shooters from other perpetrators of
mass homicides, but in reality they are similar groups (Ferguson et al., 2011). Based on a Secret
Service study released in 2002, most perpetrators had a documented history of depression, and
many had attempted suicide (Ferguson et al., 2011). The data are very similar to data available on
adult perpetrators of mass homicide. Often school shooters are differentiated from other violent
groups without any substantial rationale. While this decision is not necessarily wrong, it requires
better empirical research and reasoning (Grøndahl & Bjørkly, 2016).
Some have stated that the primary difference between school shooters and other
perpetrators of mass violence is the location of their attacks (Lankford, 2013). However, Carey
(2007) argues that school shooters are distinct from suicide terrorists because suicide terrorists
are mentally healthy and motivated purely by ideology, while school shooters are mentally
troubled and motivated by personal problems. Otherwise, school shooters tend to have
similarities with suicide terrorists, leading to conjecture that they have similar psychologies and
behaviors (Lankford, 2013).
Lankford (2013) found that school shooters were far younger than suicide terrorists,
rampage shooters, and workplace shooters. Additionally, the vast majority of perpetrators of
mass violence are male. Workplace shooters are significantly less likely to leave a suicide note
than school shooters, suicide terrorists, or rampage shooters, while workplace shooters are also
far less likely to have family problems. Workplace and school shooters are more likely to
struggle with work or school problems, with workplace shooters seemingly most different out of

59
the four types of perpetrators of mass violence (Lankford, 2013). Workplace shooters tend to be
the most “normal” psychologically, appear to have a shorter period of premeditation, and are
more likely to specifically target individuals in attacks. School shooters, on the other hand,
choose their school as a target with general familiarity but rarely target specific enemies. Even
when they feel wronged by specific individuals, they rarely seek them out during attacks
(Lankford, 2013).
Summary
Episodes of mass violence in schools are terrifying, highly publicized, and rightfully
provoke the public to call for reform. However, they remain very rare. Attackers tend to be white
males with histories of trauma, psychosis, or psychopathy. Theories provide some insight into the
motivations and purposes behind school shooters, but some attackers defy patterns and
categorization.
Existing research ranges from conclusions drawn based on case study assessments of
attacks to rigorous statistical analyses derived from meticulously coded primary sources. The
best measures for prevention consist of equipping students with the social and emotional tools to
deal with personal issues, providing a safe way for students to communicate their concerns about
themselves or another student, and building an adept threat assessment team to confront threats
quickly and effectively.

60
Chapter 4: Methods
Introduction
This chapter details the methods used in this study. This consists of a discussion of this
study’s research design, which was quantitative and descriptive as well as correlational, along
with the setting and participants relevant to this study. As secondary data analysis was used, the
setting and data collected were not determined by the researcher, but instead were established by
the researchers who collected these data initially. Next, instrumentation is discussed, with this
section including a description of the measures included in this dataset and used for analysis. The
Procedure section discusses how these data were originally collected in the various studies
whose data were compiled and used in this current study, with the process of data collection and
analysis then detailed. Finally, a summary of this chapter is presented.
Research Design
A quantitative research design was selected for use in this study, which is appropriate
when conducting hypothesis tests, that require empirical, statistical analysis of data such that the
relevant null hypotheses can either be rejected or not rejected (Edmonds & Kennedy, 2016).
Descriptive and correlational research designs were felt to be most appropriate, these being
designs that fall within the broader quantitative design (Edmonds & Kennedy, 2016). The focus
of this study was to describe the data present; namely, to describe perpetrators of school
violence, as well as other perpetrators of violence. The correlational design (Edmonds &
Kennedy, 2016) also implemented here pertains to this study’s comparisons of perpetrators of
school violence with perpetrators of other forms of violence, ACEs and their correlates, the
comparisons between the JRIC and TASSS samples, and the comparisons made between the
TASSS sample and data derived from the general population.

61
In the first set of analyses examining the JRIC dataset only, the independent variable of
interest consisted of perpetrator type; namely, whether the perpetrator was a perpetrator of school
violence or another type of violence. Within this “other” category, terrorists were most common,
with this category also including those who committed acts of violence based on their beliefs,
which included environmentalism, as well as abortion and animal rights concerns, along with
those committing violent acts on the basis of religion, race, or political beliefs.
Nearly all other measures in JRIC consisted of the dependent variables included in this
initial set of analyses. In this way, this study examined nearly all possible differences between
perpetrators of school violence and other perpetrators. Specifically, these dependent variables
consisted of the following: the year of the incident, perpetrator gender, age, and marital status,
access to weapons, military experience, law enforcement contacts, emotional distress, verbal
statements, online postings, suspicious travel, violence to the self, others, and violence from
others, property destruction, mental health diagnoses and symptoms, experiencing a loss due to a
death, the loss of a job, relationship, or residence, alcohol and drug abuse, physical or verbal
abuse, financial and relationship difficulties, vagrancy, having been radicalized by family, a
friend, or an associate, having experienced war, grievances taking the form of economic,
personal, political, racial, religious, or social, ideology (which incorporated anarchist, antigovernment, black nationalism, religious, sovereign citizen, and white supremacy ideologies),
Internet searches, extremist media consumption, the social media platform used, social media
activities, violent video consumption, the highest level of education completed, and having a
learning disability.
The following analyses examined ACEs and their correlates, both with regard to the JRIC
dataset as well as the TASSS dataset. Within the JRIC dataset, ACEs consisted of emotional

62
distress, having a mental health diagnosis, exhibiting mental health symptoms, a loss due to a
death, job, relationship, or of a residence, alcohol abuse, drug abuse, physical/verbal abuse,
financial difficulties, and relationship difficulties. Risk factors consisted of having been
radicalized by a family, friend, or associate, having experienced war, having grievances that were
economic, personal, political, racial, religious, or social, having an anarchist, anti-government,
black nationalist, religious, sovereign citizen, or white supremacist ideology, violent video
consumption, weapons access, and military experience. Finally, the byproduct of ACEs, termed
External Factors, consisted of contact with law enforcement, violence to the self, others, or from
others, and vagrancy. No Environmental Sustainable Design measures were identified within the
JRIC dataset.
With regard the analyses examining ACEs within TASSS, the ACEs examined consisted
of domestic violence, a workplace shooting, psychological issues, parents having been divorced
or separated, social stratum, significant family problems, the recent death of a relative or friend,
loss of social standing, having been dismissed from a social, political, or religious organization,
and peer aggression. Next, the External Factors analyzed consisted of a K-12 suspension or
expulsion (coded as two separate measures), a K-12 failure, a K-12 dropout, having been a
member of a street gang, having a criminal record, a victim being affiliated with a gang, and
having struck a student, teacher, administrator, or someone else, with these latter measures
consisting of four separate measures. Finally, the Environmental Sustainable Design measures
consisted of the presence of a metal detector, a school guard, and a school police officer.
In the analyses conducted comparing the JRIC and TASSS datasets, a mental health
diagnosis or symptoms in JRIC were examined alongside any psychological issues in TASSS. A
loss due to a death in JRIC was examined with the recent death of a relative or friend in TASSS,

63
while the loss of a job in JRIC was examined with the loss of social standing in TASSS. Finally,
the experience of physical or verbal abuse in JRIC was examined alongside whether the
perpetrator was a victim of peer aggression in TASSS.
The final set of analyses examined the TASSS data alongside data derived from the
general population. These analyses focused upon ACE measures identified in the TASSS dataset
along with ACE Externalizing Factors. The ACE measures examined in TASSS consisted of
parents having been divorced or separated, being in the low social stratum, and having
experienced significant family problems, with ACE Externalizing Factors consisting of K-12
expulsion rates, suspension rates, and dropout rates, along with the proportion of individuals that
were street gang members.
An alpha of .05 was selected for use in the determination of whether this study’s null
hypotheses were rejected or failed to be rejected. This is the standard within empirical research
(Wagner & Gillespie, 2018). The research questions included in this study consisted of the
following:
1. Research Question 1: Do incidents of school violence and incidents of other forms of
violence differ on the basis of year?
2. Research Question 2: Do perpetrators of school violence and perpetrators of other forms of
violence differ with regard to demographics?
3. Research Question 3: Do perpetrators of school violence and perpetrators of other forms of
violence differ with regard to weapons, the military, and law enforcement contact?
4. Research Question 4: Do perpetrators of school violence and perpetrators of other forms of
violence differ with regard to online behavior and social media use?

64
5. Research Question 5: Do perpetrators of school violence and perpetrators of other forms of
violence differ with regard to travel?
6. Research Question 6: Do perpetrators of school violence and perpetrators of other forms of
violence differ with regard to violence?
7. Research Question 7: Do perpetrators of school violence and perpetrators of other forms of
violence differ with regard to abuse?
8. Research Question 8: Do perpetrators of school violence and perpetrators of other forms of
violence differ with regard to radicalization?
9. Research Question 9: Do perpetrators of school violence and perpetrators of other forms of
violence differ with regard to grievances?
10. Research Question 10: Do perpetrators of school violence and perpetrators of other forms of
violence differ with regard to extremist media consumption?
11. Research Question 11: Do perpetrators of school violence and perpetrators of other forms of
violence differ with regard to other psychological and sociological measures?
12. Research Question 12: Are Adverse Childhood Experiences (ACEs) associated with External
Factors?
13. Research Question 13: Are Adverse Childhood Experiences (ACEs) associated with
Environmental Sustainable Design measures?
14. Research Question 14: Do the JRIC and TASSS samples differ?
15. Research Question 15: Does the TASSS sample differ from the general population?
The following null and alternative hypotheses were then created based on these 15
research questions, with each numbered null and alternative hypothesis being associated with the
same numbered research question:

65
• Hypothesis 1A: Incidents of school violence and incidents of other forms of violence
differ on the basis of year.
• Hypothesis 10: Incidents of school violence and incidents of other forms of violence do
not differ on the basis of year.
• Hypothesis 2A: Perpetrators of school violence and perpetrators of other forms of
violence differ with regard to demographics.
• Hypothesis 20: Perpetrators of school violence and perpetrators of other forms of violence
do not differ with regard to demographics.
• Hypothesis 3A: Perpetrators of school violence and perpetrators of other forms of
violence differ with regard to weapons, the military, and law enforcement contact.
• Hypothesis 30: Perpetrators of school violence and perpetrators of other forms of violence
do not differ with regard to weapons, the military, and law enforcement contact.
• Hypothesis 4A: Perpetrators of school violence and perpetrators of other forms of
violence differ with regard to online behavior and social media use.
• Hypothesis 40: Perpetrators of school violence and perpetrators of other forms of violence
do not differ with regard to online behavior and social media use.
• Hypothesis 5A: Perpetrators of school violence and perpetrators of other forms of
violence differ with regard to travel.
• Hypothesis 50: Perpetrators of school violence and perpetrators of other forms of violence
do not differ with regard to travel.
• Hypothesis 6A: Perpetrators of school violence and perpetrators of other forms of
violence differ with regard to violence.

66
• Hypothesis 60: Perpetrators of school violence and perpetrators of other forms of violence
do not differ with regard to violence.
• Hypothesis 7A: Perpetrators of school violence and perpetrators of other forms of
violence differ with regard to abuse.
• Hypothesis 70: Perpetrators of school violence and perpetrators of other forms of violence
do not differ with regard to abuse.
• Hypothesis 8A: Perpetrators of school violence and perpetrators of other forms of
violence differ with regard to radicalization.
• Hypothesis 80: Perpetrators of school violence and perpetrators of other forms of violence
do not differ with regard to radicalization.
• Hypothesis 9A: Perpetrators of school violence and perpetrators of other forms of
violence differ with regard to grievances.
• Hypothesis 90: Perpetrators of school violence and perpetrators of other forms of violence
do not differ with regard to grievances.
• Hypothesis 10A: Perpetrators of school violence and perpetrators of other forms of
violence differ with regard to extremist media consumption.
• Hypothesis 100: Perpetrators of school violence and perpetrators of other forms of
violence do not differ with regard to extremist media consumption.
• Hypothesis 11A: Perpetrators of school violence and perpetrators of other forms of
violence differ with regard to other psychological and sociological measures.
• Hypothesis 110: Perpetrators of school violence and perpetrators of other forms of
violence do not differ with regard to other psychological and sociological measures.

67
• Hypothesis 12A: Adverse Childhood Experiences (ACEs) are associated with External
Factors.
• Hypothesis 120: Adverse Childhood Experiences (ACEs) are not associated with External
Factors.
• Hypothesis 13A: Adverse Childhood Experiences (ACEs) are associated with
Environmental Sustainable Design measures.
• Hypothesis 130: Adverse Childhood Experiences (ACEs) are not associated with
Environmental Sustainable Design measures.
• Hypothesis 14A: The JRIC and TASSS samples differ.
• Hypothesis 140: The JRIC and TASSS samples do not differ.
• Hypothesis 15A: The TASSS sample differs from the general population.
• Hypothesis 150: The TASSS sample does not differ from the general population.
The inferential statistical tests described in detail later in this chapter served to test all 15 null
hypotheses included in this section.
Data Sources
This study used secondary data analysis, with the initial analyses, using the JRIC dataset,
conducted using a dataset that was compiled from a series of previous studies. Specifically, as
also detailed later in this chapter, the data analyzed were derived from the Joint Regional
Intelligence Center (JRIC), and was a compilation of a number of datasets, specifically CSISTNT, GTD, PIRUS, RAND DWTI, TIA (with this consisting of Plots, Perpetrators, and Origins
datasets), tPP (this consisting of Capital and Protests datasets), ACLED, FBI–20 Year Active
Shooter, JRIC Project Lebanon, and JRIC Targeting datasets. Due to the large number of datasets
and studies from which these data were drawn, specifics will not be discussed here with respect

68
to geographic location, if any, relevant to these specific studies, or with regard to the participant
pool. However, basic information will be presented regarding this dataset. The total sample
includes 319 perpetrators after all data cleaning was complete. This includes 13 perpetrators of
school violence and 306 perpetrators of other forms of violence.
The TASSS dataset was compiled by a group of researchers focusing on school shootings
that occurred in the U.S. between 1990 and 2016 (Freilich et al., 2021). Open-sourced, public
information, including media reports, court records, and social media, as well as other sources
were used. In addition to the date range and having taken place in the U.S., four other inclusion
criteria were present, which consisted of a criminal justice response having been present, that a
firearm must have been used, the shooting injury must have occurred at a K-12 school, and the
gun discharge must have injured or killed at least one person with a bullet wound. A total sample
of 652 shootings is documented in this dataset (Freilich et al., 2021).
In addition, as the researcher did not collect original data, no power analyses were
conducted in relation to this study, and this study’s sample size was not determined on the basis
of any power analysis conducted. Other settings and participants were present with respect to the
data collected and analyzed from the general population; however, as this encompasses a broad
range of studies, these specific details will not be presented here.
Instrumentation
The instrumentation pertaining to the JRIC data will be summarized as best possible
based on the documentation available on this dataset. The list of the full set of datasets compiled
to create the final JRIC dataset used in this study will not be repeated here, while these various
datasets had been merged as best possible in order to create the dataset used in this study.

69
This dataset incorporated a large set of categories of measures pertaining to individual
perpetrators. Demographics were present, which consisted of the gender of the perpetrator, their
age, marital status, and their highest level of education completed. Additionally, details were also
present with regard to the plot itself, including the name of the plot as well as the date at which it
occurred. Additional categories of measures pertaining to the perpetrator that were included in
these data consisted of the following: access to weapons of any kind, having a history of law
enforcement contacts or behavioral health issues, a history of possible mental illness, emotional
distress, violence experienced toward others or the self, alcohol or substance abuse, loss,
difficulties with relationships or finances, unstable living arrangements, social isolation,
grievances, verbal or written statements (i.e., “leakage”), Internet searches on grievances or
possible targets, violent or extremist media consumption, directed travel, and ideological
identification.
Of the above set of categories just described, the specific measures consisted of access to
weapons, military experience, having experienced war, contact with law enforcement, emotional
distress, verbal statements having been made, online postings, suspicious travel, violence to the
self, others, or violence from others, property destruction, a mental health diagnosis or
symptoms, a loss due to a death, a loss of a job, a loss of a relationship, and a loss of a residence,
alcohol or drug abuse, physical or verbal abuse, having experienced financial or relationship
difficulties, vagrancy, having been radicalized by a family, friend, or associate, having a
grievance, with separate measures recording whether the perpetrator had an economic, personal,
political, racial, religious, or social grievance, ideology, with separate measures recording an
anarchist, anti-government, black nationalist, religious, sovereign citizen, or white supremacist
ideology, having made Internet searches on grievances or possible targets, having consumed

70
extremist media, the social media platform used (with a total of five possible responses present),
social media activities that were engaged in (with a total of three possible responses), the
consumption of violent videos, and the presence of learning disabilities.
Regarding the TASSS dataset, a large proportion of the variables present pertained to
details relating to the shooting itself, the school, and associated content. These consisted of
whether it was a workplace shooting, the number of police documents, court documents, other
government documents, education documents, news documents, scholarly documents, websites,
and other documents, along with the total number of documents, the document reliability score,
school type, whether the shooting took place during the school day, whether the shooting was
intentional, the type of firearm used, whether a handgun was used, how the gun was obtained,
whether the gun was obtained illegally, the total number of deaths, the total number of injuries,
the total number of victims, whether there were fatalities or no fatalities, whether it was a mass
shooting (three or more victims, and four or more victims), whether the shooting occurred inside
a school building, whether the shooting occurred on a school bus, the number of students in the
attack location, the number of students in the attack area, whether a metal detector was present in
the school, whether a school guard/resource officer was present in the school, whether a school
police officer was present, the presence of barriers, accessibility in the school, whether the school
has multiple buildings, whether the school is multistoried, and whether there was a specific
target, symbolic target, and general target.
Measures pertaining to the shooter himself consisted of whether the shooter was an adult
or a juvenile, whether domestic violence was present, whether the perpetrator is a current
student, their age, gender, and race, their highest level of education completed, high school grade
level, employment status, full-time or part-time employment status, whether they have any

71
psychological issues, whether their parents are divorced or separated, their social stratum,
whether they have any significant family problems, whether they were suspended or expelled (K12), whether they experienced a K-12 failure or dropout, whether they are a street gang member,
whether they experienced the recent death of a friend or relative, whether they experienced a loss
of social standing, whether they were dismissed from a social, political, or religious organization,
whether they were a victim of peer aggression, had a criminal record, and if they struck another
student, a teacher, an administrator, or someone else. An additional measure recorded whether
the victim was gang affiliated. The remaining issue of operationalization pertains to the studies
which examined the general population; these details are presented later in the results.
Procedure
Again, as secondary data analysis was used in this study, the researcher did not collect
any original data. Due to the substantial number of studies from which this dataset was derived,
the backgrounds with regard to data collection associated with each of these various datasets will
not be reproduced here. Additionally, as secondary data analysis was relied upon in this study,
this study was deemed as being of minimal risk to individuals, and it was felt that the likelihood
and magnitude of any harm that could have been experienced by individuals was not deemed as
being any greater than that commonly encountered in everyday life.
Data Processing and Analysis
The JRIC data were derived from the Center and consisted of a compilation of a number
of datasets, these consisting of CSIS-TNT, GTD, PIRUS, RAND DWTI, TIA (with this
consisting of Plots, Perpetrators, and Origins datasets), tPP (this consisting of Capital and
Protests datasets), ACLED, FBI–20 Year Active Shooter, JRIC Project Lebanon, and JRIC
Targeting datasets. These various datasets had all been compiled, on the level of the perpetrator,

72
for use in this current study’s analyses, and the following chapter will serve to summarize these
perpetrator-level data separately with respect to perpetrator type, which was categorized as either
school-related violence perpetrator or other perpetrator.
The JRIC dataset was initially provided as an Excel format spreadsheet. This was then
comprehensively cleaned by deleting unnecessary columns and rows, with this cleaned dataset
imported into Stata 17.0 SE (StataCorp, 2021) for further data cleaning. This process of data
cleaning in Stata entailed dropping cases that had largely missing data, as well as dropping
duplicate cases that were present in the dataset. Additionally, each case identified as pertaining to
school violence was individually reviewed in order to ensure that the case in question was indeed
school violence-related for the purposes of properly conducting the proposed analyses comparing
perpetrators of school violence with other perpetrators of violence. This also entailed reviewing
all cases not identified in the dataset as being relevant to school violence to ensure that this
categorization was accurate in all cases, and that none of these cases did in fact pertain to school
violence. This process ensured that all cases identified as being related to school violence were
indeed school violence cases, and the same with regard to all non-school violence cases.
In a number of cases, duplicate cases were also identified and these duplicated cases were
either dropped from the dataset or merged together. This required numerous replacements of
individual data points so that these duplicate cases could be appropriately merged, in cases where
they were in fact merged. This process primarily entailed examining and comparing cases on the
basis of perpetrator name, while all data on perpetrator name were removed from the dataset
prior to analysis for the purposes of maintaining anonymity. Further checks were also conducted
in order to ensure that no duplicates were present in cases where no data were available on the
perpetrator’s name. Additionally, cases in which the perpetrator’s name was repeated, but in

73
which different dates were present when examining and comparing these specific instances of
violence, were not deemed as duplicate cases, but instead separate instances of violence
committed by the same individual.
Additionally, numerous discrepancies existed within this dataset that required correction
prior to any analyses being conducted. This consisted of issues, such as the same measure having
response categories of “No” and “no”, which are deemed as separate and distinct response
categories by the statistical software. For this reason, frequencies of all study measures were first
run and reviewed, and all relevant recodes made, in order to avoid this problem. Besides issues
of capitalization, this also pertained to minor discrepancies involving formatting, white space,
and so on. A number of study measures were also found to contain data from other relevant
variables; in these cases, these specific data points were recoded as missing. These two items
required several hundred lines of Stata syntax to correct.
Once the process of data cleaning was complete, analyses were conducted in Stata 17.0
SE (StataCorp, 2021). These analyses consisted of descriptive statistics, bivariate statistics, as
well as regression analysis. Nearly all of the study variables were categorical in nature.
Frequency tables were selected for use in reporting the descriptives pertaining to this study’s
categorical measures as this is an appropriate method when examining categorical measures on a
descriptive level (Nisbet et al., 2009). Of this study’s measures, only age and year were
continuous, and in these two cases, the mean and standard deviation were calculated as measures
of central tendency and variability, respectively, along with the range of these two measures, as
measures of central tendency and variability are appropriate in relation to continuous measures
(Liu, 2022).

74
The bivariate analyses conducted in this study primarily consisted of Fisher’s exact tests,
which are appropriate in situations where Pearson’s chi-square is appropriate, namely, when
examining whether a significant association exists between two categorical variables, while
being an exact test (Weinberg & Abramowitz, 2008). Fisher’s exact tests were used in order to
determine whether significant differences existed between perpetrators of school violence and
other perpetrators of violence with regard to the categorical study variables.
In addition to these Fisher’s exact tests, independent-samples t-tests were also conducted
for the continuous measures of individual age and year. The independent-samples t-test is
appropriate when examining whether the mean of some continuous outcome differs significantly
on the basis of some dichotomous independent variable; the focus, therefore, is on some group
comparison (Weinberg & Abramowitz, 2008), which was the case here. These analyses also
incorporated Levene’s tests of the equality of variance in order to determine whether this
assumption of the independent-samples t-tests had been violated (Weinberg & Abramowitz,
2008). If this assumption was violated, the calculation of the independent-samples t-test was
specified to use unequal variances in light of this assumption violation.
Finally, two linear regression analyses were also conducted on these data, with these
analyses being conducted separately on the basis of perpetrator group. These linear regression
analyses regressed a newly created measure calculated as the number of years following 2000 on
the number of incidents per year. The purpose of these analyses was to determine whether any
significant trend in the number of incidents was present over time, either with regard to an
increasing or decreasing number of incidents over time. Linear regression analysis is appropriate
when examining one continuous outcome, alongside one or more independent variables that can
take on any level of measurement (Shreve & Holland, 2018).

75
As the TASSS dataset was already in a clean state, no data cleaning was necessary with
the exception of recoding values of “-88” as missing with regard to both domestic and work
violence. All analyses examining the TASSS dataset itself consisted of Fisher’s exact tests, along
with the same descriptives as were conducted in JRIC. All analyses conducted comparing the
TASSS and JRIC samples, as well as comparing the TASSS sample with general population data,
used solely difference in proportions tests. In all inferential statistical tests conducted in this
study, a probability level of below .05 was taken to indicate statistical significance, while a
probability level of .05 or above was deemed to indicate that no significance was present, due to
this study’s use of a .05 alpha level.
Summary
This study used a quantitative research design incorporating both descriptive and
correlational elements. A series of 15 research questions and hypotheses were proposed in this
study, with a detailed description of the datasets used presented, along with a list of all variables
included in these datasets and a list of the datasets themselves that were combined into a single
dataset in the case of JRIC.
A detailed description was also presented with regard to the data cleaning process used
with regard to JRIC, and the analyses that were applied to this study’s data and the justification
behind the selection of these particular statistical tests.
The following chapter will present and discuss the results of the analyses conducted on
JRIC alone, and includes the descriptive statistics conducted, as well as the inferential statistical
tests conducted on these data.

76
Chapter 5: Results: JRIC Comparisons and Time Analyses
Introduction
In this chapter, the results of the analyses conducted solely on the JRIC dataset are
presented and discussed. Initially, these data were split into two separate subsamples, school
violence cases, as well as all other cases of violence, with these latter data mainly consisting of
incidents of terrorism. These two datasets were then analyzed separately using univariate
descriptive statistics, with the same set of descriptive statistics conducted separately on both
samples. The descriptive statistics conducted primarily consisted of frequencies being calculated
and reported on the study measures, as the vast majority of these measures were categorical in
nature. This consisted of the determination and reporting of the frequencies and percentages
associated with all response categories relevant to these measures. Several continuous measures
were also included in these data, and in these cases, the mean and standard deviation was
determined and reported, as a measure of central tendency and variability, respectively.
Additionally, bivariate analyses were conducted on the combined dataset, which
examined whether perpetrators of school violence significantly differed from other perpetrators
of violence with regard to all study measures. Nearly all of these analyses consisted of Fisher’s
exact tests, with cross-tabulation tables being presented in relation to the significant test results
in order to illustrate how these two groups differed in additional descriptive detail. Additionally,
linear regression analyses were also conducted in order to examine trends over time with regard
to incidents of school violence, as well as other incidents of violence.
School Perpetrators: Descriptive Statistics
Initially, a series of descriptive statistics were conducted on the 13 school violence
perpetrators identified in these data. Regarding the dataset in question, 10 of these 13 individuals

77
were derived from the GTD dataset, with two identified in the PIRUS dataset. Finally, a single
individual was identified in the TIA-Perpetrators dataset.
Table 5.1 presents a description of the “Main Factor” measures in relation to the school
violence perpetrators only. These measures varied slightly in their focus, while all examined the
major reasons or contributing factors behind the individual’s violent acts. The fifth measure
pertaining to this set of items is omitted from Table 5.1 as no valid cases were present with
regard to this measure among perpetrators of school violence.
To summarize the results of the descriptive statistics conducted on this set of items,
common factors found consisted of online postings (leakage), violence to others, having
displayed symptoms of mental health, and extremist media consumption. Exact frequencies and
percentages associated with each response category for all “Main Factor” measures for
perpetrators of school violence are presented in Table 5.1.
Table 5.1
Summary of Main Factors for JRIC School Violence Perpetrators (n=13)
Measure/Category Number Percent of
Valid Cases
Main Factor 1
Emotional Distress 1 10.00%
Online Postings (Leakage) 4 40.00%
Travel (Suspicious) 1 10.00%
Violence to Others 4 40.00%
Main Factor 2
Mental Health (Diagnosed condition) 2 66.67%
Mental Health (Displayed symptoms) 1 33.33%

78
Main Factor 3
Mental Health (Displayed symptoms) 4 100.00%
Main Factor 4
Physical (Perpetrator) 1 33.33%
Verbal (Perpetrator) 2 66.67%
Main Factor 6
Exposure–Conversion 1 16.67%
Grievance–Personal 2 33.33%
Grievance–Religious 2 33.33%
Ideology–White Supremacy 1 16.67%
Main Factor 7
Extremist Media Consumption 3 50.00%
Internet Searches 2 33.33%
Yes 1 16.67%
Note: The fifth measure pertaining to this set of items is omitted from Table 5.1 as no valid cases
were present with regard to this measure among perpetrators of school violence.
Regarding when these incidents occurred, dates were present with regard to 12 of these
13 cases, with the earliest case occurring on January 10, 2000, and with the final incident having
taken place on May 18, 2018. This produced a range of slightly above eight years. Next, with
respect to the individuals who perpetrated these acts of school violence, these names were
removed from the dataset prior to analysis for the purposes of anonymity. However, it will be
stated that of these 13 individuals, 11 were named, having valid data for both their first and last

79
names, while two were anonymous, with simply “juvenile” present for both their first and last
names in the dataset.
Next, regarding gender, 12 (92.31%) perpetrators were male, with a single perpetrator
(7.69%) female. Mean age was 18.29 years, with a median of 17 years. The youngest perpetrator
was 16 years of age, with the oldest 22. Regarding marital status, all seven perpetrators who had
data for marital status were found to be single (never married). With respect to whether the
perpetrator was charged in two cases out of the six that had valid data (33.33%), no charges were
filed as the perpetrator was deceased, while in the remaining four cases (66.67%), the perpetrator
was free/released. Regarding the similar measure of criminal status, of the seven cases that had
valid data for this measure, two (28.57%) were dead by suicide, one (14.29%) was deceased, and
four (57.14%) had an entry of “none.”
Several measures were thought to have potentially been associated with school violence,
with these consisting of weapons access, military experience, and law enforcement contacts.
Regarding weapons access, of the 10 perpetrators that had valid data for this measure, three
(30.00%) had a value of “unknown,” while the remaining seven (70.00%) were coded
affirmatively. With respect to military experience, only a single individual had valid data, with
this consisting of “U.S. Marines.” Contact with law enforcement was relatively common, with
two individuals (15.38%) having contact with law enforcement over a non-violent felony, and
one (7.69%) having contact with law enforcement over a violent crime. Additionally, a code of
“unknown” was associated with two cases (15.38%), while no contact was indicated in the
remaining eight cases (61.54%).
Additional measures examined items that were again theorized to be associated with
school violence. With respect to emotional distress, out of the six cases that had data for this

80
measure, four individuals had experienced emotional distress (66.67%), with this coded
negatively in a single case (16.67%). Additionally, this was unknown in one additional case
(16.67%). Regarding verbal statements, an affirmative instance was indicated in seven out of the
eight cases that had valid data for this measure (87.50%), and was unknown in the remaining
case (12.50%). Online activity was also frequent, and out of the 10 cases that had data for this
measure, six (60.00%) indicated online postings, with the remaining four (40.00%) indicating
that no online postings were made. Only one individual (12.50%) was found to have suspicious
travel in their history, with this being unknown in one case (12.50%), and the remaining six
(75.00%) coded negatively with respect to this question. Finally, four individuals (50.00%) were
found to have committed property destruction, with three having not (37.50%), and with this
being unknown in a single case (12.50%).
The following measures included in these data pertained to measures of violence;
specifically, these related to violence to the self, violence to others, as well as violence from
others. In all three cases, a total of eight individuals had valid data on these measures. Regarding
violence to the self, seven individuals (87.50%) were coded negatively in relation to this
question, while one individual (12.50%) had a code of “unknown.” Next, with respect to
violence to others, six individuals (75.00%) were coded negatively in relation to this question,
with one (12.50%) coded affirmatively, and one (12.50%) with a code of “unknown.” Finally,
with regard to violence received from others, four negative instances were found (50.00%), with
one affirmative instance (12.50%), and a code of “unknown” found in three cases (37.50%).
The following measures examined mental health and loss. Regarding mental health, both
a diagnosis of a mental health problem was measured, along with the appearance of mental
health symptoms. Out of these 13 cases, a diagnosis of a mental illness was indicated in three

81
cases (23.08%), with this being negative in nine cases (69.23%), and with the mental health
status of the individual being unknown in one case (7.69%). Additionally, the individual was
found to have exhibited mental health symptoms in five cases (38.46%) and to have not in the
remaining eight (61.54%).
The following four measures focus on different aspects of loss, this being loss from a
death, the loss of one’s job, loss based on a relationship, and the loss of a residence. In all four
cases, valid data were present for eight individuals. In examining these measures, it was found
that experiencing loss from a death was unknown in four cases (50.0%), and with four
individuals having a code of “no” (50.00%). Regarding the loss of a job, this was found to be
unknown in a total of three cases (37.50%), with negative instances found in the remaining five
cases (62.50%). Regarding a loss associated with a relationship, this was found to be unknown in
four cases (50.00%), with a negative instance found in the remaining four cases (50.00%). Next,
regarding the loss of a residence, this was found to be unknown in two cases (25.00%), with a
negative instance indicated in the remaining six cases (75.00%).
The following measures focused upon abuse, with this consisting of alcohol abuse, drug
abuse, as well as physical and verbal abuse. A total of eight individuals were found to have valid
data with respect to both alcohol and drug abuse, with all 13 school violence perpetrators found
to have valid data with respect to physical and verbal abuse. Regarding alcohol abuse, a code of
“unknown” was found in two cases (25.00%), with negative instances found in the remaining six
cases (75.00%). Regarding drug abuse, this was indicated as unknown in six cases (75.00%),
with negative instances found in the remaining two cases (25.00%). Finally, with respect to
having experienced physical or verbal abuse, this was found to be the case in two out of these 13

82
individuals (15.38%), was unknown in a single case (7.69%), and with negative instances being
indicated in the remaining 10 cases (76.92%).
The following measures examined different areas of instability, as well as radicalization.
Here, instability related to instability in the domain of personal finances, difficulties with
relationships, and instability with respect to the individual’s residence. Additionally,
radicalization pertained to whether individuals had been radicalized by family, friends, or
associates. The results of these latter measures will be described here, while these are generally
more closely associated with terrorists.
Regarding financial difficulties, this was found to be unknown in a total of two cases
(25.00%), with negative instances indicated in the remaining six cases that had valid data on this
measure (75.00%). Regarding relationship difficulties, this was found to have been present
among three out of the 13 individuals (23.08%), with this having not been the case among five
individuals (38.46%), and with this being unknown in the remaining five cases (38.46%).
Finally, regarding vagrancy, of the 13 school violence perpetrators, negative instances were
found in 11 cases (84.62%), and with this being unknown in the remaining two cases (15.38%).
With respect to whether individuals had been radicalized, this was found to have been the
case in four individuals, though in these four cases, each individual had been radicalized by a
friend, as opposed to by a family member or associate. To summarize these data, with regard to
whether these 13 individuals had been radicalized by a family member, negative instances were
indicated in 11 cases (84.62%), with this being unknown in the remaining two cases (15.38%).
As mentioned, four individuals were found to have been radicalized by a friend (30.77%), with
negative instances indicated in six cases (46.15%), and with this being unknown in the remaining
three cases (23.08%). With respect to having been radicalized by an associate, negative instances

83
were indicated in 10 cases (76.92%), and with this being unknown in the remaining three cases
(23.08%).
The following measures examined individual experience in war, as well as various
grievances, with these consisting of economic, personal, political, racial, religious, and social
grievances, with each of these six items being examined separately. Regarding having had an
experience of war, there were two affirmative instances in total, with one consisting of “Ruby
Ridge/Waco” (7.69%), and with an additional code of “other” (7.69%). Experience of war was
unknown in three cases (23.08%), with a code of “none” provided in the remaining eight cases
(61.54%).
The following measures examined the six grievances discussed above. Regarding
economic grievances, of the eight individuals who had valid data for this measure, there was a
single code of “unknown” (12.50%), and with seven negative instances (87.50%). Next,
regarding personal grievances, 10 individuals were found to have valid data on this measure, and
with six of these 10 coded affirmatively (60.00%). Additionally, three individuals were found to
be coded negatively (30.00%), and with a single individual having a code of “unknown”
(10.00%). The following measure examined political grievances, and out of the eight valid cases
associated with this measure, one individual was found to have had a political grievance
(12.50%), with this being negative in six cases (75.00%), and with a single code of “unknown”
(12.50%).
The remaining three measures examining grievances focused on racial, religious, and
social grievances. None of the eight individuals who had valid data on racial grievances were
found to have any; negative instances were indicated in seven cases (87.50%), and with this
being unknown in the remaining case (12.50%). With respect to religious grievances, affirmative

84
instances were indicated in three out of eight cases (37.50%), with this being unknown in a
single case (12.50%), and with negative instances found in the remaining four cases (50.00%).
Finally, out of the 10 individuals who had valid data on social grievances, two were found to
have positive instances with respect to this measure (20.00%). Additionally, negative instances
were indicated in six cases (60.00%), and with a code of “unknown” indicated in the remaining
two cases (20.00%).
The following measures included in these data focused upon ideology, with the specific
measures included examining anarchism, anti-government sentiment, black nationalism,
religious ideology, sovereign citizen ideology, and white nationalism. Regarding anarchist
ideology, negative instances were found in seven cases (87.50%), with a code of “unknown” in
the remaining case (12.50%). Next, the same result was found with respect to anti-government
ideology, with negative instances again found in seven cases (87.50%), and with a code of
“unknown” found in the final case (12.50%). Identical data were again found with respect to
black nationalism and sovereign citizen ideology. With respect to religious ideology, affirmative
instances were found in three out of eight individuals (37.50%), with negative instances in four
cases (50.00%), and with a code of “unknown” in the remaining case (12.50%). Finally, with
regard to white supremacy, a positive instance was found in one out of eight cases (12.50%), and
with negative instances found in the remaining seven cases (87.50%).
The final set of measures included within these data examined Internet use, media
consumption, and social media. Data were collected on the Internet searches used by these 13
individuals. Positive instances (indicating Internet searches pertaining to grievances or possible
targets) were found in eight cases (61.54%), with negative instances indicated in the remaining
five cases (38.46%). Additionally, extremist media consumption was found to be very high in

85
this sample of 13 individuals. Specifically, 10 out of these 13 persons (76.92%) were found to
have engaged in extremist media consumption, with this having been negative in two cases
(15.38%), and with Reddit specifically mentioned in one case (7.69%).
Data collected in relation to social media use consisted of both which social media
platforms were used by these individuals, if any, as well as which activities individuals were
performing using social media. Regarding the social media platform used by individuals, a total
of five separate measures were incorporated examining this one item such that a potential total of
five different platforms could be mentioned. These measures were examined measure by
measure in order to present a clear picture of which social media platforms were used by these
individuals. Regarding the first measure of social media platform use, out of the nine valid cases,
there were two mentions of Instagram (22.22%), two mentions of YouTube (22.22%), and two
mentions of “unknown” (22.22%). Additionally, single mentions were made of Facebook
(11.11%), Reddit (11.11%), and “other encrypted software” (11.11%). Fewer valid data were
present with respect to the remaining four measures of social media platform use, indicating that
based on the data available, most individuals only used a single social media platform. The
second item contained three valid cases, with one mention of Twitter (33.33%), and two
mentions of YouTube (66.67%). The third item contained only two valid cases, with both of them
referring to Facebook (100.00%), and with no valid data associated with the fourth or fifth
measures of social media platform use.
With respect to social media activities, three separate measures were included here,
similar to the five identical measures included for social media platform use. With regard to
activities, the first measure listed “consuming content” among five individuals (55.56%), with a
single case each listing communicating with extremist group members, creating propaganda,

86
distributing content, and participating in dialog. With regard to the remaining two measures,
seven valid cases were present with regard to the second measure, and with a single valid case
present with regard to the third measure. This second measure contained two mentions of
communicating with extremist group members (28.57%), with two individuals creating
propaganda (28.57%), another two individuals participating in dialog (28.57%), and one
distributing content (14.29%). The third measure only had one valid case, with this individual
participating in dialog (100.00%). Finally, with regard to violent video consumption, of the four
valid cases, two (50.00%) had affirmative instances, with the other two (50.00%) having codes
of “unknown.”
Data were also recorded on the highest level of education completed. As would be
expected based on the low average age of school violence perpetrators, these individuals
generally had a low level of education. Of the 11 cases that had valid data on this measure, a
single individual (9.09%) had a basic college degree, with two individuals (18.18%) working on
a basic college degree, while this was not yet complete. Six persons (54.55%) in total had either
a high school diploma or a GED, with one person (9.09%) having gone to vocational school, and
with one (9.09%) working on their high school diploma, with this not having yet been
completed. Finally, with regard to demographics, data were also collected on whether individuals
had a learning disability. There were only four valid cases among these 13 individuals, with all
four (100.00%) of these cases consisting of “unknown.”
Other Perpetrators: Descriptive Statistics
In addition to the descriptive statistics conducted on the 13 school violence perpetrators, a
set of descriptive statistics, identical to those completed and reported in the previous section,
were conducted for all 298 of the other perpetrators. This second set of perpetrators largely

87
included terrorists, as well as those committing acts of violence on the basis of their beliefs,
which included environmentalism as well as abortion and animal rights concerns, and those
committing violent acts on the basis of religion, race, or political beliefs. Table 5.2 summarizes
the composite JRIC data on the basis of dataset from which these data were derived. As shown,
the majority of these data were derived from the TIA-Perpetrators dataset, the GTD dataset, or
the PIRUS dataset.
Table 5.2
Summary of Other Perpetrators by Dataset
Database Number Percent
ACLED 1 0.34%
CSIS–TNT 3 1.01%
GTD 96 32.21%
JRIC 20–Year 10 3.36%
PIRUS 46 15.44%
TIA–Perpetrators 128 42.95%
TIA–Plots 5 1.68%
Other 9 3.02%
Total 298 100.00%
Table 5.3 presents a description of the “Main Factor” measures when examining all other
perpetrators. As stated previously, these measures vary slightly in their focus, and serve to
examine the primary reasons or contributing factors behind the violent acts committed by these
individuals. To summarize these findings, common factors behind the violence committed by
other perpetrators consisted of online postings (leakage), violence against property, violence to

88
others, weapons access, mental health (both diagnosed conditions and having displayed
symptoms), having perpetrated physical violence, financial and relationship factors, vagrancy,
religious and social grievances, extremist media consumption, and Internet searches. The precise
frequencies and percentages associated with all response categories for these “Main Factor”
measures are presented in Table 5.3.
Table 5.3
Summary of Main Factors for JRIC Other Perpetrators (n=298)
Measure/Category Number Percent of Valid Cases
Main Factor 1
Emotional Distress 10 4.20%
Jail/Prison 2 0.84%
Law Enforcement Contacts 4 1.68%
Online Postings (Leakage) 43 18.07%
Travel (Suspicious) 19 7.98%
Unknown 1 0.42%
Verbal Statements (Leakage) 22 9.24%
Violence Against Property 49 20.59%
Violence From Others
(Abuse) 1 0.42%
Violence to Others 47 19.75%
Weapons Access 35 14.71%
Weapons Access; Verbal
Statements 1 0.42%
Weapons Training (Military) 4 1.68%
Main Factor 2

89
Emotional Distress
5 4.24%
Loss (Residence)
5 4.24%
Loss (Death)
2 1.69%
Loss (Job)
5 4.24%
Loss (Relationship)
3 2.54%
Mental Health (Diagnosed Condition) 13 11.02%
Mental Health (Displayed Symptoms) 11 9.32%
Online Postings (Leakage)
2 1.69%
Travel (Suspicious)
8 6.78%
Travel (Suspicious); Loss
(Relationship)
1 0.85%
Verbal Statements (Leakage)
2 1.69%
Violence Against Property
9 7.63%
Violence from Others (Abuse)
6 5.08%
Violence to
Others 46 38.98%
Main Factor 3
Loss (Job)
9 11.84%
Loss (Relationship)
6 7.89%
Loss (Death)
4 5.26%
Loss (Residence)
6 7.89%
Mental Health (Diagnosed Condition) 14 18.42%
Mental Health (Displayed Symptoms) 18 23.68%
Not Applicable
2 2.63%
Unknown 17 22.37%

90
Main Factor 4
Alcohol 2 3.45%
Drugs 5 8.62%
No 3 5.17%
Not Applicable 2 3.45%
Physical (Perpetrator) 12 20.69%
Physical (Victim) 4 6.90%
Relationship; Physical
(Perpetrator) 1 1.72%
Unknown 13 22.41%
Verbal (Perpetrator) 8 13.79%
Verbal (Victim) 3 5.17%
Yes 5 8.62%
Main Factor 5
Financial 17 28.33%
No 1 1.67%
Not Applicable 1 1.67%
Relationship 13 21.67%
Unknown 16 26.67%
Vagrancy 10 16.67%
Yes 2 3.33%
Main Factor 6
Exposure–Conversion 25 10.12%
Exposure–Radical
Family/Loved One 2 0.81%

91
Exposure–Radical Friend 1 0.40%
Exposure–Religious Reawakening 20 8.10%
Exposure–Religious Reawakening; Grievance 1 0.40%
Grievance–Anti-Law
Enforcement 1 0.40%
Grievance–Religious 2 0.81%
Grievance–Economic 2 0.81%
Grievance–Personal 15 6.07%
Grievance–Political 9 3.64%
Grievance–Racial 5 2.02%
Grievance–Religious 71 28.74%
Grievance–Social 45 18.22%
Ideology–Anti-Establishment 5 2.02%
Ideology–White Supremacy 28 11.34%
No 1 0.40%
Unknown 9 3.64%
Yes 5 2.02%
Main Factor 7
Extremist Media
Consumption 75 48.70%
Internet Searches 47 30.52%
No 15 9.74%
Not Applicable 2 1.30%
Unknown 4 2.60%
Violent Video Consumption 8 5.19%

92
Yes 3 1.95%
As before, a series of descriptive statistics were first conducted on date of the incident as
well as basic demographic information pertaining to the individuals. Regarding date of the
incident, this ranged from 2000 through 2021. With respect to individual gender, 283 individuals
(91.59%) were male, with 26 (8.41%) female. As with school violence perpetrators, these data
indicate that it is an overwhelming majority of males who perpetrate violent incidents. With
respect to age, this was found to range from a minimum of 16 to a maximum of 72 years, with a
mean of 29.66 years (SD = 12.12), and a median of 26 years of age. Next, regarding marital
status, a slight majority of 150 individuals (50.34%) were single (never married), with 72
(24.16%) married, 23 (7.72%) divorced or separated, and two (.67%) widowed. Additionally,
marital status was unknown for 51 (17.11%) individuals.
Next, descriptive statistics were conducted on plot name, while data varied very widely.
The total number of cases associated with each individual plot ranged from a minimum of one to
a maximum of six, with dozens of various plots represented within these data. Due to the very
large number of individual cases with regard to plot name, these will not be individually
presented here. Similar results were found with respect to criminal status, so similarly, these
individual frequencies and percentages will not be presented here. With respect to the charges,
federal charges were present in 120 cases (48.39%), with the individual being freed or released in
52 cases (20.97%). A total of 20 individuals (8.06%) were killed during the attack, with 16
(6.45%) currently incarcerated. State charges were raised in a total of 10 cases (4.03%), with no
charges having been brought up due to the individual already being deceased in a total of eight

93
cases (3.23%). All remaining categories of response each represented a total of four cases or less,
and are not individually presented here.
Next, with respect to access to weapons, a majority of the other perpetrators, 170
individuals (58.02%) were found to have access to weapons, while this was not found to be the
case among 25 individuals (8.53%). Additionally, this was unknown in a total of 98 cases
(33.45%). Following this, with respect to military experience, it was found that among the 63
cases in which valid data were present with respect to military experience, a paramilitary militia
was indicated in a total of 24 cases (38.10%), with the U.S. Army indicated in 20 cases
(31.75%). Additionally, a terrorist training camp was specified in five cases (7.94%), with four
individuals each indicating foreign military experience, or experience in the U.S. Marines
(6.35%), with this being unknown in three cases (4.76%), and a single individual each (1.59%)
indicating the U.S. Air Force, the U.S. National Guard, and the U.S. Navy. Contact with law
enforcement was only found in a total of five cases (2.30%), with this not having been the case in
104 cases (47.93%), and being unknown in the remaining 108 cases (49.77%).
The following questions examined the mental health status of the individual as well as
their behavior, including behavior that may be considered suspicious. Regarding emotional
distress, a total of 58 individuals (19.02%) indicated emotional distress, with 150 (49.18%) not
indicating emotional distress. Additionally, this was unknown in the remaining 97 cases
(31.80%). With respect to whether verbal statements had been made, this was indicated in 180
cases (61.64%), with this failing to be indicated in 55 cases (18.84%), and unknown in the
remaining 57 cases (19.52%).
Regarding online postings, a total of 113 perpetrators (36.93%) had made online postings,
while 127 (41.50%) had not. Additionally, this was unknown among 66 individuals (21.57%).

94
With respect to having had made suspicious travel, this was indicated in 159 cases
(52.65%), was not indicated in 78 cases (25.83%), and was unknown in the remaining 65 cases
(21.52%). The following three measures examined violence, including violence to the self,
violence to others, and having received violence from others, as well as property damage.
Violence to the self was only indicated in a total of eight cases (2.65%), and was not indicated in
215 cases (71.19%). Additionally, this status was unknown in the remaining 79 cases (26.16%).
With respect to violence to others, this was found in 117 cases (38.61%), was not indicated in 88
cases (29.04%), and was unknown in the remaining 98 cases (32.34%). Having received violence
from others was found in a total of 16 cases (5.30%), was not found in 152 cases (50.33%), and
was unknown in the remaining 134 cases (44.37%). Finally, regarding the destruction of
property, this was indicated in 111 cases (36.51%), was not found in 100 cases (32.89%), and
was unknown in the remaining 93 cases (30.59%).
The following measures examined mental health status as well as loss. A total of 21
individuals (6.80%) were found to have had a mental health diagnosis, with this not having been
the case among 218 individuals (70.55%), and with their status being unknown in the remaining
70 cases (22.65%). Regarding whether mental health symptoms were exhibited, this was found
to be the case among 49 individuals (16.17%), was not found to be the case in 192 cases
(63.37%), and was unknown in 51 cases (16.83%). More specific data were associated with a
total of 11 cases, with a diagnosed condition being specified in four cases (1.32%), and mental
health symptoms having been displayed in seven cases (2.31%).
With respect to measures of loss, this included a loss relating to a death, the loss of a job,
the loss of a relationship, and the loss of a residence. A total of 16 individuals (5.30%) had
experienced a loss relating to a death, with this not indicated in 165 cases (54.64%), and the

95
remaining 121 individuals (40.07%) having a responsive “unknown” for this question. Next, 38
individuals (12.50%) had experienced the loss of a job, with this not being the case among 159
perpetrators (52.30%), and with this being unknown in the remaining 107 cases (35.20%).
Regarding loss relating to a relationship, 25 individuals (8.20%) had experienced this,
with 166 individuals (54.43%) not having experienced such a loss. Additionally, this was
unknown in the remaining 114 cases (37.38%). Regarding the loss of a residence, this was
indicated in 16 cases (5.23%), was not found among 198 individuals (64.71%), and was
unknown in the remaining 92 cases (30.07%).
The following measures examined abuse, with this consisting of alcohol abuse, drug
abuse, and physical/verbal abuse. Regarding alcohol abuse, this was found in six cases (1.99%),
was not indicated in 186 individuals (61.79%), and was unknown in the remaining 109 cases
(36.21%). With respect to drug abuse, this was more common than alcohol abuse, being
indicated in 18 cases (6.27%), not being found among 175 individuals (60.98%), and unknown in
the remaining 94 cases (32.75%). Finally, with respect to physical/verbal abuse, this was found
in 27 cases (8.85%), was not found in 161 cases (52.79%), and was unknown in the remaining
117 cases (38.36%).
The following three measures examined financial difficulties, relationship difficulties, as
well as vagrancy. Among other perpetrators, a total of 29 (9.63%) were indicated as having had
experienced financial difficulties, with negative instances indicated in 167 cases (55.48%), and
with this being unknown in the remaining 105 cases (34.88%). Regarding relationship
difficulties, positive instances were indicated in 25 cases (8.42%), with negative instances found
in 166 cases (55.89%), and instances of “unknown” indicated in 99 cases (33.33%). Additionally,
instances of “relationship” were indicated in five cases (1.68%), with instances of “financial”

96
present in two cases (0.67%). With respect to vagrancy, this was found among 35 individuals
(11.40%), was not found to be the case in 192 cases (62.54%), and was unknown in the
remaining 80 cases (26.06%).
The following set of measures examined radicalization, with the specific measures
examining whether individuals had been radicalized by a family member, by a friend, or by an
associate. A total of 12 individuals (3.91%) had been radicalized by a family member, with
negative instances in 173 cases (56.35%), and with this being unknown in the remaining 122
cases (39.74%). Next, a large proportion of other perpetrators, 95 individuals (32.76%), had been
radicalized by a friend, with negative instances indicated in 92 cases (31.72%), and with this
being unknown in the remaining 103 cases (35.52%). Finally, another substantial proportion, 108
individuals (35.18%) had been radicalized by an associate, with negative instances indicated in
81 cases (26.38%), and with this being unknown in the remaining 118 cases (38.44%).
Regarding experience fighting a war, a wide variety of data were provided here among
this set of individuals. Focusing upon the most common instances, “none” was indicated in 164
cases (56.55%), with “unknown” present in 65 cases (22.41%). All remaining categories of
response each pertained to seven individuals or fewer. With respect to grievances, these consisted
of economic, personal, racial, religious, and social grievances. With regard to having an
economic grievance, this was found to be present among 15 individuals (4.98%), was unknown
in 66 cases (21.93%), and with negative instances with respect to the remaining 220 individuals
(73.09%). With respect to a personal grievance, this was present in 60 cases (20.00%), was
unknown among 57 individuals (19.00%), and with negative instances found in the remaining
183 cases (61.00%). More common than either economic or personal grievances were political
grievances. A political grievance was indicated in a total of 109 cases (36.21%), with this being

97
unknown in 66 cases (21.93%), and not indicated in the remaining 126 cases (41.86%). Racial
grievances were not very common, with a grievance being indicated in 43 cases (14.10%),
unknown in 46 cases (15.08%), and with negative instances indicated in the remaining 216 cases
(70.82%). Religious grievances were indicated frequently, with this found among 134 individuals
(44.37%), unknown in 38 cases (12.58%), and not indicated in the remaining 130 cases
(43.05%). Finally, a social grievance was indicated in 79 cases (25.65%), with this being
unknown in 56 cases (18.18%), and with negative instances indicated in 173 cases (56.17%).
The following measures focused upon ideology, specifically having an anarchist
ideology, anti-government, as well as a black nationalist, religious, sovereign citizen, or a white
supremacist ideology. Having an anarchist ideology was found to be fairly uncommon, only
having been present among 20 individuals (6.58%), with this being unknown in 24 cases
(7.89%), and with negative instances indicated in the remaining 260 cases (85.53%). Regarding
anti-government ideology, this was indicated in 27 individuals (9.34%), was “unknown” or
“other” in 46 cases (15.92%), and with negative instances indicated in the remaining 216 cases
(74.74%). Black nationalism was also found to be uncommon, only indicated among 12
individuals (4.00%), unknown in 13 cases (4.33%), and with negative instances indicated in the
remaining 275 cases (91.67%). Religious ideology was common, and present in a majority of
other perpetrators, pertaining to 157 individuals in total (52.16%). This was unknown among 31
individuals (10.30%), and was not found in the remaining 113 cases (37.54%). Sovereign citizen
ideology was very rare, only present in 10 cases (3.27%), unknown in 13 cases (4.25%), and not
present in the remaining 283 cases (92.48%). Finally, white supremacist ideology was found
among 56 individuals (18.54%), unknown in 11 cases (3.64%), and was not present in the
remaining 235 individuals (77.81%).

98
Next, regarding Internet searches, a positive instance, indicating one or more searches
relevant to grievances or possible targets, was found in 120 individuals (39.34%), while this was
negative in 132 cases (43.28%), and was unknown in the remaining 53 cases (17.38%). The
following set of measures pertained to media and social media consumption and use. Extremist
media consumption was found in 154 cases (51.68%), was not indicated in 104 individuals
(34.90%), and was unknown in the remaining 40 cases (13.42%). Regarding social media
platforms used, data varied widely and only the most common data will be discussed here. As
stated earlier, five separate measures were present to record social media platforms used, with a
maximum of five options per individual, with three measures pertaining to social media
activities. Regarding the first measure of social media activities, data composing at least 10% of
the sample consisted of “unknown” (78 individuals, 52.00%), Facebook (19 individuals,
12.67%), and other encrypted software (15 individuals, 10.00%). In the second through fifth
measures, no response category was found to have incorporated 10% of this sample or more.
With respect to social media activities, among other perpetrators, communicating with
extremist groups was common (17 individuals in the first measure, or 8.81%), along with
consuming content (57 individuals, 29.53%), with smaller percentages of individuals creating
propaganda (nine individuals, or 4.66%), distributing content (13 individuals, or 6.74%), and
participating in dialog (11 individuals, or 5.70%). The latter two measures primarily contained
positive instances pertaining to communication with extremist groups, distributing content, and
participating in dialog.
The final measures examined violent video consumption, level of education, and the
presence of learning disabilities. Violent video consumption was fairly common, with 44
individuals (19.04%) having positive instances for this measure, with 118 individuals having

99
negative instances (51.08%), and with this being unknown in the remaining 69 cases (29.87%).
Regarding the highest level of education completed, the most common instance here consisted of
having a high school diploma or GED (93 individuals, or 40.79%), followed by “university” (52
individuals, or 22.81%), “incomplete” (27 individuals, or 11.84%), “university, post-graduate”
(20 individuals, or 8.77%), and unknown (19 individuals, or 8.33%). All remaining categories of
response each composed less than 5%, and will not be discussed individually here. Finally,
learning disabilities were fairly rare, with only 17 individuals (6.20%) having some learning
disability, with this not indicated in 139 cases (50.73%), and with this being unknown in the
remaining 118 cases (43.07%).
Bivariate Analyses
Next, an independent-samples t-test was conducted in order to determine whether any
significant differences were present in the meantime at which incidents were conducted by
school shooters versus other perpetrators. A Levene’s test for the equality of variances was
conducted in order to determine whether this assumption of the independent-samples t-test was
violated. A significant result was found, indicating significant differences in the variance of year
on the basis of group membership, and that this assumption had been violated, Levene’s F(1,
299) = 4.273, p < .05. For this reason, the independent-samples t-test was conducted specifying
the use of separate variances for each perpetrator group. This t-test failed to achieve statistical
significance, indicating no significant difference in mean time of the incident based on
perpetrator group, t(11.537) = 1.152, p = .273. With respect to these data, school violence
incidents were found to range from 2000 to 2018 (M = 2011.083), while other incidents were
found to range from 2000 to 2021 (M = 2008.291).

100
A series of bivariate analyses were conducted in order to compare the 13 school violence
perpetrators with all other perpetrators. These analyses almost entirely consisted of two-sided
Fisher’s exact tests, as most of these analyses served to determine whether there was any
significant association between group membership, which was categorized as either school
shooters or other perpetrators, and the remaining data included in this dataset, which were
primarily categorical. These analyses examined nearly all measures included in the dataset,
including perpetrator demographics.
Analyses were conducted in order to determine whether there were any significant
differences between perpetrators of school violence and other perpetrators with regard to the
“Main Factor” measures, which examined key elements behind or associated with the
individual’s act of violence. Of these analyses, significant differences were not indicated with
respect to the first measure, Fisher’s exact p = .296, or the second, Fisher’s exact p = .330, third,
Fisher’s exact p = .281, fourth, Fisher’s exact p = .715, sixth, Fisher’s exact p = .616, or seventh
measures, Fisher’s exact p = .466. No test was conducted in relation to the fifth “Main Factor”
measure as no valid data were present for this measure among perpetrators of school violence.
The results indicate no significant difference between these groups with regard to these
measures.
The following analyses tested whether there were any significant associations between
individual demographics and group membership, with group membership again categorized as
either being a perpetrator of school violence or a perpetrator of some other type of violence.
Regarding gender, no significant association was found between individual gender and group
membership, Fisher’s exact test p = 1.000. A cross-tabulation of these measures indicated that in
both cases, approximately 10% of the sample was female. Among school violence perpetrators,

101
only a single individual was female out of the total of 13 (7.69%), while the 12 remaining
individuals (92.31%) were male. Among other perpetrators, of the 306 cases, 26 were female
(8.50%), with 280 male (91.50%).
Following this, an independent-samples t-test was conducted in order to examine whether
mean age significantly differed on the basis of group membership. Initially, Levene’s test for the
equality of variances was conducted in order to determine whether the assumption of the equality
of variances was violated in these data. A significant result was indicated, indicating the violation
of this assumption, Levene’s F(1, 297) = 5.197, p < .05. As this assumption was violated, the
independent samples t-test was conducted specifying unequal variances. This analysis achieved
significance, finding that other perpetrators (M = 29.627, SD = 12.083) were significantly older
than perpetrators of school violence (M = 18.286, SD = 2.628), t(13.552) = 9.302, p < .001.
These results are in line with what was expected based on the typical demographics of
perpetrators of school violence.
Following this, the association between group membership and marital status was
examined using Fisher’s exact test. This analysis failed to achieve statistical significance,
indicating no significant difference between these two groups with regard to marital status,
Fisher’s exact p = .187.
The following set of analyses examined measures thought likely to be associated with
perpetrating violence, with this specifically consisting of access to weapons, military experience,
and contact with law enforcement. Regarding access to weapons, no significant difference was
found between perpetrators of school violence and perpetrators of other violence with respect to
weapons access, Fisher’s exact p = .792. Regarding military experience, no significant difference
was found between these two groups of individuals and this measure, Fisher’s exact p = .312.

102
Finally, regarding contact with law enforcement, this analysis did, in fact, achieve
statistical significance, indicating a significant difference between perpetrators of school violence
and others regarding law enforcement contact, Fisher’s exact p < .001. Specifically, as compared
with other perpetrators, perpetrators of school violence were more likely to have had contact
with law enforcement due to a non-violent felony or violent crime. Additionally, perpetrators of
school violence were more likely to have had no contact with law enforcement as compared with
others, while law enforcement contact was more likely to be unknown among other perpetrators,
and other perpetrators were also more likely to have an instance of “yes.” These results are
presented in further detail in Table 5.4.
Table 5.4
Cross-tabulation of Law Enforcement Contacts with Perpetrator Group
Law
Enforcement
Contacts
Group Total
Other School Violence
Yes 5 0 5
100.00% 0.00% 100.00%
2.33% 0.00% 2.19%
No 103 8 111
92.79% 7.21% 100.00%
47.91% 61.54% 48.68%
Non-Violent
Felony 0 2 2
0.00% 100.00% 100.00%
0.00% 15.38% 0.88%
Violent Crime 0 1 1

103
0.00% 100.00% 100.00%
0.00% 7.69% 0.44%
Unknown 107 2 109
98.17% 1.83% 100.00%
49.77% 15.38% 47.81%
Total 215 13 228
94.30% 5.70% 100.00%
100.00% 100.00% 100.00%
The following set of measures examined emotional distress, as well as individual
behavior, with this consisting of verbal statements, online behavior, and suspicious travel.
Regarding emotional distress, a significant difference was found in this measure on the basis of
group membership. Specifically, as compared with other perpetrators, perpetrators of school
violence are more likely to indicate emotional distress, Fisher’s exact test p < .05. With regard to
verbal statements, no significant association was found with group membership, Fisher’s exact
test p = .464. Similarly, the Fisher’s exact test conducted with online postings also failed to
achieve statistical significance, Fisher’s exact test p = .176.

104
Table 5.5
Cross-tabulation of Emotional Distress with Perpetrator Group
Emotional
Distress Group Total
Other School Violence
Yes 58 4 62
93.55% 6.45% 100.00%
19.21% 66.67% 20.13%
No 150 1 151
99.34% 0.66% 100.00%
49.67% 16.67% 49.03%
Unknown 94 1 95
98.95% 1.05% 100.00%
31.13% 16.67% 30.84%
Total 302 6 308
98.05% 1.95% 100.00%
100.00% 100.00% 100.00%
With regard to suspicious travel, a significant difference in this behavior was indicated
between perpetrators of school violence and others, Fisher’s exact test p < .01. Specifically,
perpetrators of school violence were less likely to have engaged in suspicious travel as compared
with other perpetrators, which would be expected on the basis of the atypical travel patterns
exhibited by terrorists.

105
Table 5.6
Cross-tabulation of Suspicious Travel with Perpetrator Group
Suspicious
Travel Group Total
Other School Violence
Yes 158 1 159
99.37% 0.63% 100.00%
52.84% 12.50% 51.79%
No 76 6 82
92.68% 7.32% 100.00%
25.42% 75.00% 26.71%
Unknown 65 1 66
98.48% 1.52% 100.00%
21.74% 12.50% 21.50%
Total 299 8 307
97.39% 2.61% 100.00%
100.00% 100.00% 100.00%
The following tests examined group differences with respect to violence, with these
measures again consisting of violence to the self, violence to others, violence from others, and
property destruction. Regarding violence to the self, no significant associations were indicated
between this measure and group membership, Fisher’s exact test p = .745. However, significant
group differences were found with respect to violence to others, Fisher’s exact test p < .05.
Specifically, it was indicated that perpetrators of school violence were less likely to have
exhibited violence toward others as compared with other perpetrators. In addition, these data

106
were also more likely to be unknown with respect to other perpetrators as compared with
perpetrators of school violence. Additionally, significance was not indicated in the analysis
conducted with violence from others, Fisher’s exact test p = .464, or with regard to property
destruction, Fisher’s exact test p = .604.
Table 5.7
Cross-tabulation of Violence to Others with Perpetrator Group
Violence to
Others Group Total
Other School Violence
Yes 117 1 118
99.15% 0.85% 100.00%
39.00% 12.50% 38.31%
No 88 6 94
93.62% 6.38% 100.00%
29.33% 75.00% 30.52%
Unknown 95 1 96
98.96% 1.04% 100.00%
31.67% 12.50% 31.17%
Total 300 8 308
97.40% 2.60% 100.00%
100.00% 100.00% 100.00%
The following set of measures examined mental health as well as having experienced a
personal loss. No significant associations were indicated between group membership and having
had a mental health diagnosis, Fisher’s exact test p = .080. Additionally, statistical significance

107
again was not achieved when examining the association between group membership and having
exhibited mental health symptoms, Fisher’s exact test p = .191. The following measures
examining loss consisted of measures pertaining to the loss experienced through another’s death,
the loss of a job, the loss of a relationship, as well as the loss of a residence. A significant
association between group membership and these four loss measures were not indicated with
respect to loss associated with a death, Fisher’s exact test p = .821, the loss of a job, Fisher’s
exact test p = .777, the loss of a relationship, Fisher’s exact test p = .854, or the loss of a
residence, Fisher’s exact test p = 1.000.
The following measures examined abuse, with these measures consisting of alcohol
abuse, drug abuse, as well as physical or verbal abuse. A significant relationship between group
membership and alcohol abuse was not found, Fisher’s exact test p = .758, while a significant
association was found between drug abuse and perpetrator group, Fisher’s exact test p < .05.
Specifically, school violence perpetrators were more likely to have an instance of “unknown”
with regard to drug abuse, and are also less likely to have an instance of “no.” Next, with respect
to the incidence of physical or verbal abuse, a significant difference between these two groups of
perpetrators was not found, Fisher’s exact test p = .055.
Table 5.8
Cross-tabulation of Drug Abuse with Perpetrator Group
Drug Abuse Group Total
Other School Violence
Yes 18 0 18
100.00% 0.00% 100.00%
6.34% 0.00% 6.16%

108
No 175 2 177
98.87% 1.13% 100.00%
61.62% 25.00% 60.62%
Unknown 91 6 97
93.81% 6.19% 100.00%
32.04% 75.00% 33.22%
Total 284 8 292
97.26% 2.74% 100.00%
100.00% 100.00% 100.00%
The following measures examined various difficulties experienced by individuals, with
these measures specifically focusing upon financial difficulties, relationship difficulties, and
vagrancy. No significant group differences in financial difficulties were indicated, Fisher’s exact
test p = .654, while significant differences with respect to relationship difficulties were also not
indicated, Fisher’s exact test p = .295. Regarding vagrancy, this associated analysis also failed to
achieve statistical significance, Fisher’s exact test p = .389.
The following measures focus upon radicalization, with the data collected pertaining to
radicalization by family, a friend, or an associate. Significant group differences were not
indicated with respect to having been radicalized by family, Fisher’s exact test p = .173, or by a
friend, Fisher’s exact test p = .490. However, significant differences between school violence
perpetrators and other perpetrators were indicated with respect to having been radicalized by an
associate, Fisher’s exact test p < .001. Here, perpetrators of school violence were less likely than
others to have been radicalized by an associate, instead being more likely to have a negative
instance. Additionally, other perpetrators were more likely to have an instance of “unknown.”

109
Table 5.9
Cross-tabulation of Radicalized by Associate with Perpetrator Group
Radicalized by
Associate Group Total
Other School Violence
Yes 107 0 107
100.00% 0.00% 100.00%
35.20% 0.00% 33.75%
No 80 10 90
88.89% 11.11% 100.00%
26.32% 76.92% 28.39%
Unknown 117 3 120
97.50% 2.50% 100.00%
38.49% 23.08% 37.85
Total 304 13 317
95.90% 4.10% 100.00%
100.00% 100.00% 100.00%
Experience of war was coded as the particular war or conflict that individuals participated
in, if any. This analysis failed to find any significant difference between groups with respect to
their military experience, Fisher’s exact test p = .591. With respect to grievances, a significant
association was only found with the measure examining personal grievances, Fisher’s exact test
p < .05. Specifically, perpetrators of school violence were more likely to have had a personal
grievance as compared with other perpetrators. Additionally, other perpetrators are more likely to
have had no personal grievance or an instance of “unknown” with respect to this question.

110
Table 5.10
Cross-tabulation of Personal Grievances with Perpetrator Group
Personal
Grievances Group Total
Other School Violence
Yes 59 6 65
90.77% 9.23% 100.00%
19.87% 60.00% 21.17%
No 182 3 185
98.38% 1.62% 100.00%
61.28% 30.00% 60.26%
Unknown 56 1 57
98.25% 1.75% 100.00%
18.86% 10.00% 18.57%
Total 297 10 307
96.74% 3.26% 100.00%
100.00% 100.00% 100.00%
No significant association was found between perpetrator category and having had an
economic grievance, Fisher’s exact test p = .792, a political grievance, Fisher’s exact test p =
.234, a racial grievance, Fisher’s exact test p = .846, a religious grievance, Fisher’s exact test p =
.888, or a social grievance, Fisher’s exact test p = 1.000. The following measures examined
ideology. No significant relationship was indicated between perpetrator category and anarchist
ideology, Fisher’s exact test p = .721, anti-government ideology, Fisher’s exact test p = 1.000,
black nationalism, Fisher’s exact test p = .514, religious ideology, Fisher’s exact test p = .568,

111
sovereign citizen ideology, Fisher’s exact test p = .478, or white supremacist ideology, Fisher’s
exact test p = 1.000.
The following measures examined Internet use, with these measures consisting of
Internet searches, extremist media consumption, social media platforms used, as well as social
media activities. No association was found between perpetrator category and Internet searches,
Fisher’s exact test p = .151. However, a significant association was found with extremist media
consumption, Fisher’s exact test p < .05. These data indicated that extremist media consumption
was significantly more likely among perpetrators of school violence as compared with others.
Additionally, in the case of others, an instance of “unknown” was more likely as compared with
perpetrators of school violence.
Table 5.11
Cross-tabulation of Extremist Media Consumption with Perpetrator Group
Extremist Media Group Total
Other School Violence
Yes 142 10 152
93.42% 6.58% 100.00%
48.14% 76.92% 49.35%
No 104 2 106
98.11% 1.89% 100.00%
35.25% 15.38% 34.42%
Extremist Media
Consumption 5 0 5
100.00% 0.00% 100.00%
1.69% 0.00% 1.62%

112
Facebook 2 0 2
100.00% 0.00% 100.00%
0.68% 0.00% 0.65%
Internet Searches 2 0 2
100.00% 0.00% 100.00%
0.68% 0.00% 0.65%
Reddit 0 1 1
0.00% 100.00% 100.00%
0.00% 7.69% 0.32%
Unknown 40 0 40
100.00% 0.00% 100.00%
13.56% 0.00% 12.99%
Total 295 13 308
95.78% 4.22% 100.00%
100.00% 100.00% 100.00%
Significance was also indicated with regard to the relationship between perpetrator group
and the social media platform used, if any, Fisher’s exact test p < .01. The potential of up to five
different options were allowed with regard to this item, with data collected consisting of specific
social media platforms used. Primarily, the data indicate that other perpetrators were more likely
to have an instance of “unknown” as compared with perpetrators of school violence. No other
clear pattern was indicated in these data.

113
Table 5.12
Cross-tabulation of Social Media Platform Used with Perpetrator Group
Social Media Group Total
Other School Violence
4chan/8chan 1 0 1
100.00% 0.00% 100.00%
0.68% 0.00% 0.64%
Facebook 19 1 20
95.00% 5.00% 100.00%
12.93% 11.11% 12.82%
Facebook;
Personal blog 2 0 2
100.00% 0.00% 100.00%
1.36% 0.00% 1.28%
Instagram 1 2 3
33.33% 66.67% 100.00%
0.68% 22.22% 1.92%
MySpace 1 0 1
100.00% 0.00% 100.00%
0.68% 0.00% 0.64%
Other Encrypted
Software 15 1 16
93.75% 6.25% 100.00%
10.20% 11.11% 10.26%
Other NonEncrypted
Software
4 0 4

114
100.00% 0.00% 100.00%
2.72% 0.00% 2.56%
Personal Blog 12 0 12
100.00% 0.00% 100.00%
8.16% 0.00% 7.69%
Reddit 0 1 1
0.00% 100.00% 100.00%
0.00% 11.11% 0.64%
Snapchat 1 0 1
100.00% 0.00% 100.00%
0.68% 0.00% 0.64%
Twitter 5 0 5
100.00% 0.00% 100.00%
3.40% 0.00% 3.21%
Unknown 76 2 78
97.44% 2.56% 100.00%
51.70% 22.22% 50.00%
Yahoo Chat Room 1 0 1
100.00% 0.00% 100.00%
0.68% 0.00% 0.64%
YouTube 4 2 6
66.67% 33.33% 100.00%
2.72% 22.22% 3.85%

115
Yes 5 0 5
100.00% 0.00% 100.00%
3.40% 0.00% 3.21%
Total 147 9 156
94.23% 5.77% 100.00%
100.00% 100.00% 100.00%
Of the remaining measures, significance was not indicated with regard to the second
measure of social media platform use, Fisher’s exact test p = .193, the third measure, Fisher’s
exact test p = .092, or the fifth measure, Fisher’s exact test p = .333. Only other perpetrators were
found to have valid data with regard to the fourth measure of the social media platform used, and
so a statistical test could not be conducted in this case. Regarding the analyses conducted with
social media activities, significance was not found with regard to the first measure, Fisher’s exact
test p = .060, or the third measure, Fisher’s exact test p = .417, with significance indicated with
respect to the second measure, Fisher’s exact test p < .05. In this final case, no clear association
was indicated between perpetrator group and social media activities, and so this table will not be
reproduced here.
The final three bivariate analyses examined violent video consumption, level of
education, and learning disabilities. No significant difference between groups was found with
regard to violent video consumption, Fisher’s exact test p = .066, or learning disabilities, Fisher’s
exact test p = .102, while significance was indicated with respect to level of education, Fisher’s
exact test p < .01. Here, level of education was more likely to be unknown for other perpetrators,
with school violence perpetrators more likely to have a high school diploma or GED.

116
Table 5.13
Cross-tabulation of Level of Education with Perpetrator Group
Level of Education Group Total
Other School Violence
Basic College Degree 2 1 3
66.67% 33.33% 100.00%
0.89% 9.09% 1.27%
Basic College Degree
(Incomplete) 2 2 4
50.00% 50.00% 100.00%
0.89% 18.18% 1.69%
Community
College/Trade 7 0 7
100.00% 0.00% 100.00%
3.11% 0.00% 2.97%
Doctoral/Professional
Degree 1 0 1
100.00% 0.00% 100.00%
0.44% 0.00% 0.42%
High School
(Incomplete) 4 1 5
80.00% 20.00% 100.00%
1.78% 9.09% 2.12%
High School/GED 91 6 97
93.81% 6.19% 100.00%
40.44% 54.55% 41.10%
Incomplete 27 0 27

117
100.00% 0.00% 100.00%
12.00% 0.00% 11.44%
Master’s Degree 1 0 1
100.00% 0.00% 100.00%
0.44% 0.00% 0.42%
University 52 0 52
100.00% 0.00% 100.00%
23.11% 0.00% 22.03%
University, PostGraduate 20 0 20
100.00% 0.00% 100.00%
8.89% 0.00% 8.47%
Unknown 18 0 18
100.00% 0.00% 100.00%
8.00% 0.00% 7.63%
Vocational School 0 1 1
0.00% 100.00% 100.00%
0.00% 9.09% 0.42%
Total 225 11 236
95.34% 4.66% 100.00%
100.00% 100.00% 100.00%

118
Analysis of Trends over Time
Linear regression was used in order to determine whether there were any significant
trends over time with regard to the number of incidents per year. In both analyses, the measure
for year included in the analyses was calculated as 2000 subtracted from the actual year; this was
done in order to include in the analyses, a measure for year which started from zero, ranging
from zero to 21, as opposed to ranging from 2000 to 2021.
Two linear regressions were conducted in total, with these analyses having been
conducted separately on the basis of perpetrator group. With regard to perpetrators of school
violence, a significant effect was not indicated with respect to year, suggesting no significant
trends over time with respect to these data. This result found that the coefficient for year was not
significantly different from zero. However, it should be noted that this analysis suffered from a
very small number of observations, producing a resulting low statistical power, which would
increase the difficulty of finding a significant result in this analysis. Regarding the constant, this
was also not significantly different from zero. The coefficient of .462 indicated that the mean
number of school-related incidents involving violence per year was .462, with the associated
95% confidence interval ranging from essentially zero incidents to 1.540 incidents per year. This
latter result indicates that based on these data and this model, 95% of years will be associated
with between none and approximately two incidents, with the remaining 5% of years being
associated with three or more incidents.

119
Table 5.14
Regression Analysis of Incidents on Adjusted Year for School Violence Perpetrators
Variable Coefficient
(SE) t 95%
Confidence Interval
Lower Upper Year Constant
.008 (.042) .19 -.080 .096
.462 (.517) .90 -.615 1.540
Note. *p<.05, **p<.01, ***p<.001; N = 22; F(1, 20) = .04, p = .8530; R2 = .0018, Adjusted R2 = -
.0482.
Figure 5.1 plots this relationship. As shown, a low to no association was indicated on the basis of
this figure.
Figure 5.1
The Relationship between the Number of Incidents of School Violence and Year
The same regression analysis was then conducted focusing on other perpetrators, with
these results presented in Table 5.15. Significance was, in fact, found for year in this analysis, a
significant trend over time with regard to the number of incidents unrelated to school violence.

120
Specifically, this coefficient was found to be significant and negative, indicating that the
incidence of other violent incidents, based on these data, are trending downward over time, while
this was not found to be the case with regard to incidents of school violence. In addition, the
constant was also found to achieve statistical significance, indicating that the mean number of
other incidents was significantly different from zero. A mean of 20.708 other incidents per year
was found, with the associated 95% confidence interval ranging from 12.264 to 29.151 incidents
per year. The results pertaining to this confidence interval indicate that based on these data and
this regression model, 95% of years will be associated with between approximately 12 and 29
other incidents of violence, with the remaining 5% of years either having 11 or fewer other
incidents of violence, or 30 or more incidents.
Table 5.15
Regression Analysis of Incidents on Adjusted Year for Other Perpetrators
Variable
Coefficient (SE) t
95% Confidence
Interval
Lower Upper Year Constant
-.721 (.330) -2.19* -1.409 -.033
20.708 (4.048) 5.12*** 12.264 29.151
Note. *p<.05, **p<.01, ***p<.001; N = 22; F(1, 20) = 4.78, p = .0409; R2 = .1927, Adjusted R2 =
.1524.
Figure 5.2 plots the relationship between this predictor and this outcome. A negative
association between number of incidents of other forms of violence and year was also indicated
here.

121
Figure 5.2
The Relationship between the Number of Incidents of Other Violence and Year
Summary
In this chapter, the results of the analyses of the JRIC data were presented and discussed.
These analyses consisted of descriptive statistics conducted separately for school violence
perpetrators and other perpetrators, and with bivariate analyses conducted examining whether
significant differences were present in this study’s variables when comparing these two groups of
perpetrators. Linear regressions were conducted in order to examine the presence of significant
trends over time. The bivariate analyses found significant differences between perpetrators of
school violence and other perpetrators in a number of cases. Specifically, these results found
other perpetrators to be significantly more likely to:
• Be older

122
• Engage is suspicious travel
• Exhibit violence toward others
• Have been radicalized by an associate
Additionally, perpetrators of school violence were significantly more likely to:
• Have had contact with law enforcement for a non-violent felony or violent crime
• Have had no contact with law enforcement
• Experience emotional distress
• Have a drug abuse status of unknown or a negative instance
• Have had a personal grievance
• Exhibit extremist media consumption
• Have a high school diploma or GED
No trend was found with regard to the number of incidents of school violence on the basis of
these data, while a significant negative trend was found with regard to other incidents of
violence.
The following chapter will present and discuss the results of the analyses conducted on
ACEs and their correlates within the JRIC dataset.

123
Chapter 6: Results: ACEs in JRIC
Introduction
In this chapter, the results of the analyses conducted examining the relationship between
ACEs and risk factors, alongside External Factors in the JRIC dataset, are reported and
discussed. All analyses consisted of Fisher’s exact tests, with one-sided tests conducted in all
cases where possible, as a direct association was posited between these two sets of measures.
ACEs consisted of the following: emotional distress, having a mental health diagnosis,
exhibiting mental health symptoms, a loss due to a death, job, relationship, or of a residence,
alcohol abuse, drug abuse, physical/verbal abuse, financial difficulties, and relationship
difficulties.
Risk factors consisted of having been radicalized by a family, friend, or associate, having
experienced war, having grievances that were economic, personal, political, racial, religious, or
social, having an anarchist, anti-government, black nationalist, religious, sovereign citizen, or
white supremacist ideology, violent video consumption, weapons access, and military
experience.
Finally, the byproduct of ACEs, termed external factors, that were examined in these
analyses consisted of contact with law enforcement, violence to the self, others, or from others,
and vagrancy.
Law Enforcement Contacts
The first set of analyses examined associations with law enforcement contacts (Table
6.1). In this set of analyses, significant associations were not found with the loss of a
relationship, Fisher’s exact p = .962, the loss of a residence, Fisher’s exact p = .979, emotional
distress, Fisher’s exact p = .719, financial difficulties, Fisher’s exact p = .149, having been

124
radicalized by a family member, Fisher’s exact p = .765, or an associate, Fisher’s exact p = .375,
having a personal grievance, Fisher’s exact p = .399, political grievance, Fisher’s exact p = .716,
racial grievance, Fisher’s exact p = .940, religious grievance, Fisher’s exact p = .374, social
grievance, Fisher’s exact p = .342, black nationalist ideology, Fisher’s exact p = .823, religious
ideology, Fisher’s exact p = .575, or military experience, Fisher’s exact p = .170.
There were also several cases where Fisher’s exact test could not be conducted; in the
case of having had a mental health diagnosis, all valid cases of having had law enforcement
contacts were coded “no,” precluding the conducting of this test.
This same pattern was also found in the cases of drug abuse, physical or verbal abuse,
relationship difficulties, having been radicalized by a friend, having had the experience of war,
anti-government ideology, and sovereign citizen ideology, where only cases of “no” were present
for sovereign citizen ideology, violent video consumption, and weapons access.
Table 6.1 presents all of the significant cross-tabulations. Significant associations were
found between law enforcement contacts and mental health symptoms, Fisher’s exact p < .001, a
loss due to a death, Fisher’s exact p = .005, the loss of a job, Fisher’s exact p < .001, alcohol
abuse, Fisher’s exact p = .022, an economic grievance, Fisher’s exact p = .042, having an
anarchist ideology, Fisher’s exact p = .048, and a white supremacist ideology, Fisher’s exact p =
.001. A significantly higher likelihood of contact with law enforcement was found among those
exhibiting mental health symptoms (30.77% vs. 0.00%), those who experienced a loss defined as
a death (33.33% vs. 0.00%), those who experienced the loss of a job (50.00% vs. 0.00%), those
who abused alcohol (100.00% vs. 1.12%), those with an economic grievance (25.00% vs.
0.00%), those with an anarchist ideology (20.00% vs. 0.00%), and those with a white
supremacist ideology (27.27% vs. 0.00%).

125
Table 6.1
Cross-tabulations with Law Enforcement Contacts: Significant Fisher’s Exact Tests
Measure Law Enforcement Contacts Total
No Yes
Mental Health
Symptoms
No 90 0 90
100.00% 0.00% 100.00%
90.91% 0.00% 87.38%
Yes 9 4 13
69.23% 30.77% 100.00%
9.09% 100.00% 12.62%
Total 99 4 103
96.12% 3.88% 100.00%
100.00% 100.00% 100.00%
Loss: Death
No 74 0 74
100.00% 0.00% 100.00%
94.87% 0.00% 92.50%
Yes 4 2 6
66.67% 33.33% 100.00%
5.13% 100.00% 7.50%
Total 78 2 80
97.50% 2.50% 100.00%

126
100.00% 100.00% 100.00%
Loss: Job
No 74 0 74
100.00% 0.00% 100.00%
93.67% 0.00% 88.10%
Yes 5 5 10
50.00% 50.00% 100.00%
6.33% 100.00% 11.90%
Total 79 5 84
94.05% 5.95% 100.00%
100.00% 100.00% 100.00%
Alcohol Abuse
No 88 1 89
98.88% 1.12% 100.00%
100.00% 50.00% 98.89%
Yes 0 1 1
0.00% 100.00% 100.00%
0.00% 50.00% 1.11%
Total 88 2 90
97.78% 2.22% 100.00%
100.00% 100.00% 100.00%
Economic
Grievance
No 92 0 92

127
100.00% 0.00% 100.00%
96.84% 0.00% 95.83%
Yes 3 1 4
75.00% 25.00% 100.00%
3.16% 100.00% 4.17%
Total 95 1 96
98.96% 1.04% 100.00%
100.00% 100.00% 100.00%
Anarchist
Ideology
No 99 0 99
100.00% 0.00% 100.00%
96.12% 0.00% 95.19%
Yes 4 1 5
80.00% 20.00% 100.00%
3.88% 100.00% 4.81%
Total 103 1 104
99.04% 0.96% 100.00%
100.00% 100.00% 100.00%
White
Supremacist
Ideology
No 95 0 95
100.00% 0.00% 100.00%
92.23% 0.00% 89.62%

128
Yes 8 3 11
72.73% 27.27% 100.00%
7.77% 100.00% 10.38%
Total 103 3 106
97.17% 2.83% 100.00%
100.00% 100.00% 100.00%
Violence to Self
The following set of analyses examined associations with violence to the self, or selfharm. In all cases, Fisher’s exact test could be conducted. Significant associations were not found
with emotional distress, Fisher’s exact p = .138, a mental health diagnosis, Fisher’s exact p =
.112, the loss of a relationship, Fisher’s exact p = .286, or residence, Fisher’s exact p = .811,
alcohol abuse, Fisher’s exact p = .094, drug abuse, Fisher’s exact p = .816, having been
radicalized by family, Fisher’s exact p = .803, friends, Fisher’s exact p = .264, or an associate,
Fisher’s exact p = .295, having had the experience of war, Fisher’s exact p = .396, personal
grievances, Fisher’s exact p = .677, political grievances, Fisher’s exact p = .333, racial
grievances, Fisher’s exact p = .356, religious grievances, Fisher’s exact p = .427, social
grievances, Fisher’s exact p = .486, being anti-government, Fisher’s exact p = .243, a black
nationalist, Fisher’s exact p = .660, having a religious ideology, Fisher’s exact p = .533, a
sovereign citizen ideology, Fisher’s exact p = .063, violent video consumption, Fisher’s exact p =
.249, weapons access, Fisher’s exact p = .632, or military experience, Fisher’s exact p = .161.
Several significant associations with violence to self were found. These consisted of the
associations with mental health symptoms, Fisher’s exact p < .001, a loss due to a death, Fisher’s

129
exact p = .024, the loss of a job, Fisher’s exact p = .005, physical or verbal abuse, Fisher’s exact
p = .024, financial difficulties, Fisher’s exact p = .002, relationship difficulties, Fisher’s exact p <
.001, an economic grievance, Fisher’s exact p = .001, having an anarchist ideology, Fisher’s
exact p = .002, and having a white supremacist ideology, Fisher’s exact p < .001.
The results of these associated cross-tabulations are presented in Table 6.2. These results
indicated that a significantly higher likelihood of having self-harmed was associated with mental
health symptoms (20.00% vs. 0.00%), a loss due to a death (16.67% vs. 1.21%), the loss of a job
(14.81% vs. 1.29%), the experience of physical or verbal abuse (14.29% vs. 1.97%), financial
difficulties (19.05% vs. 1.23%), relationship difficulties (21.05% vs. 0.61%), having an
economic grievance (30.00% vs. 1.05%), an anarchist ideology (20.00% vs. 0.99%), and finally,
a white supremacist ideology (24.00% vs. 1.03%).
Table 6.2
Cross-tabulations with Violence to Self: Significant Fisher’s Exact Tests
Measure Violence to Self Total
No Yes
Mental Health
Symptoms
No 179 0 179
100.00% 0.00% 100.00%
86.47% 0.00% 83.64%
Yes 28 7 35
80.00% 20.00% 100.00%
13.53% 100.00% 16.36%
Total 207 7 214

130
96.73% 3.27% 100.00%
100.00% 100.00% 100.00%
Loss: Death
No 163 2 165
98.79% 1.21% 100.00%
94.22% 50.00% 93.22%
Yes 10 2 12
83.33% 16.67% 100.00%
5.78% 50.00% 6.78%
Total 173 4 177
97.74% 2.26% 100.00%
100.00% 100.00% 100.00%
Loss: Job
No 153 2 155
98.71% 1.29% 100.00%
86.93% 33.33% 85.16 %
Yes 23 4 27
85.19% 14.81% 100.00%
13.07% 66.67% 14.84%
Total 176 6 182
96.70% 3.30% 100.00%
100.00% 100.00% 100.00%
Physical/Verbal
Abuse

131
No 149 3 152
98.03% 1.97% 100.00%
89.22% 50.00% 87.86%
Yes 18 3 21
85.71% 14.29% 100.00%
10.78% 50.00% 12.14%
Total 167 6 173
96.53% 3.47% 100.00%
100.00% 100.00% 100.00%
Financial
Difficulties
No 161 2 163
98.77% 1.23% 100.00%
90.45% 33.33% 88.59%
Yes 17 4 21
80.95% 19.05% 100.00%
9.55% 66.67% 11.41%
Total 178 6 184
96.74% 3.26% 100.00%
100.00% 100.00% 100.00%
Relationship
Difficulties
No 162 1 163
99.39% 0.61% 100.00%
91.53% 20.00% 89.56%

132
Yes 15 4 19
78.95% 21.05% 100.00%
8.47% 80.00% 10.44%
Total 177 5 182
97.25% 2.75% 100.00%
100.00% 100.00% 100.00%
Economic
Grievance
No 189 2 191
96.43% 40.00% 95.02
Yes 7 3 10
70.00% 30.00% 100.00%
3.57% 60.00% 4.98
Total 196 5 201
97.51% 2.49% 100.00%
100.00% 100.00% 100.00%
Anarchist
Ideology
No 201 2 203
99.01% 0.99% 100.00%
94.37% 40.00% 93.12%
Yes 12 3 15
80.00% 20.00% 100.00%
5.63% 60.00% 6.88%
Total 213 5 218

133
97.71% 2.29% 100.00%
100.00% 100.00% 100.00%
White
Supremacist
Ideology
No 193 2 195
98.97% 1.03% 100.00%
91.04% 25.00% 88.64%
Yes 19 6 25
76.00% 24.00% 100.00%
8.96% 75.00% 11.36%
Total 212 8 220
96.36% 3.64% 100.00%
100.00% 100.00% 100.00%
Violence to Others
Violence to others was then examined. In all cases, Fisher’s exact test could be computed.
No significant association was found between violence to others and loss from a death, Fisher’s
exact p = .327, the loss of a job, Fisher’s exact p = .419, the loss of a relationship, Fisher’s exact
p = .109, the loss of a residence, Fisher’s exact p = .108, alcohol abuse, Fisher’s exact p = .302,
drug abuse, Fisher’s exact p = .141, financial difficulties, Fisher’s exact p = .332, having been
radicalized by family, Fisher’s exact p = .203, an economic grievance, Fisher’s exact p = .502, a
personal grievance, Fisher’s exact p = .453, a religious grievance, Fisher’s exact p = .261, having
an anti-government ideology, Fisher’s exact p = .512, a black nationalist ideology, Fisher’s exact

134
p = .317, a sovereign citizen ideology, Fisher’s exact p = .183, or violent video consumption,
Fisher’s exact p = .235.
A substantial set of Fisher’s exact tests were found to achieve statistical significance.
Specifically, these consisted of the relationship between violence to others and having exhibited
mental health symptoms, Fisher’s exact p < .001, emotional distress, Fisher’s exact p = .008,
having had a mental health diagnosis, Fisher’s exact p = .029, physical or verbal abuse, Fisher’s
exact p < .001, relationship difficulties, Fisher’s exact p = .005, having been radicalized by a
friend, Fisher’s exact p < .001, having been radicalized by an associate, Fisher’s exact p = .046,
having had the experience of war, Fisher’s exact p = .007, having a political grievance, Fisher’s
exact p = .001, having a racial grievance, Fisher’s exact p < .001, a social grievance, Fisher’s
exact p < .001, having an anarchist ideology, Fisher’s exact p < .001, a religious ideology,
Fisher’s exact p < .001, a white supremacist ideology, Fisher’s exact p < .001, access to weapons,
Fisher’s exact p = .003, and military experience, Fisher’s exact p = .001.
As indicated in Table 6.3, violence to others was significantly more likely among those
experiencing emotional distress (73.81% vs. 51.49%), having a mental health diagnosis (80.00%
vs. 51.48%), exhibiting mental health symptoms (80.00% vs. 48.72%), having experienced
physical or verbal abuse (91.67% vs. 56.08%), having experienced relationship difficulties
(88.89% vs. 55.77%), having been radicalized by a friend (31.33% vs. 84.81%), having been
radicalized by an associate (51.46% vs. 66.13%), having experienced war (72.34% vs. 50.38%),
having a political grievance (64.95% vs. 41.57%), a racial grievance (90.91% vs. 50.31%), a
religious ideology (71.15% vs. 36.78%), a white supremacist ideology (90.91% vs. 52.57%),
having access to weapons (61.97% vs. 29.17%), and military experience (77.55% vs. 50.65%).
Additionally, the likelihood of violence to others was, in fact, lower among those who had a

135
social grievance (23.88% vs. 72.27%), as well as those who had an anarchist ideology (7.14% vs.
60.22%).
Table 6.3
Cross-tabulations with Violence to Others: Significant Fisher’s Exact Tests
Measure Violence to Others Total
No Yes
Emotional
Distress
No 65 69 134
48.51% 51.49% 100.00%
85.53% 69.00% 76.14%
Yes 11 31 42
26.19% 73.81% 100.00%
14.47% 31.00% 23.86%
Total 76 100 176
43.18% 56.82% 100.00%
100.00% 100.00% 100.00%
Mental Health
Diagnosis
No 82 87 169
48.52% 51.48% 100.00%
96.47% 87.88% 91.85%
Yes 3 12 15
20.00% 80.00% 100.00%
3.53% 12.12% 8.15%

136
Total 85 99 184
46.20% 53.80% 100.00%
100.00% 100.00% 100.00%
Mental Health
Symptoms
No 80 76 156
51.28% 48.72% 100.00%
90.91% 70.37% 79.59%
Yes 8 32 40
20.00% 80.00% 100.00%
9.09% 29.63% 20.41%
Total 88 108 196
44.90% 55.10% 100.00%
100.00% 100.00% 100.00%
Physical/Verbal
Abuse
No 65 83 148
43.92% 56.08% 100.00%
97.01% 79.05% 86.05%
Yes 2 22 24
8.33% 91.67% 100.00%
2.99% 20.95% 13.95%
Total 67 105 172
38.95% 61.05% 100.00%
100.00% 100.00% 100.00%

137
Relationship
Difficulties
No 69 87 156
44.23% 55.77% 100.00%
97.18% 84.47% 89.66%
Yes 2 16 18
11.11% 88.89% 100.00%
2.82% 15.53% 10.34%
Total 71 103 174
40.80% 59.20% 100.00%
100.00% 100.00% 100.00%
Radicalized by
Friend
No 12 67 79
15.19% 84.81% 100.00%
17.39% 72.04% 48.77%
Yes 57 26 83
68.67% 31.33% 100.00%
82.61% 27.96% 51.23%
Total 69 93 162
42.59% 57.41% 100.00%
100.00% 100.00% 100.00%
Radicalized by
Associate
No 21 41 62
33.87% 66.13% 100.00%

138
29.58% 43.62% 37.58%
Yes 50 53 103
48.54% 51.46% 100.00%
70.42% 56.38% 62.42%
Total 71 94 165
43.03% 56.97% 100.00%
100.00% 100.00% 100.00%
Experience of
War
No 65 66 131
49.62% 50.38% 100.00%
83.33% 66.00% 73.60%
Yes 13 34 47
27.66% 72.34% 100.00%
16.67% 34.00% 26.40%
Total 78 100 178
43.82% 56.18% 100.00%
100.00% 100.00% 100.00%
Political
Grievances
No 52 37 89
58.43% 41.57% 100.00%
60.47% 37.00% 47.85%
Yes 34 63 97
35.05% 64.95% 100.00%

139
39.53% 63.00% 52.15%
Total 86 100 186
46.24% 53.76% 100.00%
100.00% 100.00% 100.00%
Racial
Grievances
No 81 82 163
49.69% 50.31% 100.00%
97.59% 80.39% 88.11%
Yes 2 20 22
9.09% 90.91% 100.00%
2.41% 19.61% 11.89%
Total 83 102 185
44.86% 55.14% 100.00%
100.00% 100.00% 100.00%
Social
Grievances
No 33 86 119
27.73% 72.27% 100.00%
39.29% 84.31% 63.98%
Yes 51 16 67
76.12% 23.88% 100.00%
60.71% 15.69% 36.02%
Total 84 102 186
45.16% 54.84% 100.00%

140
100.00% 100.00% 100.00%
Anarchist
Ideology
No 72 109 181
39.78% 60.22% 100.00%
84.71% 99.09% 92.82%
Yes 13 100 14
92.86% 7.14% 100.00%
15.29% 0.91% 7.18%
Total 85 110 195
43.59% 56.41% 100.00%
100.00% 100.00% 100.00%
Religious
Ideology
No 55 32 87
63.22% 36.78% 100.00%
64.71% 30.19% 45.55%
Yes 30 74 104
28.85% 71.15% 100.00%
35.29% 69.81% 54.45%
Total 85 106 191
44.50% 55.50% 100.00%
100.00% 100.00% 100.00%
White
Supremacist
Ideology

141
No 83 92 175
47.43% 52.57% 100.00%
97.65% 82.14% 88.83%
Yes 2 20 22
9.09% 90.91% 100.00%
2.35% 17.86% 11.17%
Total 85 112 197
43.15% 56.85% 100.00%
100.00% 100.00% 100.00%
Weapons Access
No 17 7 24
70.83% 29.17% 100.00%
23.94% 7.37% 14.46%
Yes 54 88 142
38.03% 61.97% 100.00%
76.06% 92.63% 85.54%
Total 71 95 166
42.77% 57.23% 100.00%
100.00% 100.00% 100.00%
Military
Experience
No 76 78 154
49.35% 50.65% 100.00%
87.36% 67.24% 75.86%

142
Yes 11 38 49
22.45% 77.55% 100.00%
12.64% 32.76% 24.14%
Total 87 116 203
42.86% 57.14% 100.00%
100.00% 100.00% 100.00%
Violence from Others
The following set of analyses examined associations with violence from others.
Significance was not indicated with respect to the associations with having a mental health
diagnosis: Fisher’s exact p = .545, the loss of a relationship, Fisher’s exact p = .100, the loss of a
residence, Fisher’s exact p = .098, relationship difficulties, Fisher’s exact p = .060, having been
radicalized by family, Fisher’s exact p = .493, a friend, Fisher’s exact p = .329, or an associate,
Fisher’s exact p = .112, having experienced war, Fisher’s exact p = .480, having a personal
grievance, Fisher’s exact p = .310, a political grievance, Fisher’s exact p = .119, a religious
grievance, Fisher’s exact p = .138, a social grievance, Fisher’s exact p = .095, having an
anarchist ideology, Fisher’s exact p = .330, an anti-government ideology, Fisher’s exact p = .626,
a black nationalist ideology, Fisher’s exact p = .138, a religious ideology, Fisher’s exact p = .444,
a sovereign citizen ideology, Fisher’s exact p = .172, violent video consumption, Fisher’s exact p
= .243, weapons access, Fisher’s exact p = .188, or military experience, Fisher’s exact p = .587.
Within these analyses, alcohol abuse could not be examined as all cases were coded “no”
for this measure.

143
Significance in the Fisher’s exact tests was found with regard to the comparisons made
with emotional distress: Fisher’s exact p < .001, mental health symptoms, Fisher’s exact p <
.001, a loss due to a death, Fisher’s exact p = .028, the loss of a job, Fisher’s exact p < .001, drug
abuse, Fisher’s exact p = .001, physical or verbal abuse, Fisher’s exact p < .001, financial
difficulties, Fisher’s exact p < .001, having an economic grievance, Fisher’s exact p = .034, a
racial grievance, Fisher’s exact p < .001, or a white supremacist ideology, Fisher’s exact p <
.001.
As indicated in Table 6.4, a significantly greater likelihood of receiving violence from
others was associated with those who experienced emotional distress (32.43% vs. 0.79%), those
who had mental health symptoms (33.33% vs. 4.96%), those who experienced a loss due to a
death (30.00% vs. 5.88%), those who experienced the loss of a job (34.78% vs. 4.90%), those
who abused drugs (44.44% vs. 4.73%), those who experienced physical or verbal abuse (57.89%
vs. 2.11%), emotional distress (32.43% vs. 0.79%), or financial difficulties (41.18% vs. 4.86%),
those who had an economic grievance (33.33% vs. 7.38%), a racial grievance (46.67% vs.
3.50%), or those who had an ideology of white supremacy (50.00% vs. 4.64%).
Table 6.4
Cross-tabulations with Violence from Others: Significant Fisher’s Exact Tests
Measure Violence from Others Total
No Yes
Emotional
Distress
No 125 1 126
99.21% 0.79% 100.00%
83.33% 7.69% 77.30%

144
Yes 25 12 37
67.57% 32.43% 100.00%
16.67% 92.31% 22.70%
Total 150 13 163
92.02% 7.98% 100.00%
100.00% 100.00% 100.00%
Mental Health
(Symptoms)
No 134 7 141
95.04% 4.96% 100.00%
88.16% 43.75% 83.93%
Yes 18 9 27
66.67% 33.33% 100.00%
11.84% 56.25% 16.07%
Total 152 16 168
90.48% 9.52% 100.00%
100.00% 100.00% 100.00%
Loss: Death
No 144 9 153
94.12% 5.88% 100.00%
95.36% 75.00% 93.87%
Yes 7 3 10
70.00% 30.00% 100.00%

145
4.64% 25.00% 6.13%
Total 151 12 163
92.64% 7.36% 100.00%
100.00% 100.00% 100.00%
Loss: Job
No 136 7 143
95.10% 4.90% 100.00%
90.07% 46.67% 86.14%
Yes 15 8 23
65.22% 34.78% 100.00%
9.93% 53.33% 13.86%
Total 151 15 166
90.96% 9.04% 100.00%
100.00% 100.00% 100.00%
Drug Abuse
No 141 7 148
95.27% 4.73% 100.00%
96.58% 63.64% 94.27%
Yes 5 4 9
55.56% 44.44% 100.00%
3.42% 36.36% 5.73%
Total 146 11 157
92.99% 7.01% 100.00%

146
100.00% 100.00% 100.00%
Physical/Verbal
Abuse
No 139 3 142
97.89% 2.11% 100.00%
94.56% 21.43% 88.20%
Yes 8 11 19
42.11% 57.89% 100.00%
5.44% 78.57% 11.80%
Total 147 14 161
91.30% 8.70% 100.00%
100.00% 100.00% 100.00%
Emotional
Distress
No 125 1 126
99.21% 0.79% 100.00%
83.33% 7.69% 77.30%
Yes 25 12 37
67.57% 32.43% 100.00%
16.67% 92.31% 22.70%
Total 150 13 163
92.02% 7.98% 100.00%
100.00% 100.00% 100.00%
Financial
Difficulties

147
No 137 7 144
95.14% 4.86% 100.00%
93.20% 50.00% 89.44%
Yes 10 7 17
58.82% 41.18% 100.00%
6.80% 50.00% 10.56%
Total 147 14 161
91.30% 8.70% 100.00%
100.00% 100.00% 100.00%
Economic
Grievance
No 138 11 149
92.62% 7.38% 100.00%
95.83% 78.57% 94.30%
Yes 6 3 9
66.67% 33.33% 100.00%
4.17% 21.43% 5.70%
Total 144 14 158
91.14% 8.86% 100.00%
100.00% 100.00% 100.00%
Racial
Grievance
No 138 5 143
96.50% 3.50% 100.00%
94.52% 41.67% 90.51%

148
Yes 8 7 15
53.33% 46.67% 100.00%
5.48% 58.33% 9.49%
Total 146 12 158
92.41% 7.59% 100.00%
100.00% 100.00% 100.00%
Ideology. White
Supremacy
No 144 7 151
95.36% 4.64% 100.00%
94.74% 46.67% 90.42%
Yes 8 8 16
50.00% 50.00% 100.00%
5.26% 53.33% 9.58%
Total 152 15 167
91.02% 8.98% 100.00%
100.00% 100.00% 100.00%
Vagrancy
Finally, vagrancy was examined. In these analyses, significant associations were not
found with a loss due to a death: Fisher’s exact p = .538, the loss of a relationship, Fisher’s exact
p = .092, physical or verbal abuse, Fisher’s exact p = .079, relationship difficulties, Fisher’s exact
p = .050, having been radicalized by family, Fisher’s exact p = .621, a friend, Fisher’s exact p =
.203, an associate, Fisher’s exact p = .061, having experienced a war, Fisher’s exact p = .126,
having a political grievance, Fisher’s exact p = .227, a religious grievance, Fisher’s exact p =

149
.162, a social grievance, Fisher’s exact p = .085, having an anarchist ideology, Fisher’s exact p =
.208, an anti-government ideology, Fisher’s exact p = .095, a black nationalist ideology, Fisher’s
exact p = .182, a religious ideology, Fisher’s exact p = .527, a white supremacist ideology,
Fisher’s exact p = .129, violent video consumption, Fisher’s exact p = .099, access to weapons,
Fisher’s exact p = .209, or military experience, Fisher’s exact p = .277.
Among these analyses, significance was found with regard to emotional distress: Fisher’s
exact p < .001, having a mental health diagnosis, Fisher’s exact p = .011, mental health
symptoms, Fisher’s exact p < .001, the loss of a job, Fisher’s exact p < .001, the loss of a
residence, Fisher’s exact p = .001, alcohol abuse, Fisher’s exact p = .009, drug abuse, Fisher’s
exact p < .001, financial difficulties, Fisher’s exact p < .001, an economic grievance, Fisher’s
exact p = .002, a personal grievance, Fisher’s exact p = .022, a racial grievance, Fisher’s exact p
= .006, and having a sovereign citizen ideology, Fisher’s exact p = .001.
As shown in Table 6.5, the likelihood of vagrancy was significantly increased among
those with emotional distress (33.33% vs. 9.35%), those who had a mental health diagnosis
(36.84% vs. 12.43%), those who exhibited mental health symptoms (36.59% vs. 9.52%), those
who experienced the loss of a job (42.31% vs. 10.90%), the loss of a residence (68.75% vs.
10.16%), those who abused alcohol (75.00% vs. 12.78%), those who abused drugs (53.33% vs.
11.11%), experienced financial difficulties (42.86% vs. 9.70%), those who had an economic
grievance (55.56% vs. 10.94%), a personal grievance (22.92% vs. 9.87%), a racial grievance
(30.56% vs. 11.49%), and finally, those who had a sovereign citizen ideology (71.43% vs.
13.62%).

150
Table 6.5
Cross-tabulations with Vagrancy: Significant Fisher’s Exact Tests
Measure Vagrancy Total
No Yes
Emotional
Distress
No 126 13 139
90.65% 9.35% 100.00%
78.75% 43.33% 73.16%
Yes 34 17 51
66.67% 33.33% 100.00%
21.25% 56.67% 26.84%
Total 160 30 190
84.21% 15.79% 100.00%
100.00% 100.00% 100.00%
Mental Health
Diagnosis
No 162 23 185
87.57% 12.43% 100.00%
93.10% 76.67% 90.69%
Yes 12 7 19
63.16% 36.84% 100.00%
6.90% 23.33% 9.31%
Total 174 30 204
85.29% 14.71% 100.00%

151
100.00% 100.00% 100.00%
Mental Health
Symptoms
No 152 16 168
90.48% 9.52% 100.00%
85.39% 51.61% 80.38%
Yes 26 15 41
63.41% 36.59% 100.00%
14.61% 48.39% 19.62%
Total 178 31 209
85.17% 14.83% 100.00%
100.00% 100.00% 100.00%
Loss: Job
No 139 17 156
89.10% 10.90% 100.00%
90.26% 60.71% 85.71%
Yes 15 11 26
57.69% 42.31% 100.00%
9.74% 39.29% 14.29%
Total 154 28 182
84.62% 15.38% 100.00%
100.00% 100.00% 100.00%
Loss: Residence
No 168 19 187

152
97.11% 63.33% 92.12%
Yes 5 11 16
31.25% 68.75% 100.00%
2.89% 36.67% 7.88%
Total 173 30 203
85.22% 14.78% 100.00%
100.00% 100.00% 100.00%
Alcohol Abuse
No 157 23 180
87.22% 12.78% 100.00%
99.37% 88.46% 97.83%
Yes 1 3 4
25.00% 75.00% 100.00%
0.63% 11.54% 2.17%
Total 158 26 184
85.87% 14.13% 100.00%
100.00% 100.00% 100.00%
Drug Abuse
No 152 19 171
88.89% 11.11% 100.00%
95.60% 70.37% 91.94%
Yes 7 8 15

153
46.67% 53.33% 100.00%
4.40% 29.63% 8.06%
Total 159 27 186
85.48% 14.52% 100.00%
100.00% 100.00% 100.00%
Financial
Difficulties
No 149 16 165
90.30% 9.70% 100.00%
92.55% 64.00% 88.71%
Yes 12 9 21
57.14% 42.86% 100.00%
7.45% 36.00% 11.29%
Total 161 25 186
86.56% 13.44% 100.00%
100.00% 100.00% 100.00%
Economic
Grievances
No 171 21 192
89.06% 10.94% 100.00%
97.71% 80.77% 95.52%
Yes 4 5 9
44.44% 55.56% 100.00%
2.29% 19.23% 4.48%
Total 175 26 201

154
87.06% 12.94% 100.00%
100.00% 100.00% 100.00%
Personal
Grievances
No 137 15 152
90.13% 9.87% 100.00%
78.74% 57.69% 76.00%
Yes 37 11 48
77.08% 22.92% 100.00%
21.26% 42.31% 24.00%
Total 174 26 200
87.00% 13.00% 100.00%
100.00% 100.00% 100.00%
Racial
Grievances
No 154 20 174
88.51% 11.49% 100.00%
86.03% 64.52% 82.86%
Yes 25 11 36
69.44% 30.56% 100.00%
13.97% 35.48% 17.14%
Total 179 31 210
85.24% 14.76% 100.00%
100.00% 100.00% 100.00%
Sovereign
Citizen

155
No 184 29 213
86.38% 13.62% 100.00%
98.92% 85.29% 96.82%
Yes 2 5 7
28.57% 71.43% 100.00%
1.08% 14.71% 3.18%
Total 186 34 220
84.55% 15.45% 100.00%
100.00% 100.00% 100.00%
Conclusion
In this chapter, the results of the analyses conducted examining the association between
the ACE measures and risk factors included within the JRIC dataset, alongside External Factors,
were presented and discussed. All analyses took the form of Fisher’s exact tests, with significant
associations found in numerous cases.
Specifically, a significantly greater likelihood of contact with law enforcement was found
among those exhibiting mental health symptoms, those who experienced a loss through a death,
the loss of a job, those who abused alcohol, those having an economic grievance, an anarchist
ideology, or a white supremacist ideology.
Violence to the self was significantly associated with mental health symptoms, a loss due
to a death, the loss of a job, having experienced physical or verbal abuse, financial difficulties,
relationship difficulties, having an economic grievance, an anarchist ideology, or a white
supremacist ideology.

156
Next, violence to others was significantly more likely among those experiencing
emotional distress, having a mental health diagnosis, exhibiting mental health symptoms, having
experienced physical or verbal abuse, relationship difficulties, having been radicalized by a
friend or associate, having experienced war, having a racial grievance, a religious ideology, a
white supremacist ideology, access to weapons, or military experience, with a lower likelihood of
violence found among those with social grievances or those having an anarchist ideology.
Violence from others was found to be significantly more likely among those with emotional
distress, having mental health symptoms, experiencing a loss due to a death, the loss of a job,
drug abuse, having experienced physical or verbal abuse, emotional distress, financial
difficulties, having an economic grievance, a racial grievance, or an ideology of white
supremacy. Finally, vagrancy was significantly associated with emotional distress, a mental
health diagnosis, exhibiting mental health symptoms, the loss of a job, the loss of a residence,
alcohol abuse, drug abuse, financial difficulties, having an economic grievance, a personal
grievance, a racial grievance, or a sovereign citizen ideology.
The following chapter presents and discusses the results of the analyses conducted
examining ACEs and their correlates in the TASSS dataset.

157
Chapter 7: Results: ACEs in TASSS
Introduction
In this chapter, the results of the analyses conducted on the TASSS dataset are reported
and discussed. These analyses consisted of descriptive statistics and Fisher’s exact tests, serving
to examine the association between ACEs, which focus primarily on child maltreatment, and
peer victimization, which is thought to be the byproduct of ACEs, External Factors, and
Environmental Sustainable Design measures.
Descriptive Statistics
The ACEs examined in this study consisted of the following: domestic violence, a
workplace shooting, psychological issues, parents having been divorced or separated, social
stratum, significant family problems, the recent death of a relative or friend, loss of social
standing, having been dismissed from a social, political, or religious organization, and peer
aggression.
Frequencies and percentages are reported for all study measures of interest. Descriptive
statistics on all ACE measures are presented in Table 7.1, except for social stratum, which was
not dichotomous. With regard to social stratum, 50 individuals (43.48%) were of low social
standing, with 54 (46.96%) middle class and 11 (9.57%) of high social standing. With regard to
the remaining measures, focusing on the subsample that had affirmative instances on these items,
28 individuals (27.45%) were found to have had domestic violence, with seven (6.86%) having
had work violence. A majority of the sample had psychological issues (N = 93, 57.06%), as well
as the divorce or separation of their parents (N = 52, 51.49%), and significant family problems
(N = 80, 64.00%). A total of 22 individuals (23.66%) had experienced the recent death of a
family member or friend, with 40 (37.38%) having had the loss of social standing. Nine

158
individuals (7.96%) were dismissed from a social, political, or religious organization, with a
slight majority of 77 individuals (52.74%) having had experienced peer aggression.
Table 7.1
Frequencies of ACE Measures
Level of
Education No Yes
Number Percent Number Percent
Domestic Violence 74 72.55% 28 27.45%
Workplace
Shooting 95 93.14% 7 6.86%
Psychological
Issues 70 42.94% 93 57.06%
Divorced/Separated 49 48.51% 52 51.49%
Significant Family
Problems 45 36.00% 80 64.00%
Recent Death 71 76.34% 22 23.66%
Loss of Social
Standing 67 62.62% 40 37.38%
Dismissed from
Organization 104 92.04% 9 7.96%
Peer Aggression 69 47.26% 77 52.74%
The External Factors analyzed consisted of a K-12 suspension or expulsion (coded as two
separate measures), a K-12 failure, a K-12 dropout, having been a member of a street gang,
having had a criminal record, a victim being affiliated with a gang, and having struck a student,
teacher, administrator, or someone else, with these latter measures consisting of four separate
measures.
The results of the descriptive statistics associated with these External Factors are
summarized in Table 7.2. As shown, 41 individuals (33.33%) had a suspension, with 24
(17.78%) having had an expulsion. A total of 32 individuals (27.83%) experienced a failure, with

159
24 (15.29%) having dropped out. Sixty-eight (26.88%) were members of a street gang, with 115,
a slight majority (51.80%) having a criminal record. Additionally, in 33 cases (15.28%), a victim
was affiliated with a gang. The final four measures pertain to whether the individual had struck a
student, teacher, administrator, or someone else, with this being indicated in 201 cases with
respect to having struck a student (58.60%), in 46 cases when focusing on teachers (13.41%), 27
cases in which administrators were struck (7.87%), and 112 cases in which other individuals had
been struck by the perpetrator (32.56%).
Table 7.2
Frequencies of ACE External Factors
Level of
Education No Yes
Number Percent Number Percent
K-12 Suspension 82 66.67% 41 33.33%
K-12 Expulsion 111 82.22% 24 17.78%
K-12 Failure 83 72.17% 32 27.83%
K-12 Dropout 133 84.71% 24 15.29%
Street Gang
Member 185 73.12% 68 26.88%
Criminal Record 117 48.20% 115 51.80%
Victim Gang
Affiliated 183 84.72% 33 15.28%
Struck a Student 142 41.40% 201 58.60%
Struck a Teacher 297 86.59% 46 13.41%
Struck an
Administrator 316 92.13% 27 7.87%
Struck Someone
Else 232 67.44% 112 32.56%

160
Finally, Table 7.3 presents the results of the descriptive statistics conducted on the
Environmental Sustainable Design measures, which consisted of the presence of a metal
detector, a school guard, and a school police officer. A metal detector was present in 47 cases
(30.72%), with a school guard on site in 114 cases (68.67%). Additionally, a school police officer
was on site with respect to 114 cases (69.09%).
Table 7.3
Frequencies of Environmental Sustainable Design Measures
Level of
Education
No Yes
Number Percent Number Percent
Metal Detector 106 69.28% 47 30.72%
School
Guard/Resource
Officer
52 31.33% 114 68.67%
School Police
Officer
51 30.91% 114 69.09%
Fisher’s Exact Tests
In all cases where relevant, one-sided probability values were reported as it was
hypothesized that ACEs and ACE External Factors would be associated with a greater likelihood
of adverse outcomes across all measures examined in this study. These analyses examined each
ACE individually with all outcomes, followed by analyses examining ACE External Factors with
all outcomes.
With regard to domestic violence, significant associations were not indicated with K-12
suspension, Fisher’s exact p = .600, K-12 expulsion, Fisher’s exact p = .371, K-12 failure,
Fisher’s exact p = .750, a victim being gang affiliated, Fisher’s exact p = .168, having struck a
teacher, Fisher’s exact p = .168, an administrator, Fisher’s exact p = .215, or someone else,
Fisher’s exact p = .385. However, significant associations were found with K-12 dropout,

161
Fisher’s exact p = .026, being a street gang member, Fisher’s exact p < .001, having a criminal
record, Fisher’s exact p = .040, and having struck a student, Fisher’s exact p = .007. A crosstabulation table illustrating the association between domestic violence and a K-12 dropout is
presented in Table 7.4. As shown, individuals who had indicated domestic violence were, in fact,
significantly less likely to have dropped out (0.00%) as compared with those who had not
experienced domestic violence (50.00%).
Table 7.4
Cross-tabulation of Domestic Violence with K-12 Dropout
Domestic
Violence K-12 Dropout Total
No Yes
No 8 8 16
50.00% 50.00% 100.00%
53.33% 100.00% 69.57%
Yes 7 0 7
100.00% 0.00% 100.00%
46.67% 0.00% 30.43%
Total 15 8 23
65.22% 34.78% 100.00%
100.00% 100.00% 100.00%
Table 7.5 presents the cross-tabulation between domestic violence and having been a
gang member. These results also indicate that those who had experienced domestic violence were
significantly less likely to have become gang members (0.00%) as compared with those who had
not experienced domestic violence (37.21%).

162
Table 7.5
Cross-tabulation of Domestic Violence with Street Gang Member
Domestic
Violence Street Gang Member Total
No Yes
No 27 16 43
62.79% 37.21% 100.00%
52.94% 100.00% 64.18%
Yes 24 0 24
100.00% 0.00% 100.00%
47.06% 0.00% 35.82%
Total 51 16 67
76.12% 23.88% 100.00%
100.00% 100.00% 100.00%
Table 7.6 presents the cross-tabulation conducted between domestic violence and having
a criminal record. These results indicate that individuals who had experienced domestic violence
were significantly less likely to have also had a criminal record (36.84%) as compared with those
who had not experienced domestic violence (64.44%).

163
Table 7.6
Cross-tabulation of Domestic Violence with Criminal Record
Domestic
Violence Criminal Record Total
No Yes
No 16 29 45
35.56% 64.44% 100.00%
57.14% 80.56% 70.31%
Yes 12 7 19
63.16% 36.84% 100.00%
42.86% 19.44% 29.69%
Total 28 36 64
43.75% 56.25% 100.00%
100.00% 100.00% 100.00
The cross-tabulation between having experienced domestic violence and having struck a
student is presented in Table 7.7. These results indicate a significantly lower likelihood of having
struck a student if the individual had experienced domestic violence (10.71%) as compared with
those who had not experienced domestic violence (37.14%).

164
Table 7.7
Cross-tabulation of Domestic Violence with Struck a Student
Domestic
Violence Struck a Student Total
No Yes
No 44 26 70
62.86% 37.14% 100.00%
63.77% 89.66% 71.43%
Yes 25 3 28
89.29% 10.71% 100.00%
36.23% 10.34% 28.57%
Total 69 29 98
70.41% 29.59% 100.00%
100.00% 100.00% 100.00%
With respect to a workplace shooting, significant associations were not found with K-12
suspension, Fisher’s exact p = .850, K-12 expulsion, Fisher’s exact p = .729, K-12 failure,
Fisher’s exact p = .750, K-12 dropout, Fisher’s exact p = .257, being a street gang member,
Fisher’s exact p = .180, having a criminal record, Fisher’s exact p = .123, a victim being gang
affiliated, Fisher’s exact p = .654, having struck a student, Fisher’s exact p = .078, or a teacher,
Fisher’s exact p = .673. Significance was found in the relationship with having struck an
administrator, Fisher’s exact p < .001, or someone else, Fisher’s exact p = .025. Table 7.8
presents the cross-tabulation conducted between work violence and having struck an
administrator. These results indicate a significantly higher likelihood of having struck an

165
administrator if the individual had experienced work violence (85.71%) as compared with cases
where they had not (12.09%).
Table 7.8
Cross-tabulation of Workplace Shooting with Struck an Administrator
Work Violence Struck an
Administrator Total
No Yes
No 80 11 91
87.91% 12.09% 100.00%
98.77% 64.71% 92.86%
Yes 1 6 7
14.29% 85.71% 100.00%
1.23% 35.29% 7.14%
Total 81 17 98
82.65% 17.35% 100.00%
100.00% 100.00% 100.00
Table 7.9 presents the cross-tabulation conducted between work violence and having
struck someone else. The data presented here indicate a significantly lower likelihood of having
struck someone else if the individual had experienced work violence (14.29%) as compared with
those who had not experienced work violence (59.78%).

166
Table 7.9
Cross-tabulation of Workplace Shooting with Struck Someone Else
Work Violence Struck Someone Else Total
No Yes
No 37 55 92
40.22% 59.78% 100.00%
86.05% 98.21% 92.93%
Yes 6 1 7
85.71% 14.29% 100.00%
13.95% 1.79% 7.07%
Total 43 56 99
43.43% 56.57% 100.00%
100.00% 100.00% 100.00%
The following set of analyses examined psychological issues. These results indicated no
significant association with K-12 suspension, Fisher’s exact p = .267, K-12 expulsion, Fisher’s
exact p = .448, K-12 failure, Fisher’s exact p = .468, K-12 dropout, Fisher’s exact p = .120,
having a criminal record, Fisher’s exact p = .160, having struck a student, Fisher’s exact p =
.107, or someone else, Fisher’s exact p = .198. Significance was found in the associations
between psychological issues and being a street gang member, Fisher’s exact p = .005; a victim
being gang affiliated, Fisher’s exact p = .007; having struck a teacher, Fisher’s exact p < .001; or
an administrator, Fisher’s exact p = .046. Table 7.10 presents the cross-tabulation associated with
the first significant association, consisting of the relationship between having a psychological
issue and being a gang member. These data indicate that having had a psychological issue was

167
associated with a significantly reduced likelihood of also having been a gang member (6.10%) as
compared with those who did not have a psychological issue (22.22%).
Table 7.10
Cross-tabulation of Psychological Issues with Street Gang Member
Psychological
Issue Street Gang Member Total
No Yes
No 49 14 63
77.78% 22.22% 100.00%
38.89% 73.68% 43.45%
Yes 77 5 82
93.90% 6.10% 100.00%
61.11% 26.32% 56.55%
Total 126 19 145
86.90% 13.10% 100.00%
100.00% 100.00% 100.00
The cross-tabulation between having a psychological issue and a victim having been gang
affiliated is presented in Table 7.11. A significantly lower likelihood of a victim having been
gang affiliated was found to be associated with the perpetrator having a psychological issue
(2.74%) as compared with cases in which the perpetrator did not have a psychological issue
(17.78%).

168
Table 7.11
Cross-tabulation of Psychological Issues with Victim Gang Affiliated
Psychological
Issue Victim Gang-Affiliated Total
No Yes
No 37 8 45
82.22% 17.78% 100.00%
34.26% 80.00% 38.14%
Yes 71 2 73
97.26% 2.74% 100.00%
65.74% 20.00% 61.86%
Total 108 10 118
91.53% 8.47% 100.00%
100.00% 100.00% 100.00%
Table 7.12 presents the results of the cross-tabulation conducted between having a
psychological issue and having struck a teacher. These results indicate a significantly higher
likelihood of having struck a teacher if the perpetrator had experienced a psychological issue
(31.18%) as compared with cases in which the individual did not have a psychological issue
(8.57%).

169
Table 7.12
Cross-tabulation of Psychological Issues with Struck a Teacher
Psychological
Issue Struck a Teacher Total
No Yes
No 64 6 70
91.43% 8.57% 100.00%
50.00% 17.14% 42.94%
Yes 64 29 93
68.82% 31.18% 100.00%
50.00% 82.86% 57.06%
Total 128 35 163
78.53% 21.47% 100.00%
100.00% 100.00% 100.00%
Next, Table 7.13 presents the data illustrating the relationship between having a
psychological issue and having struck an administrator. These data show that the likelihood of
having struck an administrator was significantly higher in cases where the perpetrator had
experienced a psychological issue (17.20%) as compared with cases in which no psychological
issue was indicated (7.14%).

170
Table 7.13
Cross-tabulation of Psychological Issues with Struck an Administrator
Psychological Issue Struck an Administrator Total
No Yes
No 65 5 70
92.86% 7.14% 100.00%
45.77% 23.81% 42.94%
Yes 77 16 93
82.80% 17.20% 100.00%
54.23% 76.19% 57.06%
Total 142 21 163
87.12% 12.88% 100.00%
100.00% 100.00% 100.00%
Next, parents having been divorced or separated was analyzed. These analyses found no
significant association between parents having been divorced or separated and K-12 suspension,
Fisher’s exact p = .203; K-12 expulsion, Fisher’s exact p = .432; K-12 dropout, Fisher’s exact p
= .271; and being a street gang member, Fisher’s exact p = .515; having a criminal record,
Fisher’s exact p = .056; a victim being gang affiliated, Fisher’s exact p = .307; having struck a
student, Fisher’s exact p = .198; a teacher, Fisher’s exact p = .323; an administrator, Fisher’s
exact p = .423; and someone else, Fisher’s exact p = .239. Significance was found with K-12
failure, Fisher’s exact p = .011. Table 7.14 presents the results of this associated crosstabulation.
These data indicate that the likelihood of a K-12 failure was significantly increased among those

171
whose parents were divorced or separated (28.95%) as compared with those whose parents were
not (5.88%).
Table 7.14
Cross-tabulation of Divorced/Separated with K-12 Failure
Social Stratum Divorced/Separated with K-12
Failure Total
No Yes
No 32 2 34
94.12% 5.88% 100.00%
54.24% 15.38% 47.22%
Yes 27 11 38
71.05% 28.95% 100.00%
45.76% 84.62% 52.78%
Total 59 13 72
81.94% 18.06% 100.00%
100.00% 100.00% 100.00%
The analyses conducted with social stratum found no significant association with respect
to K-12 suspension, Fisher’s exact p = .076; K-12 expulsion, Fisher’s exact p = .319; K-12
dropout, Fisher’s exact p = .769; having a criminal record, Fisher’s exact p = .227; having struck
a student, Fisher’s exact p = .167; a teacher, Fisher’s exact p = .187; an administrator, Fisher’s
exact p = .440; and someone else, Fisher’s exact p = .496. However, significance was found
regarding K-12 failure, Fisher’s exact p = .015; being a street gang member, Fisher’s exact p =
.043; and a victim being gang affiliated, Fisher’s exact p = .048. Table 7.15 presents the results of
the cross-tabulation associated with this first significant finding, that being between social

172
stratum and K-12 failure. These results show that the lowest likelihood of a K-12 failure was
associated with individuals in the middle class (5.88%). This likelihood increased among those
of a high social stratum (12.50%) and was highest among those of low social class (33.33%).
These results indicate a nonlinear relationship between social stratum and K-12 failure,
indicating a relationship in which those of low social stratum were at the highest risk of K-12
failure, followed by those of high social standing.
Table 7.15
Cross-tabulation of Social Stratum with K-12 Failure
Social Stratum K-12 Failure Total
No Yes
Low 20 10 30
66.67% 33.33% 100.00%
33.90% 76.92% 41.67%
Middle 32 2 34
94.12% 5.88% 100.00%
54.24% 15.38% 47.22%
High 7 1 8
87.50% 12.50% 100.00%
11.86% 7.69% 11.11%
Total 59 13 72
81.94% 18.06% 100.00%
100.00% 100.00% 100.00%

173
Table 7.16 presents the data illustrating the relationship between social stratum and
having been a gang member. These results show that individuals in the middle class were least
likely to have been gang members (3.92%). This likelihood increased among those of high social
standing (10.00%) and was highest among those with low social standing (19.51%). Similarly,
here, a linear relationship is not indicated, but instead one in which those of low social standing
were at greatest risk, followed by those of high social standing.
Table 7.16
Cross-tabulation of Social Stratum with Street Gang Member
Social Stratum Street Gang Member Total
No Yes
Low 33 8 41
80.49% 19.51% 100.00%
36.26% 72.73% 40.20%
Middle 49 2 51
96.08% 3.92% 100.00%
53.85% 18.18% 50.00%
High 9 1 10
90.00% 10.00% 100.00%
9.89% 9.09% 9.80%
Total 91 11 102
89.22% 10.78% 100.00%
100.00% 100.00% 100.00%

174
The association between social stratum and a victim having been gang affiliated is
presented in Table 7.17. These results show that with regard to perpetrators in the middle class,
the likelihood of a victim being gang affiliated was lowest (0.00%). Risk was increased and
similar among perpetrators with low social standing (10.53%) and those of high social stratum
(11.11%). These results similarly indicate non-linearity but instead indicate low risk among those
perpetrators in the middle class, with higher risk for those of low and high social strata.
Table 7.17
Cross-tabulation of Social Stratum with Victim Gang Affiliated
Social Stratum Victim Gang-Affiliated Total
No Yes
Low 34 4 38
89.47% 10.53% 100.00%
39.53% 80.00% 41.76%
Middle 44 0 44
100.00% 0.00% 100.00%
51.16% 0.00% 48.35%
High 8 1 9
88.89% 11.11% 100.00%
9.30% 20.00% 9.89%
Total 86 5 91
94.51% 5.49% 100.00%
100.00% 100.00% 100.00%

175
The following analyses examined significant family problems. Significant associations
were not indicated with respect to K-12 suspension, Fisher’s exact p = .107; K-12 expulsion,
Fisher’s exact p = .331; K-12 failure, Fisher’s exact p = .100; K-12 dropout, Fisher’s exact p =
.351; being a street gang member, Fisher’s exact p = .379; a victim being gang affiliated, Fisher’s
exact p = .587; having struck a student, Fisher’s exact p = .333; a teacher, Fisher’s exact p =
.377; or an administrator, Fisher’s exact p = .567. However, significance was found with respect
to having a criminal record, Fisher’s exact p = .042; and having struck someone else, Fisher’s
exact p = .007. Table 7.18 presents the results of this first crosstabulation. As shown, among
those with family problems, the likelihood of having a criminal record was significantly higher
(44.29%) as compared with those who had not experienced family problems (26.19%).
Table 7.18
Cross-tabulation of Significant Family Problems with Criminal Record
Family
Problems Criminal Record Total
No Yes
No 31 11 42
73.81% 26.19% 100.00%
44.29% 26.19% 37.50%
Yes 39 31 70
55.71% 44.29% 100.00%
55.71% 73.81% 62.50%
Total 70 42 112
62.50% 37.50% 100.00%
100.00% 100.00% 100.00%

176
The results of the cross-tabulation between having experienced significant family
problems and having struck someone else are presented in Table 7.19. As shown, the likelihood
of having struck someone else is significantly higher among those who also experienced family
problems (28.75%) as compared with those who had not experienced family problems (8.89%).
Table 7.19
Cross-tabulation of Significant Family Problems with Struck Someone Else
Family
Problems Struck Someone Else Total
No Yes
No 41 4 45
91.11% 8.89% 100.00%
41.84% 14.81% 36.00%
Yes 57 23 80
71.25% 28.75% 100.00%
58.16% 85.19% 64.00%
Total 98 27 125
78.40% 21.60% 100.00%
100.00% 100.00% 100.00%
The following set of analyses examined having experienced a recent death. These
analyses did not find any significant association with K-12 suspension, Fisher’s exact p = .394;
K-12 expulsion, Fisher’s exact p = .708; K-12 dropout, Fisher’s exact p = .166; being a street
gang member, Fisher’s exact p = .058; having a criminal record, Fisher’s exact p = .259; a victim
being gang affiliated, Fisher’s exact p = .058; having struck a teacher, Fisher’s exact p = .589; or
someone else, Fisher’s exact p = .303. Significance was indicated with respect to K-12 failure,

177
Fisher’s exact p = .010; having struck a student, Fisher’s exact p = .015; and an administrator,
Fisher’s exact p = .012. Table 7.20 presents the cross-tabulation conducted between having
experienced a recent death of a friend or family member and K-12 failure. As shown, the
likelihood of having experienced K-12 failure was significantly increased with respect to those
who had experienced a recent death (46.67%) as compared with those who had not (12.77%).
Table 7.20
Cross-tabulation of Recent Death with K-12 Failure
Recent Death K-12 Failure Total
No Yes
No 41 6 47
87.23% 12.77% 100.00%
83.67% 46.15% 75.81%
Yes 8 7 15
53.33% 46.67% 100.00%
16.33% 53.85% 24.19%
Total 49 13 62
79.03% 20.97% 100.00%
100.00% 100.00% 100.00%
Table 7.21 presents the results of the cross-tabulation conducted between those who
experienced the recent death of a friend or family member and having struck a student. Similar to
the previous results, risk increased among those who had experienced a recent death, with the
likelihood of having struck a student found to be 86.36% among those who had experienced a

178
recent death of a friend or family member, and with this figure being 59.15% among those who
had not.
Table 7.21
Cross-tabulation of Recent Death with Struck a Student
Recent Death Struck a Student Total
No Yes
No 29 42 71
40.85% 59.15% 100.00%
90.62% 68.85% 76.34%
Yes 3 19 22
13.64% 86.36% 100.00%
9.38% 31.15% 23.66%
Total 32 61 93
34.41% 65.59% 100.00%
100.00% 100.00% 100.00%
The cross-tabulation conducted between having experienced a recent death and having
struck an administrator is presented in Table 7.22. As shown, results differed here, with none of
the perpetrators who had experienced a recent death having also struck an administrator (0.00%).
Among those who had not experienced a recent death, 21.13% had also struck an administrator.

179
Table 7.22
Cross-tabulation of Recent Death with Struck an Administrator
Recent Death Struck an Administrator Total
No Yes
No 56 15 71
78.87% 21.13% 100.00%
71.79% 100.00% 76.34%
Yes 22 0 22
100.00% 0.00% 100.00%
28.21% 0.00% 23.66%
Total 78 15 93
83.87% 16.13% 100.00%
100.00% 100.00% 100.00%
Loss of social standing was examined next, with no significant association found with K12 suspension, Fisher’s exact p = .105; K-12 expulsion, Fisher’s exact p = .100; K-12 failure,
Fisher’s exact p = .556; K-12 dropout, Fisher’s exact p = .115; being a street gang member,
Fisher’s exact p = .121; having a criminal record, Fisher’s exact p = .515; a victim being gang
affiliated, Fisher’s exact p = .738; having struck a student, Fisher’s exact p = .119; a teacher,
Fisher’s exact p = .349; an administrator, Fisher’s exact p = .121; and someone else, Fisher’s
exact p = .355. These results indicate no significant association between loss of social standing
and these other study measures.

180
The following set of analyses examined having been dismissed from an organization,
with no significant association found with K-12 suspension, Fisher’s exact p = .317; K-12
expulsion, Fisher’s exact p = .169; K-12 failure, Fisher’s exact p = .400; K-12 dropout, Fisher’s
exact p = .677; being a street gang member, Fisher’s exact p = .281; having a criminal record,
Fisher’s exact p = .424; a victim being gang affiliated, Fisher’s exact p = .470; having struck a
teacher, Fisher’s exact p = .444; or someone else, Fisher’s exact p = .242. Significance was,
however, found in the associations between having been dismissed from an organization and
having struck a student, Fisher’s exact p = .008; and an administrator, Fisher’s exact p = .017.
Table 7.23 presents the results of the cross-tabulation conducted between having
experienced a loss of social standing and having struck a student. The data indicate that the
likelihood of having struck a student was significantly reduced among those who had
experienced a loss of social standing (22.22%) as compared with those who had not (69.23%).
Table 7.23
Cross-tabulation of Loss of Social Standing with Struck a Student
Social Standing
Loss Struck a Student Total
No Yes
No 32 72 104
30.77% 69.23% 100.00%
82.05% 97.30% 92.04%
Yes 7 2 9
77.78% 22.22% 100.00%
17.95% 2.70% 7.96%
Total 39 74 113

181
34.51% 65.49% 100.00%
100.00% 100.00% 100.00%
The results of the cross-tabulation conducted between having experienced a loss of social
standing and having struck an administrator are presented in Table 7.24. These data indicate a
significantly increased likelihood of having struck an administrator among those who had also
experienced a loss of social standing (44.44%) as compared with those who had not (10.58%).
Table 7.24
Cross-tabulation of Loss of Social Standing with Struck an Administrator
Social Standing
Loss Struck an Administrator Total
No Yes
No 93 11 104
89.42% 10.58% 100.00%
94.90% 73.33% 92.04%
Yes 5 4 9
55.56% 44.44% 100.00%
5.10% 26.67% 7.96%
Total 98 15 113
86.73% 13.27% 100.00%
100.00% 100.00% 100.00%
With respect to peer aggression, no significant association was found with K-12
expulsion, Fisher’s exact p = .117; K-12 dropout, Fisher’s exact p = .381; having a criminal
record, Fisher’s exact p = .223; or having struck someone else, Fisher’s exact p = .543. However,

182
significance was indicated with respect to the associations with K-12 suspension, Fisher’s exact
p = .035; K-12 failure, Fisher’s exact p = .034; being a street gang member, Fisher’s exact p =
.008; a victim being gang affiliated, Fisher’s exact p = .001; having struck a teacher, Fisher’s
exact p = .024; a student, Fisher’s exact p = .000; and an administrator, Fisher’s exact p = .036.
Table 7.25 presents the results of the cross-tabulation associated with this first significant
result, which consists of the association between having experienced peer aggression and a K-12
suspension. As shown, the likelihood of having experienced a K-12 suspension was significantly
increased among those who had also experienced peer aggression (41.67%) as compared with
those who had not (22.45%).
Table 7.25
Cross-tabulation of Peer Aggression with K-12 Suspension
Peer Aggression K-12 Suspension Total
No Yes
No 38 11 49
77.55% 22.45% 100.00%
57.58% 35.48% 50.52%
Yes 28 20 48
58.33% 41.67% 100.00%
42.42% 64.52% 49.48%
Total 66 31 97
68.04% 31.96% 100.00%
100.00% 100.00% 100.00%

183
Table 7.26 illustrates the results of the cross-tabulation conducted between having
experienced peer aggression and a K-12 failure. As shown, the likelihood of having experienced
a K-12 failure was significantly increased among those who had experienced peer aggression
(28.26%) as compared with those who had not (10.26%).
Table 7.26
Cross-tabulation of Peer Aggression with K-12 Failure
Peer Aggression K-12 Failure Total
No Yes
No 35 4 39
89.74% 10.26% 100.00%
51.47% 23.53% 45.88%
Yes 33 13 46
71.74% 28.26% 100.00%
48.53% 76.47% 54.12%
Total 68 17 85
80.00% 20.00% 100.00%
100.00% 100.00% 100.00%
Table 7.27 presents the results of the cross-tabulation conducted between having
experienced peer aggression and having been a gang member. These results show that the
likelihood of having been a gang member was significantly higher among those who also
experienced peer aggression (20.29%) as compared with those who had not (4.84%).

184
Table 7.27
Cross-tabulation of Peer Aggression with Street Gang Member
Peer Aggression Street Gang
Member Total
No Yes
No 59 3 62
95.16% 4.84% 100.00%
51.75% 17.65% 47.33%
Yes 55 14 69
79.71% 20.29% 100.00%
48.25% 82.35% 52.67%
Total 114 17 131
87.02% 12.98% 100.00%
100.00% 100.00% 100.00%
The data associated with the cross-tabulation conducted between having expressed peer
aggression and a victim having been gang affiliated is presented in Table 7.28. As shown, the
likelihood of a victim having been gang affiliated was significantly increased in those cases in
which the perpetrator had also experienced peer aggression (17.65%) as compared with those
who had not (0.00%).

185
Table 7.28
Cross-tabulation of Peer Aggression with Victim Gang-Affiliated
Peer Aggression Victim Gang-Affiliated Total
No Yes
No 52 0 52
100.00% 0.00% 100.00%
55.32% 0.00% 50.49%
Yes 42 9 51
82.35% 17.65% 100.00%
44.68% 100.00% 49.51%
Total 94 9 103
91.26% 8.74% 100.00%
100.00% 100.00% 100.00%
Table 7.29 illustrates the results of the cross-tabulation conducted between having
experienced peer aggression and the perpetrator having also struck a teacher. As shown, the
likelihood of the perpetrator having struck a teacher was significantly reduced in cases where the
perpetrator also experienced peer aggression (14.29%) as compared with cases in which they had
not (28.99%).

186
Table 7.29
Cross-tabulation of Peer Aggression with Struck a Teacher
Peer Aggression Struck a Teacher Total
No Yes
No 49 20 69
71.01% 28.99% 100.00%
42.61% 64.52% 47.26%
Yes 66 11 77
85.71% 14.29% 100.00%
57.39% 35.48% 52.74%
Total 115 31 146
78.77% 21.23% 100.00%
100.00% 100.00% 100.00%
The cross-tabulation between having experienced peer aggression and having struck a
student is presented in Table 7.30. As shown, the likelihood of having struck a student was
significantly increased among those who had also experienced peer aggression (79.22%) as
compared with those who had not experienced peer aggression (52.17%).

187
Table 7.30
Cross-tabulation of Peer Aggression with Struck a Student
Peer Aggression Struck a Student Total
No Yes
No 33 36 69
47.83% 52.17% 100.00%
67.35% 37.11% 47.26%
Yes 16 61 77
20.78% 79.22% 100.00%
32.65% 62.89% 52.74%
Total 49 97 146
33.56% 66.44% 100.00%
100.00% 100.00% 100.00%
The cross-tabulation associated with the final significant Fisher’s exact test in this series,
that being between having experienced peer aggression and having struck an administrator, is
presented in Table 7.31. As shown, those who had experienced peer aggression were
significantly less likely to have also struck an administrator (6.49%) as compared with those who
had not (17.39%).

188
Table 7.31
Cross-tabulation of Peer Aggression with Struck an Administrator
Peer Aggression Struck an Administrator Total
No Yes
No 57 12 69
82.61% 17.39% 100.00%
44.19% 70.59% 47.26%
Yes 72 5 77
93.51% 6.49% 100.00%
55.81% 29.41% 52.74%
Total 129 17 146
88.36% 11.64% 100.00%
100.00% 100.00% 100.00%
Finally, Environmental Sustainable Design factors were examined alongside the ACE
measures to determine whether significant associations were present between these two groups
of variables. Initially, as the Environmental Sustainable Design factors were viewed as
moderators, analyses were conducted using the phi coefficient, with phi coefficients calculated
on the entire sample and then separately on the basis of the level of the Environmental
Sustainable Design factor in question, which was dichotomized as either “yes” or “no,”
indicating the presence or absence of the Design factor in question, respectively. However, these
analyses could not be conducted due to empty cells or missing rows or columns when running
these analyses separately on the basis of the Environmental Sustainable Design factor level. This
situation precluded the possibility of making comparisons of the strengths of these associations,

189
if any, on the basis of the level of these Environmental Sustainable Design factors. For this
reason, these factors were instead examined as additional predictors of these adverse outcomes
among individuals. These Environmental Sustainable Design factors focus on improving
individuals’ health and comfort when occupying buildings, while simultaneously reducing the
negative impact these buildings may have on the environment. This set of analyses focused on
whether these factors reduce the incidence of these adverse outcomes, including that of violence.
The TASSS dataset incorporated three Environmental Sustainable Design factors: the
presence of a metal detector, the presence of a school guard/resource officer, and the presence of
a school police officer. Regarding these measures, the presence of a metal detector was examined
first; significance was found with respect to K-12 expulsion, Fisher’s exact p = .018; K-12
dropout, Fisher’s exact p = .046; having struck a student, Fisher’s exact p = .032; and an
administrator, Fisher’s exact p = .029. Significance was not found with regard to the measures of
K-12 suspension, Fisher’s exact p = .153; K-12 failure, Fisher’s exact p = .175; being a street
gang member, Fisher’s exact p = .153; having a criminal record, Fisher’s exact p = .501; a victim
being gang affiliated, Fisher’s exact p = .571; having struck a teacher, Fisher’s exact p = .170; or
someone else, Fisher’s exact p = .566.
The results of the cross-tabulation associated with the first significant Fisher’s exact test,
pertaining to the relationship between the presence of a metal detector in the school and a K-12
expulsion, are presented in Table 7.32. These results indicate that the likelihood of having
experienced a K-12 expulsion was significantly increased among those cases in which a metal
detector was present (40.00%) as compared with those cases in which a metal detector was not
present (14.29%).

190
Table 7.32
Cross-tabulation of Metal Detector with K-12 Expulsion
Metal Detector K-12 Expulsion Total
No Yes
No 54 9 63
85.71% 14.29% 100.00%
81.82% 52.94% 75.90%
Yes 12 8 20
60.00% 40.00% 100.00%
18.18% 47.06% 24.10%
Total 66 17 83
79.52% 20.48% 100.00%
100.00% 100.00% 100.00%
Table 7.33 presents the results of the cross-tabulation conducted between the presence of
a metal detector in the school and a K-12 dropout. These results show that the likelihood of a K12 dropout was significantly increased among those cases in which a metal detector was present
in the school (26.92%) compared with cases in which a metal detector was not present (10.14%).

191
Table 7.33
Cross-tabulation of Metal Detector with K-12 Dropout
Metal Detector K-12 Dropout Total
No Yes
No 62 7 69
89.86% 10.14% 100.00%
76.54% 50.00% 72.63%
Yes 19 7 26
73.08% 26.92% 100.00%
23.46% 50.00% 27.37%
Total 81 14 95
85.26% 14.74% 100.00%
100.00% 100.00% 100.00%
Following this, Table.34 presents the results of the cross-tabulation conducted between
the presence of a metal detector in the school and whether or not the perpetrator struck a student.
These data indicate that the likelihood of having struck a student was significantly increased in
cases where a metal detector was present in the school (84.44%) as compared with cases in
which a metal detector was not present (68.57%).

192
Table 7.34
Cross-tabulation of Metal Detector with Struck a Student
Metal Detector Struck a
Student Total
No Yes
No 33 72 105
31.43% 68.57% 100.00%
82.50% 65.45% 70.00%
Yes 7 38 45
15.56% 84.44% 100.00%
17.50% 34.55% 30.00%
Total 40 110 150
26.67% 73.33% 100.00%
100.00% 100.00% 100.00%
The data showing the association between a metal detector in the school and the
perpetrator having struck an administrator are presented in Table 7.35. Again, the likelihood of
having struck an administrator was significantly increased in cases where a metal detector was
present in the school (84.44%) compared with cases where one was not (68.57%).

193
Table 7.35
Cross-tabulation of Metal Detector with Struck an Administrator
Metal Detector Struck an Administrator Total
No Yes
No 33 72 105
31.43% 68.57% 100.00%
82.50% 65.45% 70.00%
Yes 7 38 45
15.56% 84.44% 100.00%
17.50% 34.55% 30.00%
Total 40 110 150
26.67% 73.33% 100.00%
100.00% 100.00% 100.00%
The following analyses focused on whether a guard was present within the school and
whether this variable had significant associations with ACE External Factors. With regard to a
school guard/resource officer being present, significance was only indicated with respect to
having a criminal record, Fisher’s exact p = .012, and having struck an administrator, Fisher’s
exact p = .017. Significance was not found with respect to the associations between the presence
of a school guard/resource officer and K-12 suspension, Fisher’s exact p = .556; K-12 expulsion,
Fisher’s exact p = .614; K-12 failure, Fisher’s exact p = .287; K-12 dropout, Fisher’s exact p =
.144; being a street gang member, Fisher’s exact p = .288; a victim being gang affiliated, Fisher’s
exact p = .415; having struck a student, Fisher’s exact p = .147; a teacher, Fisher’s exact p =
.138; or someone else, Fisher’s exact p = .239.

194
Table 7.36 presents the results of the cross-tabulation associated with the first significant
finding, that between the presence of a guard in the school and the perpetrator having a criminal
record. These results indicate that the likelihood of the perpetrator having had a criminal record
was significantly increased in cases where a school guard was present (58.97%) compared with
cases where one was not present (35.71%).
Table 7.36
Cross-tabulation of School Guard/Resource Officer with Criminal Record
School Guard Criminal Record Total
No Yes
No 27 15 42
64.29% 35.71% 100.00%
45.76% 24.59% 35.00%
Yes 32 46 78
41.03% 58.97% 100.00%
54.24% 75.41% 65.00%
Total 59 61 120
49.17% 50.83% 100.00%
100.00% 100.00% 100.00%
Table 7.37 presents the results of the cross-tabulation conducted between the presence of
a guard within the school and the perpetrator having struck an administrator. These results show
a significantly lower likelihood of the perpetrator having struck an administrator was present in
cases where a school guard was present within the school (8.04%) as compared with cases in
which a guard was not present within the school (21.57%).

195
Table 7.37
Cross-tabulation of School Guard/Resource Officer with Struck an Administrator
School Guard Struck an Administrator Total
No Yes
No 40 11 51
78.43% 21.57% 100.00%
27.97% 55.00% 31.29%
Yes 103 9 112
91.96% 8.04% 100.00%
72.03% 45.00% 68.71%
Total 143 20 163
87.73% 12.27% 100.00%
100.00% 100.00% 100.00%
Finally, with regard to the presence of a school police officer, significance was indicated
in the associations with K-12 suspension, Fisher’s exact p = .048; being a street gang member,
Fisher’s exact p = .012; having a criminal record, Fisher’s exact p = .015; having struck a
teacher, Fisher’s exact p = .046; and an administrator, Fisher’s exact p = .041. Significance was
not found with respect to K-12 expulsion, Fisher’s exact p = .471; K-12 failure, Fisher’s exact p
= .414; K-12 dropout, Fisher’s exact p = .323; a victim being gang affiliated, Fisher’s exact p =
.333; having struck a student, Fisher’s exact p = .090; or someone else, Fisher’s exact p = .468
Table 7.38 presents the results of the cross-tabulation conducted with respect to this first
significant finding, that being the association between the presence of a police officer within the
school and a K-12 suspension. These descriptive statistics indicate that there was a significantly

196
higher likelihood of a K-12 suspension in cases where a police officer was present within the
school (44.44%) as compared with cases in which one was not (22.22%).
Table 7.38
Cross-tabulation of School Police Officer with K-12 Suspension
Police Officer K-12 Suspension Total
No Yes
No 21 6 27
77.78% 22.22% 100.00%
45.65% 23.08% 37.50%
Yes 25 20 45
55.56% 44.44% 100.00%
54.35% 76.92% 62.50%
Total 46 26 72
63.89% 36.11% 100.00%
100.00% 100.00% 100.00%
The results of the cross-tabulation conducted between the presence of a police officer
within the school and the perpetrator having been a gang member are presented in Table 7.39.
These results also indicate that risk increases in cases where a police officer was present within
the school. Specifically, the likelihood of the perpetrator being a gang member was 29.67% in
cases where a police officer was present within the school, with this percentage being reduced to
10.26% in cases where a police officer was not present.

197
Table 7.39
Cross-tabulation of School Police Officer with Street Gang Member
Police Officer Street Gang Member Total
No Yes
No 35 4 39
89.74% 10.26% 100.00%
35.35% 12.90% 30.00%
Yes 64 27 91
70.33% 29.67% 100.00%
64.65% 87.10% 70.00%
Total 99 31 130
76.15% 23.85% 100.00%
100.00% 100.00% 100.00%
Table 7.40 illustrates the results of the cross-tabulation conducted between the presence
of a police officer in the school and the perpetrator having a criminal record. These data show
that the likelihood of having a criminal record was higher in cases where a police officer was
present in the school, with this being 57.50%, and reduced to 34.21% in cases where a police
officer was not present in the school.

198
Table 7.40
Cross-tabulation of School Police Officer with Criminal Record
Police Officer Criminal Record Total
No Yes
No 25 13 38
65.79% 34.21% 100.00%
42.37% 22.03% 32.20%
Yes 34 46 80
42.50% 57.50% 100.00%
57.63% 77.97% 67.80%
Total 59 59 118
50.00% 50.00% 100.00%
100.00% 100.00% 100.00%
Next, Table 7.41 presents the results of the cross-tabulation between the presence of a
police officer in the school and the perpetrator having struck a teacher. These results show that
the likelihood of having struck a teacher was reduced in cases where a police officer was present
in the school (10.71%) as compared to cases where a police officer was not present (22.45%).

199
Table 7.41
Cross-tabulation of School Police Officer with Struck a Teacher
Police Officer Struck a
Teacher Total
No Yes
No 38 11 49
77.55% 22.45% 100.00%
27.54% 47.83% 30.43%
Yes 100 12 112
89.29% 10.71% 100.00%
72.46% 52.17% 69.57%
Total 138 23 161
85.71% 14.29% 100.00%
100.00% 100.00% 100.00%
Finally, the results of the cross-tabulation conducted between the presence of a police
officer in the school and the perpetrator having struck an administrator are presented in Table
7.42. These results show that the likelihood of a perpetrator having struck an administrator was
significantly reduced in cases where a police officer was present in the school (8.93%) as
compared with cases in which one was not (20.41%).

200
Table 7.42
Cross-tabulation of School Police Officer with Struck an Administrator
Police Officer Struck an Administrator Total
No Yes
No 39 10 49
79.59% 20.41% 100.00%
27.66% 50.00% 30.43%
Yes 102 10 112
91.07% 8.93% 100.00%
72.34% 50.00% 69.57%
Total 141 20 161
87.58% 12.42% 100.00%
100.00% 100.00% 100.00%
Summary
In sum, the association between ACEs, which focus primarily on child maltreatment and
peer victimization, the byproduct of ACEs, External Factors, and Environmental Sustainable
Design measures were examined in the TASSS dataset and the results presented in this chapter.
Specifically, these analyses examined the relationship between each ACE individually and all
ACE External Factors, followed by analyses examining the relationship between each ACE and
all Environmental Sustainable Design measures.
A large number of significant results were found in this set of analyses. Regarding the
ACE External Factors, negative associations were, in fact, found between domestic violence and
K-12 dropout, being a street gang member, having a criminal record, and having struck a student.

201
A positive association was found between a workplace shooting and having struck an
administrator, with a negative association found with having struck someone else. Next,
psychological issues were negatively associated with being a street gang member and a victim
being gang affiliated, with a positive association found with having struck a teacher or an
administrator. Having parents that were divorced or separated was positively associated with K12 failure, with social stratum having a non-linear association with K-12 failure, being a street
gang member, and a victim being gang affiliated. Next, significant family problems were
positively associated with having a criminal record and having struck someone else. Following
this, having experienced a recent death was positively associated with K-12 failure, having
struck a student, and negatively associated with having struck an administrator. Next, loss of
social standing was positively associated with having been dismissed from an organization and
having struck a student or an administrator. Peer aggression was positively associated with K-12
suspension, failure, being a street gang member, a victim being gang affiliated, and having struck
a teacher, a student, or an administrator.
The following Fisher’s exact tests examined Environmental Sustainable Design factors
alongside the ACE measures. These analyses found a positive association between the presence
of a metal detector and K-12 expulsion, dropout, and having struck a student or an administrator,
with the presence of a guard within the school positively associated with having a criminal
record and having struck an administrator. Finally, the presence of a school police officer was
positively associated with K-12 suspension, being a street gang member, having a criminal
record, and having struck a teacher or an administrator. The following chapter presents and
discusses the results of the difference in proportions tests conducted comparing the JRIC and
TASSS samples.

202
Chapter 8: Results: Comparing School Shooters in TASSS
with Other Perpetrators in JRIC
A series of difference in proportions tests were conducted comparing the JRIC and
TASSS samples on similar items. Difference in proportions tests were conducted between a
mental health diagnosis and mental health symptoms (JRIC) and psychological issues (TASSS).
Psychological issues were found to be significantly more common in TASSS (n = 163,
57.060%), as compared with both a mental health diagnosis among JRIC “other” perpetrators (n
= 238, 8.824%), z = 10.518, p < .0001, and having exhibited mental health symptoms (n = 250,
23.600%), z = 6.892, p < .0001. An additional analysis was conducted in which the mental health
diagnosis and mental health symptoms measures were combined; individuals were categorized as
having a diagnosis and/or symptoms, or neither. This analysis again found a significantly higher
proportion of mental illness in the TASSS sample; mental health diagnosis/symptoms (n = 250,
23.600%), z = 6.892, p < .0001.
The following analysis examined the difference in the proportion of those who
experienced a loss between the JRIC and TASSS datasets. In JRIC, this asks whether a loss,
defined as a death, was experienced by the perpetrator, while in TASSS, this is defined as the
recent death of a relative or friend. This likelihood was found to be significantly higher in
TASSS (n = 93, 23.660%) as compared with JRIC (n = 181, 8.840%), z = 3.361, p < .001.
The theme of loss was examined further, focusing on the loss of a job in the case of JRIC
and the loss of social standing in the case of TASSS. While not identical, these two measures
were believed to be similar enough to warrant a statistical comparison. This analysis again found
a significantly higher proportion in TASSS (n = 107, 37.380%) as compared with JRIC (n = 197,
19.289%), z = 3.449, p < .001.

203
The final analysis examined abuse. In the case of JRIC, data were available on whether
the perpetrator experienced physical or verbal abuse. In contrast, in the case of TASSS, data were
available on whether the perpetrator was a victim of peer aggression. In this final analysis, the
same pattern was identified in which the proportion was significantly greater in the TASSS
sample (n = 146, 52.740%) as compared with JRIC (n = 188, 14.362%), z = 7.513, p < .0001.
The following and final results chapter presents and discusses the results of the difference in
proportions tests conducted comparing the TASSS sample with data derived from the general
population.
These analyses consisted of a series of difference in proportions tests comparing the JRIC
data, focusing on perpetrators of other forms of violence only, and the TASSS data, with these
analyses examining the following pairs of measures of risk factors: mental health diagnosis and
mental health symptoms (JRIC) and psychological issues (TASSS); having experienced a loss,
defined as a death in JRIC, and the recent death of a relative or friend in TASSS; the loss of a job
in JRIC and the loss of social standing in TASSS; and having experienced physical or verbal
abuse in JRIC, and the perpetrator being a victim of peer aggression in TASSS. In all cases, a
significantly higher proportion of individuals was affected in TASSS as compared with the
perpetrators of other forms of violence in JRIC.

204
Chapter 9: Results: Comparing TASSS with the General Population
Additional analyses were conducted comparing the TASSS sample with relevant data
derived from the general population, as possible, focusing on the ACE measures identified in the
TASSS dataset and ACE Externalizing Factors. Metrics for some of these measures could be
found among the general population very quickly, while others could not be located. The
analyses presented here consist of difference in proportions tests testing whether a significant
difference in the proportions of these identified measures was present between TASSS and
identical or similar measures found in the general population or as representative a sample as
possible.
The first study, conducted by Crouch et al. (2019), included the metrics of parents having
been divorced, having experienced economic hardship, and having been exposed to violence in
the home or the neighborhood. This study used a nationally representative of 45,287 children in
the United States from the 2016 National Survey of Children’s Health (NSCH). The relevant
measures from TASSS consisted of parents having been divorced or separated, being in the low
social stratum, and having experienced significant family problems, respectively. Significance
was indicated with respect to all three tests conducted. Individuals were significantly more likely
to have had their parents get divorced or be separated in the TASSS sample (51.49%, Total n =
101) compared to national rates(21.90%), z = 7.175, p < .0001. Individuals were also
significantly more likely to be in the low social stratum in the TASSS dataset (43.48%, Total n =
115) compared with the national sample (22.50%), z = 5.377, p < .0001. Finally, individuals in
the TASSS dataset were also significantly more likely to have experienced significant family
problems (64.00%, Total n = 125), as compared with the national sample (14.80%), z = 15.411, p
< .0001.

205
A report by Sacks and Murphey (2018) also analyzed data from the 2016 National Survey
of Children’s Health (NSCH) including the percentage of individuals whose parents had been
divorced or separated, as well as the percentage that had experienced economic hardship
“somewhat often” or “very often,” which was analyzed alongside the percentage of individuals
in the low social stratum in TASSS. Both analyses were found to achieve statistical significance.
The percentage of those whose parents were divorced or separated was significantly higher in
TASSS (51.49%, Total n = 101) as compared with the national sample (25%, Total n = 50,212), z
= 6.138, p < .0001. Additionally, the percentage of those in the low social stratum was also
significantly higher among TASSS (43.48%, Total n = 115) as compared with the national
sample (25%, Total n = 50,212), z = 4.569, p < .0001.
The second relevant study was that by Merrick et al. (2018) using a nationally
representative telephone survey of 214,157 adults from 23 U.S. states during 2011 to 2014. The
relevant data from this study consisted of parents having been divorced or separated, with the
analysis conducted finding a significantly higher proportion of individuals with parents that were
divorced or separated among TASSS (51.49%, Total n = 101) as compared with the national
sample (27.63%), z = 5.361, p < .0001.
The following set of analyses examined ACE Externalizing Factors. As before, efforts
were made to locate studies containing comparable data as was in TASSS, which either randomly
sampled from the general population or used a sample that was as close as possible to a
representative sample. The first study, by Fabes et al. (2021), examined K-12 expulsion rates
using the U.S. Department of Education’s 2017-2018 Civil Rights Data Collection (CRDC) with
data from 28,222 reporting public schools. Expulsion rates were found to be significantly higher

206
in TASSS (17.78%, Total n = 135), as compared with the national sample (2.97%), z = 9.998, p <
.0001.
An earlier study by the National Center for Education Statistics (2017) also examined K12 expulsions and suspensions using the 2013-2014 Civil Rights Data Collection (CRDC). These
difference in proportions tests conducted found a significantly higher proportion of suspensions
among TASSS (33.33%, Total n = 123) as compared with the study data (5.43%, Total n =
50,044,522), z = 13.655, p < .0001. The same pattern was found with regard to expulsions, with a
significantly higher proportion found among TASSS (17.78%, Total n = 135), as compared with
the study data (0.22%, Total n = 50,044,522), z = 43.542, p < .0001.
Next, the National Center for Education Statistics (2023) examined the K-12 dropout rate
using U.S. Census data for 2021. A significantly higher proportion of students dropping out was
found in TASSS (15.29%, Total n = 157) as compared with the study data (5.20%, Total n =
49,433,092), z = 5.694, p < .0001. Finally, Pyrooz and Sweeten (2015) used the National
Longitudinal Survey of Youth 1997 to examine gang membership among individuals between the
ages of five and 17. A significantly higher proportion of gang members are found in the TASSS
dataset (26.88%, Total n = 253) as compared with the national sample (2.0%, Total n = 7,335), z
= 23.465, p < .0001.
These analyses conducted consisted of difference in proportions tests comparing some of
the TASSS measures with similar data identified in the general population. The measures
examined consisted of the following: parents having been divorced (TASSS: parents having been
divorced or separated), having experienced economic hardship (TASSS: being in the low social
stratum), and having been exposed to violence in the home or the neighborhood (TASSS: having
experienced significant family problems), those who had experienced economic hardship

207
“somewhat often” or “very often” (TASSS: individuals in the low social stratum). Additional
analyses were conducted examining ACE Externalizing Factors. These analyses examined K-12
expulsion rates, suspension rates, and dropout rates. For every factor, prevalence was found to be
significantly greater in the TASSS data as compared with the general population.
The following chapter will discuss this study’s results, compare and contrast what was
found in this study with the findings of previous literature, and also discuss how this study’s
results speak to theory relevant to the area. Recommendations will be made, focusing on how
this study’s findings could be most effectively applied in order to reduce the incidence of school
violence, as well as to reduce its severity. The limitations of this study are presented, along with
possibilities for future research. Finally, conclusions are made based on this study’s findings.

208
Chapter 10: Discussion and Conclusion
Introduction
In the past few decades, mass shootings in the United States have not only become more
frequent but with a greater number of deaths found to have occurred from such incidents
(Peterson, 2021). Of the 167 mass shootings that have taken place in the last 50 years in the
United States, 20% have occurred in the last five years. Additionally, more than half of these 167
mass shootings took place since 2000, and one-third of these occurred since 2010. The death toll
per shooting has seen a concomitant increase, with this being eight per year in the 1970s and
increasing over the decades to 51 per year presently (Peterson, 2021).
Mass shootings have also been found to take place within schools, colleges, and
universities frequently. Of the 167 mass shootings examined by Peterson (2021) that have
occurred in the past 50 years, 7.6% of these occurred in kindergarten through 12th grade
elementary, middle, and high schools, and 5.3% of these mass shootings took place in colleges or
universities (Peterson, 2021). While researchers have not found success in creating a distinct
profile of school shooters, commonalities and factors among this set of individuals have been
indicated. Mass shooters, as a group, have been consistently associated with a particular set of
risk factors, including severe childhood trauma, being in crisis, mental health disorders
(especially psychosis), and leaking plans in advance (Peterson, 2021).
Discussion
JRIC: School Versus Other Perpetrators
Various analyses were conducted using the JRIC and TASSS datasets and data derived
from the general population. For the JRIC dataset, bivariate tests were used to compare
perpetrators of school violence with perpetrators of other forms of violence. These analyses also

209
consisted of linear regressions conducted to examine the presence of significant trends over time.
The bivariate analyses found significant differences between perpetrators of school violence and
perpetrators of other forms of violence in several cases, specifically with regard to the measures
of age, contact with law enforcement, emotional distress, suspicious travel, violence to others,
drug abuse, having been radicalized by an associate, having had a personal grievance, extremist
media consumption, social media platform used and social media activities, and level of
education. No significant trends over time were detected.
Specifically, other perpetrators were significantly older than perpetrators of school
violence. Also, compared with other perpetrators, school violence perpetrators were more likely
to have had contact with law enforcement for a non-violent felony or violent crime. They were
also more likely to have had no contact with law enforcement. Perpetrators of school violence
were more likely to indicate emotional distress, were less likely to have engaged in suspicious
travel, and were also less likely to have exhibited violence toward others. School violence
perpetrators were also more likely to have a drug abuse status that was unknown or a negative
instance, less likely to have been radicalized by an associate, though were more likely to have
had a personal grievance. Perpetrators of school violence were also more likely to have exhibited
extremist media consumption and to have a high school diploma or GED.
Previous literature has also found distinctions between perpetrators of school violence
and perpetrators of other forms of violence. While some have stated that the main difference
between these two groups of individuals is the location of their attacks (Lankford, 2013), others
have mentioned how suicide terrorists tend to be mentally healthy and are motivated by their
ideology, while school shooters tend to be mentally troubled and are more motivated by their
own personal problems (Carey, 2007). It has been found that school shooters are more likely to

210
be struggling with school problems, and while they choose their school as a target, they rarely
target specific individuals, even when they feel wronged by certain persons (Lankford, 2013).
This present study explored in extensive detail the differences and distinctions between
perpetrators of school violence and perpetrators of other forms of violence. It has been clearly
shown that these two groups of individuals were not very similar but instead can be categorized
on a vast number of factors on which they differ significantly and substantially.
JRIC: Adverse Childhood Experiences
Further analyses were conducted focusing on Adverse Childhood Experiences (ACEs),
External Factors, and Environmental Sustainable Design measures, both with regard to the JRIC
dataset as well as the TASSS dataset. With regard to the JRIC data, analyses were conducted to
examine the association between the ACE measures and risk factors included within the dataset
alongside External Factors. All analyses conducted consisted of Fisher’s exact tests, with
significant associations found in several cases. Specifically, a significantly greater likelihood of
contact with law enforcement was found among those exhibiting mental health symptoms, those
who experienced a loss through a death, the loss of a job, those who have abused alcohol, those
having an economic grievance, an anarchist ideology, or a white supremacist ideology. Violence
to the self was significantly and positively associated with mental health symptoms, a loss due to
a death, the loss of a job, having experienced physical or verbal abuse, financial difficulties,
relationship difficulties, having an economic grievance, an anarchist ideology, or a white
supremacist ideology. Next, violence to others was significantly more likely among those
experiencing emotional distress, having a mental health diagnosis, exhibiting mental health
symptoms, having experienced physical or verbal abuse, having relationship difficulties, having
been radicalized by a friend or associate, having experienced war, having a racial grievance, a

211
religious ideology, a white supremacist ideology, access to weapons, or military experience, with
a lower likelihood of violence found among those with social grievances or those having an
anarchist ideology. Violence from others was found to be significantly more likely among those
with emotional distress, having mental health symptoms, experiencing a loss due to a death, the
loss of a job, drug abuse, having experienced physical or verbal abuse, emotional distress,
financial difficulties, having an economic grievance, a racial grievance, or an ideology of white
supremacy. Finally, vagrancy was significantly associated with emotional distress, a mental
health diagnosis, exhibiting mental health symptoms, the loss of a job, the loss of a residence,
alcohol abuse, drug abuse, financial difficulties, having an economic grievance, a personal
grievance, a racial grievance, or a sovereign citizen ideology.
TASSS: Adverse Childhood Experiences
The association between ACEs, which focus primarily on child maltreatment and peer
victimization, the byproduct of ACEs, External Factors, and Environmental Sustainable Design
measures were then examined with regard to the TASSS dataset. Regarding these analyses, the
ACEs identified consisted of the following: domestic violence, a workplace shooting,
psychological issues, parents having been divorced or separated, social stratum, significant
family problems, the recent death of a relative or friend, the loss of social standing, having been
dismissed from a social, political, or religious organization, and peer aggression. In addition, the
External Factors focused on consisted of a K-12 suspension or expulsion (coded as two separate
measures), a K-12 failure, a K-12 dropout, having been a member of a street gang, having had a
criminal record, a victim being affiliated with a gang, and having struck a student, teacher,
administrator, or someone else, with these latter measures consisting of four separate measures.
The Environmental Sustainable Design measures consisted of the presence of a metal detector, a

212
school guard, and a school police officer, with all analyses conducted examining the associations
between these sets of measures consisting of Fisher’s exact tests. Specifically, these analyses
examined the relationship between each ACE individually and all ACE External Factors,
followed by analyses examining the relationship between each ACE and all Environmental
Sustainable Design measures.
A large number of significant results were found in this set of analyses. Regarding the
ACE External Factors, negative associations were, in fact, found between domestic violence and
K-12 dropout, being a street gang member, having a criminal record, and having struck a student.
A positive association was found between a workplace shooting and having struck an
administrator, with a negative association found with having struck someone else. Next,
psychological issues were negatively associated with being a street gang member and a victim
being gang affiliated, with a positive association found with having struck a teacher or an
administrator. Having parents that were divorced or separated was positively associated with K12 failure, with social stratum having a non-linear association with K-12 failure, being a street
gang member, and a victim being gang affiliated. Next, significant family problems were
positively associated with having a criminal record and having struck someone else. Following
this, having experienced a recent death was positively associated with K-12 failure, having
struck a student, and negatively associated with having struck an administrator. Next, loss of
social standing was positively associated with having been dismissed from an organization and
having struck a student or an administrator. Peer aggression was positively associated with K-12
suspension, failure, being a street gang member, a victim being gang affiliated, and having struck
a teacher, a student, or an administrator.

213
The following Fisher’s exact tests examined Environmental Sustainable Design factors
alongside the ACE measures. These analyses found a positive association between the presence
of a metal detector and K-12 expulsion, dropout, and having struck a student or an administrator,
with the presence of a guard within the school positively associated with having a criminal
record and having struck an administrator. Finally, the presence of a school police officer was
positively associated with K-12 suspension, being a street gang member, having a criminal
record, and having struck a teacher or an administrator.
Difference in Proportions Tests
A series of difference in proportions tests were also conducted between the JRIC,
focusing on perpetrators of other forms of violence only, and the TASSS datasets, with these
analyses examining the following pairs of measures of risk factors: mental health diagnosis and
mental health symptoms (JRIC) and psychological issues (TASSS); having experienced a loss,
defined as a death in JRIC, and the recent death of a relative or friend in TASSS; the loss of a job
in JRIC and the loss of social standing in TASSS; and having experienced physical or verbal
abuse in JRIC, and the perpetrator being a victim of peer aggression in TASSS. In all four cases,
a significantly higher proportion of individuals was affected in TASSS as compared with the
perpetrators of other forms of violence in JRIC.
TASSS Versus the General Population
The final set of analyses conducted consisted of difference in proportions tests comparing
some of the TASSS measures with similar data identified in the general population, with efforts
made to find comparable data in the general population that was sampled randomly or with their
methods selecting a sample as close as possible to that of a random sample. The measures
examined consisted of the following: parents having been divorced (TASSS: parents having been

214
divorced or separated), having experienced economic hardship (TASSS: being in the low social
stratum), and having been exposed to violence in the home or the neighborhood (TASSS: having
experienced significant family problems), those who had experienced economic hardship
“somewhat often” or “very often” (TASSS: individuals in the low social stratum). Additional
analyses were conducted examining ACE Externalizing Factors. These analyses examined K-12
expulsion rates, suspension rates, and dropout rates. For every factor, prevalence was greater
among the TASSS population than among the general population.
Hypothesis Tests
Regarding the comparative analyses conducted with the JRIC dataset, to summarize the
specific differences found between school and other perpetrators of violence, the results
indicated that as compared with other perpetrators of violence, perpetrators of school violence
were more likely to have had some contact with law enforcement specifically due to a nonviolent felony or a violent crime, and are also more likely to have had no contact with law
enforcement (Hypothesis 3). Additionally, those who had perpetrated school violence were more
likely to have indicated emotional distress (Hypothesis 11), were less likely to have engaged in
suspicious travel (Hypothesis 5), and were also less likely to have committed violence toward
others as compared with other perpetrators (Hypothesis 6). Perpetrators of school violence were
also more likely to have had a drug abuse status that was unknown, or to have a negative
instance (Hypothesis 7), and were less likely to have been radicalized by an associate
(Hypothesis 8), and more likely to have had a personal grievance as compared with other
perpetrators (Hypothesis 9). Additionally, school violence perpetrators were also more likely to
have exhibited extremist media consumption (Hypothesis 10), and to have a high school diploma
or GED as their highest level of education (Hypothesis 2).

215
When examining these results alongside this study’s research questions and hypotheses, it
was found that null Hypotheses two, three, and five through 11 were rejected on the basis of this
study’s analyses, with the remaining null hypotheses, one and four, having failed to be rejected.
Additionally, null Hypotheses 12 through 15 were also rejected, as the results found Adverse
Childhood Experiences (ACEs) to be associated with External Factors (Hypothesis 12), and
Environmental Sustainable Design measures (Hypothesis 13), with significant differences also
indicated between the JRIC and TASSS samples (Hypothesis 14) and between the TASSS sample
and the general population (Hypothesis 15).
Previous Literature and Theory
The results found in this present study generally support the findings of previous
literature and what would be expected based on the theories cited in this dissertation. Similar to
many previous studies, this study took a prevention-oriented approach to the study of school
shooters, incorporating the public health model in its conceptual and theoretical framework.
Prevention-Oriented Approach
Within the context of targeted school violence, a public health approach focuses on the
prevention of elevated threats of domestic violence (National Security Council, 2021) by
nurturing healthy individuals in a community while simultaneously seeking to maintain the civil
rights, privacy rights, and civil liberties of those in the community (Brown, 2022). In doing so, it
aims to focus on the health and well-being of the communities in which individuals who may
come to perpetrate an act of school violence live, as well as the health and well-being of
perpetrators themselves. Additionally, the public health approach also aims to understand the
policies and circumstances that might lead to an attack instead of focusing solely on the
perpetrators or potential perpetrators.

216
Previous literature has found that a prevention-oriented approach toward reducing
targeted school violence is not only efficacious but is also vital in combating this problem
(Alathari et al., 2019). As approximately half of school shootings take place in less than 15
minutes, the time required for law enforcement to appropriately respond to such situations is
simply not available. The majority of school shootings are not stopped by law enforcement, and
with this being the case, a prevention-oriented approach becomes necessary (Bonanno &
Levenson, 2014). Not only is it the case that law enforcement is typically not able to respond to
school shootings until after the incident is over, but school shootings also continue to increase in
frequency as well as with regard to the number of deaths incurred in each shooting despite the
implementation of different physical security measures within schools themselves.
Recommendations have been made for the implementation of comprehensive targeted prevention
programs within schools in order to prevent acts of school violence through the identification of
potential perpetrators of violence, the determination of the level of risk, and the implementation
of intervention strategies such that that risk can be minimized (Alathari et al., 2019). It has also
been suggested that a low threshold for intervention be implemented so that schools can
determine which students are at risk before their behavior can negatively impact the safety of
other students (Alathari et al., 2019).
Public Health Model
The public health model incorporates four phases or tiers, these consisting of the
following: Primordial (or Upstream) prevention, Primary prevention, Secondary prevention (or
intervention) and Tertiary prevention (or treatment). The initial phase or tier of the public health
model, that of primordial or upstream prevention, focuses upon the baseline policies of the
community as well as the conditioning of its well-being, as these factors impact the health of all

217
individuals in the community and foster the creation of a healthy and functioning community.
Through this aim of achieving a community that is both healthy and functioning, intervention in
potential future school shootings is possible at this stage. This initial tier, the Primordial
Prevention tier, can be viewed as reducing the risk of targeted school violence, such as school
shootings, through the promotion of relevant protective factors, which include policies for
reducing community-level drug and alcohol use, reducing adverse childhood experiences
(ACEs), as well as by promoting policies that address mental health issues and bullying.
Previous research has found that over one-third of school shooters have a history of
substance use or abuse, while negative home life factors are also common among school shooters
(Alathari et al., 2019). These results correspond strongly with those found in this present study,
both with regard to substance use/abuse being common among perpetrators of school violence,
as well as the preponderance of negative home life factors. It would be suggested that a
community that evidenced a low frequency of substance or alcohol abuse would be less likely to
exhibit targeted violence. As protective measures against drug and alcohol use, SAMSA has
suggested faith-based resources as well as, more generally, after-school activities as communitybased protective measures (Brown, 2022).
Focusing further on adverse childhood experiences (ACEs), these experiences have been
found to be an important precursor of substance and alcohol abuse. This current study found
numerous significant and positive associations between Adverse Childhood Experiences (ACEs),
External Factors, and Environmental Sustainable Design measures, suggesting the importance of
ACEs and the expected benefit in reducing violence if these ACEs can be reduced. In this vein, a
study by Campbell et al. (2016) indicated associations between behavioral symptoms common
among unhealthy communities, specific medical complaints, and ACEs. More specifically,

218
specific forms of abuse or negative experiences were found to be linked with certain adverse
outcomes; behavioral symptoms associated with ACEs included heavy drinking, smoking, and
risky behavior. Particularly strong associations were identified between verbal abuse and heavy
drinking, as well as between sexual or verbal abuse and smoking as well as risky behavior. In
addition, medical conditions that were found to be associated with ACEs consisted of diabetes,
heart problems, strokes, and depression, with sexual and verbal abuse also found to be associated
with diabetes and depression. At the primordial level, a public health approach should
incorporate policies aimed at the reduction of the prevalence of ACEs, which should thereby
reduce the prevalence of negative community behaviors, as well as medical conditions that have
been found to be associated with targeted violence.
In addition, the role of mental illness has been found to be important within the context of
targeted violence, with this being an additional primordial factor that could be managed in the
attempt to reduce the risk of targeted violence, including the risk of school shootings. Mental
illness has been found to be very common among school shooters, with over two-thirds of school
shooters having evidenced a mental health condition in the period leading up to their attack
(Alathari et al., 2021). Furthermore, 78% of perpetrators have either attempted suicide or
exhibited suicidal ideation before their attack (Ioannou et al., 2015). Within the realm of mental
illness, school shooters commonly suffer from depression (71%), with the majority having a
psychiatric history (57%) (Ioannou et al., 2015), 31% previously exhibiting violent behavior, and
27% having been arrested previously (Vossekuil et al., 2002). Furthermore, a study of 29 school
shooters in the US who committed their shooting between 1995 and 2015 found a total of 10 of
the shooters to have shown signs of psychosis or psychopathy, and eight of these shooters having
experienced hallucinations or delusions (Farr, 2018). The results of this present study support

219
those found in previous research, both with regard to mental illness, either with respect to signs
or symptoms, being common with regard to perpetrators of school violence, as well as violence
being common more generally among these individuals.
With regard to criminality, it has been suggested that school shootings are not typically a
portion of a larger body of more general criminal activity, with one study suggesting that only
7.5% of cases fit this criterion (Ioannou et al., 2015). These conditions, which may lead an
individual to be more likely to perform an act of school violence, both with regard to mental
illness and criminality, may have, as their root, a combination of both genetic predispositions as
well as severe life stressors. While genetic predispositions cannot be accounted for, stressors that
may contribute to mental illness most commonly pertain to those in the domain of the family, as
well as social issues and academic problems. Furthermore, bullying is also a relevant factor at
the primordial level, with close to half of school shooters indicating that they were bullied by
their classmates (Alathari et al., 2021). Others have found that most perpetrators experience
bullying and marginalization (Ioannou et al., 2015). Similar to leakage, it is not the case that
most bullying occurs without the knowledge of an adult. In fact, even school officials are
frequently aware of bullying that a future school shooter has experienced prior to their act of
violence. However, school officials are commonly unable or unwilling to respond to bullying
effectively (Alathari et al., 2021).
While the focus of the primordial tier is on a community’s general health, the next tier,
the primary prevention tier, focuses on policies and strategies that aim to reduce the prevalence
of some specific health issue. With regard to school shootings, the primary prevention tier
focuses on reducing the risk factors that can lead to psychological issues, which may make it
more likely for a school shooting to take place. Challenges abound with respect to the creation

220
and maintenance of a society that is resilient to the threat of school shootings. In fact, efforts to
reduce such events may increase the drivers that lead to their occurrence (Joint Regional
Intelligence Center, 2022). Attempts to reduce violent extremism could cause members of the
community to feel singled out, for example, motivating them to commit an act of violence, while
even discussions on the importance of changing gun ownership laws may lead some individuals
to commit violent acts in opposition.
Violent risk factors have been delineated by the CDC at the individual, family, social, and
community levels (CDC, 2020). At the individual level, identified risk factors for violent acts
include violent victimization, ADHD, signs of aggressive behavior during early development, a
history of drug and alcohol abuse, poor behavioral control, cognitive and IQ deficits, and the
demonstration of anti-social beliefs. Family-level risk factors include factors relating to
parenting, including overly strict or particularly lax parenting, disinterested parents, low parental
income and education, parental criminality, and parental substance use. Community-level risk
factors include poor economic opportunities, minimal social interaction, population transiency,
family disruption, and disorganized neighborhoods (CDC, 2020). It is thought that social
interactions within the community reduce the risk of school shootings through the provision of
an open environment where individuals feel connected to one another, reducing the risk of severe
exclusion among individuals, the latter having been identified in many cases of school shootings.
Additionally, by cultivating and encouraging healthy social interactions within the community,
individuals would be more likely to speak with someone about concerns they may have
regarding any specific individual (Alathari et al., 2021). Reducing violent acts may also be
achieved through the building of partnerships across various societal levels and sectors (National

221
Security Council, 2021); this highlights the relevance of social interactions at the primary and
secondary prevention tiers within the public health model as they relate to school shootings.
In the secondary prevention tier of the public health model, attempts are made to prevent
an injury or a disease from occurring after exposure to risk factors has already occurred. In the
context of school shootings, this secondary prevention or intervention tier would be relevant after
an individual has already been found to have been exposed to risk factors associated with
perpetrators of school shootings. In addition to all of the risk factors already discussed, this also
includes that of awareness within the public of a person’s grievance, awareness within the public
of a person’s ideology, leakage, and stockpiling weapons (Weine et al., 2017; see also Alathari et
al., 2019). Additional factors consist of having a history with law enforcement, as well as, more
generally, a history of disciplinary actions and an interest in violence (Alathari et al., 2019).
Contact with law enforcement has been identified as one of the top 10 risk factors among
perpetrators of domestic violent extremism, with approximately 30% of cases having a history of
law enforcement contact (Joint Regional Intelligence Center, 2022). In addition, close to onethird of school shooters had some interaction with law enforcement before their attack (Alathari
et al., 2021).
The concept of aggrieved entitlement is also important within the context of the primary
prevention tier of the public health model (del Carmen et al., 2022). With respect to the
perpetrators of school shootings, this generally takes the form of a relational issue, more
specifically, typically consisting of a grievance between the male perpetrator and a female (del
Carmen et al., 2022). In this case, attempts to stop a potential school shooting before it happens
may require asking questions to the right individuals in the school community about the previous
relationships an at-risk individual may have had.

222
Finally, regardless of the method or methods used to identify an at-risk individual, once
the individual has been identified, an intervention must take place before a violent act is
committed. Threat management involves de-escalating the potential threat by speaking with the
individual in order to convince them that an act of violence would not resolve any grievance they
might have (Joint Regional Intelligence Center, 2022).
The last tier in the public health model, the tertiary prevention or treatment tier, aims to
resolve a public health issue after it has already taken hold in an individual. Within this tier, there
is an aim to minimize any complications of the health issue or the chances of a relapse. Within
the context of school shootings, this final tier pertains to working with individuals who have
already perpetrated an act of violence so that their actions do not escalate into an act of mass
violence, such as a school shooting. This tier can also aim to help individuals not re-offend if
they have already performed an act of violence (Brown, 2022).
Other Theory
As stated earlier in this study, a particular problem facing research into school shootings
is the lack of well-developed theory, with research generally having been more descriptive and
practical as compared with theory-based. Relevant theories include psychological theories which
aim to understand cognitive, emotional developmental, and clinical factors, as these all pertain to
school shooters; social learning theory, the frustration-aggression hypothesis, as well as other
theories of regression are also relevant (Grøndahl & Bjørkly, 2016). Despite the importance of
these relevant theories, no single model has been developed that can explain which individuals
will perpetrate school shootings and which will not. However, explanatory models have been
developed that inform researchers and others as to what might cause an individual to perpetrate a
school shooting (Grøndahl & Bjørkly, 2016).

223
The results of the analyses conducted in the present study cannot specifically speak to
any particular model, but the results found are suggestive of the veracity of the models as
proposed by, e.g., Levin and Madfis (2009) and Bonanno and Levenson (2014). The Levin and
Madfis (2009) model incorporates five steps: a period of prolonged stress, strain without social
control, followed by a period of acute stress, planning the atrocity, and finally, the commission of
the violent act. In line with the risk factors stated earlier, future school shooters commonly begin
in a period of prolonged stress or chronic strain, which may be the result of bullying or
difficulties in interpersonal relationships (Bonanno & Levenson, 2014). The second stage of
strain without social control may result from the presence of few or no social connections or
meaningful relationships which would otherwise help these individuals cope with their stress. In
the following stage of acute stress, the individual experiences a catastrophic loss. As in this
study’s findings, the experience of a significant loss, whether it was a death or another type of
loss, was found to be common among perpetrators of school violence. Within the same model,
the individual then proceeds to the planning phase at least two days prior to the attack. In the
fifth and final stage, the individual carries out the attack, which requires access to and the ability
to adequately use firearms (Alathari et al., 2019; Bonanno & Levenson, 2014). Similarly, with
regard to all risk factors discussed, while many individuals experience the first three stages in
this model, some even going on to plan an attack, very few will actually carry out an attack
(Bonanno & Levenson, 2014). While many researchers cite perpetrators as being lonely,
alienated, and victims of bullying, the fact remains that many students have the same
characteristics or experience these same situations or events without ever performing any
significant act of violence (Grøndahl & Bjørkly, 2016).

224
Leakage
While leakage, in and of itself, does not generally correspond with future targeted
violence (Meloy et al., 2014), it remains the most common method for the identification of a
potential perpetrator of a future school shooting and can consist of something as simple as an
inappropriate or violent statement (FBI, 2022). It was reported by the JRIC (2022) that leakage
is the most common risk factor among violent perpetrators following social media use and is
present in nearly all cases of violence (Meloy et al., 2014). Specifically, studies of attacks have
found leakage to a peer to be present in 93% of cases, while in adult reports, concern about the
perpetrator before the attack is carried out is present in 88% of cases (Wetterneck et al., 2005).
Many cases of leakage occur offline in simple conversations with peers, though in many cases,
the peer assumes that the perpetrator was joking and does not report the incident (Abel et al.,
2022).
Researchers have suggested that students should have a way to report, for example,
threatening or disturbing social media posts to authorities (Dowdell et al., 2022). While
appropriate, as previous findings indicate that the majority of the cases of leakage occur offline,
every school should have a safe and confidential method of reporting such leakage to school
administrators expediently. While most cases of leakage do not correspond with a specific act of
future school violence, the investigation of every case of leakage should result in a vast reduction
in the incidence of school violence, including school shootings, so long as the appropriate
interventions are taken with the individual in question if they are deemed to be at risk based on
the recommendations provided by the public health model as applied to this particular problem.

225
Limitations
A number of limitations were present within the study. With regard to the data analyzed,
the JRIC dataset was not a random sample; instead, it consisted of a convenience sample. This
limits external validity and generalizability due to the fact that the results obtained in the
analyses conducted for this study cannot be generalized to a larger population. This is stated
while noting that the particular sample of interest in this study consisted of perpetrators of school
violence and, in particular, perpetrators of school shootings, along with perpetrators of other
forms of violence. For this reason, it is unlikely that any future study will incorporate random
sampling when analyzing such a specific and small population. Additionally, these data also only
incorporated perpetrators of violence or acts of violence. This makes comparisons difficult with
the general population, or cases in which individuals planned but did not carry out an act of
violence, which increases the difficulty of determining why some at-risk youth or young adults
eventually commit a violent act, while others do not.
A second limitation of the study consisted of the small sample size associated with the
JRIC dataset. The analysis of a dataset with a smaller sample size serves to reduce the statistical
power associated with any analysis conducted, all other things being equal, which serves to
increase the difficulty in finding significance with regard to any test conducted. However, the
same issue exists as pertains to the non-random sampling method used in the case of these two
datasets, that being that the focus of interest in this study consisted of a population that is very
small to begin with, and for this reason, any concomitant dataset is likely to incorporate a small
sample size.
Another notable result of the use of datasets with smaller sample sizes consists of reliance
on bivariate analysis; due to the smaller sample sizes present within these data, the application of

226
multivariate statistics, which generally require a larger sample size in order to achieve the same
level of power, was not deemed appropriate, due to the same issue of statistical power.
Additional limitations pertain to the datasets themselves. The JRIC dataset proved a
challenge as the data were derived from various studies, none of which conform to any other
with respect to the codebook, variable coding, variables in question, or anything else. This
required careful management in merging the data from all these various studies into a single
cohesive dataset, which could then be used for analysis after extensive cleaning. Additionally, the
TASSS dataset suffered from a very terse and non-descriptive codebook, which raised many
questions as to what precisely each variable was measuring. This was determined in further detail
through the examination of the final report published by the authors of the TASSS study, which
provides some additional description as to the nature of the variables themselves.
A final limitation that will be noted here consists of the fact that this study only
incorporated quantitative or categorical variables. While a qualitative or mixed-method study
would be interesting, challenges abound, in particular with regard to how important qualitative
data could be collected pertaining to school shooters, many of whom do not survive their attack,
either through suicide or by being killed by law enforcement during the attack. However sparse
the population may be, a qualitative examination of living school shooters may provide further
insight into the reasons behind their progression from more or less ordinary students to
perpetrators of an act of mass violence on their fellow students.
Future Research
The limitations discussed in the previous section point to numerous recommendations for
researchers in this area and possibilities for future research. Regarding the sampling method, it is
debatable whether a random sample could truly be obtained within this research, with the

227
exception of comparisons made with samples drawn from the general population, as was
conducted in this study. Any random sampling method focusing on perpetrators of violence,
whether that be school violence or other forms of violence, would prove most challenging.
However, studies conducted on the general population may be insightful as to why some
individuals exposed to the same risk factors proceed in a certain path while others, the vast
majority of individuals, do not. This could be achieved through a study incorporating a random
sampling method and would allow for obtained results to have greater external validity and
generalizability than studies conducted solely on perpetrators of violence. Another challenge here
may consist of sampling both perpetrators and non-perpetrators in the same study, particularly
due to the low proportion of perpetrators of violence in the general population.
Additionally, future studies could incorporate larger sample sizes, hence achieving
greater statistical power, and thereby being more likely to find significant results in the analyses
conducted. Despite this, it should be noted that a vast number of significant findings were still
obtained in this present study despite the sample size issues mentioned previously. It may be the
case that the greatest benefit of future studies incorporating larger sample sizes would be the
inclusion of larger, multivariate models in order to study the data obtained. This would provide
numerous advantages, including the ability to examine the effect of any specific predictor while
holding all other predictors in the model constant and reducing the risk of biases, such as omitted
variable bias.
Future research would also benefit greatly by collecting well-documented novel data,
with a clear and detailed codebook available for analysis. This was, in fact, a problem with
regard to both JRIC and TASSS; in both cases, codebooks or reports substituting for codebooks
were generally found to have very brief and commonly vague descriptions of the variables in

228
question. In scientific research, each variable should be clearly operationalized with a detailed
description available, leaving no possibility for doubt as to what the variable is measuring and
exactly how it is coded or recorded.
Finally, as discussed in the previous section, nearly all the research in this area appears to
be entirely quantitative based on the literature reviews conducted for this study. Similarly, this
study itself was also purely quantitative in nature. Research taking the form of interviews with
surviving perpetrators of school violence, and in particular, school shootings, may provide
valuable insight as to the reasons why these specific individuals commit such a severe act of
violence while the vast majority of others exposed to the same risk factors do not.
Recommendations
The results of the analyses conducted for this study found numerous significant
differences between perpetrators of school violence and perpetrators of other forms of violence,
significant differences between individuals in the JRIC and TASSS datasets, significant
associations between Adverse Childhood Experiences (ACEs), External Factors, and
Environmental Sustainable Design measures in both datasets, and with individuals in these
datasets being significantly more likely to suffer from these problems as compared with the
general population.
Overall, these results would suggest the appropriateness of a two-pronged approach to
reducing school violence. A very large set of risk factors were identified in this study as to
individuals who perpetrate acts of school violence. These results could be used to create a
formula for a risk score that could be determined within schools with regard to all enrolled
students. Students with a risk score above a certain threshold could then be screened further and
treated as per the recommendations provided by the public health model as applied to this

229
specific problem in order to minimize any risk that may be present with regard to their future
inclinations for committing an act of school, or other, violence. However, the results obtained,
along with the literature reviews conducted, also reveal the disturbing fact that individuals who
go on to perform an act of school violence, including school shootings specifically, are exposed
largely to the same risk factors that the majority of students and schools in the United States are
also exposed to. The results do not, therefore, suggest that the screening of students and the use
of a risk factor or similar screening method would not be beneficial in determining those
individuals who are more likely than the vast majority to go on to commit an act of school
violence; however, it does suggest that it is essentially all students that are at some level of risk
of committing school violence, simply due to the fact that it is essentially all students that are
exposed to one or more of the relevant risk factors identified in this and other studies as being
potential precursors of committing targeted violence.
This would, therefore, suggest the implementation of screening based on risk factors and
programs or procedures that are applied to all students regardless of their potential risk. This
could include the various community-building factors discussed under the public health model,
as these implementations should serve to decrease violence on the community level, as well as
any program that would reduce any of the risk factors identified either in this or previous studies,
as this should also serve to reduce the incidence of school violence. This pertains, specifically, to
the reduction of mental illness, bullying in schools, and programs or counseling that would help
students work through any interpersonal relationship problems or difficulties, and so forth, with
regard to all other risk factors mentioned. Finally, interventions, specifically for “high-risk”
individuals or, more generally, with regard to the student population, should be present at each of

230
the four tiers associated with the public health model. All interventions should also be adjusted
based on any novel results present in future research.
Researchers have also provided more specific recommendations along these lines. In
order to reduce violence more generally, Alathari et al. (2023) suggests the implementation of
community bystander reporting, concern over individuals exhibiting an unusual interest in
violent topics, workplace violence prevention plans in businesses with interventions for those
who may present a risk of violence, careful consideration of strategies for resolving interpersonal
grievances, immediate intervention in cases where an individual is sharing final communications
or committing final acts, and communities helping individuals who are managing stressful life
circumstances or mental health problems or crises. Alathari et al. (2021) also highlight the
importance of intervention by the school and other students, as well as the community, parents,
and families, as appears necessary, along with the importance of school resource officers in
preventing targeted school violence.
Others have suggested that local government agencies take specific, concrete steps to
improve communities’ capacity to prevent individuals from becoming at-risk individuals in the
first place. This could take the form of investment in community mental health services and
improving collaboration between judicial courts and mental health services (Amman et al.,
2017). School-level policies, including intense supervision, clear rules, firm discipline, and high
teacher/parent engagement, have also been recommended. The importance of school leadership
providing information on potential threats to threat managers has also been highlighted (Amman
et al., 2017). The results of the present study suggest that relevant risk factors are present among
school perpetrators of school violence at the various levels of the public health model, and
therefore do suggest the importance and potential efficacy of these recommendations.

231
Conclusions
School shootings in the United States have been increasing not only in number, but also
with regard to the number of associated deaths per school shooting. This study performed a
series of analyses focused on the JRIC and TASSS datasets, examining how perpetrators of
school violence differ from perpetrators of other forms of violence, the associations between
Adverse Childhood Experiences (ACEs), External Factors, and Environmental Sustainable
Design measures, comparisons between the JRIC and TASSS samples, and comparisons between
the TASSS dataset and data derived from the general population. Numerous differences were
found between perpetrators of school violence and perpetrators of other forms of violence, with
many significant associations found between ACEs, External Factors, and Environmental
Sustainable Design measures, and with significant differences indicated between the JRIC and
TASSS samples, and when comparing the TASSS sample with the general population.
The results found in this study provided many novel results, while also supporting the
findings of previous literature, and suggesting the validity of the theories cited in this study.
Limitations of this study were discussed, along with possibilities for future research, along with a
set of recommendations that, if implemented, should serve to substantially reduce the number of
school shootings in the United States.

232
References
Abel, M. N., Chermak, S., & Freilich, J. D. (2022). Pre-attack warning behaviors of 20
adolescent school shooters: A case study analysis. Crime Delinquency, 68(5), 786–813.
Alathari, L., Drysdale, D., & Driscoll, S. (2019). Protecting America’s schools: A U.S. Secret
Service analysis of targeted school violence | U.S. Secret Service National Threat
Assessment Center (NTAC). https://www.secretservice.gov/sites/default/files/2020-04/
Protecting_Americas_Schools.pdf
Alathari, L., Drysdale, D., & Driscoll, S. (2021). Averting targeted school violence: A U.S. Secret
Service analysis of plots against schools | U.S. Secret Service National Threat Assessment
Center (NTAC). https://www.secretservice.gov/newsroom/reports/threatassessments/schoolcampus-attacks/details-0.
Alathari, L., Drysdale, D., & Driscoll, S. (2023). Mass attacks in public spaces: 2016–2020 |
U.S. Secret Service National Threat Assessment Center (NTAC).
https://www.secretservice.gov/newsroom/reports/threat-assessments/schoolcampusattacks/details-0.
Amman, M., Bowlin, M., Buckles, L., Burton, K. C., Brunell, K. F., Gibson, K. A., Griffin, S. H.,
Kennedy, K., & Robins, C. J. (2017). Making prevention a reality: Identifying, assessing,
and managing the threat of targeted attacks. Institute for Predictive Analytics in Criminal
Justice. https://www.fbi.gov/file-repository/making-prevention-a-reality.pdf/view
Bentson, K. A. (2006). An epidemiological approach to terrorism. Air Force Institute of
Technology. https://scholar.afit.edu/etd/3288.

233
Bockler, N., Roth, V., Stetten, L., & Zick, A. (2014). Targeted school violence and the web of
causes: Risk factors and the problems of specificity. International Journal of
Developmental Science, 8(1–2), 49–51.
Bonanno, C. M., & Levenson, R. L. (2014). School shooters: History, current theoretical and
empirical findings, and strategies for prevention. SAGE Open, 4(1), 1–11.
Brown, M. (2022). Preventing targeted violence and terrorism. Center for Prevention Programs
and Partnerships (CP3).
Campbell, J. A., Walker, R. J., & Egede, L. E. (2016). Associations between adverse childhood
experiences, high-risk behaviors, and morbidity in adulthood. American Journal of
Preventive Medicine, 50(3), 344–352.
Carey, B. (2007, April 18). For rampage killers, familiar descriptions, “troubled” and “loner,” but
no profile. New York Times. http://query.nytimes.com/gst/fullpage.html?res=9406E2
Center for Prevention Programs and Partnerships (CP3). (2022). Building local prevention
frameworks. Homeland Security. Retrieved February 26, 2023, from
https://www.dhs.gov/publication/building-local-prevention-frameworks
Centers for Disease Control and Prevention (CDC). (2020). Risk and protective factors | violence
prevention | youth violence. Retrieved February 24, 2023, from
https://www.cdc.gov/violenceprevention/youthviolence/riskprotectivefactors.html#Risk
%20Factors
Crawford, N. (2002). New ways to stop bullying. Monitor on Psychology, 33(9), 64.
https://www.apa.org/monitor/oct02/bullying

234
Crouch, E., Probst, J. C., Radcliff, E., Bennett, K. J., & McKinney, S. H. (2019). Prevalence of
adverse childhood experiences (ACEs) among U.S. children. Child Abuse & Neglect, 92,
209–218.
del Carmen, A., Bowling, B., Gibson, C., Chism, K., Luo, F., Dobbs, R., Petrowski, T., & Tao, J.
(2022). An analysis of post-Columbine school shootings. Stephenville, Texas.
https://www.tarleton.edu/ipac/wp-content/uploads/sites/315/2022/12/
IPACSchoolShootingsResearch.pdf
Department of Homeland Security (DHS). (2021). Fiscal year 2021 targeted violence and
terrorism prevention (TVTP) grant program. The Department of Homeland Security.
Department of Homeland Security (DHS). (2022). If you see something, say something®
campaign information packet. https://www.dhs.gov/sites/default/files/2022-
09/S4_Campaign%20Information%20Packet%202022.pdf
Dowdell, E. B., Freitas, E., Owens, A., & Greenle, M. M. K. (2022). School shooters: Patterns of
adverse childhood experiences, bullying, and social media. Journal of Pediatric Health
Care, 36(4), 339–346.
Duplechain, R., & Morris, R. (2014). School violence: Reported school shootings and making
schools safer. Education, 135(2), 145–150.
Durkheim, E. (1897). Le suicide: Etude de sociologie. Paris: Alcan.
Dutton, D. G., White, K. R., & Fogarty, D. (2013). Paranoid thinking in mass shooters.
Aggression and Violent Behavior, 18(5), 548–553.
Edmonds, W. A., & Kennedy, T. D. (2016). An applied guide to research designs: Quantitative,
qualitative, and mixed methods. SAGE Publications.

235
Nisbet, R., Elder, J., & Miner, G. D. (2009). Handbook of statistical analysis and data mining
applications. Elsevier Science.
Fabes, R. A., Quick, M., Catherine, E., & Musgrave, A. (2021). Exclusionary discipline in U.S.
public schools: A comparative examination of use in pre-kindergarten and K-12 grades.
Educational Studies, 1–18.
Farr, K. (2018). Adolescent rampage school shootings: Responses to failing masculinity
performances by already-troubled boys. Gender Issues, 35, 73–97.
Federal Bureau of Investigation (FBI). (2022). Active shooter incidents in the United States in
2021.
Ferguson, C. J., Coulson, M., & Barnett, J. (2011). Psychological profiles of school shooters:
Positive directions and one big wrong turn. Journal of Police Crisis Negotiations, 11(2),
141–158.
Flannery, D. J., Fox, J. A., Wallace, L., Mulvey, E., & Modzeleski, W. (2021). Guns, school
shooters, and school safety: What we know and directions for change. School Psychology
Review, 50(2-3), 237–253.
Freilich, J. D., Chermak, S. M., Connell, N. M., Klein, B., Greene-Colozzi, E. (2021).
Understanding the causes of school violence using open source data. Office of Justice
Programs. https://nij.ojp.gov/library/publications/understanding-causes-school-violenceusing-open-source-data
Grøndahl, P., & Bjørkly, S. (2016). Research quality and psychological theory in publications on
school shooters with multiple victims—A systematic review of the literature. Cogent
Psychology, 3(1).

236
Hall, R. C. W., Friedman, S. H., Sorrentino, R., Lapchenko, M., Marcus, A., & Ellis, R. (2019).
The myth of school shooters and psychotropic medications. Behavioral Sciences & the
Law, 37(5), 540–558.
Hussin, S., & Gunaratna, R. (2019). The Los Angeles model in deradicalisation. In Terrorist
deradicalisation in global contexts: Success, failure and continuity (1st ed.). Taylor &
Francis Group.
Ioannou, M., Hammond, L., & Simpson, O. (2015). A model for differentiating school shooters
characteristics. Journal of Criminal Psychology, 5(3), 188–200.
Joint Regional Intelligence Center (JRIC). (2022). Fighting many foes: 20 years of extremist
violence in America.
Keatley, D. A., Mcgurk S., & Allely, C. S. (2020) Understanding school shootings with crime
script analysis. Deviant Behavior, 41(9), 1084–1096.
Landau, E. (2012, December 19). Rejection, bullying are risk factors among shooters. CNN.
https://www.cnn.com/2012/12/18/health/ctshooting-mental-illness/index.html
Langman, P. (2009). Rampage school shooters: A typology. Aggression and Violent Behavior,
14(1), 79–86.
Langman, P. (2013a). School shooters on college campuses. Journal of Campus Behavioral
Intervention, 1, 6–39.
Langman, P. (2013b). 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 (pp. 131–156). Springer Science +
Business Media.

237
Langman, P. (2016a). Expanding the sample: Five school shooters. www.schoolshooters.info,
1(6), 1–13.
Langman, P. (2016b). School shooters: Nine brief sketches. www.schoolshooters.info, 1(6), 1–7.
Langman, P., & Straub, F. (2019). A comparison of averted and completed school attacks from
the National Policing Institute Averted School Violence database. National Policing
Institute. https://www.policinginstitute.org/publication/a-comparison-of-averted-andcompleted-school-attacks-from-the-police-foundation-averted-school-violence-database/
Lankford, A. (2013). A comparative analysis of suicide terrorists and rampage, workplace, and
school shooters in the United States from 1990 to 2010. Homicide Studies, 17(3), 255–
274.
LAPD (2019). Resource guide: Mental health intervention training.
LAPD & Los Angeles County Department of Mental Health (2020). Mental evaluation unit:
Operations guide.
https://lapdonlinestrgeacc.blob.core.usgovcloudapi.net/lapdonlinemedia/2021/09/
MENTAL-EVALUATIONUNIT-LAPD-_-LA-COUNTY-DEPT-OF-MENTALHEALTH.pdf
Lenhardt, A. M. C., Farrell, M. L., & Graham, L. W. (2010). Providing anchors—reclaiming our
troubled youth: Lessons for leaders from a study of 15 targeted school shooters. The
Educational Forum, 74(2), 104–116.
Lenhardt, A. M. C., Graham, L. W., & Farrell, M. L. (2018). A framework for school safety and
risk management: Results from a study of 18 targeted school shooters. The Educational
Forum, 82, 3–20.

238
Levin, J., & Madfis, E. (2009). Mass murder at school and cumulative strain: A sequential model.
American Behavioral Scientist, 52, 1227–1245.
Liu, X. (2022). Categorical data analysis and multilevel modeling using R. SAGE Publications.
Meloy, J. R., Hoffmann, J., Roshdi, K., & Guldimann, A. (2014). Some warning behaviors
discriminate between school shooters and other students of concern. Journal of Threat
Assessment and Management, 1(3), 203–211.
Meloy, J. R., & O’Toole, M. E. (2011). The concept of leakage in threat assessment. Behavioral
Sciences and the Law, 29, 513–527.
Merrick, M. T., Ford, D. C., Ports, K. A., & Guinn, A. S. (2018). Prevalence of adverse
childhood experiences from the 2011–2014 behavioral risk factor surveillance system in
23 states. JAMA Pediatrics, 172(11), 1038–1044.
Millon, T., & Davis, R. (1998). Ten subtypes of psychopathy. In T. Millon, E. Simonsen, M.
Birket-Smith, and R. Davis (Eds.), Psychopathy: Antisocial, criminal, and violent
behavior (pp. 161–170). John Wiley and Sons.
National Center for Education Statistics (2017). Percentage of students receiving selected
disciplinary actions in public elementary and secondary schools by type of disciplinary
action, disability status, sex, and race/ethnicity: 2013–2014. Institute of Education
Sciences.
https://nces.ed.gov/programs/digest/d17/tables/dt17_233.28.asp?referer=raceindicators
National Center for Education Statistics (2023). Preprimary, elementary, and secondary
education: Status dropout rates. Institute of Education Sciences.
https://nces.ed.gov/programs/coe/indicator/coj/status-dropout-rates

239
National Security Council (NSC) (2021). National strategy for countering domestic terrorism.
https://www.whitehouse.gov/wp-content/uploads/2021/06/National-Strategy-forCountering-Domestic-Terrorism.pdf
Peterson, J. (2021). A multi-level, multi-method investigation of the psycho-social life histories of
mass shooters. National Institute of Justice.
https://www.ojp.gov/library/publications/multi-level-multi-method-investigation-psychosocial-life-histories-mass
Pfeifer, B., & Ganzevoort, R. (2016). Tell me why? Existential concerns of school shooters.
Religious Education, 112(2), 123–135.
Pressman, D. E. (2009). Risk assessment decisions for violent political extremism. Public Safety
Canada. https://www.publicsafety.gc.ca/cnt/rsrcs/pblctns/2009-02-rdv/index-en.aspx
Pyrooz, D. C., & Sweeten, G. (2015). Gang membership between ages 5 and 17 years in the
United States. Journal of Adolescent Health, 1–6.
Redlener, E. (2006). How to make school safer. Time.
http://time.com/time/nation/article/0,8599,1543803,00.html
Rothe, E. M. (2022). Inside the mind of the adolescent school shooter: Contributing factors and
prevention. World Soc Psychiatry, 4, 85–93.
RTI International (2017). Countering violent extremism: The use of assessment tools for
measuring violence risk.
https://www.dhs.gov/sites/default/files/publications/OPSR_TP_CVE-Use-AssessmentTools-Measuring-Violence-Risk_Literature-Review_March2017-508.pdf

240
Sacks, V., & Murphey, D. (2018). The prevalence of adverse childhood experiences, nationally,
by state, and by race/ethnicity. Child Trends.
https://www.childtrends.org/publications/prevalence-adverse-childhood-experiencesnationally-state-race-ethnicity
Shreve, J. N., & Holland, D. D. (2018). SAS certification prep guide: Statistical business
analysis using SAS9. SAS Institute.
Silk, J., & Mlre, E. (2017). Redesigning Care: Using design thinking to restructure the care team.
The Journal of Campus Behavioral Intervention, 5, 12–16.
Snell, L., & Volokh, A. (2005). School violence prevention: Strategies to keep school safe.
Reason Public Policy Institute.
Sommer, F., Leuschner, V., & Scheithauer, H. (2014). Bullying, romantic rejection, and conflicts
with teachers: The crucial role of social dynamics in the development of school
shootings—A systematic review. International Journal of Developmental Science, 8, 3–
24.
Southers, E. G. (2013). Homegrown violent extremism: Designing a community-based model to
reduce the risk of recruitment and radicalization. [Unpublished doctoral dissertation].
University of Southern California.
StataCorp. (2021). Stata statistical software: Release 17. StataCorp LLC.
Wagner, W. E., & Gillespie, B. J. (2018). Using and interpreting statistics in the social,
behavioral, and health sciences. SAGE Publications.
Weinberg, S. L., & Abramowitz, S. K. (2008). Statistics using SPSS: An integrative approach.
Cambridge University Press.

241
Weine, S., Younis, A., & Polutnik, C. (2017). Community policing to counter violent extremism:
A process evaluation in Los Angeles. Office of University Programs, Science and
Technology Directorate, U.S. Department of Homeland Security.
Wetterneck, C., Sass, D. A., & Davies, W. H. (2005). Perceptions of risk factors for school
violence. Journal of School Violence, 4(2), 153–166.
Widham, R. C., Hooper, L. M., & Hudson, P. E. (2005). Selected spiritual, religious, and family
factors in the prevention of school violence. Counseling and Values, 49(3), 208–16.
Wilson, A. E. V. (2010). [Review of the book Why Kids Kill: Inside the Minds of School
Shooters, by P. Langman]. Journal of the American Academy of Child & Adolescent
Psychiatry, 49(9), 969–971.
Wike, T. L., & Fraser, M. W. (2009). School shootings: Making sense of the senseless.
Aggression and Violent Behavior, 14(3), 162–169.
Vossekuil, B., Fein, R., Bornum, R., & Modzeleski, W. (2002). The final report and findings of
the safe school initiative: Implications for the prevention for school attacks in the United
States. United States Secret Service and United States Department of Education.

Abstract (if available)

Abstract The research problems focused on in this study consisted of the following questions: What factors or characteristics delineate perpetrators of school violence from perpetrators of other forms of violence? What are the associations between adverse childhood experiences (ACEs), external factors, and Environmental Sustainable Design measures? How do the JRIC and TASSS samples differ? How does the TASSS sample compare with the general population?

A total of 15 research questions were developed in order to examine these research problems. This study’s purpose was to reduce the incidence of school violence and, specifically, the incidence of school shootings, particularly in the United States, while taking a public health approach to the problem. This study examined perpetrators of school violence as an independent group, as well as the similarities and differences between perpetrators of school violence and perpetrators of other forms of violence in order to determine what factors might delineate perpetrators of school violence both independently and in comparison with other perpetrators of violence.

Further analyses examined associations between ACEs and related measures in both datasets, as well as analyses comparing the JRIC and TASSS samples, and the TASSS sample with the general population. A quantitative approach was taken in this study, incorporating both descriptive and correlational methods, with perpetrators of school violence as well as other perpetrators of violence examined on a descriptive level, and with these two groups compared using inferential statistical tests primarily consisting of Fisher’s exact tests, along with independent-samples t-tests and regression analyses. Associations between ACEs and related measures were examined using Fisher’s exact tests, while comparisons between samples took the form of difference in proportions tests.

The results of the analyses conducted for this study found significant differences when comparing perpetrators of school violence with perpetrators of other forms of violence with respect to age, level of education, contact with law enforcement, emotional distress, suspicious travel, violence to others, drug abuse, having been radicalized by an associate, having had a personal grievance, extremist media consumption, social media platforms used, and social media activities engaged in. Significant associations between ACEs and related measures were found in both samples, as well as significant differences between the JRIC and TASSS samples, as well as between the TASSS sample and the general population.

These results allowed for a better understanding of perpetrators of school violence. These results also allowed for a series of recommendations to be provided that, if implemented, should substantially reduce the incidence of school violence, and specifically, school shootings in the United States.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Building a trauma-informed community to address adverse childhood experiences

PDF

A sociocultural and developmental approach to intimate partner violence among a sample of Hispanic emerging adults

PDF

Life course implications of adverse childhood experiences: impacts on elder mistreatment, subjective cognitive decline, and caregivers' health

PDF

The role of social support in the relationship between adverse childhood experiences and addictive behaviors across adolescence and young adulthood

PDF

Long-term impacts of childhood adversity on health and human capital

PDF

The perceived effects of clean water infrastructure in Nasarawa State, Nigeria

PDF

Childhood obesity and prevention: nutrition, cooking, and gardening learning approach for an educational intervention program to increase fruit and vegetable knowledge, preference, and consumptio...

PDF

Critical factors in evaluating compliance to United Nations Security Council Resolution 1540: developing a methodology for compliance evaluation

PDF

The impact of childhood trauma on substance use and mental health during the SARS-CoV-2 pandemic among young adults

PDF

Redlining, neighborhood change, and individual outcomes: an exploration of how space shapes the landscape of inequality from past to present

PDF

An examination of the law enforcement response to active shooter incidents in U.S. cities: findings and implications for improving future response efforts

PDF

Social construction of the experience economy: the spatial ecology of outdoor advertising in Los Angeles

PDF

The institutional context of Korean philanthropy and the role of government and (quasi-) community foundations

PDF

Outcomes-based contracting through impact bonds: ties to social innovation, systems change, and international development

PDF

A miracle or a mirage? A study to evaluate the impacts of microfinance

PDF

The impact of the environment on childhood allergic rhinitis: findings from the Children’s Health Study

PDF

The effects of interlocal collaboration on local economic performance: investigation of Korean cases

PDF

Health care utilization and spending of the U.S. aging population

PDF

The causal-effect of childhood obesity on asthma in young and adolescent children

PDF

The influence of organizational ethics on job attitudes and government performance

Asset Metadata

Creator Vargas, Alejandro Jr. (author)

Core Title Unraveling the threads of tragedy: a public health exploration of adverse childhood experiences as precursors to targeted school violence and protective factors for mitigation strategies

School School of Policy, Planning and Development

Degree Doctor of Policy, Planning & Development

Degree Program Planning and Development,Policy

Degree Conferral Date 2024-05

Publication Date 04/24/2024

Defense Date 04/09/2024

Publisher Los Angeles, California (original), University of Southern California (original), University of Southern California. Libraries (digital)

Tag ACEs,adverse childhood experiences,J6 offenders,mass shootings,OAI-PMH Harvest,Public Health,school shootings,targeted school violence,TASSS,The American School Shooting Shooting Study

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Leach, William D. (committee chair), Heppenstall Heger, Astrid (committee member), Hoffman, Bruce (committee member)

Creator Email lapd31083@gmail.com,vargasal@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC113893040

Unique identifier UC113893040

Identifier etd-VargasAlej-12855.pdf (filename)

Legacy Identifier etd-VargasAlej-12855

Document Type Dissertation

Format theses (aat)

Rights Vargas, Alejandro Jr.

Internet Media Type application/pdf

Type texts

Source 20240422-usctheses-batch-1143 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu