PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

MAY; , m

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:lProj/Aoc&Pres/ClRC/DateDue.hdd

IDEOLOGICALLY-MOTIVATED HOMICIDE IN THE UNITED STATES, 1990-
2008: A COMPARISON OF FAR-RIGHT AND IDENTITY-BASED HOMICIDE
EVENTS
By

Jeffrey Allen Gruenewald

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Criminal Justice

2009

ABSTRACT

IDEOLOGICALLY-MOTIVATED HOMICIDE IN THE UNITED STATES, 1990-
2008: A COMPARISON OF FAR-RIGHT AND IDENTITY-BASED HOMICIDE
EVENTS
By

Jeffrey Allen Gruenewald

Hundreds of homicides have been perpetrated by far-right extremists in the
United States in the last three decades, and scholars, as well as the Department of
Homeland Security, have suggested that there remains a significant threat of violence
from far-right extremists to public and national security. We to date, however, know very
little about the entirety of this complex and heterogeneous form of violence, largely due
to a number of conceptual, theoretical, and methodological challenges faced by scholars.
In order to advance knowledge and understanding of far-right extremist homicide, and
respond to these issues, this dissertation examines the nature of far-right perpetrated
homicide in the United States between 1990 and 2008 from a criminal event perspective.
Homicide events are quantitatively examined based on systematically collected open-
search data and analyzed using descriptive and multivariate statistical techniques. In
addition, variable-centered theme analyses are used to firrther an in-depth understanding
of the similarities and differences in homicide subtypes in the context of interactions
between homicide actors and the situational contexts in which homicides occur. This
dissertation contributes to the scholarly literature on domestic terrorism, hate/bias crime,
and homicide. This research also contributes to the ability of the police and other criminal
justice administrators to recognize, investigate, and prosecute far-right perpetrated and

other types of identity-based homicides.

ACKNOWLEDGEMENTS

I would like to acknowledge the support of the National Consortium for the Study
of Terrorism and Responses to Terrorism (START). The funding and other resources
provided by the START Pre-doctoral Fellowship over the last two years was imperative
to the completion of my dissertation research on far-right homicides.

A number of individuals also helped to make the completion of this dissertation
possible. I would like to thank Drs. Steven Cherrnak, Joshua Freilich, and William Alex
Pridemore for helping me to develop the topic and research approach for this study. Drs.
Cherrnak and Freilich were also kind enough to share with me their data on far-right
extremist crimes from their ongoing START-funded research. A large portion of the
homicide data stem from their comprehensive database on crimes committed by the
domestic far-right in the United States.

I would also like to thank a number of researchers, whose efforts have
substantially contributed to the open-source data analyzed in this dissertation, including
William Parkin, Roberta Belli, Christina Kress, Michael Suttmoeller, David Caspi, and
Matthew Santana.

Vera Gruenewald and Roberta Belli contributed to the preparation of this
manuscript. Dr. Nicholas Corsaro provided advice on quantitative analyses. Finally, I
would like to thank my Committee members for their thoughtful comments and

constructive criticisms on my dissertation proposal.

iii

TABLE OF CONTENTS

List of Tables ............................................................................................. vii
List of Figures ........................................................................................... ix
Glossary of Terms ....................................................................................... x
Chapter 1: Introduction ................................................................................. 1
Defining Homicide Types and Subtypes ......................................................... 3
Threat of Far-right Violence ....................................................................... 4
Research Problem .................................................................................... 5
The Far-Right Extremist Homicide Project ...................................................... 7
Research Questions ................................................................................. 8
Data and Methodology ............................................................................. 9
Contribution of the Dissertation ................................................................... 14
Review of Chapters ................................................................................. 15
Chapter 2: Review of the Literature .................................................................. 19
Descriptive Studies ........................................................................... 20
Methodological Issues ........................................................................ 23
Police Reporting ........................................................................... 24
Survey Data ............................................................................... 26
Conceptualizing 1M Violence ................................................................ 27
IM Violence as Heterogeneous .......................................................... 29
Far-Right Extremist Homicide Events ................................................... 31
Offender-Based Studies ......................................................................... 36
Relative Deprivation ......................................................................... 36
Low Self-Control ............................................................................. 38
Modified Social Control Theory ............................................................. 40
Neutralization Theory ........................................................................ 41
Defended Neighborhoods Theory ............................................................ 42
Critique of Offender-Based Theories ........................................................ 43
Conclusion ....................................................................................... 45
Chapter 3: Research Approach .................................................................. 47
Situational Perspectives ...................................................................... 48
Rational Choice ........................................................................... 49
Social Interactionism ..................................................................... 51

iv

Symbolic Interactionism .................................................................. 51

Integrated Approaches .................................................................... 53
Criminal Event Perspective and Far-Right Homicide ....................................... 54
Conceptualizing IM Homicide Events ...................................................... 59
Research Questions .......................................................................... 63

Chapter 4: Data and Methodology ............................................................... 65
Identifying Far-Right Perpetrated Homicide Events ........................................ 67

Indicators of Far-Right Extremism ....................................................... 67
Identifying Other IB Homicide Events ...................................................... 71
Creating an Event-Based Open-Source Database ........................................... 73
Open-Source Searching ...................................................................... 75

Advantages of Open-Source Data Collection ............................................ 76

Responding to Challenges of Open-Source Data Collection ............................ 77
Coding Homicide Events ..................................................................... 80

1M Homicide Situations .................................................................. 81

1M Homicide Suspects .................................................................... 84
Comparative Analyses ........................................................................ 85
Multivariate Comparative Analysis ......................................................... 86
Qualitative Analyses ................................................................................ 87

Variable-Centered Theme Analysis ...................................................... 89
Conclusion ................................................................................... 94

Chapter 5: The Context of Far-Right Homicide Events ........................................ 95
Considering Context ......................................................................... 98
Geospatial Context of Far-Right Homicide Events ......................................... 99
Temporal Context of Far-Right Homicide Events .......................................... 105
A Comparison of Far-Right and “Average” Homicide Events ............................. 110
Intensity of Far-Right Affiliation ............................................................ 120
Evaluation of Open-Source Information ..................................................... 122
Summary of Findings ........................................................................ 125

Chapter 6: A Comparison of Far-Right 1M and Non-1M Homicide Events, 1990-2008. . . 128

Descriptive Findings ......................................................................... 130
Multivariate Findings ........................................................................ 139
Variable-Centered Theme Analyses ......................................................... 145
Profit-Seeking Homicide ................................................................. 146
Known Victim-Offender Relationships .................................................. 151
Indoor Location ........................................................................... 155
Discussion ..................................................................................... 158

Chapter 7: A Comparison of F ar-Right and Non Far-Right Identity-Based (1B) Homicides. 161
Descriptive Findings: Comparing Far-Right IB and Non Far-Right IB Homicide. . . . . 166

Multivariate Findings ........................................................................ 167
Variable-Centered Theme Analysis: Known Victim Offender Relationships ............. 171
Anti-GLBT Homicides as Crime Events .................................................... 175
Explaining Anti-GLBT Homicide ........................................................ 176
Chapter 8: Discussion ........................................................................... 179
A Review of the Extremist Homicide Project and Maj or Findings ........................ 179
Research Implications .............................................................................. 184
Research Limitations ......................................................................... 186
Future Research .............................................................................. 188
Appendix A ...................................................................................... 192
Appendix B ....................................................................................... 193
References ................................................................................................ 194

vi

LIST OF TABLES

Table 1.1 Step 1: Identification of Homicide Events from Multiple Sources .................. 11
Table 1.2 Step 2: Open-Source Searching of Homicide Events ................................ 12
Table 1.3 Step 3: Coding of Homicide Event Data and Quality of Information ............... 13
Table 4.1 Incident, Suspect, and Assessment of Data Codeboolm ............................. 41
Table 5.1 Percentage of Far-Right Homicide Events by State, 1990-2008 .................... 101

Table 5.2 Descriptive Findings on Far-Right and Average Homicide Events, 1990-2006. . . 112

Table 5.3 Strength of Far-Right Suspect Affiliation to the Far-Right ......................... 121
Table 5.4 Percent Missing Values by Source Type and Salience, 1990-2008 ................. 123
Table 6.1 Far-Right Homicide Motivational Circumstances ................................... 132
Table 6.2 Descriptive Findings for Far-Right IM and Non-1M Homicides .................... 133
Table 6.3 Binary-Logistic Regression Findings for Predicting Motivational Subtypes ...... 140
Table 6.4 Affiliation Type for F ar-Right Homicide Suspects with Prior Arrests ............. 142
Table 6.5 Profit-Seeking Behaviors for Far-Right Homicide .................................. 147
Table 6.6 Predicted Probabilities for F ar-Right Non-1M Homicide Events ................... 152
Table 6.7 Known Victim-Offender Relationships for Far-Right Homicide ................... 153
Table 6.8 Indoor Location for F ar-Right Homicide ............................................ 156

Table7.1 Descriptive Findings on Far-Right and Non Far-Right Identity-Based (1B)
Homicide ......................................................................................... 166

Table 7.2 Binary Logistic Regression Findings for Predicting Homicide Suspect Subtype. 168

vii

Table 7.3 Known Victim-Offender Relationships by Non Far-Right Suspects ............... 172

Table A. 1. Indicators of Ideological Motivation ................................................... 192

Table B. 1. Indicators of Far-Right Extremist Affiliation .......................................... 193

viii

LIST OF FIGURES

Figure 3.1 Types of Far-Right Homicide Events in the United States .......................... 60
Figure 3.2 Types of Identity-Based Homicide Events in the United States ..................... 62
Figure 5.1 Rate of Homicide by Type and Year, 1990-2007 .................................... 106
Figure 5.2 Far—Right Homicide by Type and Year, 1990-2007 ................................. 109
Figure 5.3 Average F ar-Right Suspect Age by Group Type .................................... 118
Figure 7.1 Identity-Based (IB) Homicide Victim Types by Suspect Affiliation ............... 171

ix

GLOSSARY OF TERMS

Domestic Far-Right: This dissertation relies on Freilich and Chermak’s (2006:32)
definition which states that the domestic far-right is “. . .composed of individuals or
groups that subscribe to aspects of the following ideals: they are fiercely nationalistic (as
opposed to universal and international in orientation), anti-global, suspicious of
centralized federal authority, reverent of individual liberty (especially their right to own
guns, be free of taxes), believe in conspiracy theories that involve a grave threat to
national sovereignty and/or personal liberty and a belief that one’s personal and/or
national “way of life” is under attack and is either already lost or that the threat is
imminent (sometimes such beliefs are amorphous and vague, but for some the threat is
from a specific ethnic, racial, or religious group), and a belief in the need to be prepared
for an attack either by participating in paramilitary preparations and training and
survivalism. It is important to note that mainstream conservative movements and the
mainstream Christian right are not included.”

Far-Right Ideological-Motivated (1M) Homicide: Homicide events that involve at least
one suspect belonging to the domestic far-right and primarily motivated by animus
toward social minority groups, governmental workers, abortion providers, and other
ideological targets.

Far-Right Non Ecologically-Motivated (non-1M) Homicide: Homicide events that
involve at least one suspect belonging to the domestic far-right and do not target
ideological or symbolic victims (e.g. social minorities, governmental workers).

Identity-Based (IB) Homicide: Homicide events primarily motivated by animus toward
the minority group status or identity of victims, including anti-race/ethnicity, anti-
gay/lesbian/bi-sexual/transgender, and anti-religious affiliation. These homicides have
also been referred to “hate crimes” or “bias crimes.” Identity-based homicides are one
type of IM homicide event and are committed by those homicide suspects affiliated with
the far-right movement in the United States as well as those suspects who are not.

CHAPTER I: INTRODUCTION

F ar-right extremistsl are responsible for a number of homicides every year in the
United States. The perpetrators’ extent of involvement may vary in far-right
organizations (e.g. militias, skinhead gangs, Ku Klux Klan) (Heitmeyer, 2003), and
homicides committed by far-right extremists represent a diverse group of events
involving complex combinations of motives and targets. Some far-right perpetrated
homicides, for example, have been motivated by anti-government sentiments. In one
high-profile case that occurred in 1997, a far-right extremist murdered two police officers
during a traffic stop and, shortly thereafter, traveled to a nearby town and fatally shot a
judge who had previously ruled against him in a land-use related case. The suspect had
previously made numerous threats against the female judge’s life, but, seemingly, had no
prior relationships with the murdered police officers (Suprynowicz, 2002). Other
homicides perpetrated by far-right extremists, however, have been motivated by bigotry
or “hatred” toward particular social minority groups (e. g. members of the
Gay/Lesbian/Bisexual/Transgender or “GLBT” community, African Americans, Jewish
persons). An example of a high-profile case motivated by animus toward homosexuals is
the 1999 murder of two male partners in their home. The perpetrators were brothers who

adhered to both anti-Semitic and anti-GLBT beliefs (Finz, 1999). In another homicide

 

I A comprehensive definition of the domestic far-right is provided on page 33, and a number of indicators
that are used to determine far-right extremist perpetrator status are provided in Appendix A. For now, it
should suffice that some far-right extremists are especially nationalistic, suspicious of centralized federal
authority, and reverent of individual liberty. Often far-right extremists believe that their way of life and that
the United States is under attack from racial, sexual, and religious minorities. Many far-right extremists
also have “hot-button” issues, such as abortion, homelessness, feminism and other lifestyle choices. Some
far-right extremists formally belong to far-right extremist groups, while others belong to informal groups or
no group at all.

that occurred in the same year, a young far-right extremist went on a three day shooting
spree across the Midwest that targeted racial and religious minorities, injuring multiple
people and killing two victims before killing himself (Washington Post, 1999). F ar-right
extremists have also committed homicides motivated by moral convictions regarding
particular lifestyle issues (i.e. abortion, homelessness), as well as homicides not directly
motivated by extremist ideology. These examples Show that motives and targets can be
varied. Unfortunately, to date, we know little about the entirety of this complex and
heterogeneous social problem, as there has been no systematic research on this topic.

To advance knowledge of this topic, this dissertation comparatively examines the
complex phenomenon of far-right extremist homicide events. A number of research
questions are asked regarding the similarities and differences in the nature of far-right
homicides and typical homicides, in far-right homicide motivational subtypes, and in far-
right hate-motivated homicide events and those events perpetrated by non far-rightists.
This study relies on the criminal event perspective that emphasizes the multiple and
interchanging dimensions of homicide events, as well as the importance of considering
the context in which these crimes occur (Sacco & Kennedy, 2002). In order to address the
questions posited in this study, an original database was constructed based on open~
source information on all far-right extremist homicide events occurring in the United
States between 1990 and 2008.2 These data were supplemented with data on “average”
homicides and “hate-motivated” homicide events perpetrated by non far-right extremists

during the same time period in the United States for purposes of comparison. Employing

 

Some years in which far-right homicide occurred are not included in particular comparative analyses due
to the unavailability of comparison data for these years. The presentations of findings for these analyses are
noted accordingly.

a mixed-methodology, quantitative and qualitative techniques were used to describe the
extent and nature of far-right homicide events compared to other forms of homicide.
Defining Homicide Ting and Subtypes

It is necessary to briefly provide some working definitions of the types of
homicide events of interest in this dissertation. In addition, a more in-depth discussion of
conceptual issues is provided in the next sections.

F ar-right extremist homicides are events involving at least one homicide suspect
linked to the far-right extremist movement in the United States. Therefore, this type of
homicide largely depends on suspects’ beliefs, ideological convictions and involvement
with organized extremist groups rather than any single type of motivation for the
homicide. Indeed, all homicides perpetrated by far-right extremist are of interest to this
study, regardless of suspects’ motivations.

In order to comparatively examine the different subtypes of homicide events
committed by far-right extremists, additional labels that capture the motivational
variations of homicide events are provided. One motivational subtype is referred to as
ideologically-motivated (1M) homicide. These events consist of homicide occurrences
primarily motivated by animus toward social minority groups, governmental targets,
abortion providers, sex offenders, and to a lesser degree, other targets linked to far-right
extremist belief systems. Conversely, non ideologically—motivated (non-1M) homicide
events are those occurrences primarily motivated by anything but animus toward such
targets. Examples of non-1M homicide events include arguments with loved ones and

acquaintances, disagreements within extremist groups, robbery, and failed drug deals.

Finally, a more specific motivational subtype of homicide, identity-based (IB)
homicide, plays a major role in the current study. Identity-based homicides are those
occurrences specifically motivated by animus toward social minority-groups. These
events are referred to as IB homicides rather than “hate crimes” or “bias crimes” because
it is assumed that it is not always possible to judge the perpetrators’ specific views of
their victims. Using the IB label also avoids the usage of “hate crimes” as a label for
these events. Avoiding the question of the ability to judge the feelings and thoughts of
suspects (see, for instance, Bruce, 2001), the definition of [B homicide rests on the
availability of ample evidence indicating suspects targeted victims because of their real or
perceived social minority statuses or identities.

Threat of Far-Right Violence

Homicide events committed by far-right extremists is an important topic of
research. Domestic terrorism remains a significant threat to national security and the
general public, especially in regards to far-right perpetrated violence. Past research has
found that domestic acts of terrorism outnumber international terrorist attacks against the
United States 7 to 1 (LaFree, Dugan, Fogg & Scott, 2006). Furthermore, far-right
extremists have claimed hundreds of lives and future terrorist attacks from far-right
extremists are likely (Hewitt, 2003; see also Carlson, 1995; Riley & Hoffman, 1995). A
survey of police organizations across the United States concludes that police continue to
view far-right extremist crime as a substantial threat (Freilich, Cherrnak, & Simone,
2009). Moreover, this topic is especially relevant in light of a recent official report

published by the Department of Homeland Security that concluded that far-right

extremist membership and activity appear to be on the rise.3 The report asserted that a
combination of factors, including the election of an African American president and
rising unemployment, have created an environment facilitative of far-right extremism.
Research Problem

Despite the threat posed by far-right extremists, knowledge is limited on these
offenders, their backgrounds, and the situational contexts in which they commit homicide
events. In regards to the current state of research on far-right extremist homicides, there is
good news and bad news (Gruenewald, Freilich, & Cherrnak, 2009). The good news is
that information on homicides targeting social minority groups, often involving far-right
extremists, is increasing (e.g. Levin & McDevitt, 1993; Perry, 2001). Homicides of this
type are often newsworthy enough to attain local and, often, national media coverage
(Colomb & Damphouse, 2004), resulting in a large amount of information on these
homicide cases. Also, some police agencies, as well as advocacy groups across the
nation, have begun to collect information on homicides targeting social minorities
(Jenness & Broad, 1997; Levin & McDevitt, 1993; Martin, 1996; Strom, 2001), including
those perpetrated by far-right extremists. In addition, information on homicides
committed by far-right extremists can many times be accessed relatively easily from the
Internet by searching extremist group sites, personal and extremist group blogs, court
documents, and other publicly available documents.

The bad news is that currently there is a lack of empirical research on IM violence

(Silke, 2001), and, specifically, far-right extremist homicides, causing our knowledge and

 

3 This Department of Homeland Security report was pulled shortly after its release due to the controversial
information it contained on the possible threat of returning war veterans on the increase in far-right
extremist activity (Hudson, 2009).

in.depth understanding of this complex problem to be sorely underdeveloped
(Gruenewald et al., 2009). There are methodological and theoretical reasons that account
for the lack of research on this topic. Indeed, to date there has generally been no way to
distinguish violent crimes committed by far-right perpetrators from those that were not in
official homicide data. Moreover, scholarly literature on forms of domestic terrorism has
largely been on left-wing extremism rather than far-right extremism, but past research has
found there to be important differences between left and right-wing domestic terrorism
(Hewitt, 2003; Smith, 1994). Therefore, there exists only anecdotal evidence, often from
journalistic accounts, in regards to the nature and extent of far-right extremist homicide in
the United States (Gruenewald et al., 2009).

In addition, theoretically, the limited research on IM violence has generally
approached the topic from offender-based perspectives (Perry, 2001). For instance,
scholars studying anti-racial violence have tested Gottfredson and Hirschi’s (1990) self-
control theory as an explanation, while others have approached this topic fi’om “defended
neighborhoods” theory (Green, Strolovitch, & Wong, 1998), and the relative deprivation
perspective (Levin & McDevitt, 1993). While these offender-based theories may be
useful in explaining limited aspects of what influences offenders to commit 1M offenses,
they speak little to issues such as how other situational factors can shape offender
decisions across a variety of subtypes of IM offending.

Therefore, what is needed is a broad theoretical and methodological research
approach that overcomes many of these issues that have stunted the advancement of a
more in-depth understanding of far-right perpetrated homicide events. How this research

overcomes these issues, theoretically, by approaching the study of far-right extremist

homicides fiom a criminal event perspective, and methodologically, by relying on open-
sources and quantitative and qualitative analyses, is discussed below.
The Far-Right Extremist Homicide Study

The current study began during a fellowship with the National Consortium for the
Study of Terrorism and Responses to Terrorism (START)4 based on an initial project
titled “F ar-Right Extremist Homicide in the United States, [990-2007 .” For the first time,
the general purpose of the research was to capture, quantitatively and qualitatively, the
nature of far-right homicides based on systematic open-source data collection.

The study of far-right extremist homicides is approached from a criminal event
perspective (CEP) (Sacco & Kennedy, 2002). The CEP conceptualizes criminal acts as
multi-dimensional events, consisting of suspects, victims, and other actors that interact in
particular situational contexts (Block, 1981). This perspective assumes that homicide
suspects are rational beings whose actions are shaped by the evolving social and physical
contexts in which homicide events occur. In this way, the research avoids an over-
reliance on offender-based theories of crime motivations and considers more holistically
how offenders make decisions to commit crimes largely based on the situations in which

they find themselves.

 

4 The National Consortium for the Study of Terrorism and Responses to Terrorism (START) is a Center of
Excellence of the US. Department of Homeland Security based at the University of Maryland. The START
Center website can be accessed at http://www.start.umd.edu/start/. This research was supported by the
United States Department of Homeland Security through the National Consortium for the Study of
Terrorism and Responses to Terrorism (START), grant number N00140510629. However, any opinions,
findings, and conclusions or recommendations in this document are those of the authors and do not
necessarily reflect views of the US. Department of Homeland Security.

This study identifies all far-right extremist homicides that occurred in the United
States between 1990 and 2008 from open-sources. Identity-based homicides perpetrated
by non far-right extremists are also identified for purposes of comparison. Open-source
information is collected on each identified homicide from approximately twenty web-
based search engines discussed more fully in Chapter 4. Data are then coded for each
homicide incident and suspect in separate but relational Microsoft Access files. The
quality of open-source information for each homicide event is also assessed in this Access
database. Quantitative and qualitative analyses are conducted to examine the similarities
and differences in far-right extremist homicides and average homicides, far-right IM and
non-1M homicide, as well as far-right IB and IB homicides perpetrated by non far-
rightists.
Research Questions

The following research questions are asked to capture the nature of far-right
extremist perpetrator backgrounds, offending patterns, and interactions with other
criminal and non-criminal actors, and to identify how far-rightists may be unique from
other types of IM homicide offenders. Questions are also asked to facilitate an
understanding of how homicide situations compare to the situational elements of other
types of 1B homicide. Research Question 1 is the focus of Chapter 5, research Question 2
is the focus of Chapter 6, and research Question 3 is the focus of Chapter 7.

Question 1: What is the nature of far-right extremist homicide event

situations in the United States between 1990 and 2008, including suspect

demographic characteristics, other background attributes, and the

situational contexts of homicide events?

A). What are the similarities and differences of the elements of

homicide situations in those far-right extremist homicides motivated
by ideology and the “average” homicide event in the United States.

8

Question 2: Within far-right extremist homicides, what are the similarities
and differences in [M and non-1M homicide event characteristics?

Question 3: What are the similarities and differences in [B homicides

perpetrated by far-right extremists and IB homicides perpetrated by non far-
right extremists?

Data and Methodology

This study systematically addresses these research questions by identifying,
collecting, and systematically coding open-source data on a set of far—right perpetrated
homicide events in the United States between 1990 and 2008. Open-source data are
collected for a set of other [B homicide events perpetrated by non far-right extremists for
the purposes of comparison. Comparison data are also collected from the publicly
available Uniform Crime Report-Supplementary Homicide Report published annually by
the Federal Bureau of Investigation (FBI) in order to collect a sample of homicides to
represent the “average” homicide events in the United States during the same time period.

This study overcomes many methodological problems associated with relying
solely on official homicide data. This set of far-right perpetrated homicide events were
identified as part of a broader project funded by START, a Center of Excellence of the
United States Department of Homeland Security based at the University of Maryland,
that collected open-search data on all crimes committed by the domestic far right. The co-
principle investigators of this project, Dr. Steven Chermak (Michigan State University)
and Dr. Joshua Freilich (John Jay College of Criminal Justice, CUNY), identified all far-
right perpetrated homicides from a number of sources, including existing terrorism
databases, extremist watch-group organizations, academic sources, and other personal

informants (Freilich and Chermak, 2007). This dissertation extracts data on a set of far-

9

right perpetrated homicide events from this project on which I have been a research
assistant with since its inception. Due to the somewhat unique and sometimes complex
methodological approach, it will be briefly described here.

Data fi'om open-sources are collected on each far-right perpetrated homicide
event. Each unique homicide is defined as a separate event by considering the temporal
and geographic characteristics of the homicide occurrence. In order to collect information
on each homicide event, approximately twenty web-based search engines are
systematically searched and all information relevant to any aspect of the homicide case is
collected and consolidated into a single open-source document. Open-sources include
court documents, advocacy group publication (e.g. Southern Poverty Law Center, Anti-
Defamation League), academic documents, media publications, as well as other sources.

The first of two additional types of data comes from the Supplementary Homicide
Reports, which consist of publicly available event-level homicide data on “all” homicides
that occur in the United States. Homicides are identified by randomly selecting
approximately 10 percent of all homicides that occurred during this time period
(n=2,840). Only a limited number of variables are available in these data, but they can be
used to conduct a number of comparisons between far-right homicides and a sample of all

homicides in the United States.

10

Table 1.1 Step 1: Identification of Homicide Events from Multiple Sources

 

 

Comparison Data Sources
Uniform Crime Report- Non Far-Right Extremist
Far-Right Extremist Open- Supplementary Homicide Identity-Based (IB) Open-
Source Homicide Data Reports Source Homicide Data
Homicide event data extracted Homicide event data extracted Homicide event data collected for
from the Freilich and Chermak’s from the annual UCR- the Extremist Homicide Project,
(2007) Extremist Crime Supplementary Homicide Report which extended the ECDB

Database (ECDB) project. Data
are based on open-sources.

 

published by the Federal Bureau
of Investigation. Data are based
on standardized police reports.

 

project by collecting data on IE
homicides not perpetrated by far-
rightists for purposes of
comparison.

 

Second, a set of IB homicides not perpetrated by far-right extremists are identified

fiom open-sources for the purposes of comparison. Motivations for this set of IB

homicides include animus toward the GLBT community, racial/ethnic minorities,

religious minorities, and the homeless. To a large extent, the same sources used by the

Chermak and Freilich (2007) project are used to identify these homicide events. In

addition, a number of advocacy groups (e. g. Human Rights Campaign, National Asian

Pacific American Legal Consortium) are searched and targeted media searches are

conducted to identify additional [3 homicides. Identifying and collecting open-source

data on this set of homicide events is a significant contribution and extension of the

research on far-right perpetrated crimes conducted by Freilich and Chermak (2007). As

the same variables are considered for both far-right homicides and the set of non far-right

IB homicides motivated by animus toward social minorities, comparative analyses are

possible for the first time that illuminate similarities and differences between these forms

of IM violence.

11

After each homicide event is identified, open-source information is gathered from
multiple sources. Key characteristics of each homicide (e. g. suspect names, date of
events) are then used as index terms in open searches using approximately twenty web-
based search engines. Two levels of search are conducted. The primary search engine
scans for information in six websites, as shown in Table 1.2, while a secondary search
engine searches in less well-known search engines for supplementary open-source
information. An open-source document is created based on the available information,

including all information gathered for that homicide.

Table 1.2 Step 2: Open-Source Searchirg of Homicide Events

 

Primary Search Engines Secondary Search Engines

 

Proquest Google Scholar

Yahoo Amazon

Google Google US. Government
Copemic Federation of American Scientists
News Library Center for the Study of Intelligence
Surf Wax

Dogpile

Mamma

Librarians’ Internet Index

Scirus

All the Web

Google News

Google Blob

Homeland Security Digital Library

 

 

Based on the open-source information, data are coded in separate codebooks for
incident and suspect homicide dimensions. Examples of homicide characteristics include:
victim-offender relationship, weapon use, location, motivation, suspect’s prior criminal
behavior, and other demographic variables. While information was collected for hundreds

of variables for each homicide event, including both victim and far-right group variables

12

(F reilich & Chermak, 2009), only a select number of variables shown in Table 1.3 were
included in this study for a number of reasons. One reason is that the focus of the current
study is on the suspect and incident (or situational) dimensions of far-right homicide
events, as well as how these dimensions interact to shape fatal outcomes. The variables
selected also represent those most emphasized in past homicide research. Moreover, these
variables are the most comparable with official homicide data collected by criminal
justice agencies, which were used for purposes of comparison in this dissertation.

In addition to homicide variables, the quality of open-source information is
assessed in a separate codebook for each homicide event. Both quantity and quality of

information is systematically evaluated.

Table 1.3 Step 3: Coding of Homicide Event Data and Quality of Information‘

 

 

Incident Codebook Suspect Codebook Assessment of

Information Codehook

Weapon type Age Number of Each Source

Circumstances Race Type Used

Victim-Offender relationship Gender Total Number of Sources

Location Group Membership Used

State Prior Criminal History

Year Drug/Alcohol Use During Incident

 

 

 

Based on the data collected in this comprehensive database, descriptive and
multivariate statistics are used to address research questions on the nature of homicide
types and subtypes. The main purpose of multivariate analysis is to identify key variables

that distinguish between homicide subtypes. To supplement these quantitative findings,

 

5 Each variable listed in Table 1.3 is also discussed in more detail in Chapter 4.

l3

systematic qualitative theme analyses are conducted for each significant variable to
examine more in depth are the homicide suspects’ interactions and attachments of
meaning to experiences in particular situations and the effect of these processes on their
decisions to commit homicide. The “variable-centered theme analysis” provides a more
comprehensive understanding of the similarities and differences of far-right perpetrated
homicide and the complex and heterogeneous phenomena of IM homicide in the United
States.
Contribution of the Dissertation

The current research makes a number of contributions to the scholarly literature
on far-right extremist and other forms of IM violence and informs criminal justice and
counter-terrorism policies. First, since the terrorist attacks of September 11‘”, scholarly
interest and research on M violence has dramatically increased. Unfortunately, research
has been devoid of original data sources and plagued by a number of other
methodological issues (Silke, 2001, 2004). Thus, scholars interested in studying IM
violence in the United States should greatly benefit from the comprehensive set of data
collected on far-right and other [M homicide incidents and participants developed in this
study. Second, this research contributes to the well-established criminological literature
on homicide by advancing knowledge of far-right perpetrated homicide events. While
homicide scholars have studied a number of different subtypes of homicide (Flewelling &
Williams, 1999), there has been no empirical research on far-right extremist homicide
events. Third, the limited body of past empirical research on IM violence has focused
almost exclusively on explaining offender motivations. By combining quantitative and

qualitative theme analysis techniques, this study facilitates a more in-depth understanding

14

of this social phenomenon by approaching far-right homicide as a multi-dimensional
event, for instance, by considering how offender motivations interact with situational
dynamics to shape fatal outcomes. Finally, this study increases the potential to study and
understand the unique dynamics of far-right extremist violence, and could, ultimately, aid
American criminal justice agencies in the prevention, detection, investigation and
prosecution of the most violent forms of domestic terrorism. The next section outlines the
remainder of this study, highlighting the purposes and main emphases of each chapter.
Review of Chapters

The next chapter, Chapter 2, is a review of prior empirical studies on [M crime in
general. Many of these studies address, to some degree, the issue of far-right extremist
violence, while others neglect to distinguish between far-right and non far-right
perpetrated violence. Both types of studies are relevant and are reviewed. In addition to
reviewing prior research, this chapter discusses the many conceptual, theoretical, and
methodological challenges facing the study of 1M violence in the United States. First,
Chapter 2 reviews descriptive studies based on official data collection and victim
surveys. The issues of defining and categorizing types of IM violence as well as the lack
of unbiased data on this type of crime are addressed. Second, Chapter 2 reviews the
limited number of empirical studies that apply or test offender-based theories of IM
offender motivations. The narrow scope of offender-based theories in providing a
comprehensive understanding of this type of crime is also discussed.

In response to the limitations of past research discussed in Chapter 2, Chapter 3
presents a comprehensive theoretical framework that conceptualizes crimes as multi-

dimensional events which occur in particular situational contexts. This chapter first

15

discusses how criminological theories have been typically applied to study IM violence,
and proposes that Sacco and Kennedy’s (2002) criminal event perspective (CEP) is the
most comprehensive and otherwise appropriate theoretical framework for studying far-
right extremist homicide events. This chapter, discusses how the CEP is applied to this
topic, and specifically, the conceptualization of far-right extremist homicide and the
disaggregation of subtypes of IM homicide in response to specific research questions.
Finally, the research questions advanced in this study are outlined.

Chapter 4 addresses the methodological approach to this study. First, indicators
that are used to distinguish far-right extremist homicide events and other IM homicide
events perpetrated by non far-right extremists are presented. Second, the process used to
identify homicides in open-source materials is presented, distinguishing between far-right
perpetrated homicide events and other IB homicides committed by non far-rightists.
Third, the open-sources used to collect information on each homicide event, the protocol
used to systematically collect open-source data, and the stages of data collection are
described. Inter-reliability between data searchers is also discussed. Fourth, the
systematic coding of data for homicide suspects and situational characteristics, specific
codebooks and variables of interest, as well as the stages of the coding process are
addressed. Finally, the plans for quantitative and qualitative analysis of far-right
extremist homicide events are outlined.

Chapter 5 has three purposes. The first purpose of this chapter is to provide the
temporal and geographic context of far-right homicide in the United States during the
time period of interest. The second purpose is to examine the nature of far-right extremist

homicide in comparison to the “average” homicide event in the United States to explore

l6

far-right homicides as, possibly, a unique phenomenon. Open-source homicide data and
official national homicide data previously collected by the Federal Bureau of
Investigation are utilized for this comparison. The third purpose of this chapter is to
assess the quality of open—source information for each far-right extremist homicide. In
doing so, it becomes possible to identify the effects of several types of bias which could
influence the interpretations of findings.

Chapter 6 extends the previous chapter by using open-source homicide data to
comparatively examine the incident and the suspect-level characteristics of two
motivational subtypes of far-right extremist homicide, IM and non-1M homicides.
Descriptive results are presented for each homicide subtype. Results of multivariate
analysis that identify the key distinguishing factors of homicide motivational subtype are
presented and discussed. In addition, variable-centered theme analyses are conducted for
each homicide characteristic found to be a significant distinguishing factor. Consistent
with the CEP, theme analyses are used to explore the operation of the key variables in the
context of the entire homicide event to advance a more in-depth understanding of the
shaping of fatal outcomes by these variables.

Chapter 7 also relies on open-source homicide data, but places focus only on IB
homicide events. Unlike Chapter 6, which analyzed motivational subtypes, this chapter
examines the similarities and differences in two suspect subtypes, far-right homicides and
those homicides not perpetrated by far-right extremists. Findings based on descriptive
and multivariate analyses are also presented and discussed, and the key factors
distinguishing between far-right IB and non far-right IB homicide are identified. Theme

analysis is then used to further explore how these homicide factors Operate.

17

Finally, Chapter 8 reiterates the purposes of the current study and summarizes the
key findings fi'om each chapter. This chapter also discusses the limitations of the current

study and the plans for future research on the topic of far-right extremist homicides.

l8

CHAPTER 2: REVIEW OF THE LITERATURE

This chapter reviews the prior literature on [M violence, including far-right
extremist homicides.6 The purpose of this chapter is to place the study of far-right
perpetrated homicides in the context of several methodological and theoretical issues
plaguing past studies on IM violence. The following review of prior literature is
presented in two sections. First, a review of primarily descriptive studies of 1M violence
is presented. Much of the past empirical research has been primarily descriptive (e. g.
Garofalo & Martin, 1993; Maxwell & Maxwell, 1995; Strom, 2001). While not applying
or testing specific theoretical explanations of 1M offending, these studies are useful for
establishing the conceptual parameters and extent of IM violence. In addition, empirical
descriptions of 1M violence can be used to extend and build more nuanced theoretical
explanations of this type of crime. Following this review of descriptive studies, two
issues, conceptualizing and quantifying IM violence, are considered in order to critically
reflect on the current knowledge of IM offending, and more specifically, far-right
homicide offending. Second, an overview of past studies that aim to explain why
offenders commit IM violence is provided. Furthermore, a critical reflection on offender-
based theories is provided, focusing on the neglect of such theories to consider issues of
situational context of crime events. How the research approach used in the current study

overcomes or addresses these issues is discussed in the next chapter.

 

6 A recent review of the research on terrorism and counter-terrorism by LaFree and Dugan (in press)
examines many of the same topics that are addressed in this chapter (e. g. offender-based theories and
methodological issues with terrorism data); however, the current chapter focuses on [M crime, in general,
rather than solely on political terrorism, and draws more so from studies relevant to this dissertation (i.e.
1M violence in the United States).

19

Descriptive Studies

This section provides a review of prior descriptive studies that have relied on
quantitative data to gauge the extent and, to some degree, the nature of IM violence. To
date, there have been no studies that have provided an empirically-based descriptive
account of far-right homicides in the United States. However, a number of scholars have
analyzed various sources of data on [B violence, or violence that is motivated by bigotry
or animus toward social minority groups, including to some extent far-right perpetrated
violence. Therefore, this literature is reviewed in order to lay the groundwork for the
current study, which focuses on far-right perpetrated homicide.

Since some police departments began collecting data on 1M violence in the late
19808, scholars have relied upon police data to describe this type of crime (Garofalo &
Martin, 1993; Levin & McDevitt, 1993; Maxwell & Maxwell, 1995). Past studies have
largely been based on official police data collected by specialized law enforcement units
at the local level and national statistics provided by the Federal Bureau of Investigation at
the federal level (Strom, 2001). In addition, a very small number of victim surveys and
general population surveys for particular types of IM violence have been conducted
(Berril, 1992; Comstock, 1991).

Some scholars have examined police-reported IM violence from large cities in the
United States, such as New York City and Baltimore (Garofalo & Martin, 1993; Martin,
1996; Maxwell & Maxwell, 1995). These studies examined the nature of IM violence and
comparatively analyzed subtypes of this form of crime by disaggregating by types of
motivation. Past research has found that IM offenders tend to be more likely to be males

than traditional crime offenders (Martin, 1996; Messner, McHugh, & Felson, 2004). In

20

addition, IM violence tends to be more likely to be committed by young perpetrators than
traditional crimes. Maxwell and Maxwell (1995), for instance, found that the proportion
of juveniles arrested for this form of crime was higher than juvenile arrests for offenses
generally, and a large proportion of IM violence was committed by those under 21 years
of age (71 percent). Ideologically-motivated violence also tended to be more likely to be
carried out by multiple perpetrators against multiple victims (Martin, 1996).

Rather than using official police data, other descriptive studies have relied on
surveys to examine particular types of IM violence, such as anti-GLBT violence. For
example, Comstock (1991) examined survey data that he and others collected from
victims of anti-GLBT crimes, which measured anti-sexual orientation violence
experienced by gay men and lesbians perpetrated by non-lesbian and non-gay men (p.
34). Comstock (1991) also examined newspaper and magazine articles for information on
1M victims and offenders. In his study, a number of IM offense and offender
characteristics were considered, including crime participant demographics and incident
characteristics. Confirming past research of IM crime based on police data, Comstock
(1991 :91) found that anti-GLBT perpetrators were more likely to be male and under 20
than common offenders. He also found that such crimes tended to be carried out by
multiple perpetrators against multiple victims, and that weapons were slightly less likely
to be used in anti-GLBT violence. The most frequent location for violence was found to
be outside “gay bars.” This varied, however, by gender as men were more likely to be
victimized in “cruising areas,” while women were more likely to be victimized in

residences.

21

Berrill (1992) also examined a number of studies based on surveys of anti-gay
and lesbian victims, providing an overview of anti-gay violence in the United States.
Survey studies of gay men and lesbian victims showed overwhelmingly that this
population was disproportionately victimized by a variety of IM offenses. Berrill’s (1992)
review also revealed there to be elevated victimization and an elevated level of fear of
victimization among the GLBT population in the United States.

More recently, Herek (2007) collected information from a United States
probability sample of gay, lesbian, and bisexual adults (N=662) using Internet surveys.
He examined various types of crimes against the GLBT population, finding strong
evidence of disproportionate victimization based on sexual orientation since age eighteen.
Herek (2007) found that the majority of anti-GLBT crimes were violent, but also
concluded that this finding varied by gender, as gay men were more likely to experience
violent and property victimization when compared to lesbians and bisexuals. In addition,
Herek (2007) found that gay men were more likely to be victims of stranger violence
compared to lesbians and bisexual men and women, primarily of bigoted heterosexual
male offenders.

Finally, one study anonymously surveyed nearly 500 non-criminal young adults
(i.e. community college students) to examine the prevalence rates of, and motivations for,
anti-gay harassment and violence (Franklin, 2000). This study found that 1 in 10 had
committed an anti-gay act of physical violence or threat of violence, while another 24
percent had participated in anti-gay name calling.

These descriptive studies serve as the most comprehensive evaluation of the

extent and nature of IM crime in the United States. Unfortunately, there remain a number

22

of methodological and substantive issues hindering the advancement of knowledge about
this type of crime (Berk, Boyd & Hamner, 1994; Gruenewald et al., 2009; Silke, 2001).
Two interrelated challenges to current research on [M violence will be discussed below.
First, the state of the data on which past studies were based is problematic. Problems
include the lack of systematic data on [M violence and the biased and limited nature of
data that are available. Second, issues remain in defining and conceptualizing types and
subtypes of IM violence, including far-right perpetrated homicides.
Methodological Issues

The first methodological issue plaguing past research centers around the issue of a
general lack of data on all forms of IM violence, including far-right perpetrated
homicides. As noted, no study has evaluated the state of the data on all forms of 1M
violence. Although research is still limited, some have analyzed the state of research on
one particular type of IM crime, such as politically-motivated terrorism (Schmid &
Jongrnann, 1988; Silke, 2001). Though largely surveying research on international
terrorism, their findings also apply to research on domestic forms of terrorism. In
particular, scholars found that, partially due to the practical problems of collecting data
on terrorists, most scholars studying this topic do not generate their own data (Schmid &
Jongrnan, 1988; Silke, 2001). Instead, scholars most often rely on data collected by others
(Schmid & Jongrnann, 1988). Silke (2001), updating Schmid and Jongrnan’s (1988) study
on the state of political terrorism research, found that the situation had not improved, as
there remained a lack of terrorism data being generated by scholars.

These issues also apply to research on forms of politically-motivated violence in

the United States. To date, original data on forms of domestic terrorism, such as far-right

23

extremist violence in the United States, is very limited. In a review of far-right extremist
literature, scholars found that almost 40 percent of studies relied upon either no external
source of information (e. g. editorial style article or opinion piece) or a single source of
information (Gruenewald, et al., 2009). Moreover, over 40 percent of studies relied on
either academic studies or media sources as primary sources of information on far-right
extremism.

In contrast to research on politically-motivated ideological violence, substantial
increases in the amount and quality of data on violence motivated by hatred of social
minority groups have been made since the 1990 passage of the Hate Crime Statistics Act
(Nolan, Akiyama, & Berhanu, 2002; Strom, 2001). Both the UCR-Hate Crime Statistics
and the NIBRS data provide national data on 1M crimes. Despite apparent progress, there
are a number of problems with these datasets, including non-reporting and under-
reporting by some states (McDevitt, Balboni, Bennett, et a1, 2000). For example,
estimates of Specific types of IM violence by advocacy agencies have been higher than
those provided by official crime databases (see Jacobs & Henry, 1996). Moreover, the
data that are available provide a limited number of variables on each crime, offering little
insight into the many unknowns about the nature of IM crimes. In addition to a general
lack of data on IM violence, another primary issue highlighted in past research includes
reporting and recording practices of 1M violence.

Police Reporting

Past studies have revealed that the methods used by police to report and

categorize IM violence is influenced largely by police organizational considerations

(Boyd, Berk, & Hamner, 1996; Haider-Markel, 2002). Police agencies vary in the extent

24

to which they devote special attention and resources to investigating this form of crime
(McDevitt et al., 2000). Some police agencies like that in New York City had organized
specialized bias crime units in existence for decades, while other agencies have a single
or no officer to specifically address IM violence. Police agencies, also, vary to the extent
to which they adopt definitions of IM violence and crime policies (Grattet & Jenness,
2001), depending on a numbers of factors, including local sociopolitical contexts in
which police operate (Hamm, 1998). While some police organizations have no way to
report this form of crime, or have ambiguous policies on the issue (Berk et al., 1994),
other agencies have a concrete system for reporting IM violence. These variations in
agency definitions of IM crime and differences in policies affect the content of official
data on IM violence, thus affecting what we know about these forms of crime.

In addition to police agency influences on IM crime reporting, police reporting of
this form of crime is also shaped by individual officer perceptions of rules, on the spot
decision-making, and challenges to accurately assess offender motives (Haider-Markel,
2002). Boyd et a1. (1996), for example, found that some patrol officers, resenting the
increased paperwork required when violence was recorded, had an ideological-
motivation, as this extra work took emphasis away from more “serious” police work.
Furthermore, they argued that some police officers viewed most IM violence as relatively
unimportant and the result of simple “human nature” (p.827).

Biases also occur because of the challenges of sorting out complex motives from
IM violence and the differing ideas of officers on what constitutes such a crime (Boyd et
al., 1996). For instance, some officers rely on a “but for” criterion in that but for the

ideological motivation the crime would not have occurred, while other officers focus on

25

the proximate motivation in that the ideological motivation must be the most proximate
cause (Berk et al., 1994:131). Martin (1996), who examined police reports of IM violence
in Baltimore County and New York City, argued that police officers faced a number of
problems in reporting IM violence, including distinguishing between partial and sole
ideological motives. Indeed, police were often responsible for assessing conflicting
reports, as well as identifying intended targets of the violence, separating multiple
statuses, and determining whether there was provocation from the victim (Martin, 1996).
In short, police officers continue to face a number of challenges in regards to
investigating IM violence that can often go beyond the usual investigative duties.

Some types of IM violence may also be underreported because victims are less
likely to report victimization to police. For instance, Perry (2003) suggested that UCR-
Hate Crime Statistics underreport some crimes because gay and lesbian victims may not
want to report victimization and consequently “out” themselves. Similarly, the homeless
and immigrants in the United States illegally may not want to report victimization in
order to avoid police attention (Perry, 2003). It is also likely that law enforcement
underreport such crimes (Berk et al., 1994: 133).

Survey Data

In addition to the reporting and recording of IM violence in official data, there
remain a number of methodological issues concerning the use of survey data on victims
and offenders of IM violence. When discussing the weaknesses of survey designs, it is
useful to think in terms of how well those individuals surveyed represent the target
population of interest (Babbie, 1990; Dillman, 1978), and identify who may be

systematically excluded from the survey sample. One weakness of surveys is that they are

26

difficult to administer to certain populations (Babbie, 1990). Therefore, systematic biases
occur during sampling because it is difficult or impossible to locate a completely
representative sample of individuals to systematically survey. For instance, surveying
victims of IM homicide is impossible, while systematically surveying IM offender
populations would likely prove difficult as many are in highly secured prisons or
deceased.

Many of the survey studies reviewed above sampled members of the GLBT
community based on member lists of GLBT organizations. Therefore, persons identifying
as GLBT who did not belong to such groups, and who potentially differed from those
included in the sample, essentially selected themselves out of the study. The same notion
is true of Internet surveys, as those GLBT community members who do not regularly
accessing the Internet or particular websites would not be sampled.

Despite these data limitations, the empirical research on IM violence reveals that
this type of crime is a heterogeneous phenomenon. Unfortunately, most past research has
been largely unable to comparatively analyze IM violence subtypes due to definitional
issues, including if to adopt standard definitions of IM crime, varying inclusion criteria,
and distinguishing between perpetrator motivations (Berk et al., 1994; Boyd et al., 1996;
Jenness & Broad, 1997; McDevitt et al., 2000; Nolan et al., 2002). Therefore, in addition
to methodological issues, challenges in conceptualizing type of IM violence like far-right
perpetrated homicides have inhibited our understanding of this form of violence.
Conceptualizing 1M Violence

Another challenge confronting the study of far—right extremist perpetrated

homicide and IM violence in general has been the multitude of competing definitions of

27

IM violence. Scholars studying IM violence that is motivated by extremist political
beliefs acknowledge that there exists no single agreed upon definition of terrorism (see
Hamm, 1998; Silke, 1996; Weinberg, Pedahzur, & Hirsch-Hoefler, 2004). In an oft-cited
Study of theoretical and methodological issues of terrorism research, Schmid and
Jongrnann (1988) surveyed scholars working in the field and found 109 existing
definitions and twenty-two unique elements characterizing such definitions. Based on
these findings, Schmid and Jongrnan (1988) incorporated sixteen definitional elements
into a comprehensive definition of terrorism. As a result, they defined terrorism as:
“. . .an anxiety-inspiring method of repeated violent action, employed by (semi)
clandestine individual, group, or state actors, for idiosyncratic, criminal, or
political reasons, whereby—in contrast to assassination—the direct targets of
violence are not the main targets. The immediate human victims of violence are
generally chosen randomly (targets of opportunity) or selectively (representative
or symbolic targets) from a target population, and serve as message generators.
Threat-and violence-based communication processes between terrorist
(organization), (imperiled) victims, and main targets are used to manipulate the
main target (audience(s)), turning it into a target of terror, a target of demands, or
a target of attention, depending on whether intimidation, coercion, or propaganda
is primarily sought” (p.28).
Scholars studying IM violence against social minority groups have also faced the
challenge of having multiple definitions of this type of violence, and many have strove to
establish a general definition of this type of violence that can be universally applied. The
first widely acknowledged definition of violence targeting social minority groups in the
United States stems from the passage of the Hate Crimes Statistics Act of 1990, which
mandated the Attorney General to collect data on crimes motivated by race, religion,
sexual orientation or ethnicity (Hamm, 1998; Perry, 2003). The FBI, who has taken

charge of collecting national statistics on bias crimes, has defined this form of violence as

“- - .a criminal offense committed against a person, property, or society that is motivated,

28

in whole or in part, by the offender’s bias against a race, religion, disability, sexual
orientation, or ethnicity/national origin” (FBI, 2004). As is the case for political terrorism
in the United States, the FBI has provided an official definition of bias crimes that is
accepted by a number of criminologists studying this form of IM violence. This
standardized definition of violence targeting social minority groups has in many ways
facilitated the empirical study and building of knowledge on this form of violence in the
United States since 1990, as a number of authoritative studies on this topic have relied on
data from the F BI’s UCR-Hate Crime Statistics collected since the passage of the 1990
HCSA.
IM Violence as Heterogeneous

Rather than striving for universal definitions of IM violence, others have
emphasized the difficulties of fitting the heterogeneous category of IM violence under a
single definition like those provided by the FBI or local police agencies. One issue has
been which target groups to include as victims. For example, there has been substantial
debate on the inclusion of anti-sexual orientation crimes as a form of IM crime (Jenness,
2003). Others have, also, discussed the controversial issue of the inclusion of crimes
against women, such as rape, as an IM crime (Wolfe & Copeland, 1994). Moreover, it is
clear, the types of crimes considered to be ideologically-motivated depends on the
definitions and working policies of local police departments and other agencies in charge
of documenting this form of 1M violence (Hamm, 1998; Jenness & Grattet, 2005).

Another definitional issue that is faced by scholars studying IB violence is
identifying such crimes as identity-based through evidence of the perpetrators’

motivations. One of the most discussed definitional issues is the critical distinction

29

between what Berk et al. (1994) have referred to as “symbolic” and “actuarial” bias
crimes. Actuarial bias-motivated crimes are crimes that are motivated not solely by bias
or hatred toward social minority groups, but, also, other often instrumental motives (e. g.
robbery). Symbolic bias-motivated crimes, in contrast, are those that are motivated solely
by a perpetrator’s ideology rooted in structural inequality and hatred of social minority
groups. Symbolic IM crimes are expressive with potentially unique patterns and causes
that are unique from actuarial crimes. The distinction between symbolic and actuarial
forms of IM crimes targeting social minorities, and the tendency to gloss over these
differences, has fueled critics of bias crime legislation (see Jacobs & Henry, 1996;
Lawrence, 1999). In particular, Berk et a1. (1994) and other critics suggest that only
symbolic IM crimes should be considered as “bias-motivated,” and not distinguishing
these purest forms of IM crime targeting social minorities from actuarial/instrumental IM
crimes may result in the overestimation of the extent of this serious phenomenon.

In addition to various conceptual obstacles facing research on IM crimes against
social minorities, some scholars studying politically-motivated terrorism do not push for
“Holy Grail” definitions but, instead, recognize that there are multiple lenses through
which terrorism scholars can conceptualize terrorism. Schmid (2004), for instance,
suggests that terrorist motives can be mixed, both criminal and political, making it
difficult to distinguish between terroristic and non-IM criminal acts (see also LaFree &
Dugan, 2004). Furthermore, other scholars have argued that to understand politically-
motivated ideological violence it is necessary to identify and study separately its many
forms. As an example, in one of the most authoritative studies on American terrorism,

Smith (1994) recognizes multiple subtypes of terrorism in the United States. In particular,

30

Smith (1994) following the FBI, disaggregated first by international and domestic
terrorists that were operating on United States soil. Within the category of domestic
terrorism, he recognized three types of terrorism: left-wing, ri t-wing, and single-issue
terrorism. Furthermore, under right-wing terrorism, the author listed a number of
different types of groups that vary in a number of important ways such as ideology,
criminal activities, and other group dynamics.
F ar-Right Extremist Homicide Events

What is evident from Smith (1994) and others is that relying only on standard or
general definitions of IM violence ignores important similarities and differences among
its different types. Moreover, potentially unique offending patterns and situational
contexts of particular [M violence subtypes are ignored. One issue in past research on IM
violence critical to the current research has been the failure of past scholars to
conceptually distinguish far-right perpetrated violence from other forms of IM violence.
Nonetheless, while there has been no empirical study of the many subtypes of far-right
violence, some scholars have begun to typify, conceptually, various categories of far-
right violent events. Indeed, Heitrneyer (2003) has identified six types of violent events
that are committed by far-right perpetrators. Typification schemes such as these are
important because they serve as initial attempts to categorize far-right extremist homicide
events based on the current state of far-right extremist literature. Heitrneyer’s (2003)
typification scheme is outlined below in detail. Unfortunately, Heitrneyer’s categorization
of far-right violent events is limited for a number of reasons, as will also be discussed

below.

31

The first type of violent event perpetrated by far-right extremists that Heitrneyer
(2003) identified was opportunity-dependent violence often committed by members of
loosely organized youth gangs or groups. This type of far-right violent event was
described by Heitrneyer (2003) as non-strategic, unfocused, and situation-dependent.
Those perpetrating this type of violence are usually young and not official members of
organized far-right extremist organizations, but still maintain or are sympathetic to far-
right extremist beliefs. Targets of this type of violent event are apparent members of
social group minorities or those who are different in some way. This opportunity-
dependent violence mirrOrs what is often referred to as hate or “bias” crime (Levin &
McDevitt, 1993; Perry, 2001). One common motivation of far-right offenders
perpetrating this type of violence is what Levin and McDevitt (1993) have referred to as
“thrill-seeking.” One common scenario of this type of violence involves groups of young
males who socialize or “hang out” together, possibly drinking or using drugs, deciding to
target members of social minority groups for a thrill. Most often, this type of violent
crime event occurs in public, such as in a street or public park.

The second type of far-right violence is subcultural violence. Like opportunity-
dependent violent events, subcultural violence is perpetrated by members of youth gangs
or groups. However, this type of violent event occurs within the context of a youth
culture (Hamm, 1993). In the United States, the most applicable far-right extremists that
commit subcultural violence are skinheads. Skinhead subculture consists of core working
class values, as well as racist, religious, and gendered beliefs about members of social
minority groups. Skinhead culture is stylistically reflected in their appearance (i.e.

tattoos, shaved head), dress (i.e. Dr. Martens boots), music preferences, and often a

32

willingness to engage in violence to defend the skinhead subcultural beliefs (Blazak,
2001; Harmn, 1993).

Often these violent acts go beyond the situation-dependent violence and extend to
purposeful and pre-mediated attacks on social minority groups or members of other far-
right groups, such as anti-racist skinheads (Hamm, 1993). These attacks are seen as a way
to fiirther the skinhead cause, as well as serve to reify the subcultural values and beliefs
of skinhead groups.

The third type of IM violence perpetrated by far-right extremists is religious—
oriented violence (Heitrneyer, 2003). Far-right anti-religious violent events can vary
widely in their nature. Some events are motivated by the beliefs of extreme Christian
fundamentalism (Aho, 1990), and possibly focus attention targeting those associated with
abortion in the United States (Nice, 1988). Other religious-oriented violent events target
members of sexual minority groups. To illustrate, one anti-sexual orientation scenario
involves the victimization of gays and lesbians as they enter or leave establishments
catering to the GLBT community (i.e. “gay bars”).

The alternative beliefs of the Christian Identity religion have fueled a number of
past religious-oriented violent events targeting abortion doctors, frequenters of gay
establishments, and other symbolic targets (see also Barkun, 1997). Eric Rudolph, who
was responsible for the 1996 Olympic bombing and a number of religious-oriented pipe
bombings in the late 1990s, is a high-profile example of a religious-oriented extremist.

The fourth type of violent events is those perpetrated by the Ku Klux Klan.
Heitrneyer (2003) grants violent events perpetrated by the Ku Klux Klan a distinct

category as they have a long history of far-right extremism and remain the poster-child

33

for bigoted violence in the United States (Maguire, 1995). Most often targeted by KKK
members are Jewish and African Americans. While the majority of KKK activities do not
involve physical violence but, instead, symbolic intimidation tactics (e. g. cross-buming),
many members associated or claiming to be affiliated with the KKK have been
responsible for murder and other violent occurrences.

The fifth type of violence is perpetrated by individuals with a background in the
far-right whose primary target is the government. Although this type of event often
challenges government authority, the perpetrators often have no ambition for attaining
political power (Heitrneyer, 2003). A number of types of far-right groups commit acts
against the government. Perpetrators of this form of violent events are often paramilitary
militias (Heitrneyer, 2003), though most militias in the United States do not actively
engage in violence against the government or members of social minority groups
(Chermak, 2002). However, some groups that espouse Christian Identity beliefs and
maintain a paramilitary structure have turned violent in the past. For example, a
paramilitary group known as The Order went on a violent crime rampage in the mid-
1980s, including killing a left-wing talk show host and committing robberies to firnd an
effort to overthrow the government (Kaplan, 1995).

Finally, the sixth type of far-right violence is perpetrated by right wing “pogroms”
(Heitrneyer, 2003). Far-right violent events committed by these offenders are not
necessarily committed by far-right extremists attached to groups, but by individuals or
small group factions who have become violent. The events are motivated by individuals

Who harbor deep resentment for social minority groups and the government. Heitrneyer

(2003) suggests that this fornr of violence is eruptive and can result in serious harm to

34

victims. Examples of this form are killings and bombings that result in mass casualties at
the hands of far-right extremists, such as the 1995 Oklahoma City bombing (Gruenewald,
2004; Michel and Herbeck, 2001).

Although Heitrneyer’s (2003) typification scheme can serve as a starting point in
the categorization of far-right violent events, substantial issues remain. In particular, the
categories are not rooted firmly in empirical research. That is, while Heitrneyer (2003)
provides overall numbers on 1M crime in the United States and across the world, the
extent and nature of the multiple subtypes of far-right extremist violence are ignored. In
fact, there exists very little empirical research on far-right violence to base the naming of
unique types of violence (Gruenewald et al., 2009). Also, Heitrneyer’s typification
scheme glosses over potentially important subtypes of violence perpetrated by far-right
extremists, such as non-1M violence fueled by traditional motivations, as well as other
types of violence fueled by both ideological and traditional motivations. How these
subtypes of far-right violence compare is virtually unknown. Finally, it is possible that
categories of violence may vary by the particular type of violence committed by far-right
extremists. If, for instance, homicide and less serious types of assaults follow the same
plotlines and maintain the same motive structures remains an unanswered empirical
question. In sum, in addition to methodological obstacles, prior descriptive research on
far-right perpetrated homicide and IM violence in general has been stunted by
disagreements and ambiguities in conceptualizing this form of crime.

Heitrneyer’s (2003) typology of far-right violence events is an important
contribution to the study of far-right extremist violence. Nonetheless, there are likely to

be types of violent events perpetrated by far-right extremists that do not fit nicely within

35

these categories, as well as types of violence that have yet to be categorized. Before
discussing how the current study overcomes these obstacles in Chapter 3, it is necessary
to review a small literature that has proposed a number of specific theories of offender
motivations. The theoretical challenges facing this literature are also addressed in the
next chapter.
Offender-Based Studies

In addition to descriptive studies, other studies have focused primarily on
explaining offender motivations for committing IM crimes. This research has assumed
that key antecedent variables are able to best explain (or predict) why offenders commit
IM violence. The associations between offender attributes and criminological behavior
have often been modeled as simple linear cause and effect relationships. To date, there
have been few studies explaining why offenders commit IM crime, and even less have
addressed those specifically committed by far-right offenders. Therefore, what follows is
a review of past studies that have primarily focused on providing explanations for [M
offending. Following a brief review of past offender-based studies, the failure of this
research to consider how social contextual factors shape the nature of IM offending is
discussed.
Relative Deprivation

One of the most important offender-based studies on 1M offending is by Levin
and McDevitt (1993), who examined [B offending in Boston. They analyzed case files
from the Community Disorders Unit of the Boston Police department for offenses with a
known offender between 1991 and 1992. All officers in this particular police unit were

trained to identify [B crimes. Based on Boston police reports, Levin and McDevitt (1993)

36

created a three-pronged typology based on their apparent motivations. The first type of IB
offender was the “thrill offender.” This type of offender acted because he or she was
bored and seeking some sort of high or increased level of excitement. The second type of
offender was the “turf defender” who viewed themselves as defending their
neighborhood, property, and way of life fiom outsiders or “others.” Finally, the third type
of offender was the “mission offender.” Mission offenders were motivated to commit [B
violence to rid the world of impurities, including racial, religious, and sexual orientation
minorities. Levin and McDevitt (1993) found that mission offenders tended to be
associated with the far-right extremist movement and adherence to far-right extremist
beliefs.
Almost a decade later, the authors reexamined 169 of the case files used in their
prior study to update their typology of IM offenders (McDevitt, Levin, & Bennett, 2002).
In their more recent study, McDevitt et a1. (2002) included retaliation as a category of
offender motivation. Retaliatory offenders were motivated by real or perceived attacks
against offender’s social group identity. However, of all the offender motivation
categories, they still found that the most common offender type was the thrill-seeker.
One theme that ran throughout Levin and McDevitt’s work is the importance of

relative deprivation as an underlying mechanism for IB offenders. As struggling, working
class young adults find themselves facing joblessness and an inability to achieve the
“American Dream,” while simultaneously perceiving certain social minorities succeeding
and receiving “undeserved” advantages, some feel personally threatened and resentful.
They suggested that either in the context of the workplace or neighborhood, social

minority groups are sometimes used as scapegoats for personal and societal economic

37

woes and general feelings of powerlessness and despair. According to Levin and
McDevitt (1993) relative deprivation also remains an underlying factor for many of those
offenders more immediately motivated by thrill seeking, defending turfs, racial
ptu'ification missions, and retaliation.

Unfortunately, Levin and McDevitt (1993) were only able to examine IB crimes
in one US. city. Thus, it is possible that their typification scheme for [M offenders may
not apply to other cities or regions of the country. Moreover, Levin and McDevitt (1993)
were not able to examine the many subtypes of far-right homicide offenses, a central aim
of the current research. While this study did recognize that “mission offenders” tended to
be far-right extremists, little is known about the complex nature of far-right extremist
homicide events.

Low Self-Control

In another study, Messner et a1. (2004) applied Gottfredson and Hirschi’s (1990)
self-control theory to test whether IB offenders were more likely to have increased self-
control compared to traditional offenders. They exarrrined anti-racial assaults data from
the National Incident-Based Reporting System to test competing models for explaining
[B offending. One model referred to as the “specialist model” maintained that offenders
were calculating, future-oriented and tended to specialize in particular forms of crimes.
Specialist offenders were driven primarily by their religious, social, and political
ideologies in their crime commission. In support of this model they drew on group
conflict theory to explain how IB crimes served as an expression of group conflict that
could send messages to competing social minority groups (Blalock, 1967). Empirical

support of the specialist model has come from other studies that have found, for instance,

38

that skinhead offenders arrested for [M crimes have generally not committed other non-
IM types of prior crimes (Blazak, 2001; Hamm, 1993).

The competing “versatility model” was based primarily on Gottfredson and
Hirschi’s (1990) self-control theory that maintains that offenders tend to not specialize in
particular forms of crime, but remain versatile in their offending behaviors. That is, while
offenders may maintain some form of prejudice against particular social minority groups,
they are not primarily motivated to commit crime by their ideological convictions.
Instead, offenders tend to be driven by other, more routine criminal motivations. This
model maintains that [M offenders remain as opportunistic as other offenders, but add
that they may be more likely to choose social minorities as worthy targets. Moreover,
bigoted convictions may in some instances serve as a facilitating or disinhibiting factor.

To test these models, Messner et a1. (2004) argued that if IB offenders used
alcohol and drugs as much as routine offenders, then a versatility model would be
supported. Interestingly, Messner et a1. (2004) concluded that because anti-racial assault
offenders were actually more likely to use drugs and alcohol rather than less, their study’s
findings supported the versatility model as opposed to a specialist model of IB offending.
Interestingly, they also found that IB offenders were more likely to injure their victims of
racial assault compared to traditional offenses (assaults). This finding supports past
descriptive research, discussed above, that has found that some types of IB violence tend
to be especially serious (Comstock, 1991). Unfortunately, Messner et a1. (2004) only
examined racial assaults, thus comparisons with other types of IB violence were not

possible. In addition, due to issues with the data used in their study, only racial assaults in

39

eleven states were considered, thus limiting opportunities for making generalized
conclusions.
Modified Social Control Theory

In an innovative study of skinhead violent offending, Hamm (1993) developed an
integrated theoretical approach to explain the process of how ordinary young working-
class males evolved into domestic terrorists. Hamm (1993) combined functionalist,
differential reinforcement, and neo-Marxist cultural theoretical orientations to ultimately
explain the domestic terrorist activities of American skinheads.

Using native and prison field work techniques for contacting skinheads, Hamm
(1993) conducted a nationwide survey on terrorist activities by skinhead group members.
Based on these observations, Hamm (1993, 1994) observed that members of skinhead
groups were part of a working class culture. That is, skinheads did not avoid hard work,
but instead valued blue-collared work. Hamm (1993, 1994) theorized that in the course of
socialization, individuals developed the idea that they were victims of injustice through
various forms of media, such as White Power rock music, books, and movies, and were
forced to form reified beliefs about how minority groups have caused this victimization.
These distorted beliefs formed the social reality of far-right domestic terrorist
subcultures, as members also began to adopt a particular skinhead style of dress and
behavior. Therefore, Hamm (1993) suggested that skinhead responses to social minority
groups were shaped by perceived victimization and the subcultural processes they
adopted and experienced as a way to deal with such victimization. In some instances,
skinheads were triggered to commit violent acts against social minorities by heavy

consumption of beer drinking during times of male bonding (aka “berserking”). Hamm

40

(1993, 1998) referred to these acts of violence as acts of domestic terrorism, as they were
messages of morality intended to reach wider audiences.

Hamm’s (1993) ethnographical study of skinheads is one of the only in-depth
analyses of [B offending that attempts to provide a detailed and nuanced theoretical
explanation of skinhead violence. However, as also indicated by Perry (2001), Hamm’s
(1993) conclusions about skinhead violence are difficult to generalize, especially
considering the multitude of individuals who may identify as a particular variation of a
skinhead (e. g. neo-Nazis, racist, non-racist skinheads). Furthermore, little remains known
about other types of violence perpetrated by other types of far-right extremists (e. g.
militia members, lone-wolf offenders).

Neutralization Theory

Byers, Crider, and Biggers (1999) applied neutralization theory to explain why
offenders committed crimes against a particular religious group, specifically the Amish.
They relied on a small number of confidential, face-to-face interviews to examine IB
violence targeting Amish in one Indiana County. The focus of this study was on whether
or not offenders participating in a form of 1B violence known as “claping,” (i.e. throwing
objects at Amish buggies or other property for purposes of harassment and intimidation)
relied on neutralization techniques to rationalize their behaviors.

They hypothesized that offenders relied upon neutralizations, or linguistic
accounts of justification to rationalize their violent behavior. Neutralization theory stems
from Sutherland’s (1947) theory of differential association, as neutralization or
rationalization was described as one learned behavior favorable to the commission of

crime. Neutralization theory was first proposed by Sykes and Matza (1957) as an

41

explanation for how offenders rationalized their deviant behaviors. Assuming that
criminal offenders share the same societal values as noncriminals, neutralization theory
suggests that offenders rely on neutralization techniques in order to temporarily distance
themselves from such values in order to commit crime. Supporting Levin and McDevitt
(1993), Byers et a1. (1999) found that thrill-seeking was a common explanation for
participating in crimes against the Amish. Supporting neutralization theory, IB offenders
tended to rely on various techniques to rationalize their violent crimes against the Amish.
As further discussed below, however, the authors failed to consider situations in which
certain contextual factors shaped potential offenders’ decisions not to commit IB violence
against the Amish. In effect, a somewhat simplistic explanation for anti-Amish violence
is offered for a potentially complex phenomenon.
Defended Neighborhoods Theory

In another study, Green, Strolovitch, & Wong (1998) proposed a “defended
neighborhood” explanation for anti-racial offending at the aggregate level. To do so, they
examined correlates of anti-racial crime at the neighborhood level in New York City
between 1987 and 1995. A number of related theoretical models were presented to
explain aggregate rates of racial offending. One model they proposed was referred to as
the “defending neighborhood” theory, which predicted that neighborhoods having a long
history of a predominantly white population and experiencing an influx of a particular
racial/ethnic minority group were more likely to experience anti-racial IB crime. The
reason for the increase in IB crime is that under certain circumstances some Whites living
in these rapidly changing neighborhoods feel as if it is necessary and appropriate to

“defend their turf” from being taken over by racial minority groups.

42

In their study, Green et a1. (1998) found support for the defended neighborhood
theory as anti-racial crimes were most common in mostly White geographic areas that
had experienced major demographic change in the form of in-migration of racial
minorities. For example, anti-racial crimes in New York City neighborhoods that had
longstanding racial minority populations had less anti-racial crime. Surprisingly, they
found that socioeconomic indicators have little influence on this particular form of [B
violence at the aggregate level. However, as further discussed below, Green et al. (1998)
failed to consider how social contextual factors like economic conditions may have
facilitated IB offending. Included in the next section are the issues of contextual factors
and offender motivations (and decision-making) interacting to shape IB offending and the
absence of these issues in offender offender-based studies.

Critigue ot Otlender-Based Theories

Due to the lack of empirical research on far-right perpetrated homicide and IM
violence in general, any attempt at theoretical explanation of IM offending may be
considered a contribution to the literature at this point in time. All works discussed thus
far have applied established theories or combinations of theoretical approaches and
skillfully applied them to some type of IM violence. However, it may also be useful for
this research to establish some guidelines or criteria to gauge productive theoretical
development.

Irnportantly, scholars’ criteria used for evaluating theoretical contributions are
likely to vary depending on the intended goals of the theory. For instance, the goal of the
broad social learning perspective is to provide a general theory of crime to explain why

offenders commit deviant acts. Akers and Jensen (2006) suggested that the criteria for

43

good theory include logical consistency, scope of theory, usefulness, and strength of
empirical support. On the other hand, another volume devoted to the goal of developing
theories of violence suggests that parsimony, originality, generalizability, level of
empirical support (or testability), and validity were the capstone criteria for good theory
(Jackson, Zahn, & Brownstein, 2004). What if, though, the goal of a particular research
project is not only to explain offender behavior but, also, to provide a more in-depth
understanding of a particular type of crime situation? Indeed, the current study intends to
advance theoretical explanations to understanding the complex and multi-dimensional
nature of far-right violence. Therefore, criteria for good theory need to reflect this
purpose.

A criterion for good theory is the consideration of the context in which criminal
behavior occur (Morash, 2006). However, many of the offender-based theories applied to
IM violence reviewed above have ignored the contexts in which IM offending occurs.
Violence always occurs within particular contexts, including macrosocial or historical
contexts, intermediate or organizational contexts, and microsocial or situational settings
(Morash, 2006; see also Short, 1997). Identifying the contextual factors can reveal how
characteristics of crime situations facilitate, hinder, or in some way shape criminal
outcomes.

In addition to ignoring situational contexts, past studies advocating for particular
theories of 1M offender motivations have most often deemphasized or ignored offender
agency, or the ability of criminal actor to act in a self-directed way (Morash, 2006).
Theories of violence that assume that people do not have agency can be considered

deterministic and tend to present an over-socialized conception of individuals. Theories

44

that consider the role of human agency reject overly deterministic explanations of why
crimes do or do not occur while also considering the various constraining factors that
operate to limit and Shape the decisions and behaviors of crime participants.

Past studies attempting to explain IM offender motivations have largely failed to
consider the context of offending and the ability of participants to act in a self-directed
way in particular situations. For instance, Green et al.’s (1998) aggregate analysis of the
effect of economic conditions on IB offending ignored how varying economic conditions
affect individuals’ decisions to commit [B offenses in particular situations. Their study
did little to advance an understanding of the different Situations where socioeconomic
factors facilitate and hinder IM offending. Similarly, Byers et al. (1999) focused
exclusively on applying neutralization theory as an explanation for IB offending against
the Amish, and failed to discuss how certain situational contextual factors could cause the
young offenders to decide not to commit such violence. Considering offender motivations
in combinations with these other factors would lead to more holistic understandings of
the entirety of IM violent events.

Conclusion

This chapter reviewed the limited body of past research on 1M offending. While
this research focuses on homicides perpetrated by individuals and groups associated with
the far-right extremist movement, there is currently no empirical research on this serious
form of violence. Therefore, the majority of past research was on IM offending in
general, as far-right extremist homicide offending is one subtype of IM offending. First,
descriptive studies of IM violence were reviewed. While data on IM crime is increasing,

it is clear that empirical research has been inhibited by difficulties in conceptualizing 1M

45

violence and an inability to distinguish far—right perpetrated violence from other types of
IM violence. Second, some scholars have attempted to explain offender motivations for
IM offending. This chapter revealed the limitations of offender-based theories that
neglect issues of crime contexts, crime participant agency, and the multi-dimensionality
of crime situations and participant statuses. The next chapter introduces the research
approach for the current study. In particular, the criminal event perspective is proposed as
a broad theoretical and methodological approach to studying far-right homicide events

that overcomes many of the issues faced by past scholars.

46

CHAPTER 3: RESEARCH APPROACH

The current research focuses primarily on one type of IM violence, specifically
homicide events committed by perpetrators associated with the far-right extremist
movement in the United States between 1990 and 2008. A set of other IB homicide
events not perpetrated by far-right extremists and a random sample of all homicide events
(representing typical or “average” homicides) are also examined for comparative
purposes. As previously stated, the general purpose of the current study is to facilitate an
in-depth understanding of the complex and heterogeneous nature of far-ri ght extremist
homicide. Chapter 2 revealed that there remains a number of methodological, theoretical,
and other substantive issues that have thus far hindered the study of this topic. In
particular, offender-based theories have often ignored the situational contexts that shape
and determine the outcomes of criminal events. In addition, offender-based theories have
failed to consider interactions that occur between crime participants that also shape fatal
outcomes. Therefore, in this chapter, a different research approach for studying the most
serious form of far-right extremist violence is proposed. This chapter introduces a broad
theoretical perspective that conceptualizes [M homicide as a multi-dimensional event and
emphasizes the dynamic interchanges between criminal actors in particular situational
contexts (Block, 1981; Sacco & Kennedy, 2002; Meier et al., 2001). First, situational
theories of crime that conceptualize violence such as far-right homicide as
multidimensional events are outlined. Second, how these situational theories make up a
broader criminal event perspective (CEP) and are utilized in this study to advance

knowledge and facilitate an in-depth understanding of far-right perpetrated and other IM

47

homicide events is discussed. Finally, general and specific research questions that guide
the current research are presented.
Situational Perspectives

This study approaches the study of far-right extremist homicides from a
situational crime perspective, or a theoretical perspective that emphasizes the multi-
dimensional nature of crime events (Block, 1981). While criminologists have in the past
studied homicides from a situational perspective, to date there has been no study of far-
right extremist homicides or other form of [M homicide from a situational perspective
(see also LaFree & Dugan, in press). Criminologists have historically acknowledged that
crimes occur within situational contexts and many have considered crime situations
worthy of scholarly research (Birkbeck & LaFree, 1993). In fact, the pioneering
criminologist Edwin Sutherland (1947) suggested decades ago that situational
explanations may be more important in explaining crime than offender-based theories.
Moreover, in their advancement of the popular self-control theory of crime, Gottfredson
and Hirschi (1990) suggested that while some offenders may be more or less predisposed
to criminal behaviors, situational elements are pivotal for providing the opportunities to
commit crime. However, to date, few criminologists have directly incorporated
situational factors into their theoretical explanations of crime (Birkbeck & LaFree, 1993).
For example, Sutherland’s differential association theory and Gottfredson & Hirschi’s
(1990) self-control theory, despite both recognizing the value of situational factors, focus
primarily on uni-dimensional theories to explain why offenders commit crime.

The situational crime perspective is not a specific theory of criminal behavior, but

a broad framework, and a number of more specific theories of crime and criminality can

48

fit under this umbrella (Sacco & Kennedy, 2002). Thus far, there have been four types of
situation-based theories or families of theories that conceptualize crimes as multi-
dimensional situations or events (see Meier et a1, 2001; Meithe & Regoeczi, 2004). First,
rational choice theories have conceptualized crime situationally rather than as uni-
dimensional offender acts, and have emphasized the rational decision-making abilities of
criminal offenders in particular situations (Clark and F elson, 1993). Second, the social
interactionist approach has focused primarily on the interactions between crime offenders
and victims (Tedeschi and Felson, 1993). Third, symbolic interactionist theory has been
relied upon by not only to explain how victim-offender interactions shape crime
outcomes but, also, how crime actors’ subjective interpretations of situational dynamics
can allot meaning to crime event processes and outcomes (Birkbeck & LaFree, 1993;
Goffrnan, 1967; Katz, 1988). Finally, there have been a limited number of attempts in
past research to integrate routine activities, social interactionist, and symbolic
interactionist theoretical perspectives to provide a more comprehensive explanation of
crime events (Miethe & Meier, 1994). Each of these theories or combinations of theories
alone provide only a limited explanation of criminal events. Each one of these situational
crime perspectives and how they do and do not address the multiple dimensions of
criminal events is briefly reviewed below.
Rational Choice

The rational choice perspective assumes that criminal actors make rational
decisions to avoid or participate in crime based on the situations in which they find
themselves in order to avoid negative costs and increase positive benefits for themselves

(Clarke and Felson, 1993). Other theoretical developments operating at the individual

49

level, such as opportunity and life-style theories (Hindelang, Gottfredson, & Garofalo,
1978), extend rational choice theories and explain how situational factors increase and
decrease the likelihood that crime events will occur by shaping how people spend their
time, energy, and money (see Sacco & Kennedy, 2002:37). Other neighborhood—level
theoretical explanations have focused on how physical settings and social contexts help to
explain neighborhood crime rates (e.g. broken windows theory), and effective responses
to crime hot-spots (Sherman, Gartin, & Buerger, 1989).

Routine activities theory is one of the most popular explanations of crime
occurrences rooted in the rational choice perspective (Cohen & Felson, 1979). This
theory maintains that the likelihood of crime occurring increases when one or more of
three conditions is absent in a particular time and place: 1) presence of capable guardians
2) motivated offenders and 3) suitable targets. While this theory has become increasingly
applied to a number of crime types, scholars advocating for a Situational perspective have
suggested that routine activities theory is limited in its ability to provide comprehensive
explanations for a number of reasons. One of the most crucial limitations of this theory is
that it tends to focus exclusively on the victim dimension of crime events, or the
situational factors that make victimization more likely to occur, thus ignoring variations
in offender motivations and opportunities (Birkbeck & LaFree, 1993). Related, routine
activities and other rational choice theories have been criticized for their tendency to
assume that offenders are constantly motivated (or passive) in regards to committing

crime given the opportunity and right circumstances (Jensen & Brownfield, 1986).

50

Social Interactionism

Other scholars have focused less on the victim dimension of crime situations, and
more on the interaction or interchanges that occur between victims and offenders.
Transactions are the “. . .interactions between individuals that have led to the outcome of
crime” (Sacco & Kennedy, 2002: 61). Social interaction theories combine offender-based
theories that can help explain what actors bring into particular situations and routine
activities theory that help explain how violent outcomes are linked to the routine
activities of victims and offenders (T edeschi & Felson, 1993). Social interaction theories
build on rational choice and routine activities theories by considering not only the
offender or victim dimensions of violent events in isolation but, also, the shaping of
criminal outcomes by dynamic transactions between victims and offenders through the
cost-benefit decision-making of criminal actors. Scholars advocating for more
comprehensive situational explanations of crime events, however, have argued that social
interactionist theories of violent events only consider the interaction between victims and
offenders and neglect to consider the subjective interpretations or meanings attached by
victims, offenders, and other actors to their experiences that, also, shape the outcomes of
violent events (Birkbeck & LaFree, 1993).
Symbolic Interactionism

In addition to considering the social interactions of criminal actors, other theories
are more adept at considering the meanings that actors attach to the interactional elements
of crime situations and how these meanings shape crime outcomes. When offenders are
motivated, victims are suitable, and the contextual factors are bent toward crime, criminal

behavior is more likely to occur (Cohen & Felson, 1979). Furthermore, criminal actors

51

are more likely to decide to commit crime when it benefits them in some way (Clark and
Felson, 1993). However, violent events such as homicide are ultimately social events, as
they involve interactions between people. Therefore, offenders are not determined to
commit violence even when situational elements facilitate such behavior.

To better explain how actors’ interpretations of event circumstances shape their
behavior, Some scholars have advocated the theory of symbolic interactionism to theorize
how offenders’ subjective interpretations of situations shape the nature of crimes (Athens,
1980; Blumer, 1969; Lofland, 1969). Like other interactionist theories of crime, symbolic
interactionism assumes criminal behavior occurs in situations by rational actors, but
places more emphasis on the offenders as active interpreters of situational conditions and
the important role these interpretations can play in shaping violent crime occurrences. In
one study, Athens (1980) examined homicide events from a symbolic interactionist
perspective and argued that past positivist perspectives on this topic failed to
conceptualize actors as rational and active interpreters of their experiences. Athens
interviewed fifty-eight prison inmates convicted of homicide to examine their
interpretations of homicide events based on their memory. Based on these interviews,
Athens developed a typology of homicide suspects and events.

Symbolic interactionism, however, is not without its problems. In particular,
Birkbeck and LaFree (19932113) have suggested that symbolic interaction theory
provides no explanatory links between the important concepts of opportunities for crime,
motivations for crime, and criminal outcomes. In addition, some of the most important
examples of criminological applications of symbolic interactionism to crime occurrences

have been labeled as anecdotal and non-generalizable (Katz, 1988), and providing only

52

descriptive and almost trivial explanations for criminal occurrences (Athens, 1980; see
Birkbeck & LaFeee, 1993).
Integrated Approaches

A limited number of scholars have also attempted to combine elements of
situation-based theories in order to provide more comprehensive explanations of crime
events that overcome shortcomings of each separate approach. Miethe and Meier (1994),
as an example, provided an integrated model that assumed that routine activity patterns
and lifestyles often increased exposure of victims to motivated offenders and risky
situations. Moreover, they argued that subjective values and interpretations of actors and
the level of guardianship determine choice of particular crime targets (p.27). In this way,
Miethe and Meier (1994) overcame many of the limitations of each individual situational
theory by integrating key aspects of each.

Luckenbill and Doyle (1989) also advanced an integrated perspective by
combining notions of culture (i.e. “culture of honor” or subculture theory), social
structural position, situational settings (i.e. public versus private settings), and finally,
symbolic interactionism, as violence was theorized to be more likely when the offender’s
“self” was attacked. They argued that there was variation in individuals’ associations
with the culture of violence and that occupants of varying structural positions differed in
levels of “disputatiousness” and aggressiveness. Luckenbill and Doyle (1989) argued that
these differences interacted and tended to be most pronounced when the self was attacked
in a public setting.

Sacco and Kennedy (2002) have suggested that each of these different theoretical

theories or combinations of theories remain useful for explaining and furthering

53

understanding of crime events. However, as indicated above, each theoretical strain is
associated with specific shortcomings that in some ways inhibit them from providing a
comprehensive account of the entire criminal event. Recently, scholars like Sacco and
. Kennedy (2002) have promoted a broad research approach referred to as the criminal
event perspective (CEP). The CEP synthesizes and accommodates these and other
theoretical perspectives in order to establish a framework for providing more holistic
descriptions and explanations for complex criminal events occurring at particular times in
particular places (see Meier et al., 2001). The current study relies on this perspective as a
theoretical and methodological tool for examining far-right extremist homicide events.
Criminal Event Perspective (CEP) and Far-right Homicide

Drawing from situation-based theories, the criminal event perspective (CEP)
serves as a broad theoretical approach that conceptualizes crime as multi-dimensional
events (Block, 1981; Sacco & Kennedy, 2002). Criminal events are constituted by the
dynamic interdependencies between criminal actors and the situational contexts in which
they interact. It is assumed that criminal events occur in a particular time and place, and
these geographic and temporal factors are assumed to be intricately connected to criminal
events (Wilkinson & Fagan, 2001). The CEP also maintains that criminal events develop
through stages and are constituted of precursor factors, dynamic interactions between
victims, offenders, and other actors, as well as an event aftermath (Sacco & Kennedy,
2002). Precursors to criminal events, for example, consist of physical settings and other
situational variables that shape criminal outcomes. Sacco and Kennedy (2002:309) have
also suggested that “[a]s social events, crimes involve not only offenders but also other

participants, including victims, the police, witnesses, and bystanders, all of whom act and

54

react within particular social and physical circumstances.” In short, this perspective
assumes that crimes cannot be separated from their social and physical settings
(Brantingham & Brantingham, 1993; Miethe & Meier, 1994; Sacco & Kennedy, 2002),
and both the social aspects of interactions and the symbolic meanings attached to
situational elements are considered important factors that shape criminal outcomes.

The current research applies the CEP to the study of far-right extremist homicides
to provide a more holistic understanding of the contextual factors and the processes
involved in how far-right extremists target victims in particular situational contexts. In
this way, the CEP directly addresses the criticisms of offender-based theories by
assuming that there is more to crime than uni-dimensional offender behaviors (Sacco &
Kennedy, 2002:309).

Every far-right extremist homicide is considered a unique situation or event. In
every situation, far-right extremist offenders bring their predispositions which may
enhance or inhibit the likelihood of engagement in criminal behavior. In this way,
offender-based theories remain an important consideration within the CEP. Moreover,
victim behaviors are shaped by routine activities and lifestyle choices that decrease or
increase their chances of being victimized in various situations. The interactions between
far-right perpetrators, victims, and others involved in the homicidal situation are shaped
by a number of situational factors, such as physical and social settings, the presence of
weapons, crime participants’ alcohol and drug use during the event, and many other
factors that facilitate or hinder criminal outcomes. Homicide event actors are not
considered passive agents but, instead, remain active interpreters of the many situational

elements in which they interact. The subjective interpretations of actors’ demeanors and

55

behaviors affect the behaviors of other actors as all involved attempt to accomplish a
number of goals (e. g. saving face, establishing masculinity, etc.) (Athens, 1980, Birkbeck
& LaFree, 1993; Katz, 1988; Polk, 1994), which depend on the immediate needs of each
actor (Tedeschi & Felson, 1993).

There are a number of advantages to examining far-right extremist homicide fi'om
the criminal event perspective. One major advantage is that the CEP considers far-r1 ght
extremist homicide events to be essentially social constructions. A number of scholars
have already examined the social construction of IM crimes by focusing on the
relationships between historical social movements, social movement organizations,
legislation, court rulings, and police decision-making (e. g. Grattet & Jenness, 2001;
Jenness & Broad, 1997; Jenness & Grattet, 2005). This research has shown how the
involvements of these institutions in the definitional processes have resulted in IE crimes
being a “new” and evolving category of crime. Thus, while these homicides have very
“real” consequences, how they are defined, reported, and studied is a product of the
actions of number of claimsmakers who have an interest in how IM homicide is defined.
Recognition of the socially constructed nature of IM homicide, and the various subtypes
of IM homicide, suggests that scholars should not rely only on official data provided by
police agencies or definitions promoted by any single government agency, advocacy
group, or other claimsmaker. Instead, scholars should be free to seek out multiple sources
of information on IM homicide events and define and categorize far-right perpetrated
homicides and other forms of homicide in ways that promote more comprehensive

understanding of this problem’s actuality in response to more nuanced research questions.

56

Because the CEP does not necessarily rely on legal definitions of crimes, it is
possible to disaggregate socially constructed crimes, such as far-right extremist homicide
in order to comparatively analyze the different subtypes of this type of violence based on
specific research questions important in advancing a comprehensive understanding of this
phenomenon. Past research on traditional forms of homicide has suggested that
comparatively analyzing homicide subtypes is desirable because it increases our
understanding of homicide, as different types of homicide have different patterns,
explanations, and situational contexts (Decker, 1993; F lewelling & Williams, 1999,
Pizarro, 2008). Another reason to disaggregate is the risk of serious policy consequences
for over-generalizing the effects of various factors in terms of all homicides (Flewelling
& Williams, 1999; Maxfield, 1989). In short, different forms of homicide and homicide
offenders may be associated with very different characteristics, making it difficult to
respond to all homicides in the same way.

There have been a number of ways that homicides have been disaggregated in
past criminological research. In particular, a number of homicide studies have
disaggregated homicide events by victim-offender relationships (Smith & Parker, 1980),
motive (Decker, 1993; Pizarro, 2008; Riedel, 1987), victim types, such as patricide,
infanticide (see Riedel & Welsh, 2001), circumstances, such as gang involvement
(Maxson, Gordon, & Klein, 1985; Decker & Curry, 2002; Pizarro & McGloin, 2006),
race of participants (Parker, 2001) and age of homicide participants (Lee & Bartkowski,
2004). From a cursory review of the literature, it appears that homicides are typically
disaggregated by victim-offender relationship and homicide circumstance categories (see

also Maxfield, 1989).

57

Despite all of the discussion on various ways to disaggregate and examine
subtypes of homicide, it is also evident that there is no “Holy Grail” of classifying
homicide types (F lewelling & Williams, 1999). All categories mask differences, and
there are a number of issues related to the quality of homicide data that affect the ways in
which forms of homicide can be comparatively analyzed. Loftin, Kindley, Norris, &
Wiersema (1987), for instance, have argued that there is a lack of data to disaggregate
due to incomplete police data and incomplete and overlapping categories capturing the
reality of homicide. Moreover, research categorizing homicides will always encounter
homicides that do not fit neatly within categories, or are not easily explained by theories
attached to particular types of homicide (Decker, 1996). Nonetheless, the issues discussed
by Loftin et a1. (1987) and others should not be viewed as reasons to halt comparative
analysis of homicide subtypes. Instead, their recognition of the many issues in comparing
homicide characteristics serves as a charge for the current research. For instance, as
Decker (1996) suggests, the “deviant” homicides that do not fit within particular
designated categories, and which are qualitatively different from other occurrences within
the same category should become the object of study. For instance, racially-motivated
ideological homicides, an expressive homicide occurring between strangers, should not
be treated as simply an exception, but, instead, as a unique form of homicide worthy of
exploring (Decker, 1996). Moreover, this finding should indicate that new categorization
schemes need to be explored in order to identify how homicide occurrences compare.

Indeed, rather than relying solely on the categorization schemes of past research,
it is necessary to explore new ways to comparatively examine deviant homicides, such as

various subtypes of IM homicides. Categorization schemes should depend on the research

58

questions being asked and the ultimate goals of the study (Flewelling & Williams, 1999).
Therefore, the current study disaggregates IM homicides by the perpetrators affiliation or
non affiliation with the far-right extremist movement, as well as by the social minority
group targeted. In order to further understanding of this type of violence, it is necessary
to develop new categorization schemes, new forms of homicide data, and new research
designs that align with the research questions being asked.
Conceptualizing IM Homicide Events

This study conceptualizes IM homicide as a multidimensional phenomenon,
crossing boundaries of homicide motivated by ideologies rooted in extreme political
beliefs to ideologies based on social group identities (see Garofalo, 1997; Hamm, 1998).
There are two types of offenders that commit [M homicide events considered in this
study. Below, Figure 3.1 only outlines the different homicide subtypes committed by far-
right extremists. Homicide events perpetrated by far-right extremists include any
occurrences that were committed by an individual or group that has past or present ties to
the far-right extremist movement (see Smith, 1994 for discussion of far-right wing

terrorism in the United States).

59

Fi ure 3.1 es of Far-r1 ht Homicide Events in the United States

Type of
Homicide
Suspect

 

   
   
   
 

  

Far-Right
Homicide
Offenders

 

 

Type Of Non IM
Horrucrde Homicide lM Homicide
Event

 

 

 

Based
Anti- Anti- Ami- Ethnicity 223T
HOmeless Race Religion / Nation
'
. Anti-
Christian

  

Antl-
Government
and other
Political
Targets

  
    

 

 
   
   

 

of Origin

 

 

Identity
Targets

 

The essential difference between a far-right homicide and other type of homicide event is
the tie that one or more of the perpetrators has to the far-right extremist movement.
Nonetheless, defining what is or what is not far-right related is a difficult task, as the
definition of a far-right extremist group has often been unclear or conflicting in past
research. For the sake of clarity, this study relies on Freilich and Chermak’s (2006)
definition of the domestic far-right extremist movement developed from their systematic
review of the extant research:
“The domestic far-right is composed of individuals or groups that subscribe to
aspects of the following ideals: they are fiercely nationalistic (as opposed to
universal and international in orientation), anti-global, suspicious of centralized
federal authority, reverent of individual liberty (especially their right to own guns,
be free of taxes), believe in conspiracy theories that involve a grave threat to
national sovereignty and/or personal liberty and a belief that one’s personal and/or
national “way of life” is under attack and is either already lost or that the threat is

imminent (sometimes such beliefs are amorphous and vague, but for some the

60

threat is from a specific ethnic, racial, or religious group), and a belief in the need

to be prepared for an attack either by participating in paramilitary preparations

and training and survivalism. It is important to note that mainstream conservative
movements and the mainstream Christian right are not included” (Freilich and

Chermak, 2006:32; see also Gruenewald et al., 2009).

It is possible that a far-right perpetrator will not be tied to a particular far-right group but
still fit the profile of a far-right extremist perpetrator. For example, the Oklahoma City
bomber, Timothy McVeigh, is considered far-right wing perpetrators despite not having
direct ties to a particular far-right group.

As outlined in Figure 3.1, there are two general types of homicide events that can
be committed by far-right extremist homicide suspects, including those motivated by
ideology and those that are not. In other words, it is possible for far-right extremists to
commit homicides that are fueled by traditional crime motives. Furthermore, it is possible
to disaggregate those homicides that are motivated by ideology by specific motivational
subtypes. In particular, some IB homicides, often referred to as “bias crimes” or “hate
crimes” refer to those instances of homicide motivated by animus towards particular
social minority groups (i.e. racial animus, anti-sexual orientation sentiments, and anti-
religious sentiments) (Perry, 2001). Other types of [M homicide are not based on animus
towards social minority groups but, instead, include acts of political extremism that
target, for example, government employees as well as abortion clinics and clinic
employees (Freilich & Pridemore, 2007; Kaplan, 1995; Nice, 1988). In this study,
homicide motivated by extreme political ideology and hatred toward social minority
groups is considered under the umbrella term of “[M homicide.”

It is also possible for [B homicide events to involve no suspects with links to the

far-right extremist movement in the United States. The diagram presented in Figure 3.2 is

61

similar to Figure 3.1 with the differences begin that only [B homicide events are
considered and the subtypes are defined by the suspects’ affiliations (or lack thereof) with

the far-right extremist movement.

Fi ure 3.2 T es of Identi -Based B Homicide Events in the United States

 

   
   
 

 

 

  
 

  
 

    
  

 

  
 
 
 

 

Type of Identity-
I-Iornicide Based
Event Homicide
Events

Type of
Horrricide { N0" Far— Far-Right

Right Extremists
Suspe t

c Extremists
Ethnicity

  

 

 

 

   

- . Et nicity Antl- Anti- ' Anti- _
Antl- Anti- . Reli l /Natron
Nation . gon
Race Religion ($0an GLBT Race of Origin

 

Homicide events not perpetrated by far-right extremists refer to all IB incidents
that do not involve homicide perpetrators with known ties to the far-right extremist
movement. These incidents can also be conceptualized as bias or hate-motivated
homicides that do not involve far-right extremist perpetrators (Levin & McDevitt, 1993;
Perry, 2001). The FBI’s definition of hate-motivated crime is “a criminal offense
committed against a person, property or society which is motivated, in whole or in part,

by the offender’s bias against a race, religion, disability, sexual orientation, or

62

ethnicity/national origin” (FBI, 2004:65).7 This study applies this definition to IE
homicide events, though incidents against homeless individuals are also considered when
the victim’s status as a homeless person is the primary reason he or she was targeted.
Research Questions
Based on this conceptual scheme of far-right perpetrated and other IB homicide
events, this study is guided by a number of general and more specific research questions
that address the incident and suspect dimensions of far-right extremist homicide. The
research questions are:
Question 1: What is the nature of far-right extremist homicide event
situations in the United States between 1990 and 2008, including suspect
demographic characteristics, other background attributes, and the
situational contexts of homicide events?
A). What are the similarities and differences of the elements of
homicide situations in those far-right extremist homicides motivated
by ideology and the “average” homicide event in the United States.

Question 2: Within far-right extremist homicides, what are the similarities
and differences in [M and non-1M homicide event characteristics?

Question 3: What are the similarities and differences in [B homicides

perpetrated by far-right extremists and IB homicides perpetrated by non far-
right extremists?

In conclusion, this chapter extended academic debate on IM violence by
discussing how the CEP is used in the current study to examine the largely unexplored
topic of far-right extremist homicide. The CEP advances a more comprehensive
understanding of this topic by conceptualizing far-right extremist homicides as multi-

dimensional events constituted by interactions among suspects, victims, and other actors

 

7 More recent information on “hate crimes” gathered by the FBI can also be found at
http://www.fbi.gov/ucr/hc2007/abouthcs.htm.

63

in particular situational contexts that shape the outcomes of fatal situations. A number of
general and specific research questions were posed that focus the examination of this
phenomenon in regards to two key dimensions of far-right extremist homicide events,
homicide suspects and situational contexts. In the next chapter, how this study addresses

these research questions will be discussed.

64

CHAPTER 4: DATA AND METHODOLOGY

This chapter outlines and describes the data and method for this study. A limited
number of scholars have already discussed the methodological pitfalls of past studies of
identity-based (IB) crime, many of which are grounded in issues of data quality and
availability for forms of IM violence such as terrorism (LaFree et al., 2006; Silke, 2001).
One major problem is the lack of an existing official source of data on crimes associated
with domestic terrorism (LaFree et al., 2006). Therefore, it is not possible to identify this
type of crime, involving far-right perpetrated homicide, in official crime data sources
such as the annual Uniform Crime Report-Supplementary Homicide Report (SHR) (FBI,
2004). Being unable to extract homicide information from official homicide data sources
has impeded scholars’ abilities to identify the similarities and differences in [M homicide
subtypes. Also, past terrorism studies have relied heavily on anecdotal media accounts of
terrorist occurrences (Silke, 2001). Thus, knowledge about the entirety of IM violence,
including far-right perpetrated homicide, remains limited.

This study addresses these issues by collecting data on a set of far-right homicides
in the United States between 1990 and 2008, as well as by collecting open-source data on
other IB homicide events perpetrated in the United States by non far-right extremists for
purposes of comparison during the same time period. These homicide events were
identified and information was gathered based on a number of open-sources, including
existing terrorism databases, official criminal justice data sources, academic and
journalistic publications, documents fiom extremist group watchdogs, advocacy groups
(e. g Southern Poverty Law Center, Anti-Defamation League), and other non-profit

organizations, as well as from systematic media searches. Additional information from

65

police or court records was also sought in particular homicide cases on a case-by-case
basis. Data on far-right perpetrated and other IB homicide events were comparatively
analyzed quantitatively based on a comprehensive open-source database, as well as
qualitatively analyzed using a variable-centered theme analysis approach for a set of far-
right homicide events. In addition to identifying IM homicide from open-sources, all
homicides occurring in the United States were randomly sampled from the SHR to
represent the “average” homicide. F ar-ri ght homicides were compared to average
homicides to identify the similarities and differences of far-right homicide and the
common homicide event in the United States. The SHR homicide data were downloaded
from the Inter-University Consortium for Political Science Research website.

This chapter consists of a number of sections. First, the indicators used to
distinguish between far-right extremist homicide events and other IM homicide events
perpetrated by non far-right extremists are presented. Second, specifically how homicides
were identified in open-source materials is discussed, describing the specific sources and
processes for identifying the set of far-right perpetrated homicides, and a set of other IB
homicides. Third, the open-sources used to collect information on each homicide event,
the protocol or instrument used to systematically collect open-source data, and the stages
of data collection are described. Fourth, the systematic coding of data for homicide
suspects and situational characteristics, specific codebooks and variables of interest, as
well as the stages of the coding process are addressed. Finally, the quantitative and

variable-centered theme analyses of far-right extremist homicide events are outlined.

66

Identif Far-R1 ht Pe etrated Homicide Events

This study identified a set of far-right perpetrated homicide events that occurred
in the United States between 1990 and 2008 (n=276). Far-right perpetrated homicide
events involve three general types of criminal motivations: ideologically-motivated (1M)
homicides, non ideologically-motivated (non-1M) homicides, and homicides primarily
motivated by traditional crime motives. As LaFree and Dugan (2004:68) have
recognized, past research has failed to “. . .step back and recognize that criminal events
and terrorist attacks look much the same regardless of the motivation behind them: both
are events that can be counted and both display non-random temporal and spatial patterns
that are likely associated with endogenous and exogenous characteristics of offenders,
targets, and situations.” Therefore, homicides that are fueled by ideological and non-
ideological motives may look very similar or different in ways that are yet to be
systematically determined. As this research comparatively examines subtypes of far-right
homicide, all of these homicide subtypes are of interest. Importantly, the only criteria for
inclusion are that a homicide event was a criminal homicide and that at least one
homicide suspect was affiliated with the far-right extremist movement in the United
States at the time of the homicide occurrence or prior to it. Following the Federal Bureau
of Investigation definition of homicide, in this research, criminal homicide events include
both murders and non-negligent manslaughter, and exclude negligent manslaughter and
justifiable homicide (FBI, 2004).
Indicators of F ar-Right Extremism

A number of scholars have provided partial descriptions of the multi-faceted

domestic far-right movement (Aho, 1990; Barkun, 1997; Berlet & Lyons, 2000; Kaplan,

67

1995). As presented in Chapter 3, one of the most inclusive descriptions of the domestic
far-right movement has emerged from a comprehensive and systematic review of the far-
right extremist literature (see Freilich and Chermak, 2007; Gruenewald etal., 2009).
Based on this definition, homicide suspect association with the far-right extremist
movement was determined by a number of indicators. Drawing from the domestic far-
right extremism literature, a number of concrete indicators of far-right extremist ideology
emerged (Freilich and Chermak, 2007). These indicators are used in the current research
as inclusion criteria for the set of all far-right perpetrated homicides in the United States
between 1990 and 2008 (see Appendix A).

First, far-right homicide offenders often claim to be associated with the broad far-
right wing movement or specific far-right extremist groups operating in the United States.
Therefore self-admission or other evidence of association with the broader movement or
one or more far-right extremist groups indicates that the homicide should be examined in
this research. There are a number of types of far-right organizations, some formally
organized (e. g. KKK, Aryan Nation, The Order, and Aryan Brotherhood), while others
are only loosely-affiliated groups of individuals (Heitrneyer, 2003; Kaplan, 1995). There
are also a number of ways that a homicide offender can be associated with a far-right
group. For example, an offender could have committed the homicide in order to attain
membership or increase his or her status in the group. Other offenders may claim to be a
part of the wider far-right extremist movement but claim no membership to specific
organizations. Still, others can be affiliates seeking membership or ex-members of far-
right organizations. All homicide offenders associated with past or present affiliation of

any sort with specific far-right groups or the general far-right movement are examined in

68

this study. In addition, the intensity of suspect affiliation to the far-right movement is
systematically coded and explored.

Second, far-right homicide offenders often communicate their extremist beliefs
through verbal comments and written statements. Verbal communication of extremist
beliefs may occur during the commission of the homicide, as well as following homicide
occurrences in the context of confessions of killings to friends, media outlets, or criminal
justice officials. Other times, far—right movement leaders may verbally communicate by
giving Speeches on various issues that indicate their extremist ideologies. Written
statements of far-right extremist beliefs can exist as private diary entries and letters to
friends and family or public manifestos sent to media. These types of written and verbal
indicators of far-right extremist beliefs serve as criteria for inclusion of homicide events
in this research.

Third, indicators of far-right extremist affiliation also include markings and
accessories found on the body of the homicide offender. In particular, many far-right
extremist homicide offenders have extremist tattoos on their body. Examples include
tattooed words such as “White power” or other symbols such as swastikas. Other
indicators are accessories, such as Doc Martin boots historically worn by members of
skinhead organizations (Hamm, 1994). All serve as criteria for inclusion.

Fourth, another indicator of far—right affiliation is the presence of far-right
extremist literature and other symbols at the homicide scene or in the offender’s
possession (e. g. in residence or vehicle). Examples of far-right literature may include
specific extremist books (e.g. Mein Kempfl Turner Diaries) or general literature, such as

pipe bomb-making and survivalist manuals and other pieces of conspiracy literature.

69

Lyrics of songs or actual music in the form of music CDs in the possession of a homicide
perpetrator that can be linked to the White power music scene are other indicators of far-
right extremist affiliation (Cotter, 1999; Hamm, 1993). Homicide events are not included
simply because there is evidence that a perpetrator(s) was in possession of such materials.
Indeed, it is necessary that the literature be either placed symbolically by the perpetrator
to be found during the investigation, or that other indicators are also present. In these
instances, possession of literature would bolster the evidence of far-right affiliation.

Fifth, an additional indicator of far-right affiliation is when the association of
perpetrators to the far-right extremist movement is made known during the process of
police investigation or criminal trial. For example, the prior [B offending history of a
homicide perpetrator often reported in the media and in court transcripts during criminal
trials. While a history of 1B offending does not necessarily indicate far-right affiliation,
the testimony of witnesses and others prior to and during homicide trials often indicates
perpetrators’ adherence to far-right groups or general extremist beliefs. Also, in
combination, prior offenses and witness testimony in regards to perpetrators’ affiliations
with the far—right movement that are made known in open-source documents could serve
as indicators or criteria for inclusion.

The final general indicator of far-right affiliation is participation in the broader
far-right extremist movement. There are a number of ways offenders can be involved in
the far-right movement. One way is to be involved in media, such as the hosting of
websites, radio or television shows that promote far-right extremism. Another way is to
be interviewed by various media outlets. Subjects may also participate in the far-right

movement by organizing or attending different types of extremist meetings, such as

70

organizational meetings, rallies, festivals, and protests. A final way to participate in the
far-right movement is leafleting, a common activity that can often indicate extremist
ideologies.

It was a necessary criterion of inclusion for this study that a homicide offender
either adhered to the far-right extremist movement and/or could be linked to a specific
far-right extremist group. The indicators of far-right affiliation were used as evidence of
this adherence and/or membership to the far-right extremist movement in the United
States.

Identify_ipg Other IB Homicide Events

This dissertation also identified a set of other IB homicide events not perpetrated
by far-right extremists that occurred in the United States following 1990. Only those
types of [B homicide defined by the FBI (2004) as “bias” crimes were considered.
Specifically, homicides motivated by animus toward racial/ethnic minorities, sexual-
orientation minorities, religious minorities, and those with a foreign national origin were
considered. The indicators used as criteria for inclusion in the sample resembled those
relied upon to identify far-right extremist homicides in the first portion of this study (see
Appendix B). However, importantly, the criterion that at least one perpetrator was
associated with the far-right extremist movement did not apply to this set of homicide
events.

Past research provides some direction on indicators of IE crime that aid in
identifying IB homicide events. In addition, this study relied on definitions of “bias
crime” indicators, or indicators that a homicide was motivated by animus toward as social

minority group, that had been generally used by police agencies. An example of such a

71

definition is that used in the Massachusetts Model Protocol for Bias Crime Investigation,
which defined bias crime indicators as “[o]bjective facts, circumstances, or patterns
attending a criminal act(s), which, standing alone or in conjunction with other facts or
circumstances, suggest that the offender’s actions were motivated, in whole or in part, by
any form of bias.” A number of indicators were used in this study to identify IB
homicides (Holmes, 1992).

In particular, verbal comments (c. g. racial/ethnic slurs) made by the perpetrators
during the homicide incident or statements made by perpetrators before or following the
homicide serve as an indicator of an identity-based motivation. Other symbols, drawings,
and markings also served as indicators. For instance, actual or representations of hanging
nooses or hate graffiti near the homicide indicated 1B motivations. Real or perceived
differences between the offender and victim of the homicide also served as indicators of
[B motivation. These differences include race, ethnicity, sexual orientation, religion,
political ideology or beliefs. Victim and witness perceptions of the motivation also served
as key indicators of identity-based motivations. Because many indicators could be present
only prior to police becoming involved in the homicide investigation, testimony from
victims and witnesses could be important to determining suspect motive.

Where a homicide occurred was also an indicator of [B motivation. For example,
the location of the homicide could be the setting for prior IB crimes. The homicide
location could also be a place where the victim had been previously targeted, as well as
serve as a general symbolic location for an IB homicide (e. g. religious institution, “gay

bar”/establishment, and governmental office).

72

Homicide perpetrator and victim characteristics also served as indicators. In
particular, offenders’ histories of prior IB offending was an important indicator. Also, if
the victim was in the company of a targeted group or was perceived as breaking social
norms in some way (e. g. two men holding hands) could be an important indicator of IE
motivation.

In sum, indicators of homicide events involving far-right extremists and indicators
of other [B homicide events in this study required unique, but overlapping, selection
criteria that have been summarized in Appendix A and Appendix B. Far-right affiliation
indicators emerged from a body of literature describing the values and ideologies of far-
right extremists in the United States (Gruenewald et al., 2009). Other selection criteria for
IE homicides were largely based on the indicators developed by police bias crime
investigation training manuals, such as the one utilized in Massachusetts (Holmes, 1992;
see also Federal Law Enforcement Training Center, 2005).

Creating an Event-Based Open-Source Database

The homicide events that were examined in this study were identified by
systematically searching a number of open-sources, including existing terrorism
databases (e. g. RAND-MIPT, official sources (e. g. congressional hearings, FBI terrorism
reports), scholarly and journalistic accounts, watchdog reports (e. g. SPLC-Intelligence
Report, ADL), and systematic media searches. Dr. Steven Chermak (Michigan State
University, School of Criminal Justice) and Dr. Joshua Freilich (City University of New
York, John Jay College of Criminal Justice) have worked to compile an account of all
crimes committed by the far-right since 1990. They have been funded by the Department

of Homeland Security to construct a comprehensive database on these events known as

73

the Extremist Crime Database (ECDB). Procedures for identifying homicide occurrences,
data searching protocols, and coding procedures were all borrowed, though in some
instances modified, from Freilich and Cherrnak’s far-right crimes project. The F reilich
and Chermak (2007) data collection and coding strategies were of great benefit to the
current research, as these procedures have endured a long process of peer-review, and
have been successfully implemented in an ongoing project.

First, homicides have been identified from a number of academic sources
systematically reviewed and critiqued (see Gruenewald et al., 2009). Second, Chermak
and Freilich have extracted far-right homicides from large terrorism databases that are
publicly available (e. g. Global Terrorism Database),8 developing a large network of key
informants through past work on domestic terrorism projects and their affiliation with
START. Third, homicide incidents were extracted from official sources of data, such as
the FBI’s annual Terrorism in the United States. Fourth, Dr. Freilich and Dr. Chermak
have identified many far-right homicides from archives of extremist watchgroup
publications, such as the Anti-Defamation League, Southern Poverty Law Center, and a
special radicalization collection housed at the main library at the Michigan State
University.9 In addition to extremist group watchdog organizations, other non-profit
organizations like the Center for Homicide Research (CHR) in Minneapolis have
provided documentation on particular types of [B homicide events that are used in this

study. Indeed, last fall I traveled to the CHR in Minneapolis and met with Dallas Drake,

 

8 The Global Terrorism Database is a publicly-available open-source database on terrorist events around
the world from 1970 to 2007 and can be accessed at http://wwwstartumdedu/atdfl

9 Information on the Michigan State University’s special collections can be found at
hgp://spgcialcollections.lib.msuedu/html/mpterials/collections/radicalism c0112.jsp.

74

the director of the Center. While there, I secured access to a comprehensive database on
anti-GLBT homicides that have occurred in the United States. This database was useful,
as it is being used in this study to cross-check the accuracy of the anti-GLBT homicide
events collected for this project and to aid in identifying additional anti-GLBT homicide
events that had yet to be identified.

In sum, a set of homicide events have been identified and compiled into a
comprehensive list or set of all IM homicides perpetrated by far-right extremists and a set
of other IB homicide events that have occurred in the United States not perpetrated by
far-right extremists. The research conducted for this study has been granted approval by
the Internal Revenue Board at Michigan State University. ‘0
Open-Source Searching

Some of the key information on far-right homicide events has been gathered
during the identification process of each event from the sources described above.
However, to supplement this information, an open-search protocol was developed and
used to search for information on each homicide. There were a number of steps to the
open-search process. First, keywords based on details from each identified homicide
event (e. g. perpetrator and victim names, location of homicide, etc.) were used as
keywords to search major newspaper indexes, such as Lexis-Nexis and Proquest, and
information on each homicide was collected.

Second, a number of extremist group watchdog organizations (e. g. SPLC, ADL,
Human Rights Watch), human rights advocacy groups, and other non-profit organization

were thoroughly searched for additional information concerning each homicide event.

 

'0 [RB # 06-110, Approval # i028245
75

For homicides that occurred in more recent years, information was often available from
these organizations on the Internet. For homicides occurring early in the time period of
interest (i.e. 1990-1995), it was necessary to search the hard copies of their periodicals
and other reports.

Third, the same or similar keywords were used to search for information on
homicides from major search engines, such as Google and Yahoo. A number of less
known search engines (e.g. GoogleScholar, Amazon, Dogpile, Mamma, Scirus, etc.) were
also searched using similar keywords in order to generate and collect all available
information for each homicide and homicide participants. Fourth, in order to retrieve
information about specific extremist homicides from local news media coverage,
keywords were used to search for information from Newslibrary, an online newspaper
index of various regional and local newspapers around the country. As discussed in the
open-search protocol, all open-source information retrieved on each case was gathered in
a single document and prepared for coding.

Advantages of Open-Source Data Collection

As an alternative to official data, some scholars have collected open-source data

 

on IM violent events (e. g. LaFree etal., 2006; Mickolus, Sandler, Murdock, & Fleming,
1993; Schmid & Jongrnan, 1988),11 though none have collected information on all crimes
related to the far-right until the far-right crimes project led by Freilich and Chermak
(2007). As stated previously, important questions concerning the nature of IM crimes

cannot be answered based on officially-collected crime or domestic terrorism data.

 

11 Interestingly, a study by Freilich, Chermak, and Simone (2009) also found that law enforcement relies
on open-sources just as much as non open-sources.

76

Therefore, the most obvious advantage to open-source data collection is that is allows for
the collection of data on topics on which there exists little or no official data.

Another advantage of open-source data collection is that it allows researchers to
adopt their own definitions of the events that are of most interest, as opposed to the
sometimes very limited definitions used by official sources (LaFree etal., 2006). For
instance, it is possible in this study to define far-right homicide as any homicide
committed by an individual or group with past or present ties to the far-right movement.
Adopting broad definitions of events allows researchers of domestic terrorism to examine
different subgroups of events, perpetrators, victims, and extremist groups. As the goal is
to examine similarities and differences across these homicide subtypes, open-source data
collection is a very amenable method for this research. Also, a more practical advantage
of open source data collection is its facility for the simultaneous collection of data on
multiple events by different individuals. For example, multiple undergraduate and
graduate research assistants working on Freilich and Chermak’s far-right crimes database
(2007) or LaFree et al.’s (2006) Global Terrorism Database were able to search and code
open-source data on multiple occurrences independently and simultaneously. Therefore,
it is possible to collect a significant amount of data on incidents, victims, and offenders in
a relatively short period of time.

Responding to Challenges 01 Open-Source Data Collection

There are, however, disadvantages to open-source data collection, including

 

threats to the reliability and validity of data. Though having multiple data collectors
involved in open-source data collection is advantageous in terms of efficiency, there may

be a cause for concern if the inter-reliability amongst data collectors is low. Reliability

77

amongst data collectors could be low, seemingly, if one or more data collectors neglect to
consider particular sources of data. For this study there were a number of ways to ensure
consistency of open-source data collection for all homicide cases. First, all data collectors
were trained identically in order to ensure that each searches for event data in the same
manner. All research assistants collecting open-search data also went through a
probationary period in which the quality of each of the homicide cases they were
assigned was evaluated by project leaders. Second, a protocol was created that directed
data collectors on how to systematically search for open-search materials. Third, the same
Internet search engines were searched for each homicide event, ensuring that each case
was searched in an identical fashion. Fourth, collection of open-source data for this study
occurred in stages, which increased the chances that all available information from open-
sources was collected. In particular, coders first searched for open-source information
based on homicide identifiers (e. g. suspect and victim names), and secondly, data
collectors conducted target searches based on information uncovered during the initial
search. Finally, data collection was continuously being monitored in regards to inter-
reliability concerns and the quality of data collection across data collectors, as the content
of the data collection for a percentage of homicide cases was being compared across
research assistants (LaFree et al., 2006). These types of inter-reliability checks are not
different in nature from the sorts of checks necessary in survey research (Babbie, 1990).
Two other threats to the quality of the open-source data involve biases introduced
by the lack of information or “misinformation” presented by particular sources, often
times media sources (LaFree et al., 2006:24; Noble, 2004). For some events there is little

or no available information from open-sources, resulting in missing data for those events.

78

Fortunately, for some events, like homicide and terrorism events, the seriousness of the
event ensures that there is typically a substantial amount of information on the event from
a variety of open-source materials. In this way, it is possible to triangulate information by
particular sources in order to ensure their accuracy. On the other hand, the nature of cases
that receive much attention may be substantially different than less celebrated cases.
Thus, it is necessary to systematically compare the nature of cases that received
substantial attention with those that did not to identity potential biases. Misinfonnation,
or event details that are distorted or non-factual, present another type of challenge to open
source data collection. Misinforrnation may also be a problem as the result of particular
biases inherent in the originators of the data or due to unintentional errors (LaFree et al.,
2006:24). There are ways to reduce the possible effects on misinformation on data
quality. For instance, it is necessary to assess the historical accuracy of a particular data
source by examining its past coverage of other relevant issues, or by comparing its
coverage of a particular event to the coverage of other known reliable sources of data
(Noble, 2004). Noble (2004) suggested that those collecting open-source data should
assess the quality of each data source and keep logs of such assessments to increase the
reliability of open source data over time.

For this study, there were independent sources of information (e. g. police reports,
court transcripts, other media sources) that were available for use in the comparison of
data sources and reliability assessments. The quality of open-search data was evaluated
for each homicide event examined. A number of factors were considered in this
assessment, including the total number of different open-search materials, the number of

each individual type of collected open-source materials (e.g. court documents, watchdog

79

reports, etc.), and the consistently and reliability of the information in regards to case
details. Open-ended issues, such as having a redundancy or a lack of open-source
materials on a homicide event, were also recorded.

In this study, the effect of the quality and quantity of open-source information was
assessed in regards to the effects on missing values for all homicide variables. First, this
study compared homicide incident and suspect characteristics for cases that received
three or less sources of information with those that received four or more. Second,
focusing more on the quality of open-source information, far-right homicides having
news reports as their only source of information were compared with those homicides
having news reports in addition to other sources of information.

Coding Homicide Events

The collected open-search materials produced a wealth of qualitative data for each
homicide event. In order to systematically examine the homicide events of interest,
information on homicide suspects and situational contexts were coded in two separate but
relational codebook interfaces (hereto after referred to simply as codebooks) created in
Microsoft Access. Data were extracted for a number of suspect-related and situational
variables for each homicide event.

Research assistants coded and inputted data into the codebooks that were used for
this study. There were a number of ways in which this study ensured successful inter-
reliability among data coders. First, all went through extensive training on the coding
process for each variable. Each new coder entered into a probationary status in which he
or she coded a number of duplicate cases as seasoned coders. Second, each codebook

includes a variable to capture specific coders’ names. Therefore, this study was able to

80

explore the data for abnormalities across data coders. Third, in similarity with searching
for open-source data, coding of homicide events essentially occurred in successive stages.
In the first stage, coders entered data for variables based on all available open-source
information. Because many of these homicide cases were not closed and continue to
evolve, potentially missing observations were continuously filled in based on newly
gathered open-source information. This presented the continued opportunity for coders to
recheck their past work, as well as the work of fellow open-source coders.

In addition, if different open-sources contained conflicting information about a
homicide event then, greater weight was given to the more “trusted” sources. For
instance, court documents, reports regarding court proceedings, information from sources
with direct access to case information would be considered a more trustworthy or reliable
sources than, for instance, another second-hand source (see Sageman, 2004).

To date, this is the only known comprehensive database consisting of data on all
homicide events perpetrated by far-right extremists, as well as a number of other IB
homicide events not perpetrated by far-right extremists. Examples of the collected
variables are briefly described below.

IM HomicidL Situations

As shown in Table 4.1, there were two codebooks in addition to the codebook for
the quality and quantity of information, including an 1) IM homicide incident/situational
codebook and an 2) IM homicide suspect codebook. There were a number of homicide
incident-based attributes that were coded for based on open-source data. Many of the
variables for which data were collected were similar to those used by other major

homicide databases, while other variables were unique in their capturing of various

81

aspects of far-right extremism and other ideological beliefs and values. Codebook l,

pertaining to far-right homicide situational characteristics, contains variables that capture

basic descriptive details for each homicide occurrence, including location, number of

offenders, weapon(s) used, circumstances, and homicide victim—offender relationship.

Table 4.1 Incident, Suspect, and Assessment of Data Codebooks

 

Codebook 1. Incident Codebook 2. Primary Sumt Codebook 3. Assessment of Data
Number of Offenders Age Quantity of Information
Single=0 Adult ( l9 and above)=0 3 or less sources=0
Multiple=1 Juvenile (1-18)=1 4 or more sources=1
Weaponi Age Group Quality of Information
Firearm=0 1-18=0 News Only=0
Non-Firearm=l 19-28=1 News and other sources=l
29-3 8=2
Circumstance 39-48=3
Not for profit=0 49-58=4
Profit-related=1 59 and up=5
Victim-Offender Relationshipd: Criminal History
Unknown=0 No Priors=0
Known=l Priors=l
Location Drug/Alcohol Use During Incident
Outdoor=0 No=0
Indoor=l Yes=1
Relevance Affiliation to the far—right
Far-right IM=0 Low=0
Far-right non-[M=l High=l
Non Far-right IB=2
Group

Specific Issue/Motive
Social Minority=l
Anti-GovernmenFZ
Other Ideological=3
Non-1M Movement=4
Non-1M Non Movement=5

 

Non Group Related=0
Group Related=l

 

 

1 Categories were collapsed for multivariate analyses.

82

Other situational variables in Codebook 1 address more directly the aspects of far-
right extremism. III particular, the homicide situational codebook codes for the relevance
of each homicide, as it was possible for far-right extremists to commit both [M and non-
IM types of homicide. The relevance of the homicide event was measured as 0) far-right
ideological 1) far-right non-ideological and 2) far-right identity-based. Ideological
homicides are those motivated primarily by hatred of social minority groups,
government, or by some other lifestyle issue. Non-ideological homicide incidents include
a number of homicide subtypes, including preparatory homicides which were those that
were committed in preparation for a planned, future terroristic event. Also included in
homicides deemed as other criminal activity, but movement related, were those
committed to furthering the far-right movement or a particular far-right organization.
Finally, non-ideological homicides also include homicides categorized as other criminal
activity. Not related to the far-right extremist movement, however, were those homicide
events committed by far-right extremists as a result of traditional crime motives. In
addition, a variable measuring the specific motivation (e. g. anti-govemment, anti-social
minority, anti-abortion provider) is included.

There were coding issues that require further elaboration. One important issue to
consider is the identification of separate homicide occurrences when homicides appear to
be related. In this study, homicide incidents were distinct from one another if they
occurred at different times, in different locations, and involved different victims.
Moreover, if a single perpetrator killed multiple victims in a single location in a short

period of time, the homicides were coded as a single homicide occurrence. However,

83

during the shooting spree, if the murders occurred at different locations, then the each of
the related homicides were coded as separate incidents.
1M Homicide Sum

Once a homicide event was identified and it was concluded that it should be
included, attributes of each perpetrator charged with the homicide were captured in
Codebook 2. The homicide situation codebook (Codebook l) and the suspect codebook
(Codebook 2) both include a ”case identifier” that was used to link each homicide suspect
to a particular homicide event. Some suspect variables were included to examine the
extent and nature of the ideological makeup of the homicide offender. In particular, this
study determined if the offender was involved with the far-right group at the time of the
incident. Other variables systematically explored various aspects of suspect
demographics and backgrounds (e. g. age, race, and gender).12 Also related to past
behavior, this study measured whether illegal substances were used during the homicide.

Each perpetrator was the subject of a unique codebook. Moreover, if a single
perpetrator was responsible for multiple homicides, they were the subject of a different
codebook for each homicide incident, as it may be possible that some of the attributes
(e. g. prior arrests, age, etc.) changed in the time lapsing between homicides. In this way,
each homicide incident can be linked to the involved perpetrator(s) for a particular time

and location.

 

12 Although suspect race and gender information was collected and coded, these variables are not included

in the analyses due to a lack of variation. Over 95 percent of homicide suspects were White males between
1990 and 2008 in the United States.

84

Comparative Analyses

This study relied on both quantitative and qualitative analyses to further
knowledge and understanding of far-right perpetrated homicide events. Data from
Microsoft Access codebooks were exported into the quantitative SPSS data program and
NViv08 for qualitative analyses. One type of analysis conducted was descriptive statistics
on the extent and nature of a set of far-right homicides. Specifically, the number and
proportion of observations associated with each category of all variables were considered.
Based on the systematically collected open-source data, these analyses addressed
questions regarding the most prevalent characteristics of the suspect and situational
dimensions of homicide events for the first time in comparison to the average homicide
event.

In addition, comparative analyses were conducted on particular subgroups of far-
right homicide and other IB homicide situational attributes by disaggregating by
homicide motivation type and by suspect type. First, within the set of far-right
perpetrated homicide events, comparative analyses of IM and non-1M homicides were
undertaken. Second, focusing only on IB homicides, thus excluding homicide cases
perpetrated by far-right extremists fueled by traditional motives, comparative analyses
were conducted on subgroups defined by the perpetrators of homicide events being far—
right or non far-right extremists. Comparative subgroup analyses provided an important
initial assessment of the potentially diverse nature of homicide suspects and the

situational contexts in which this type of offending occurred.

85

Multivariate Comparative Analysis

Following descriptive analyses of homicide events, multivariate comparative
subgroup analyses of far-right perpetrated and other [B homicide events were conducted.
Writing in regards to the use of multivariate analyses to study 1M crime, Silke (2001)
argued that terrorism researchers have not taken advantage of statistical analysis
compared to other disciplines, such as journalism and criminology. Consequently,
terrorism studies have been less likely to be able to draw generalizable findings related to
domestic-terrorism related crimes. In response to Silke’s diagnosis of the research on
terrorism, this study for the first time statistically compared the relative natures of 1M and
non-1M homicides perpetrated by far-right extremists in the United States since 1990. In
addition, the study statistically compared those [B homicides perpetrated by far-right
extremists and those that were not.

The purpose of the multivariate comparative analyses was to statistically identify
the most important situational and suspect-related factors that distinguish between the
homicide subgroups of interest. For example, in order to test the hypothesis that far-right
perpetrated homicides are significantly more likely to be committed in groups compared
to other IB homicide events, this study established type of homicide as the dependent
variable and the number of homicide suspects as the independent variable. An
appropriate statistical technique was used to examine the relationship between the
independent and dependent variables, while controlling for other potentially important
variables (e. g. location of event, offender demographics).

In order to statistically compare these homicide subgroups, this study relied on

binary-logistic regression. Binary-logistic regression is an appropriate regression tool for

86

this study as the respective dependent variables (i.e. type of homicide, type of homicide
suspect) are measured with dichotomous variables (Long, 1997). Moreover, past scholars
have relied on logistic regression to comparatively analyze homicide subtypes (LaFree &
Birkbeck, 1991). This particular statistical technique made it possible to identify which
particular homicide attributes were the most important for distinguishing between
homicide types and subtypes.

Qualitative Analyses

The quantitative comparative analyses identified the nature and extent of far-right
extremist homicide, as well as the key factors that distinguished between homicide
subtypes in the United States. However, quantitative analyses were limited in their ability
to provide an in-depth understanding of how situational contexts interact with homicide
actors’ decisions to shape fatal outcomes, a key aspect of the CEP. Categories of
homicide events based only on quantitative descriptions of homicide only partially
capture these dynamics.

In the past, a limited number of scholars have examined homicide events
qualitatively. For instance, recently some scholars have attempted to identify how
combinations (or configurations) of victim, offender, and incident characteristics result in
(binary) homicide outcomes, such as instrumental (e.g. robbery homicide) or expressive
(e. g. argument-related) fatalities (Meithe & Regoezci, 2004). Meithe and Regoezci
(2004) combined qualitative and quantitative methods using a computer software
program called Qualitative Comparative Analysis (QCA) to systematically generate

profiles of particular types of binary homicide outcomes (Drass & Ragin, 1989).

87

Despite the novelty of QCA, there remain a number of issues with this program
that limit its ability to develop meaningful categories of homicide. First, it is only
possible to focus simultaneously on a small number of homicide factors that might shape
homicide outcomes. Second, when using the QCA program it is only possible to compare
the profiles of two predetermined subtypes of homicide, thus limiting its ability to
explore multiple comparisons simultaneously. In short, the QCA software is a useful tool
for exploring the heterogeneity of predetermined homicide subtypes, but it is limited in
its ability to develop categories of homicide based on themes that capture the dynamic
nature of far-right perpetrated homicide events.

In contrast, other scholars have qualitatively examined homicide narratives or
scenarios of homicide events (Athens, 1980; Katz, 1988, Luckenbill, 1977, Polk, 1994).
Although the specific purposes for analyzing scenarios have not always been the same,
these studies have focused on describing the dynamic interchanges between homicide
actors and situational contexts in order to more aptly categorize homicide events.
Scholars have argued that quantitative examinations have often ignored the
developmental stages that comprise the “social occasions” of [M homicides and more
localized or “situated transactions” between homicide actors (Luckenbill, 1977).
Moreover, quantitative analytical techniques fail to consider the subjective interpretations
of actors’ situational experiences that undoubtedly shape homicidal outcomes, and thus
how homicides are categorized (Athens, 1980). While some past examinations have been
criticized as more haphazard than others (e. g. Katz, 1988), a few scholars have
systematically described their deductive and inductive identification of homicide

subtypes or categories, and themes within each category, in order to capture the reflection

88

of homicide subtypes’ unique interactions between homicide actors in particular social
contexts (Luckenbill, 197 7; Polk, 1994). Themes are essentially constructs that capture
behavioral and cognitive patterns of interactions between homicide actors and situational
contexts. Their work has suggested that systematically identifying common themes based
on the interaction of homicide actors in particular contexts can facilitate knowledge about
why offenders commit homicide, as well as advance a more holistic understanding of
these violent events.
Variable-Centered Theme Analysis

This study qualitatively analyzed open-source information on homicide events to
identify variable-centered themes of homicide subtypes in the United States between
1990 and 2008. The purpose of the variable-centered theme analysis was to build on the
quantitative findings by exploring and elaborating on the ways in which key homicide
variables found to be important in distinguishing far—right homicide subtypes, operate in
interactions between homicide actors and situational contexts. Understanding how
situational contexts influence criminal behavior is a key component of the criminal events
perspective (Sacco & Kennedy, 2002). Methods of qualitative research are useful for
examining complex social phenomenon, such as far-right extremist homicide, and for
providing more in-depth descriptions of these events (Merriam, 1998; Strauss & Corbin,
1998). Qualitative methods can facilitate the exploration of heterogeneity within this
topic, and can help to identify homicide cases that do not fit apparent modal patterns.
Finally, qualitative research methods are useful when criminal actors’ interpretations of
situations affect the outcomes of these situations (Merriam, 1998; Strauss & Corbin,

1998)

89

This study can be described as mixed-methods research, as it relied on qualitative
methods to better explain and elaborate on the quantitative findings (Cresswell, Clark, &
Plano, 2007). Thus far, discussed in detail has been the identification of far-right
perpetrated homicide events. From open-sources of far-right perpetrated homicide events,
the collection and systematic coding of data, and the use of quantitative analyses, were
used to describe the nature of far-right homicides. In addition, quantitative multivariate
analysis was described as one way to assess similarities and differences between
homicide subtypes (e. g. IM vs. non-IM, anti-government vs. anti-social minority
motivations). However, identifying important distinguishing variables across homicide
motivational and suspect subtypes does little to explain how these variables act in the
context of other key variables, or illustrate instances in which seemingly important
variables fail to distinguish between homicide subtypes. While theme analysis does not
necessarily aid in drawing generalizable conclusions, it does facilitate in-depth
understandings of the similarities and differences in homicide subtypes.

When it was determined from quantitative analysis that a particular homicide
variable (e. g. number of perpetrators) distinguishes between two far-right perpetrated
homicide subtypes (i.e. IB and non-1B), a variable-centered theme analysis was
conducted to explore how this variable’s interaction with other situational variables
shaped these particular types of fatal outcomes. The question asked when identifying
themes from the homicide event scenarios was “How did key distinguishing variables
operate in the context of dynamic far-right perpetrated homicide situations for each
homicide subtype?” and “What were the most common variable-centered themes?” As an

example, suppose that quantitative analyses indicated that far-right perpetrated IB

90

homicides tend to be more associated with suspects offending in groups (compared to
single offenders) when compared to far-right perpetrated non-1B homicides. Theme
analysis made it possible to explore the open-source information for each homicide in
which 1B homicides were perpetrated by multiple suspects to provide more. detailed
descriptions of the various ways these deadly transactions occurred. Examples of themes
could include “attack while intoxicated on acquaintance” or “single-offender White
supremacist mission offending spree on random victims,” both of which draw from prior
literature on far-right violence (Hamm, 1993; Heitrneyer, 2003; Levin & McDevitt,
1993), and capture the interactions between homicide actors in particular situations. It
was then possible to provide more detailed descriptions of each theme found to be
prevalent in specific homicide subtypes. In this way, theme analyses were used to build
more complex understandings of far-right homicide events.

The variable-centered theme analysis identified how certain variables shaped
certain subtypes of homicide. However, the purpose of this study is to not only develop a
profile of certain types of homicide but, also, to identify and explore homicide cases in
which these variables shape all far-right homicides, including those cases that may not fit
the typical pattern. Simply because a variable, such as profit circumstances, is most
associated with non-1M homicides, it does not necessarily hold that IM homicides are not
also shaped by profit-seeking behaviors in some instances. Although useful in some
cases, there may be danger in developing profiles of particular homicide subtypes.
Discussing perhaps less common patterns of how variables shape other subtypes less
associated with it can further understanding of this complex phenomenon by illuminating

the ways in which homicide subtypes compare in nature and process.

91

The preparation for the variable-centered theme analysis occurred in a number of
sequential steps. First, the quantitative values for each variable coded in Microsoft Access
were imported into a Microsoft Word document. All of the original open-source
information was combined with the imported quantitative data. Each case was captured in
a separate document. Second, each homicide case was imported into the qualitative
software N Vivo8. Third, each category of every homicide incident and suspect was
“autocoded” into a “node.” For instance, the incident variable weapon use nodes were
created for guns, knives, bodily weapons, blunt objects, and other weapon types. Coding
each variable in this way made it possible to qualitatively examine each case that fit
certain criteria. For instance, it was possible to systematically and efficiently examine the
interaction of each weapon type used by suspects across far-right IM and non-IM
homicides (or other homicide subtypes) by examining the information available for each
case that fits those particular criteria.

By identifying particular cases that fit the criteria of interest, it was then possible
to isolate homicide cases to begin identification of themes in certain variable interactions
in the broader context of the homicide situation. Continuing with the example, one
possible theme for weapon type could be “skinheads using knives in drunken brawls.”
This theme could provide context to understanding why far-right ideological homicides
were more likely to involve non-guns.

Finally, the number of times that a particular theme was identified for each far-
right perpetrated homicide subtype was tallied. This made it possible to make more

generalizable conclusions about the relative frequency of identified themes for particular

92

' homicide categories. Tracking the prevalence of themes also indicated that particular
themes were dominant across multiple subtypes of homicide events.

As discussed previously, it is recognized that current typification schemes based
on strictly quantitative descriptions of homicide motive or victim-offender relationship
variables may not sufficiently capture the diverse nature of far-right homicide events (see
Polk, 1994:19-21). Therefore, it was necessary to explore how key homicide variables
(e. g. number of suspects) operated in the context of the dynamic homicide situations to
distinguish between homicide subtypes. An advantage to relying on open-source data for
systematically analyzing far-right homicide case narratives was that information on each
homicide reflects a number of sources of data. In this way, it was possible to more aptly
capture both the successive stages of the homicide event and the interpretations of actors
as experienced by suspects, witnesses, police, as well as interpretations of homicide
events by media, extremist watchgroups, personal blogs and others. Relying on open-
sources provides for a comprehensive or holistic description of the homicide event as it
was experienced by all involved homicide actors (see also Luckenbill, 1977). Apparent
suspect motives, though often complex and dynamic, were described by incorporating
suspect accounts, as well as interpretations of motives by witnesses, police, and others
with an interest or stake in the homicide case. Undoubtedly, each data source presents
biases. For instance, suspects may have been more likely to describe the homicide as
justified based on the victim’s behavior, while witnesses may have perceived the
homicide as unprovoked. Considering multiple sources of information neutralized the
inherent biases of particular sources by capturing multiple interpretations of the unfolding

of each homicide event (Luckenbill, 1977).

93

Conclusion

The preceding chapter proposed a method for a quantitative and qualitative
examination of far-right extremist homicide events that have occurred in the United
States since 1990. It is recognized that systematic research on this topic is virtually non-
existent, and it is expected that the current study will make substantial contributions in
the overlapping research areas of homicide studies, general terrorism studies, and far-

right extremism.

94

CHAPTER 5: THE CONTEXT OF FAR-RIGHT HOMICIDE EVENTS

The primary purpose of this chapter is to address the first research question posed
in this dissertation regarding the elements of homicide events (i.e. incident and suspect
characteristics) and their comparison across far-right extremist homicide and the
“average” homicide that has occurred in the United States between 1990 and 2008. To
address the uniqueness of far-right homicides, a select number of far-right homicide
characteristics are compared with the characteristics of a random sample of “all”
homicide events that occurred in the United States based on official homicide data
collected by the Federal Bureau of Investigation. Through comparative analysis, this
chapter begins to address how findings support and contradict many of the claims made
about violence committed by far-right extremists in past research as well as popular
culture.

In addition, one of the advantages of relying on the criminal event perspective
(CEP) to examine this phenomenon is the emphasis placed on describing the situational
contexts of criminal events (Sacco & Kennedy, 2002). In this chapter, this is achieved in
part by exploring the geospatial and temporal variations in far-right extremist homicide
during this time period. When and where homicide events have occurred likely influence
the nature of far-right homicide events and how they are perceived by the public,
policymakers, and the media. As will be shown, far-rightists do not commit homicide in
every state. Identifying which states have no or few homicide occurrences, and which
states have substantially more, is one step toward providing an explanation for the

dispersion of homicide events across the nation. Moreover, focusing on the geospatial

95

distribution of these events allows for speculation on how the locations of far-right
homicides can shape the nature of the particular events examined in this study.

In addition to being geospatially distributed, far-right homicides are also
temporally distributed and these occurrences have fluctuated over time. How such
fluctuations of far-right homicide correspond to the occurrence rate of the average
homicide in the United States remains unknown. It is also unclear how different subtypes
of far-right homicide fluctuate over time. For instance, are far-right homicides that are
primarily motivated by ideology similar to or different from non-ideologically-motivated
events in regards to temporal distribution? Examining these fluctuations comparatively
can begin to show if the occurrence rate of far-right homicide follows or diverges from
the trajectory of the average homicide, as well as how the occurrence rate of far-right
motivational subtypes compare, over time. Answers to these questions influence how far-
right homicide is examined in this study by determining which homicide subtypes are
compared.

As described previously, the current study relies on the CEP and emphasizes the
need to collect and analyze original data that are able to provide answers to specific
research questions of interest. This study relies on open-source data to examine the
phenomenon of far-right homicide and other IB homicide events not perpetrated by far-
rightists. Although the advantages and disadvantages of open-source data were discussed
in the last chapter, the specific limitations of the data used in this study needs to be
assessed. Also, because relying on open-source data to study homicide events is relatively
uncommon in criminological research, it is important to identify potential biases in the

data and understand how they may influence the interpretations of the subsequent

96

findings. One potential source of bias explored in this chapter is a lack of open-source
information for certain event characteristics due to a lack of attention given to the
homicide by open-sources. Thus, the extent of missing values for homicide
characteristics for cases in which substantial open-source data are available and those
events in which only little information is available are compared. Another potential
source of bias involves relying solely on media, as media may either present only select
details of incidents, neglect to cover incidents entirely, or flame incidents in some other
biased manner. Moreover, media may be a source of “misinformation” (F alkenrath, 2001;
LaFree et al., 2006). In order to address this type of bias, the extent of missing values for
select homicide characteristics are compared for those homicides in which news media
are the only source of information and those events that have multiple sources of
information available.

The current chapter is presented in four sections which reflect each of the
aforementioned stated purposes. First, a table capturing the extent of far-right homicides,
and respective motivational subtypes, for every state and the District of Columbia is
presented. Expected and unexpected findings about the geospatial distribution of this
form of homicide, and the implications of these findings, are discussed. Second, simple
comparative time-lines juxtaposing the rate of far-right homicide occurrences and the rate
of the average homicide, as well as far-right IM and far-right non-1M homicides, are
presented. In addition to actual changes in the occurrence of homicide, the extent across
time, to which homicide is shaped by external factors, is considered. Also, the effect,
over time, on homicide occurrence trajectories by changes in the recording and reporting

of 1B crimes following high-profile homicide events are discussed. Third, after

97

considering the geospatial and temporal contexts of homicide events across the country, a
number of homicide incident-level and suspect-level characteristics are comparatively
examined across far-right and average homicide events, including number of homicide
offenders, weapon type, circumstances, victim-offender relationship, suspect age, suspect
race, and suspect gender. Fourth, multiple assessments of the quality and quantity of
open-source information available for each far-right extremist homicide are conducted in
order to identify potential biases inherent in the data, including the number and type of
sources of information.
Considering Context

A key component of the CEP is the context of the criminal event (Sacco &
Kennedy, 2002), including the spatial and temporal components of each event. By doing
so, it becomes increasingly possible to speculate how the dimensions of homicide suspect
and incident intersect with macro-level factors to shape the empirical nature of these
fatal outcomes, thus leading to more meaningful interpretations of findings. ‘3

The fact that the nature of far-right homicide represented in this dissertation is
shaped by an array of social forces has not been obscured. Although occurrences of IM
violence undoubtedly occurred prior to 1990, the phenomena of “hate crimes” and
“domestic terrorism” are relatively new. Scholars have traced the social construction
process of 1M violence, which has included the influential activities of social movement
groups, modern advocacy groups, lawmakers, and other claimsmakers (see J enness &

Broad, 1997). Others have studied how claimsmakers continue to shape if particular

 

'3 Although past research has found important macro-level factors that affect rates of particular forms of
ideological crimes, the data used in the current the study do not permit such an analysis at the present time.

98

crimes are reported as ideologically-motivated or not (Hamm, 1998; Jacobs & Henry,
1998; Jenness & Grattet, 2005).

This study has attempted to create a database that captures the actuality of this
violence by not relying solely on those crimes determined to be ideologically-motivated
by officials. Open-source data on far-right homicides and other [B homicides were
collected on homicide when and if it was determined that events were committed by far-
rightists or were motivated by animus toward social minorities based on predetermined
lists of indicators of far-right extremism and [B crime. In this way, it was hoped that the
data in this study would reflect the actuality of the phenomena as closely as possible.

Nonetheless, it would be a mistake to neglect to consider how social forces have
influenced the “actual” nature of far-right homicide Since 1990, since crime and the
reporting of crime do not occur in a vacuum. It would also be a mistake to reject studying
the conditions of far-right and other IB homicide because, as “strict constructionists”
have suggested (Spector & Kitsuse, 1977), such conditions are social constructions that
do not exist in reality. Instead, the homicide events included in this study are viewed as
actual events that have devastating consequences on victims and victims’ communities,
as well as events that are shaped by an array of social factors that should be thoughtfully
considered (Best, 1995; Loseke, 2003). Considering the possible geospatial and temporal
effects on the nature of homicide events is one way this study attempts to accomplish this
task.

Geospatial Context of Far-Right Homicide Events
Based on a review of the far-right extremist literature (Gruenewald et al., 2009),

there is evidence that far-right extremist activities have been associated with certain

99

portions of the nation (e. g. the Pacific Northwest) (Neiwert, 1999), as well as particular
states (e.g. Idaho, Montana, Michigan) (O’Brien & Haider-Markel, 1998), and where
high-profile far-right groups have based their headquarters (see also Cook & Kelly,
1999). However, it is clear from this review that scholars have not adequately considered
state-level variations in homicide occurrences committed by all far-right extremists in the
United States. '4 In addition to advancing a more in-depth understanding of this
phenomenon, examining far-right homicides in this way can aid law enforcement and
policymakers in assessing the risk of serious far-right violence in their respective states.
Table 5.1 indicates that far-right extremist homicides have not been identified in
every state. Since 1990, thirteen states (over 25% of states) and the District of Columbia
have had no such homicides identified. Some of the most intriguing findings involve
these states and those that have very few homicides attributed to them. Though Montana
may have a relatively small population, Michigan (n=3) has a relatively large population,
which makes the number of far-right homicides especially surprising considering the
elevated far-right activity in this state. One possible conclusion is that states which
remain breeding grounds for extremist activity, such as Michigan (i.e. paramilitary
militias), may not be hot-beds for extremist crime (see Freilich & Pridemore, 2006).
Another explanation is that there may be a reluctance to report the far-right affiliations of
homicide offenders to reporting agencies in Michigan and other states (e. g. SPLC, ADL,
media outlets). Indeed, scholars have found that the reporting of IM crimes varies by

sociopolitical contexts and police practices (Berk et al., 1994; Hamm, 1998).

 

'4 This is not to say that other types of crimes involving the domestic far-right have not been studied at the
national level (see, for example, Freilich & Pridemore, 2005, 2007; O’Brien & Haider-Markel,l998;
Pridemore & Freilich, 2005).

100

Table 5.1 Percentage of Far-Right Homicide Events by State, 1990-2008 (N-274):t

 

Percentage 2000 Census 2000 Population

State N Population Rankdzt
California 53 19.3 33,871,648 1
Texas 33 12.0 20,851,820 2
Florida 15 5.5 15,982,378 4
Ohio 14 5.1 11,353,140 7
Arizona 14 5.1 5,130,632 14
Pennsylvania 1 1 4.0 12,281,054 6
Idaho 9 3.3 1,293,953 39
Washington 9 3.3 5,894,121 13
Colorado 8 2.9 4,301,261 22
Oregon 8 2.9 3,421,399 27
New York 7 2.5 18,976,457 3
Massachusetts 7 2.5 6,349,097 1 5
Oklahoma 7 2.5 3,450,654 28
Nevada 6 2.2 1,998,257 35
Arkansas 5 1.8 2,673,400 32
Illinois 5 1.8 12,419,293 5
Missouri 5 1.8 5,595,211 18
Indiana 4 1.5 6,080,485 16
Minnesota 4 1 .5 4,919,479 21
North Carolina 4 1.5 8,049,313 10
Tennessee 4 1.5 5,689,283 17
Utah 4 1.5 2,233,169 34
Georgia 3 1.1 8,186,453 9
Kansas 3 1.1 2,668,418 33
Louisiana 3 1 . 1 4,468,976 25
Michigan 3 1.1 9,938,444 8
Mississippi 3 1.1 2,844,658 31
New Jersey 3 1.1 8,414,350 1 1
New Mexico 3 1.1 1,819,046 36
Maryland 2 0.7 5,296,486 19
New Hampshire 2 0.7 1,235,786 41
Alaska 1 0.4 626,932 47
Nebraska 1 0.4 1,71 1,263 38
South Carolina 1 0.4 4,012,012 24
Virginia 1 0.4 7,078,515 12
Wisconsin 1 0.4 5,363,675 20
Connecticut 0 --- --- «-
Delaware 0 -- -- ---
Hawaii 0 --- --- «-
Iowa 0 --- --- ---
Kentucky 0 --

Table 5.1 continued

Maine 0 --- --- «-
Montana 0
North Dakota 0
Rhode Island 0 m m "-
South Dakota 0 --
Vermont 0 "' --- ---
West Virginia 0 '" --- ---
Wyorrgg 0 '” --- ---

 

iState of occurrence is unclear (and thus missing) for two horrricide events.
ttNumbers based on 2000 Census state population makings where “1” represents the state with the largest
population during this year.

Also intriguing, relatively few far-right homicides were identified as occurring in
the South during this time period, despite historical racial tensions marked by anti-Black
lynching in this region of the nation. In particular, states such as Georgia and Louisiana
(n=3) have fewer far-right homicides than might be expected, especially considering
Georgia’s relatively large population. One possible explanation for these findings is that
homicides committed by far-right extremists or, more generally crimes motivated by
ideology, are more likely to be reported in locations where there are active social
movements advocating for the recording of such events (McVeigh, Welch, & Bjamason,
2003). Also, suspects’ beliefs or affiliations with the far-right movement are less likely to
be emphasized in open-sources when these events do occur in these states.

It is also evident from Table 5.1 that some states have experienced considerably
more homicides which were committed by far-right extremists than others. Some
explanations for these findings are worth exploring. One explanation is that some states
with a relatively large number of far-right homicides also have a relatively large
population, increasing the likelihood of such a crime occurring. In particular, California

(n=50) and Texas (n=33) have the most homicides as well as the largest state populations.

102

Florida (n=14) and Ohio (n=14) are also states with large populations and a relatively
large number of far-right homicides, though there are exceptions (e. g. New York).
Another explanation is that states which house large state prisons known for experiencing
substantial White supremacist group activity are plausibly more likely to have a higher
number of far-right homicides in their state. Far-right extremist homicides that occur in
prison are included in the current study and, as will be discussed later, it is often the case
that White supremacist gangs in the prison affect crime outside prison walls. One of the
most notable examples includes San Quentin State Prison in California where the Aryan
Brotherhood originated in the late 1960s (Pelz, Marqaurt, & Pelz. 1991). The United
States Penitentiary Administrative Maximum Facility (ADX) in Florence, Colorado is
another example of a state prison that likely affects the salience of far-right extremist
homicide in this particular state, as Colorado is not a state with a large population. Texas
is also a state in which practically the entire prison system is plagued with White
supremacist (e. g. Texas Aryan Brotherhood) criminal activities that often extend beyond
the prison walls (Pelz, Marqaurt, & Pelz. 1991). In addition, states that are
disproportionately populated by cities that have a historical presence of counterculture
groups or gangs are likely to experience an elevated number of homicides involving far-
right extremists. Examples may include Portland, Oregon and Seattle, Washington, which
have had relatively elevated skinhead populations since the early 1990s (Langer, 2003).
Each of these states has a relatively small population and an elevated number of
occurrences of far-right homicide. Idaho is another state with a small population that has
experienced a disproportionate amount of far-right homicide. Interestingly, almost 90

percent of the far-right homicides in this state have been non ideologically-motivated,

103

suggesting that there may be a substantial far-right presence in this state that commits
more traditional forms of violence. Idaho has been home to a number of extremist groups,
such as the Aryan Nation, which have a history of producing violent White supremacists
that commit traditional crimes of violence and, as will be shown later, often against one
another. Alabama may also fall under the category of states with relatively small
populations and an elevated number of homicides. One explanation is the historical
presence of the Ku Klux Klan (KKK) in this state. However, other Southern states have
experienced a large KKK presence and Alabama appears to not fit with the pattern of low
occurrence (or low reporting) of far-right homicide in the South since 1990. An
alternative explanation, therefore, is that the presence of the largest watch- group for far-
right violence, the Southern Poverty Law Center, is headquartered in this state. In other
words, far-right homicides occurring in Alabama may be more likely to be reported due
to the proximity of the homicide occurrences to this watchdog organization. Finally,
Massachusetts is a state that has a disproportionately large number of far-right events
relative to its population, which cannot be explained by any of the aforementioned
explanations. The majority of the homicides, nearly 90 percent, in this state were
ideologically-motivated. One explanation for the Massachusetts finding is that this state
has been responsible for drafting some of the earliest “hate crimes” legislation, training
law enforcement to investigate IM crimes, and for organizing some of the first organized
bias crime units. These activities have likely all promoted the reporting of violence
perpetrated by far-right extremists (J enness & Broad, 1997).

In short, this is the most systematic examination of far-right homicide in the

United States to date. While this analysis does not provide complete explanation for state-

104

level variation of far-right homicide, it does provide a context for interpreting the
findings regarding situational and suspect-level homicide characteristics in subsequent
chapters. Below further discussion of the context of far-right extremist homicides is
provided by visual representation of the year when events have occurred since 1990.
Temporal Context of Far-Right Homicide Events

In addition to being spatially distributed across the United States, far-right
extremist homicides are temporally distributed in patterned ways. Since 1990, there have
been apparent spikes in far-right group memberships and activity. For instance, the SPLC
Intelligence Report reported that the number of far-right extremist groups increased to a
record number (SPLC, 1999). In addition, in 1998 the SPLC reported the number of
extremist group websites had hit an all-time high. The Department of Homeland Security
also recently reported that a number of social and other incidental factors have
culminated to likely affect the far-right extremist sentiments and interest in far-r1 ght
group membership (Hudson, 2009).ls Such factors included the economic collapse, rising
home foreclosures, and the election of the fust African American president. Although
there have recently been some high-profile far-right homicides, it is unknown if these
factors will result in an upswing in far-right extremist violence. Nonetheless, the

distribution over time, of past far-right extremist homicide occurrences, can be examined.

 

15 This report was later removed by the Department of Homeland Security due to the controversy it caused
over claiming that returning war veterans posed a risk to increasing far-right extremism. The report was to
be reworked in a more useful and accurate manner (Hudson, 2009).

105

 

 

 

 

Figure 5.1 Rate of Homicide by Type and Year, 1990-2007 (N-257):l: i

P 10 _,___,_ i

e . i

. 8 i "

3 1" x / \ A ’

C 6 I l' A‘ ~ “I ‘ 1 \ 7 \ f "— l

n : 1

t 2 r T

o i 7“ f‘"”I'"“1’""”'l’"w’TT"”"T'"“”1"””Ti . I I 1 1""1 I 1' r ”'1 '

H N m in r~ on H N m in h i

Sssaaafaa§§sss§s§8 i

Year ’

 

 

iThe year of the homicide event cannot be determined for 17 events (6.2%).

In Figure 5.1 the percentage of far-right homicide that occurred each year
between 1990 and 2007 is plotted. This line is based on the homicide events identified in
open-sources as described in the last chapter. The number of annual far-right extremist
homicides in the United States is small relative to the number of total homicides that
occur in the United States. This could easily lead some to question the importance of this
topic. However, as discussed previously, the potential effects of such an event on victims
and their respective communities can be highly consequential (Berrill, 1992).
Ideologically-motivated homicides have in the past opened windows and fueled
homeland security and bias crime policy discussion in ways that traditional homicides
have not (Colomb & Damphousse, 2004; Jenness & Broad, 1997). In addition, homicide
scholars have effectively argued that studying “deviant” forms of homicide can greatly
further our understanding of and add nuance to explanations of all homicide (Decker,

1996).

106

While the rate of the average homicide in the United States was relatively high in
the early 1990s and steadily leveled off over time (Blumstein, Rivara, & Rosenfeld,
2000), the rate of far-right homicide has been more erratic and, thus, is more difficult to
interpret. The year with the most far-right extremist homicides since 1990 was 1999
(n=23), while in 1991 and 2003 there were only nine far-right extremist homicides
identified in the United States. There was a relatively dramatic rise in far-right extremist
homicide occurrences in the early 1990s and, then again, in the late 1990s. A few
possible partial explanations for the apparent peaks and valleys, and apparent differences
from the average homicide can be posited.

The increase in the early 19905 far-right extremist homicide could be due to the
increase in reporting of IB crimes that occurred during this time following the signing of
the Hate Crimes Statistics Act in 1990 (Jacobs & Henry, 1996; Jenness and Broad, 1997).
Some have suggested that this legislation increased reporting of this type of crime and
accounted for the seeming “hate crime epidemic” that occurred in the 19905 (Jacobs &
Henry, 1996). In other words, some have suggested the increase was socially constructed.
However, it is clear that the early peak in the 1990s was not the only, or most dramatic,
spike in the rate of homicide committed by far-right extremists. Thus, another
explanation for this early increase could be an actual increase in violence committed by
extremists, such as skinheads and other White supremacists (Hamm, 1993).

During the mid-19905 there was an apparent decrease in far-right extremist
homicide. Some scholars have discussed a crackdown by law enforcement on far-right
extremists during this time, especially following the Oklahoma City bombing in 1995

(Chermak, 2002). This is one possible explanation for the decline. Some scholars have

107

suggested that far-right extremist groups disbanded or lost members following the very
high-profile terrorist event, as many far-rightists did not want to be associated with
domestic terrorism and wanted to avoid the added scrutiny by law enforcement
(Chermak, 2002).

Conversely, the most dramatic climb of the far-right homicide rate occurred in the
late 1990s, peaking in 1999. This peak followed a particular year, 1998, in which two
major high-profile IB homicides occurred. The first incident was the brutal anti-gay
murder of college student Matthew Shepard near Laramie, Wyoming, while the second
case was the anti-Black murder of James Byrd, Jr. in Jasper, Texas. Supporting Colomb
and Damphousse (2004), it appears that these events may have been able to set off moral
panics in which reporting and publicity of extremist violence increased despite whether
or not the actual occurrences increased. Further support is found in Figure 5.2 which
shows that far-right IM events peaked during this time period, while non-1M events

decreased.

108

 

 

 

 

 

 

 

 

 

 

Figure 5.2 Far-Right Homicide by Type and Year, 1990-2007 (N-257)=l:*
18 arr“:
16 .. .-----.-_-.-.a~-.- UK... "um.--”-Mfl-----_-----.-,__---a... ' l
I i
m 10 __ as; , . I" “ Non—
b 8 W... W, x ,1 i w. , - ,A Ideological l
t . s
.. 6 , I I “wk-“1;“. “4“ m . Ideological i
i r )1, ‘1
: _. z x s ..
' 4 \ ‘ ’I‘\ l
l 2 ‘4
l o . . . . . . . . .,
O F‘ N M V V) \O I‘ 00 O\ '— M V) I"
assaaaaaaa§s§ s§s§e

 

:ltThe year of the homicide event cannot be determined for 17 events (6.2%). ’Far-right homicide
data are not included due to the unavailability of comparison data.

It is possible that more attention was being paid to those homicides targeting
social minority groups during this time. Both the Shepard and Byrd homicides resulted in
public outcry, legislative reform, and the victims remain symbolic victims of the gay
rights and recent transformations of the civil‘rights movement, respectively (J enness &
Broad, 1997).

Following these high-profile homicides and the September 11th terrorist attach on
New York City and the Pentagon, the United States experienced the steepest decline in
the rate of far-right extremist homicide since 1990. Nonetheless, there is indication that
far-right extremism has been steadily on the rise, and recent examples of far-right
extremist homicide indicate that this phenomenon is unlikely to dissipate in the near
future (Holthouse, 2009). Importantly, though, Figure 5.2 shows that not all types of far-

right homicide are on the rise. Instead, it appears that 1M and non-1M homicides have

109

been diverging in recent years. While IM and non-1M homicide events have fluctuated
since 1990, it appears that non-IM events are becoming more of a threat in the United
States. Chapter 6 firrther examines how non-1M homicide compare to IM occurrences.

Understanding the time and place that far-right homicides occur advances a more
in—depth understanding of this phenomenon. Juxtaposing the occurrence rate of this
violent crime with the average homicide in the United States over time reveals that there
may be important differences in the nature of far-right homicide from all homicide. It is
now necessary to systematically compare far-right homicide incident and suspect
characteristics with the average homicide occurring in the United States since 1990,
exploring these similarities and differences across these homicide subtypes.
A Comparison of Far-Right and “Average” Homicide Events

As previously stated, two major purposes of this chapter are to present
preliminary descriptive findings on a select number of far-right homicide characteristics,
and, then, to examine how these homicides may be unique from the average homicide in
the United States. The far-right homicide data presented in this chapter were extracted
from the Extremist Crime Database (Freilich & Chermak, 2006). The comparison
homicide data were compiled fiom a random sample of homicides fi‘om the Uniform
Crime Reports-Supplementary Homicide Reports as described in Chapter 4. Table 5.2
compares the nature of far-right incident and suspect characteristics for far-right

homicides and the average United States homicide.l6

 

‘6 It is important to note here that far-r1 ght homicides are not being compared to a random sample of all
non far-right homicides, but instead, a random sample of all homicides (including far-right homicide).

110

In addition to plotting the rate and occurrence of homicide events over time, open-
source and official data can be integrated for purposes of comparing the nature of
homicide. Previous chapters have described how relying solely on official data sources
says little about the nature of far-right homicide.‘7 Original data collection based on
open-sources allows researchers to define and typify homicide events addressing specific
research questions rather than based on what data are available from secondary sources.
Another advantage to creating an original database is that is possible to design the data to
be comparable with other sources of homicide data such as the SHR. By doing so, it
becomes possible to build on the strengths of the already available data sources and to
contribute to what is known about the nature of homicide in the United States.

The remaining debate in regard to the uniqueness of [B crimes from traditional
crime in the United States adds significance to this study. For instance, one Senate
Republican Policy Committee report argues that [B homicides are not unique
(Republican Policy Committee, 2000). This report also states that IB crimes are not
particularly vicious or terrifying, or especially harmful to the community. However,
others who study this form of violence have claimed there to be substantial differences,
for instance, in the psychological harm imbued on victims and the social harm on
communities (Lawrence, 1999; Levin, 1999; Perry, 2001; Perry & Iganski, 2009).

Bivariate statistical significance tests (i.e. Chi-square) indicate that far-right
homicides are statistically unique from the average homicide in the United States.

Although these findings are only preliminary in that they do not reveal precisely how

17
A recent 2009 report by the National Counterterrorism Center also included a section written by START

Center Director Dr. Gary LaFree discussing the issue of the unavailability of such data and the advantages
and disadvantages of open-source data (NCC, 2009)

111

these variables are different from one another, they do show that the proportional
distributions of these variables are statistically different. A number of findings from this
analysis are worth highlighting from this analysis.

First, similar to the average homicide that occurs in the United States, far-right
homicides are primarily committed with firearms. As firearms are predominately the
most prominent weapon used in homicides committed in the United States, this finding is
not particularly surprising. However, as shown in Table 5.2, a higher proportion of far-

right offenders rely on non-firearm weapons compared to the average homicide in the

United States.

Table 5.2 Descriptive Findinggn Far-Right and Average Homicide Events, 1990-2006: ta
Far-Right Homicide All Homicide

N Percentage N Percentage_

Incident Characteristics
Number of offenders (N=235, 2840)""*

Multiple 133 56.6 485 15.8
Single 102 43.4 2588 84.2
Weapon (N=242, 2840)"“
Gun 135 55.8 1940 68.3
Knife 50 20.7 389 13.7
Blunt 26 10.7 119 4.2
Bodily 22 9.1 154 5.4
Other 9 3.7 238 8.4
Circumstance (N=254, 1916)*"*
Not for profit 196 77.2 1644 85.8
Profit-related 58 22.8 272 14.2
Victim-Offender Relationship (N=223, 1560)""
Strangers 124 55.6 443 28.4
Friends/Acquaintances 85 35. 1 636 40.8
Domestic Partners 8 3.6 282 18.0
Other family 6 2.7 199 12.8

Primary Suspect Characteristicsn:
Age (N=246, 1730)'

112

ob

51

ill

b.

hi

Table 5.2 continued

 

1-18 28 11.4 281 15.8
19-28 109 44.3 734 41.2
29-38 67 27.2 383 21.5
39-48 28 11.4 225 12.6
49-58 9 3.7 81 4.6
59 and up 5 2.0 76 4.3
Race (N=226, 1877)*"*
White 223 98.7 863 46.0
Black 3 1.3 1014 54.0
Sex (N=250, 1941)“*"'
Male 246 98.4 1737 89.5
Female 4 1.6 204 10.5

 

""pSDOl ”‘ps.05"ps.01"ps.01
:tDue to the availability of SHR data, only homicides occurring between 1990 and 2006 are considered.
ttOnly one far-right homicide suspect for each homicide incident is considered.

One of the most substantial differences across homicide types is in the use of blunt
objects and, to a lesser degree, other non-firearm weapons. This finding evokes images of
the use of “billy clubs” and other makeshift blunt object weapons used by the young
skinhead offenders described in Hamm’s research (1993, 1994, 1998).

A second finding of interest is the relatively high proportion of far-right
homicides that involved multiple offenders compared to the average United States
homicide. Over 50 percent of far-right homicides involved multiple offenders. Combined
with findings about the use of more intimate weaponryI8 by far-rightists, it appears that
beatings of victims by multiple suspects may be one common theme of far-right extremist
homicides that deserves further elaboration. In order to explore this far-right homicide

subtype further, all events with multiple suspects that also involved blunt objects were

 

l8 . . . . . . . .
Intrmate weaponry refers to weapons that require relatively mtrmate physrcal Interaction between
suspect and victim.

113

reviewed. There were two modal themes discovered among these events: 1) homeless
beatings'9 and 2) within-group killings.

The beating murders of homeless persons (n=16) were some of the most brutal
homicides examined in this study. The majority of these homicides involved multiple
suspects and all involved non-firearm weapons, including combinations of knives, sticks,
boots, and other blunt and bodily weapons (e. g. fists). Common circumstances involved
groups of youth, often loosely organized groups of skinheads or neo-Nazis, who preyed
on homeless in packs. Beatings and stabbings took place outdoors in public or semi-
public locations. As evidenced in open-source information on these homicides,
perpetrators often verbalized their disdain for homeless as lazy as well as their
perceptions of homeless persons as drains on society (and as easy targets). Some of the
homicides were at least partially motivated by the desire of suspects to win favor with
hate groups.

Another type of group-related homicide, often involving blunt objects, were
within-group killings that were committed to further group goals, such as for internal
group discipline and for ridding the group of real or perceived police informants. As is
discussed more in the next chapter, individuals disrespectful of the group were often
victims of far-right extremist homicides and were often killed with blunt objects and
other non-firearm weapons.

Although statistically significant, far-right homicide and the average homicide are

surprisingly proportionately similar across circumstance (profit-seeking behaviors or not)

 

19 While research on [B homicides that target homeless is limited, one exception is a recent study by
Waccholz (2009) that examines the victirrrization of homeless persons.

114

categories. Both official data on homicide and open-source data on far-right homicide
include information on profit-related circumstances involved in each homicide event.
Approximately 14 percent of average homicides and 21 percent of far-right homicides
involved profit-seeking circumstances. Thus, there were proportionately slightly more
profit-seeking homicides in the far-right homicide sub-category when compared to the
average homicide in the United States. This is important because past research has tended
to associate far-right extremist offending with ideological “mission offenses,” or offenses
driven strictly by ideology (Levin & McDevitt, 1993). In addition, others have suggested
that domestic terrorists, such as far-right extremists tend to specialize in IM offending
rather than be versatile criminals (Gottfredson & Hirschi, 1990). Past research, however,
has examined only IB crimes reported by police or focused on select incidents of IB
offending. Because this study considers all homicides committed by far-right extremists,
not only 1M crimes, it becomes clear that far-right extremists commit more than 1M types
of violence and that these crimes of violence may involve profit-seeking behaviors. In
subsequent chapters, these motivational and suspect subtypes are more fully explored.

The findings also Show, expectedly, that the largest proportion of far-1i ght
extremist homicides involve offenders and victims unknown to each other, while
conversely, the average homicide is more likely to involve offenders and victims known
to each other (often intimately). This finding evokes the image of far-r1 ghtists as domestic
terrorists seeking out symbolic targets randomly based on their minority statuses. This
study found that this type of homicide storyline has been played out in many homicide
events. The extent that far-rightists commit homicides against minorities will be

discussed below and in subsequent chapters.

115

More unexpectedly, this study shows that the proportion of homicides which
involved offenders and victims known to each other, as fiiends or acquaintances, was
comparable across homicide categories. That is, while far-right homicides are much less
likely to involve intimates, they appear to be just as likely to involve known victims. This
suggests that far-right extremists not only seek out random and symbolic targets, but also
tend to target a number of more familiar victims. The vast majority of far-right known
victims were classified as acquaintances rather than friends or family. Indeed, as will be
shown in the next two chapters, acquaintance homicides were evenly distributed into IM
homicides and non-IM events. Most often these victims and offenders were never on
fiiendly terms despite knowing each other and, in some instances, victims and offenders
knew each other only briefly prior to the fatal transaction. These homicides, which have
not been examined in past research, included murders of disloyal affiliates, loved ones,
and other victim types, such as GLBT community members, which are also explored in
more in-depth in the next chapter.

In addition to incident-level characteristics, suspect-level characteristics were
considered and the findings are presented in Table 5.2. As expected from past research,
the vast majority of suspects involved in far-right extremist homicides were White males.
In this way, far-right suspects are unique from the average homicide offender in the
United States who continue to be disproportionately African American and, relatively,
more likely to be female. While some past studies have examined female participation in
far-right extremist activities (Gilmore, 1996), the far-right movement in the United States
remains dominated by males and actively promotes traditional notions of masculinity,

largely through their discourse (Ferber, 1998). The overwhelming presence of White

116

males in the far-r1 ght extremist movement has factored into some explanations of why IM
crimes occur. For example, Perry (2001) hypothesized that one reason offenders commit
IB violence is to “do difference.” Extending the “doing gender” theory, Perry suggested
that social structural factors have in many ways denigrated the statuses or identities of
working-class White males. Particularly in the United States, deindustrialization in the
last thirty years has left blue-collar workers and agricultural workers out of work. At the
same time, the increase of racial/ethnic minorities in the United States has threatened
White racial hegemony. Additionally, the increase in women’s rights and participation in
the labor force has caused White males to feel threatened on a number of fronts. In
response, Perry (2001) suggested that White males have used violence against social
minorities to reify their various identities, and the racial, sexual, socioeconomic status
lines they view as being blurred by the forces of modernity. In this way, the use of
violence has been a way to create or “do difference.” More discussion on the doing
gender and related doing difference theories is provided in Chapter 7, specifically on how
IB offenders may use anti-GLBT violence to reify their sexual identities in specific
homicide situations.

Conversely, some findings, such as those regarding offender age, did provide
somewhat more surprising results. In his research, Smith (1994) found that domestic
terrorists in the United States tended to be older than the comparable traditional offender.
However, the findings from this study reveal that this pattern may not apply to all types
of far-right extremist offending. While far-rightists who committed homicide were
slightly less likely to be categorized as juveniles, and slightly higher in the 29-38 age

group compared to traditional homicides, as past research would suggest, far-r1 ghtists

117

were represented higher in the 19-28 year-old age group compared to traditional
homicide offenders. Specifically, this finding is contrary to Smith’s finding that far-right
extremists tended to be in their mid to late 305.

There are a number of possible explanations for the apparent similarities in far-
right homicide offenders age to average offenders’ age, and the discrepancies in this
study’s findings compared to Smith’s (1994) research. One partial explanation is that
Smith (1994) examined more than homicide offenders. Therefore, far-rightists who

commit the most extreme violence may be dissimilar in nature from other far-right

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

criminals.
Figure 5.3 Average Far-Right SuspectAge by Group Type (N-276)=l: ;
45 i
40 m l
l
M 35 ”" l
C :1 1%. l
a 30 r..- ,.'~ ‘3 l
A 20 ~--~ 3: EF- 1
C 15 A...» 3 w, _',..____..____,._.._._ ‘
'3 -' Fj- l Primary Suspect

10 ,__- i”: = 3
5. L. i
0 .s.. " .. intuit __-.,_-w,___-z-_w.( '

of 80%

of

i
J

 

 

tOnly one far-right homicide suspect for each homicide incident is considered.

118

Figure 5.3, above, shows that those of the skinhead and neo-Nazi type, a common type of
far-right homicide offender, who has rarely been prosecuted at the federal level, has a
lower mean age than other types of far-right extremists.

The significant differences in far-right homicide and the average homicide events,
as well as some of the surprising similarities in some of the characteristics, have a
number of implications. One implication is the justification of the argument made in this
study that far-right homicides should be considered as a separate subtype of homicide
event. Indeed, one of the fundamental premises of this research is the need to
systematically examine homicide perpetrated by far-right extremists, as there is reason to
expect that the nature of these events is unique. Related, these findings of difference
suggest that the indicators of far-r1 ght extremism are useful for typifying a category of
homicide events based on suspects’ affiliation with the far-right movement in the United
States. It also follows that such affiliations plausibly affect the nature of these far-right
homicide events in ways that can be empirically examined. Nevertheless, upon further
inspection of these variables, some of the findings, despite being statistically different,
were surprisingly similar. To illustrate, over 20 percent of far-right homicides involved
profit-seeking behaviors. Far-right homicides also were surprisingly likely to involve
known victims. Findings such as these suggest that far-rightists may be a more
heterogonous group, and that these homicides may take on many motivational forms,
than prior research has suggested or has been able to capture. In response to these
findings, the next chapters directly examine motivational and suspect subtypes of far-
right homicide events. Before moving on to these analyses, however, this chapter further

explores far-right suspect differences by systematically examining their affiliations with

119

the far-right movement and how differences in these affiliations affect the nature of

homicide events.

Intensig of Far-Right Affiliation

Thus far, far-right homicides have been treated as a single type of homicide in
which any homicide involving at least one suspect affiliated with the movement since
1990 is included in this study. One question not addressed, however, is if far-rightists
differ in the intensity of their affiliation to this movement. It is possible that homicides
involving suspects that are highly involved in the far-right movement may differ from
those that may be less intensely affiliated. Comparatively examining the nature of
homicide events across far-rightists who vary by level of intensity presents an

opportunity to examine how suspect and situational factors combine to shape fatal
outcomes. In this study, the nature of far-1i ght homicides committed by suspects with
Strong affiliation to the far-r1 ght movement were compared with the nature of those
events committed by suspects with weak affiliation to the movement. How a suspects’
affiliation to the far-right was coded depended on two criteria. The fnst criterion was the
number of indicators present in open-source data that indicated affiliation to the far—right
movement. The second criterion was the number of indicators or bits of evidence that
Stood contrary to suspect affiliation with the far-right. Examples of contrary evidence
include, but were not limited to, verbal or written communication of non-adherence to
ffir-~1'ight extremism by suspects, witnesses, court actors, and others. Affiliation to the far-

right movement was measured on a 4—point scale as illustrated in Table 5.3.

120

Table 5.3 Strength of Far-Right Suspect Affiliation to the Far-Right

Category

4=Undisputed established present or past adherence
to far-right ideology and/or far-right specific or
general group with multiple indicators present.

3-Clear established present or past adherence to far-
right ideology and/or far-right specific or general
group with a single indicator present.

2-Disputed present or past adherence to far-right
ideology and/or far-right specific or general group
with multiple indicators present.

l-Disputed established present or past adherence to
far-right ideology and/or far-right specific or general
group with a single indicator present.

 

Criteria

1) Multiple (2 or more) far-right indicators found 2)
No evidence found contrary to far-right affiliation.

1) Only single far-right indicator found 2) No
evidence found contrary to far-right affiliation.

1) Multiple (2 or more) far-right indicators found 2)
Evidence found contrary to far-right affiliation.

1) Only single far-right indicator found 2) Evidence
found contrary to far-right affiliation.

A quantitative analysis was conducted to compare the nature of homicide events

perpetrated by suspects with highly intense affiliations to the movement with events

perpetrated by suspects less intensely affiliated to the movement. It was found that

homicides committed by suspects with strong affiliation to the far-r1 ght movement

(scored 3 or 4, 87%) were proportionately more likely to be committed by multiple

offenders, to target known victims, and were proportionately more likely to be committed

by adult suspects rather than juveniles when compared to those homicides involving

suspect less intensely affiliated to the movement (scored 1 or 2, 13%). In other words, 13

percent of the homicides involved suspects with low affiliations to the far-right

movement, and those particular events involved single offenders, and juvenile suspects

that targeted unknown offenders. As will become more evident in the next chapter, cases

inVolving suspects with low affiliations to the movement generally align with the far-

r1'ght IM homicide events. The suspects involved in these homicides may be fueled by the

ideas propagated by the far-right movement but have, for instance, decided not to join

121

formal groups or take part in other organizational or subcultural aspects of the movement.
One explanation is that these offenders are particularly violent or extreme in their
rhetoric, which alienates them from other far-rightists. In any case, the nature of these
events may be unique from the rest of far-right [M homicides that involve perpetrators
with more intense affiliations. Therefore, in the next chapter, the multivariate models will
be run both with these cases included and not included in order to detect potential
influences (or biases) on results.

Evaluation of Open-Source Information

The final purpose of the current chapter is to evaluate the open-source homicide
data in order to identify potential biases that could shape how findings are interpreted. In
Chapter 4, a number of types of potential biases associated with open-source data
collection were discussed (see also LaFree et al., 2006). One potential bias is
newsworthiness bias in which some homicides receive substantial attention compared to
others due to their extraordinary nature, possibly differentially affecting the findings.
Before examining far-right extremist homicide subtypes in the next chapters, it is
necessary to examine potential differences in the extent of missing data for incident and
suspect characteristics for those homicides in which only a minimal amount of
information is available compared to those cases for which a more substantial amount of
information is available.
Another potential bias is one associated with relying solely on media sources

(F alkenrath, 2001; LaFree et al., 2009). Media may not include some information that
other sources do, such as victim-offender relationships or suspects’ race. Therefore, it is

also necessary to comparatively examine rates of missing data for those homicide events

122

in which only news print media information was available and those cases in which
multiple types of information was available. The results of these comparisons are shown
in Table 5.4.

The first comparison to consider involves homicides in which there were three or
less sources of open-source information and those in which there were four or more
sources of open-source information. The number of sources for each category was chosen
because very few homicides involved only a single source of information and less than 8
percent of homicides were identified in which there were only two sources of open-
source information available. While it is still necessary to be cautious in interpreting the
descriptive findings in Table 5.4, choosing three or less sources of information as the cut-
off point assured that enough homicides would be in each category for meaningful

comparisons.

Table 5.4 Percent Missing by Source Type and Salience, 1990-2008:

 

 

 

 

Quality of Information Quantity of Information
News Only News and Other 3 sources or 4 or more sources

_ (n-56) (n-219) less (n-39) (n=232)
_ N % N % N % N %

Incident Characteristics

Number of offenders 56 0.0 219 0.0 39 0.0 232 0.0

Weapon 54 3.6 209 4.6 37 5.1 224 3.4

Circumstance 56 0.0 219 0.0 39 0.0 232 0.0

Victim-Offender Relationship 47 16.1 190 13.2 27 30.8 206 1 1.2

Suspect Characteristics

Age 55 1.8 212 3.2 39 0.0 225 3.0

Race 56 0.0 219 0.0 39 0.0 232 0.0
fix 56 0.0 219 0.0 39 0.0 232 0.0

 

my one far-right homicide suspect for each homicide incident is considered.

123

Of the variables that were considered, victim-offender relationship and offender
race variables experienced increased amounts of missing values. It is likely that coders in
many instances were not able to ascertain this information when the available information
for these homicides was limited. Nonetheless, compared to other national homicide
databases like the UCR-Supplementary Homicide Reports, the amount of missing data
for certain variables is very low (FBI, 2004). Indeed, a recent volume of Homicide
Studies was devoted to the causes and responses to missing data issues (e. g. low
clearance rates) in the SHR (Homicide Studies, 2004). Fortunately, all far-right
homicides received some type of attention from media, advocacy groups, criminal justice

agencies, and most cases had multiple sources of information. Therefore, although a

 

limited number of far-right homicides have a higher rate of missing values for variables
like victim-offender relationships, the extent of missing data is relatively low compared
to other homicide data sources.
Supporting the findings for the quantity of information comparison, the quality of

imformation revealed that victim-offender relationship and suspect race had relatively
increased amounts of missing values for homicide cases in which only news media
sources were available. Often, media do not report suspect race as a matter of policy.
Thus, it appears that relying only on news media sources may introduce bias due to a lack
of information. Fortunately, only one-fourth of all homicides had media as their only
source of information. Moreover, the potential bias introduced is likely distributed evenly
across the motivational subtypes compared in the next chapter. More specifically, victim-
offender relationship and suspects’ age had relatively more missing values for non-1M

homicides, while weapon use and the number of offenders had relatively more missing

124

data for [M homicides events. Thus, missing values for homicide variables of interest to
the current study were not skewed toward a particular motivational subtype. In sum,
while missing data remains a problem in general for homicide research (Homicide
Studies, 2004), the far-right homicide data do not introduce additional issues of missing
data because they originate from open-sources.

Summagy of Findings

This chapter provided an introduction to the rest of the analysis presented on the
examination of far-right extremist homicide events. There were three primary purposes of
this chapter including describing the context of far-right homicide in the United States
since 1990, quantitatively comparing far-r1 ght and “average” homicide events, and
assessing the quality of open-source data on far-right homicide. Achieving these goals
guides the analysis of far-right homicide and aids in the interpretation of the findings
presented in the next two chapters.

The consideration of the geospatial and temporal dimensions of far-right homicide
events presented a number of interesting findings. For the first time, the extent of far-
right homicide and its subtypes was presented, revealing that not all states have
experienced this type of homicide occurrence. Moreover, some states appear to have a
disproportionately large or small amount of far-right homicide events in relation to the
size of their population. Past research and different aspects of the far-right movement
were used to speculate on these variations, including the presence of state prisons
infamous for housing White supremacist prison gangs, or a historical presence of a
particular type of far-right group in particular cities (e. g. skinheads in Portland). The

temporal context of far-r1 ght homicide was also examined by explaining how state and

125

regions within states vary in terms of sociopolitical contexts that affect reporting of [M
violence (Hamm, 1998). The trajectory of the rate of far-right homicide was compared to
the rate of “all” homicide in the United States, as well as how far—right 1M and non-1M
compared over time. These analyses showed that far-right homicide did not decrease over
time as did the “average” homicide in the United States, and experienced a number of
peaks and valleys. External factors, such as the passing of the 1990 HCSA and high-
profile IB crimes and responses to these crimes were posited as partial explanations for
socially constructed increases in far-right homicide. Stated more concisely, it is believed
that a number of factors have influenced the temporal nature of far-right homicide that
can possibly aid in describing this complex phenomenon and in suggesting how the
nature of this phenomenon may differ based on the time period examined.

One finding important in regards to the motivational subgroup analysis conducted
in the next chapter is the seeming increase in far-right non-1M homicides and the decline
in far-right IM homicides. Therefore, examining the similarities and differences in these
homicide subtypes is crucial for understanding the current threat of far-right homicides in
the United States.

The comparative analysis between far-r1 ght homicide and the “average” homicide
in the United States was the first attempt to combine open-source and official homicide
data in this fashion. The variables examined were limited to what was available in the
SHR as a secondary data source. Each of the homicide characteristics were significantly
different when compared using bivariate Chi-square analysis, suggesting important

differences in the nature of far-right homicide.

126

Some of the findings were especially surprising. For instance, there were
proportionately more homicides involving profit-seeking behaviors compared to the
average homicide in the United States. Another surprising finding was the large
proportion of those events targeting known victims. These findings were surprising
because IM violence has been portrayed anecdotally as targeting unknown victims based
solely on ideological beliefs. These and other important findings indicate a need for
further analysis of the variety of far-r1 ght offending patterns.

Finally, this chapter assessed the quality of information available for far-right
homicide based on open-source information. Particularly, this chapter attempted to
identify the potential effects of newsworthy bias and media bias by examining how both
the quantity and quality of available data sources affected the extent of missing data for
each variable. Findings revealed that there was more missing data for variables such as
victim-offender relationship and select offender demographics. However, it was also
concluded that “missing” values have been a historical problem for these variables, as
information is often unknown to investigators, and the extent of missing values was
relatively low for far-right homicides compared to the same variables in official homicide

data sources.

127

CHAPTER 6: A COMPARISON OF F AR-RIGHT [M AND NON-[M
HOMICIDE EVENTS 1990-2008

 

The last chapter made a number of interesting conclusions based on the findings
of a comparative analysis that juxtaposed the nature of far-right homicide with the
“average” homicide in the United States. However, a single profile of a “far-right
homicide” did not emerge from these preliminary findings. For example, far-right
homicides often involved victims known to far-right suspects and over 20 percent of
these cases actually involved profit-seeking behaviors. In this way, far-right homicides
appeared similar to and less distinguishable from the majority of homicides that
commonly occur in the United States. On the other hand, some findings aligned with how
far-right perpetrated violence has been described in past research and popular culture,
including the involvement of non-firearm weapons and multiple perpetrators. These
similarities and differences in the findings presented thus far suggest that far-rightists
commit multiple types of homicide events, and indicate a remaining need to take into
account varying motivational circumstances (or motivational subtypes) within this type of
homicide event.

One of the primary contributions of this dissertation is its examination of all types
of homicide events committed by far-right extremists, regardless of the nature of the
particular motivational circumstances. To date, there have been no available sources of
data that allow for the examination of far-right homicide suspects and situational
characteristics. As discussed in previous chapters, available homicide and terrorism
databases are not able to distinguish far-right affiliations of suspects or do not include
these cases because suspects were not charged with terrorism but, instead, with a myriad

of other criminal charges (NCC, 2009; Smith, 1994).The current study, however, is based

128

on an open-source database that facilitates the comparative examination of far-right
homicide suspects and situational characteristics across motivational subtypes. Indeed,
one of the key pieces of data gathered from the open-source information for each
homicide was the motivational circumstances of each event, including both ideologically-
motivated (1M) and non ideologically-motivated (non-1M) homicides.

The wealth of data on each far-right homicide from this original open-source
database allows this study to seek answers to a number of research questions. The far-
right extremists commitment of other types of homicide that are not primarily motivated
by ideological beliefs are of particular interest. Are non-ideologically motivated far-right
extremist homicides distinguishable from those that are primarily motivated by ideology?
And what makes these homicide events unique? Is it the situational characteristics or
suspect characteristics, or both? In addition, how do important distinguishing variables
operate in the context of the entire homicide event?

This chapter also illustrates the advantages of approaching the study of far-right
homicide from a mixed-method approach (Cresswell & Clark, 2007). Bivariate statistics
are used to identify proportional similarities and differences in far-right IM and non-1M
event characteristics. Multivariate statistics are used to identify the most significant
predictors of homicide motivational subtype, or to identify those characteristics most
significantly related to each subtype. The CEP, however, also recognizes that quantitative
methods alone cannot capture the complexities and nuances of a particular phenomenon,
such as far-right homicide. Therefore, variable-centered theme analysis is used to
qualitatively explore in more depth how each variable found to be statistically important

operates in the context of the homicide event. In this way, variable-centered themes are

129

identified to more fully capture the variety of ways the particular variable combines with
other situational factors to shape these fatal outcomes.

Furthering the general understanding of far-right extremist homicide is a worthy
academic exercise and has Significant policy implications. This specific examination
directly addresses the debate over if, and if so, to what extent, [M violence is unique
from traditional violence, and also extends this discussion and adds nuance to it by
placing focus on homicide perpetrated by far-right extremists. These findings have
relevance and contribute to the discussion of prosecuting IB crimes and domestic
terrorism, as well as sentencing these types of offenders. Second, developing profiles of
far-right extremist homicides, in general and by motivational subtypes, can aid in police
investigations of this type of violence (Parker, 2009; see also Chermak, Freilich, &
Shemtob, 2009). Homicide event profiles can advance law enforcement’s ability to
identify the general qualities and assess the risk of far-right offenders and, also, can
answer questions regarding how these types of events occur.

Descriptive Findings

A number of suspect-level and situational far-right homicide characteristics are
examined in this chapter, including those discussed in the last chapter. Also, examined
are attributes only available for far-right homicide events (e. g. group type, location, drug
and alcohol use). This chapter includes a multivariate analysis that identifies which
characteristics best distinguish between motivational subtypes, but it is first necessary to
describe the differences in these motivational subtypes in more detail.

There are two motivational subtypes explored in far-right homicide events. The

first type is ideologically-motivated (IM) homicide, which includes those homicides

130

targeting social minorities (based on religious, race/ethnic, sexual orientation, immigrant,
nationality statuses or identities, and homeless statuses or identities) and other
ideologically-based targets, such as abortion providers and sex offenders. The second
subtype of far-right homicide explored in this chapter is non ideologically-motivated
(non-1M) homicide, which includes events that do not target social minorities as the
primary motive. Moreover, these events may or may not have been committed to further
the far-right extremist movement.

One key finding shown in Table 6.1 is that far-r1 ght extremist homicide offenders
have committed proportionately more non-[M homicide events (n=56.7%) than [M
homicides since 1990 in the United States. Because past studies have only anecdotally
considered the IM offending of far-rightists, the extent of non-[M offending was virtually
unknown. This finding, again, highlights the benefits of examining not only the [M
homicide offending patterns of far-rightists, but all of the other types of homicide
offending by far-right extremists as well. Table 6.1 shows that there were 117 IM far-
right homicides that have occurred since 1990 in the United States. The majority of these
homicides targeted members of social minority groups, while the next modal category
was homicide events that targeted govemment-related victims. Other IM homicide
motives included, for instance, those motivated by animus toward abortion providers, sex

offenders, and other leftist and feminist groups.

131

Table 6.1 Far-Right Homicide Motivational Circumstances

 

N (270) Percent

Ideological (n-l 17, 43.3%)

Anti-Social Minority Groups 91 77.8

Anti—Government 1 8 l 5.4

Other Ideological 8 6.8
Non-Ideological (n-153, 56.7%)

Movement Related 52 34.0

Non-Movement Related 101 66.0

 

The remainder of the homicide events were not ideologically-motivated. A
relatively small proportion of these non-1M homicides were related to the far-right
movement in some way. An example of this type of homicide would be a far-right
extremist murdering a fellow extremist group member or affiliate because of his
disrespect toward the group or alleged “snitching” on group criminal activities. Another
example would be when a far-right extremist commits a string of robberies in order to
fuel activities aimed at furthering their beliefs or group cause. The majority of non-[M
homicides, however, had no apparent ties to the far-right movement. Instead, this
subcategory of homicide consisted of domestic homicides, drug-related homicides, and
other felony homicide events.

The next step is to examine the similarities and differences in the natures of far-
right [M and non-1M homicide events. In Table 6.2 the findings for the comparative
analysis of suspect and incident-level variables for each homicide motivational subtype

are presented.

132

Table 6.2 Descriptive Findings Far-Right IM and Non-1M Homicides (N-270) :l:

 

Non Ideological (N=153) Ideological (N=ll7)

 

N Percent N Percent
Incident Characteristics
Number of offenders (n=141, 107)
Multiple 90 63.8 53 49.5
Single 51 36.2 54 50.5
Weapon (n=148, lll)
Gun 86 58.1 61 55.0
Knife 32 21.6 20 18.0
Blunt 12 8.1 14 12.6
Bodily 1 1 7.4 13 l 1.7
Other 7 4.8 3 2.7
Circumstance (n=153, 117)
Not for profit 1 l 1 72.5 100 85.5
Profit-related 42 27.5 17 14.5
Victim-Offender Relationship (n=129, 106)
Strangers 48 37.2 85 80.2
Friends/Acquaintances 67 51.9 21 19.8
Domestic Partners 8 6.2 0 0.0
Other family 6 4.7 0 0.0
Group (n=125, 105) ,
Group Related 77 61.6 47 44.8
Non Group Related 48 38.4 58 55.2
Location (n=140, 113)
Outdoor 49 35.0 66 58.4
Indoor 91 65.0 47 41.6
Primary Suspect Characteristics
Age (n=147, 115)
1-18 18 12.2 10 8.7
19-28 64 43.5 51 44.3
29-38 43 29.3 33 28.7
39-48 17 1 1.6 11 9.6
49-58 3 2.0 7 6.1
59 and up 2 1.4 3 2.6
Criminal History (n=153, 1 l7)
Priors 8 1 52.9 51 43 .6
No Priors 72 47.1 66 56.4
Drug/Alcohol Use During Incident (n=153, 117)
Yes 20 13.1 25 21.4
No 133 86.9 92 78.6

 

:tOnly one far-right homicide suspect for each homicide incident is considered.

133

As shown in Table 6.2 the majority of non-[M far-right homicides having
occurred in the United States Since 1990 involved multiple offenders, while [M events
were more evenly distributed across single and multiple offender subtypes. This finding
was somewhat unexpected because it was assumed that far-right non-1M homicides
would tend to resemble traditional homicide offenses more so than [M events. Scholars
relying on various situational perspectives, including symbolic interaction, have
discussed how multiple offenders and bystanders can shape the evolvement of violent
encounters into homicidal events (Luckenbill, 1977; Meier et al., 2001). For instance, as
discussed further below, one common way that additional offenders can facilitate extreme
violence is by deriding the offender who may not want to “lose face” into confronting an
adversary or participating in a killing. Often suspects felt as if they had no choice but to
participate in the homicide or face retribution from their accomplices.

Indeed, it is likely that homicides committed by formal far-right extremist groups
that often involve extensive crime syndicates are influencing the findings for the “number
of homicide offenders” variable. Approximately 80 percent of non-[M homicides
involving formal groups also involve multiple offenders. As was shown above, a
substantial proportion of non-[M far-right motives included within-group violence (e. g.
internal discipline). These cases involved multiple offenders, as group leaders made a
decision to have the victim murdered, and then ordered a lower-level member or
individual desiring to gain admission into the group. This explanation is supported by the
finding that over 60 percent of non-[M homicides involved suspects linked to formal
groups, as compared to IM homicides which are mostly non-group related. The

seriousness of this type of violence became apparent after a large number of Aryan

134

Brotherhood offenders were federally indicted on racketeering and other charges that
included sixteen homicides that stemmed back to the 1970s (Coverson, 2006). The court
records on this case showed how this particular crime syndicate crossed state boundaries
and committed criminal activities within and outside prison walls and included attempted
homicides, drug homicides, and various conspiracy charges. This case also revealed the
ways in which group leaders routinely ordered group members and affiliates to murder
fellow members who were thought to have disrespected or had been disloyal to the Aryan
Brotherhood.

In addition to the number of offenders involved in each far-right homicide,
another intriguing finding was that non-1M homicides were proportionately more likely
to involve profit-seeking behaviors when compared to [M homicides. Profit-seeking
behaviors included most commonly were acts of robbery and burglary. Therefore, it is
clear that internal discipline among far-right groups was not the sole circumstance of
non-1M far-right extremist homicides. More specific examples of non-[M homicides
involving profit-seeking behaviors are discussed in the next section. Other far-right non-
IM homicide characteristics, these variables appear relatively more Similar to the average
homicide in the United States in a number of ways. First, far-right non-1M homicides
were more likely to involve the use of firearms and knives as primary weapons compared
to the other motivational homicide subtype. In this way, IM homicides have been more
associated with intimate weaponry compared to their non-[M counterpart. Non
ideologically-motivated homicides, on the other hand, involved more efficient and lethal

modes of killing (i.e. firearms).

135

Second, non-[M homicides were proportionately more likely to involve victims
known to offenders rather than random strangers when compared to [M events similar to
the average homicide in the United States. Thus, it is clear, based on these findings, that
[M homicides were more likely to target strangers, which aligns with what has been
written anecdotally in regards to the symbolic attacks on random victims by far-right
extremists.

Considering circumstance, victim-offender relationship, and weapon type, in
combination, reveals that far-right extremist homicides tend to differ from the average
homicide in the United States. Average homicides involving profit-seeking behaviors (i.e.
instrumental or felony homicides) tend to be associated with unknown victim-offender
relationships, while expressive homicide events tend to involve arguments with known
victims (Block, 1981; Decker, 1993; see also Miethe & Drass, 1999; Salfati, 2000).
Indeed, far-right non-1M homicides, compared to far-right IM events, are proportionately
more likely to involve profit-seeking behaviors and known victims. In addition, in
average homicides, intimate weaponry tends to be more associated with known victims.
For far-right homicides, however, intimate weaponry is associated with [M homicide
events that are also proportionately more related to unknown victim-offender
relationships.

Other remaining homicide characteristics cannot be as easily compared to
homicide data collected at the national level by criminal justice agencies. Nonetheless,
these characteristics are relevant to the current study as they may lead to a more nuanced

understanding of far-right extremist homicide. These variables include homicide location,

136

as well as suspects’ criminal history and if far-right suspects were under the influence of
drugs or alcohol during the commission of the homicide.

It is initially clear that far-right non-[M homicides were more commonly
committed indoors in comparison to [M far-right homicides. This, too, is inconsistent
with past research that has tended to characterize far-right homicides as events that occur
in public settings. As discussed further below, non-[M homicides often occurred indoors
in more common locations such as residences. In contrast, far-right [M homicides fit the
portrayal of [M homicide as a public act targeting random victims based on their social
minority statuses. Possibly the most high-profile event of this type was the cross-state
shooting murder by World Church of the Creator member Benjamin Smith on July 4,
1999. Smith began his murderous crime spree near Chicago killing an Afiican American
man and injuring a number of Orthodox Jews returning home from worship. On his two-
day rampage, Smith crossed into Indiana, killing an Asian graduate student outside his
church just off the campus of Indiana University. Smith’s case is relatively unique in that
be targeted multiple social minority groups in one homicide event.

In addition to indoor locations, non-1M homicides, more often than not, involved
suspects with past criminal records. This is predictable as far-rightists that commit non-
IM homicides would likely not be those that tended to specialize in [M offending.
Moreover, non-[M homicides involved suspects that were, proportionately, less likely to
be under the influence when compared to IM homicide offenders. This suggests that non-
[M offenders may be somewhat more premeditated than [M offenders. Importantly,
Messner et a1. (2004) also found use of illegal substances to be associated with [M

offending. They used this evidence as support for the hypothesis that [M offenders do not

137

necessarily specialize in ideological offending, but are more or less “versatile” criminals
(Gottfredson & Hirschi, 1990). This dissertation builds on their research by identifying
similarities and differences in [M and non-[M in regards to the use of drugs and alcohol
for far-right extremists. The findings on prior criminal history and alcohol/drug use
during the homicide event also indicate that it is not always appropriate to use these two
variables as similar indicators of suspects’ deviant behaviors. That is, it would not be
unreasonable to hypothesize that suspects, with prior arrest histories, involved in the
average homicide would also be those that tend to abuse drugs and alcohol. This,
however, does not seem to be the case when considering the different subtypes of far-
right homicide events. The findings on far-right extremist homicides support in part what
Hamm (1993) found concerning skinheads in the United States. Indeed, his research
found that skinheads rarely had serious prior arrest histories, but that offenders often
drank alcohol (i.e. beer) during their criminal activities.

The last homicide characteristic to discuss before addressing the multivariate
analysis findings are those associated with the age of far-right homicide suspects. The age
distribution across far-right [M and non-[M homicide events were very similar, so it does
not appear that [M offenses involved younger offenders, which was one plausible
explanation proposed in the last chapter to explain the similarities in far-right and the
average homicide offender’s age since 1990. The findings on age also suggest that it is
unlikely that far-right suspects commit different types of violence as they progress into
older age groups. Instead, it appears that far-right extremists are more or less versatile
criminals, or those who commit both [M and non-[M offenses (Gottfredson & Hirschi,

1990), and do so mostly at a relatively young age.

138

Thus far, a number of proportional similarities and differences between IM and
non-[M homicides committed by far-r1 ght extremists have been discussed based on
descriptive findings that are presented in Table 6.2. Motivational subtypes of far-right
homicides were comparatively examined in response to the descriptive findings provided
in the last chapter which suggested that far-right homicides may be a heterogeneous
phenomenon. The findings in this chapter have so far reinforced the notion that there are
at least two motivational subtypes within far-right homicide that can be identified,
although there remain some similarities between them. As past research on other forms of
homicide have suggested (Flewelling & Williams, 1999; Pizarro, 2009), disaggregating
homicide events in meaningful ways can advance our understanding of these complex
forms of violence. However, it is still unknown based on descriptive findings if there are
statistically significant predictors that can distinguish between these types of homicide
motivational subtypes committed by far-r1 ght extremists. The next section addresses this
research question.

Multivariate Findings

Seven binary-coded incident and suspect-level homicide characteristics were
entered into a multivariate binary-logistic regression as independent variables to
determine whether each variable could distinguish between the motivational subtypes of
homicide (dependent variable), net the effects of the other homicide factors. These
variables were selected because they have been found to be important in past homicide
research and are linked to substantive questions and debates in the hate crimes and
domestic terrorism literatures. These variables also represent theoretically-relevant

factors in which substantial data were available. Indeed, one of the advantages of the

139

open-source database is that it allows for the analysis of variables for which data on the

national level are limited (e. g. alcohol/drug use, criminal history). Binary-logistic

regression was used because it remains an appropriate statistical method to use when the

dependent variable is binary—coded ( l=far-right 1M, 0=far~right non-1M). The

multivariate findings shown in Table 6.3 reveal that there are three incident-level and a

single suspect-level homicide factor that serve as significant predictors of homicide

motivational subtype.

Table 6.3 Binary-Logistic Regression Findings for Predicting Motivational Subtypes:

 

 

(N-270)
B (SE) Odds Ratio (OR)
Incident Characteristics
Multiple Offenders -.557 (.406) 0.573
Gun-Related .532(.440) 1 .702
Profit-Motivated -1.587(.500)"”""‘ 0.205
Known Victim -2.024(.408)*"'” 0.132
Multiple Deaths -.407(.528) 0.666
Group-Related -.478(.440) 0.620
Occurred Indoors -.941 (.394)""' 0.390
Primary Suspect Characteristics
Juvenile -.322(.639) 0.725
Had Priors -.625(.385)"‘ 0.535
Under the Influence During .266(.519) 1.305
Constant 2.1 18(.460)"" --
Pseudo R2 .309 m
-2 Log Likelihood 181.111 «-

 

TPS- I 1111,3055!!! *pS.01""pS.001

tOnly one far-right homicide suspect for each homicide incident is considered.

The only suspect-level homicide characteristic that distinguished between

homicide motivational subtypes is evidence of suspects’ prior arrests found in the

140

available open-source information. In particular, homicides involving suspects who had
known priors were significantly (p50. 10)20 related to non-1M homicides compared to the
reference category (suspects with no evidence of priors). The type of arrest, such as a
violent crime, is not captured in this variable. While it is often possible to attain criminal
histories on state prisoners from open-sources,2| information on prior arrests that
occurred when suspects were younger than eighteen is nearly impossible to attain from
open or closed sources. Therefore, this variable is used as a way to capture the extent of
criminal behavior rather than the quality of these crimes. The descriptive finding found in
Tale 6.2 that both [M and non-1M far-right homicides have over 40 percent of primary
suspects with evidence of prior arrests is an important finding in itself, as this type of
information on prior arrests (i.e. any type of prior arrest) has not been measured in past
research. As stated above, past research has found that certain types of [M offenders,
such as skinheads, tend to not have serious criminal histories (Hamm, 1993; see also
Levin & McDevitt, 1993).22 The multivariate findings shown in Table 6.2 are also useful
in understanding the behavior of far-rightists, as they reveal that prior arrests are
significantly related to far-right non-[M homicides. While the quantitative findings
indicate a statistical relationship between suspects’ past criminal behavior and subtype
(non-1M) of far-right homicide, it remains unclear what types of far-right extremists are

involved in these non-1M homicides. Understanding how prior arrests operate to shape

 

20 As this is the first empirical study of far-right extremist homicide, and this Study is in many ways
exploratory, variables found to be significant at the .10 level will be considered. Future research will extend
this study by increasing the sample size (2008 and 2009 homicide data) and by evaluating the utility of
articular homicide measures using stricter significance levels.
1 An example of an open-source in which information about some far-right suspects can be found at
http://www.theinmggelocatorcomh

22 It should be noted that Hamm’s data is based on a limited number self-report survey data and regards
only prior felony arrest.

141

fatal outcomes in the context of the homicide event could aid in understanding the
similarities and differences in far-right IM and non-1M homicide events.

In order to identify the different homicide subtypes of far-r1 ghtists and elucidate
the different situational circumstances that occurred, a variable—centered theme analysis
was conducted for the variable of prior arrest history. Information on each homicide
meeting these criteria was reviewed and open-source information was entered into the
qualitative data software package N Vivo8 which made identifying homicides that were

both non-[M and profit-seeking more efficient.

Table 6.4 Affiliation Type for Far-Right Homicide Suspects with Prior Arrests:
Non-[M [M

N (81) Percentage N (48) Percentage

White Supremacist Group Member 39 48.1 15 31.3
Non-Prison 28 71.8 12 80.0
Prison 1 1 28. 2 3 20.0

Self-styled White Supremacist 6 7.4 6 12.5

Skinhead 16 19.8 13 27.1

Neo—Nazi 10 12.3 12 25.0

Anti-Government Extremist 8 9.9 --- ---

Anti-Abortion Provider 0 0.0 2 4.1

Survivalist 2 2.5 -- «-

 

:EOnly one far-right homicide suspect for each homicide incident is considered.

The findings in Table 6.4 show that far-right extremist homicides involving suspects with
evidence of priors are proportionately likely to be White supremacists tied to formal
organizations. A large proportion of these events were perpetrated by the prison gang
Aryan Brotherhood. Nonetheless, of all [M homicides involving perpetrators with priors,
20 percent were committed in prison, revealing that violence has been a problem

experienced substantially outside of prison walls. This is an important finding as past

142

research on far-right extremist violence has largely ignored these offenders. Moreover, it
is difficult to categorize these offenders who appear to be the most criminally active of all
far-right extremists within existing typifications of far-right extremist violence. White
supremacist group members could possibly fit within Heitrneyer (2003) category of far-
right “subcultural violence.” While subcultural violence could easily apply to many of the
homicides perpetrated by skinheads and neo-Nazis (Hamm, 1993), the offending of
White supremacists in the United states is largely non ideologically-motivated, and when
motivated by ideology tends to target government officials and social minority groups.
Therefore, it may be useful to consider White supremacists as a unique category of far-
right homicide violence.

Table 6.4 also reveals that skinheads and neo-Nazis with priors were,
proportionately, more likely to commit [M homicide compared to non-1M events. The
most common type of homicide perpetrated by these types of offenders in which there
was evidence of prior arrests for suspects was anti-race (primarily anti-Afiican
Americans) homicide. The majority of these suspects were described as having links to a
skinhead or neo-Nazi group, though a formal group was not identified for these cases. In
this study, a suspect was coded as a particular type of far-rightist (e. g. “skinhead”) if they
were labeled as so in the open-source materials, criminal justice officials, themselves, or
some other source. It is possible that a group of young males simply called themselves a
skinhead gang despite not having any other organizational characteristics of a cohesive

group (e. g. leader, organizational goals). However, when a formal group was mentioned,

143

the Aryan Nations,23 arguably the largest and most violent neo-Nazi organization in the
United States, was the most common group linked to the homicide.

One homicide subtype that is not shared across motivational categories is “anti-
govemment extremist.” While, as has been discussed, different types of far-rightists share
similar beliefs, anti-government extremists are those that have as their primary issue
government corruption and often perceived government conspiracies (e. g. FBI, ATF,
IRS) against themselves. Seven of the eight events, involving anti-government extremists
with priors, happened to involve murders of law enforcement officers. TheSe suspects
usually had long histories with government authorities in the shape of filing bogus
lawsuits, failing to pay taxes, and making threats to local officials. In other words, these
far-rightists could be described as “ticking time bombs,” and when confronted by law
enforcement, go off. In these instances police operated at a disadvantage as suspects were
heavily armed and prepared to do battle against who they perceived as representatives of
a government posing a grave threat to their livelihood, constitutional rights, and personal
welfare.

As alluded to above, the statistical effects of suspect prior arrests on far-right
homicide motivation subtype are moderate to weak. One way this study tested the
robustness of this variable as a predictor of homicide subtype was to rerun the regression
model after sorting out those more “extreme” cases that involved homicide suspects with
weak affiliations to the far-right. As discussed in the last chapter, approximately 12
percent of the homicides involved suspects that held only weak affiliations to the far-right

movement. These homicide events tended to involve lone-wolf ideologically-motivated

 

2 . . . .
3 More Information about the Aryan Nations can be found at htrp:flwwwaryan-natrons.org/.

144

offenders who targeted strangers in outdoor locations. As expected, when the model was
rerun again without these homicide events included, the variable of prior offenses became
less Significant. Therefore, it is likely that these “mission offenses” were influencing the
relationship between this variable and homicide subtype. In addition, when these
particular homicide events were removed from the model, group affiliation became a
significant predictor. This aligns with the other descriptive findings shown in Table 6.3
which revealed that formal White supremacist gangs were the most prominent type of
non-[M offenders. Importantly, the situational variables remained significantly related to
homicide motivational subtype when the new model was run without the homicides
involving less intensely affiliated offenders.
Variable-Centered Theme Analyses

The remainder of this chapter focuses on exploring themes that capture the
situational variations in far-right extremist homicide. What follows is a discussion of the
findings stemming from the significant quantitative findings, as well as the variable-
centered theme analysis conducted for each of the situational findings found to be
significant. The purposes of the remainder of this chapter are to identify key
distinguishing characteristics of far-right homicide subtypes and to further the
understanding of this form of violence by qualitatively describing how important
situational homicide factors operate to shape these fatal outcomes. This is accomplished
by identifying themes capturing the influence of situational variables on non-[M events,
as well as Show how these variables shape IM homicides in order to better understand the

similarities and differences in these motivational subtypes.

145

Pro tt—Seekin Homicides

The first key situational finding, as shown in Table 6.3, is that homicide events
that involved profit-seeking behaviors were significantly more likely to be associated
with non-[M far-right homicides, net the effects of the other homicide characteristics
considered. This finding is consistent with the bivariate results discussed previously.
Therefore, it is clear that while far-rightists do commit homicides involving profit-
seeking behaviors, they tend to be motivated primarily by non-1M circumstances. The
question remains as to how profit-seeking behaviors by offenders operate in the context
of the entire homicide event. What are the common themes or types of situations for
profit—seeking non-[M homicides? That is, were these profit-seeking behaviors also
associated with [M homicide offenses as well, and if so, how? As will be shown, the
ability to address questions such as these is one of the advantages of approaching this
topic from a criminal event perspective and from a mixed-method approach.

A variable-centered theme analysis was conducted to identify how this particular
characteristic shaped the fatal outcomes of interest. Seven variable-centered themes were
identified revealing the multiple ways profit-seeking circumstances shape far-right
homicide events. Table 6.5 shows that robbery of persons was the most salient subtype of

non-[M far-right homicide.

146

Table 6.5 Profit-Seeking Behaviors for Far-Right Homicide

 

 

Non-[M [M

Variable-Centered Theme N (30) Percentafl N (14) Percentage
Robbery of person 1 1 36.6 14 100.0
Robbery of business 5 16.7 --- ---
Robberies Committed in Preparation for “War” 4 13.3 --- ---
Within-Group Dispute and Robbery 3 10.0 --- ---
Revenge Killing and Robbery 3 10.0 -- ---
Police Murder During Escape of Robbery 2 6.7 --- ---
Drug Deals Gone Awry and Robbery 2 6.7 --- ---

 

The majority of non-[M robberies of persons had little or nothing to do with their far-
right extremist beliefs or affiliations, as the forceful taking of money or property was the
primary objective of the event. Robbery was the primary objective of all of the business
robberies as well, the second most common variable-centered theme, although it was
clear that in at least two of these homicide events, offenders found it easier to murder the
business employees because of their nationality and ethnicity. Therefore, while robbery
was the primary motivating factor, ideological beliefs shaped these fatal events to some
degree. Other types of non-[M homicides were directly related to the far-r1 ght movement.
In particular, a number of homicides were carried out in the process of robberies
committed to gain money for what suspects viewed as an upcoming racial struggle or
“race war.” These involved crime sprees, and in some instances, involved multiple
murders. One high-profile example was the murder of a family by far-right extremist
Chevie Kehoe and one of his accomplices. Kehoe became well-known for his non-fatal
shoot-out with police that was caught on police camera after a traffic stop. Prior to the
shoot-out, Kehoe had robbed a family known to him as family acquaintances who also

happened to be gun dealers. Kehoe robbed the family of their guns, money, and other

147

valuables, and then with his accomplice murdered the family, including a father, mother,
and young child. The purpose of the non-1M robbery-murder was to gather weapons and
ftmds to ignite what the Kehoe’s viewed as an inevitable struggle or war between the
races. The Kehoe’s beliefs had been shaped by their father’s strict Christian Identity
teachings. Before being caught and sent to prison, Kehoe would be charged with two
other murders, neither one of which targeted social minority groups or other ideological
targets.

Another theme involving movement-related circumstances that emerged from the
open-source data on these events was within-group disputes, or arguments between far-
right extremist members of the same extremist organization. 24 Arguments included
disputes over membership status and disloyalty to the group, as well as other types of
disputes. In these instances, robbery tended to be a secondary motive. Moreover, the rest
of the homicides captured in the revenge and police murder themes involved robbery as a
non-primary circumstance.

Although profit-seeking behaviors were mostly associated with non-[M
homicides, some [M events also involved profit-seeking behaviors. As shown in Table
6.4, all of the seventeen far-right [M homicides involving profit-seeking behaviors were
robbery of persons. Only five of these events involved victims known to the suspects and
nearly half of the events targeted homeless individuals. 111 one way, these motivational
subtypes were similar in regards to prominent themes, as robbery of persons is the modal

category of non-[M events. However, by looking more closely at the different profit-

 

24 See Phillips (2009) who found that many “bias crimes” tended to be primarily motivated by
interpersonal conflicts rather than “hate.”

148

seeking circumstances involved in far-right [M homicide, two unique themes emerged.
What was unique about these events was how robbery was used as a way to firrther
victimize the target rather than used as a primary motive. In addition, robbery also served
the purposes of concealing identities of victims and for the collection of “souvenirs”25 in
addition to serving as a way to further victimize the target beyond the physical attack.
Ideological profit-seeking homicides often involved brutal beatings and stabbings of
victims known to be homosexuals or racial minorities, or homeless males. These attacks
were less likely to involve the use of firearms, and often involved instances of overkill.26
For instance, one case involved the violent murder of a Vietnamese student who was
stabbed twenty-two times in the heart, in addition to other wounds. The victim had no
money to give the perpetrator, although it was reported that he did offer his car keys at
some point during the attack. In another case, an African American woman was stabbed
nineteen times in addition to having her skull crushed, while another male victim was
burned on top of a mound of used tires because of his homosexuality in another brutal
homicide. As many of these events targeted the homeless, substantial monetary gain was
unlikely a primary factor in these cases. In sum, profit-seeking homicides were most
associated with non-[M homicides. Moreover, this subtype of homicide most commonly
involved robbery of persons in which robbery was a primary motive. However, variable-

centered theme analysis revealed that other non-1M homicides involved robbery as a

secondary motive, such as those homicides committed to further the far-right movement

 

25 See also Hamm (2007) in regards to far-rightists collection of evidence as “toterns.”

26 The term “overkill” is a subjective term used here to describe the use of torture (physical, sexual, and
psychological) against victims, as well as the unnecessary infliction of fatal and non-fatal injuries before
and after the death of the victim.

149

in ways such as preparing for race wars and maintaining group norms through fatal acts
of group-related discipline.

Examining the homicides as situational events also revealed how [M and non-1M
homicides were similar in a limited number of instances, as robbery of persons occurred
in both event subtypes. To illustrate, theme analysis revealed that all [M profit-seeking
homicides involved robbery as a secondary motive. These events were unique from non-
IM events as vicious stabbings and beatings were secondary to firearm-related homicides.

The findings have a number of implications for the use of the criminal event
perspective. In particular, the CEP illustrated the limitations of relying solely on simple
and predetermined categories of homicides (Decker, 1996; Flewelling & Williams,
1999). Profit-seeking homicides are not a homogeneous subtype of events. Some
involved robbery as a primary motive, while others involved robbery as a secondary
motive. Therefore, making conclusions solely on the quantitative distribution of profit-
seeking events would be misleading. Past research has suggested that there is a need to
exclude those [M events not primarily motivated by ideology (Berk et al., 1994), but the
advantages of comparatively examining these types of events is important. Moreover,
basing categorization schemes of homicide events on specific research questions Should
remain a priority. Analyzing closely related far-right homicide subtypes illuminates the
nature of those events of most interest to the current study and leads to more nuanced
explanations of these fatal occurrences. Relying on theme analysis also revealed a
number of findings regarding those homicides that did not fit the general patterns.
Focusing attention on “negative cases” or deviant homicides (Decker, 1996) advanced a

more nuanced understanding of this complex phenomenon. Indeed, examining how key

150

variables operated in the context of the entire event showed that profit-seeking IM
homicides more often involved intimate weaponry, outdoor locations, and robbery as a
secondary motive. These contextual variables helped shape the brutality of these events.
Focusing on the entire event for [M and non-[M homicide in this way also facilitated
conclusions about the uniqueness of [M events. It became clear that the [M profit-seeking
homicides tended to be much more brutal attacks that included sexual torture and
mutilation of victims. While both ended in death, some have suggested that the
viciousness of these attacks inevitably affects victims’ families, local communities, and
respective minority group communities.
Known Victim-Offender Relationships

The second statistically significant finding involves the relationships between the
variable “victim-offender relationships” and homicide motivational subtypes. In Chapter
5 it was found that far—right homicides in general tended to involve proportionately more
unknown victim-offender relationships relative to the average homicides that occur in the
United States during the same time period. Moreover, as Shown in Table 6.3, this study
found that some far-right homicides have, in fact, involved known victim-offender
relationships, however, these events were more likely to be associated with the non-1M
motivational homicide subtype when compared to [M homicides. Further statistical
analyses also indicated that in addition to being an important predictor of the far-right
non-[M homicide category, victim-offender relationship is the strongest predictor of this
type of homicide. By transforming the binary logistic regression coefficients into
“predicted probabilities” it is possible to examine how particular homicide variables

compare in regards to their predictive power of homicide motivational subtype relative to

151

each other when all other variables are set to their mean values.27 As shown in Table 6.6,
far-right homicides were most likely to be non ideologically-motivated when victims

were known to suspects (63%).

Table 6.6 Predicted Probabilities for Far Right Non-1M Homicide Events

 

 

Predicted Probabilities
Prior Arrests .49
Profit-Seeking .50
Known Victim-Offender Relationships .63
Indoor Locations .54

 

Questions remain as to how these types of homicides have occurred. Also, how do
the dynamic plotlines for non-[M events compare to [M events in which victims are
known? How do victim-offender relationships operate in the context of the entire event?
Non-ideological homicides committed by far-right extremists have never been
systematically examined, so studying situations where victim-offender relationships are
known can increase understanding of these events. Similar to the variable-centered theme
analysis conducted for the statistically significant finding of profit-seeking behaviors, a
variable-centered theme analysis was conducted to examine how known victim-offender
relationships have operated to shape fatal outcomes.

The most reoccurring theme of non-[M homicides in which victim-offender
relationship was known involves reactions by far-rightists to real or perceived disrespect

to the extremist group. While a limited number of homicides represented by this theme

 

27 See Liao (1994) and Spohn and Holleran (2001)
152

were reactions to disrespect toward the group’s belief system (e. g. verbally disrespecting
offender’s German heritage), the vast majority of these situations entailed disrespect to

group norms or rules, and, as was shown in the last section, only a few of these instances

involved robbery.

Table 6.7 Known Victim-Offender Relationships for Far-Right Homicide-1:
Non-1M [M

N (L4) Percentgge N (12) Percentage_

Murder of Disrespectful Group Affiliates 32 43.2 -- ---
Murder of Intimates and Family 12 16.2 -- «-
Petty Arguments with Acquaintances 9 12.1 --- ---
Anti-Police 6 8. l --- ---
Anti-Gay Victims 6 8.1 6 50.0
Victims of Convenience 4 5.4 -- ---
Anti-Black Victims 3 4.1 4 33.4
Anti-Abortion Providers 1 1.4 1 8.3
Anti-Sex Offenders 1 1.4 1 8.3

 

:tOnly one far-right homicide suspect for each homicide incident is considered.

Common themes included the murdering of group members or affiliates thought
(or otherwise known) to be informants. Other common themes included the murder of
group members who were thought to cheat the group financially or members who desired

to leave the group. In these cases, it appears that loyalty to the group was more important

than prior camaraderie.

Nonetheless, the next most common theme was very different from the first and
included the murder of love ones, intimate partners and other family members. These
homicides were exclusively committed outside of prison walls. Interestingly, a number of
these intimate homicides were part of a larger killing spree by the extremists. For

instance, three of these events involved young males “snapping” and killing their parents

153

or guardians, and suspects then going on to commit large-scale murders (e. g. school
shootings).

Other known victim-offender relationships involved police homicides and anti-
gay homicides in which offenders encountered the victims in their normal daily routine.
Homicides targeting police officers who had prior relations with suspects, for example,
occurred six times since 1990, highlighting the explosive threat that far-rightists pose to
law enforcement (National Consortium for the Study of Terrorism and Responses to
Terrorism, 2009). One theme that emerged from a review of these cases is that all were
initiated by the police officers who were responding to complaints about the suspects’
behaviors.28 For instance, in two of the cases, local police officers were called to the
home of the suspect because they were in their yards carrying large firearms. In all of
these cases, the responding officers were local law enforcement officers who had
previous, often frequent, contact with the suspect(s). Nonetheless, the suspects each time
perceived the arrival of the officers as substantial threats. As discussed earlier, far-
rightists often dislike police as they are viewed as the flout-men for the government they
view as corrupt (see also Freilich & Chermak, 2009; Pitcavage, 1998). Responding to the
perceived threat, officers were ambushed by the suspect’s gunfire. From these findings it
appears that officers who are known to suspects are not necessarily viewed as less
threatening to suspects than those who are not. It could be that an officer’s frequent
reappearance could trigger suspect’s suspicions of being watched by authorities (Freilich

& Chermak, 2009; Pitcavage, 1998).

 

28 Because there was no evidence in open-source materials that these anti-police homicides were primarily
motivated by the suspect’s anti-government sentiments, homicide events were coded as non-1M.

154

By analyzing non-1M homicides in which victims were lmown, this study was
able to illuminate this theme of anti-police homicide which highlights the danger that far-
rightists continue to pose to law enforcement, even when officers who fall victim are
known to suspects. Relying on the CEP and mixed method approach also allowed this
study to identify some of the anecdotal factors and to speculate on the suspects’
interpretations of the situations and how these interpretations, in combination with other
variables (e. g. location, presence of weapons) shaped these events.

Thus far findings have revealed that non-[M far-right homicides were more likely
to involve suspects with prior criminal histories, and involved situations in which profit-
seeking behaviors and known victim—offender relationships were relatively more
common. In addition, Table 6.3, shows that another situational variable, homicide
location, also successfully distinguishes between far-right homicide motivational
subtypes.

Indoor Location

Homicides that occurred indoors were also more significantly associated with
non-[M far-right homicide events relative to outdoor homicide events. This finding is
consistent with the portrait of far-right homicides emerging from the findings on other
significant predictors of homicide subtypes presented in this chapter, such as known
victim-offender relationships. That is, homicides occurring indoors are also more likely
to involve victims and offenders known to each other. However, it is still unknown what
types of indoor locations are the most frequent indoor location types for these non-[M
homicides, and the different ways in which the homicide location interacts with the other

homicide circumstances for both non-1M and non-[M events. Therefore, a final variable-

155

centered theme analysis is used to explore how this variable operates in the context of the
entire homicide event for far-right homicide motivational subtypes.

As shown in Table 6.8, the most salient indoor location for non-1M homicide
events was the residence. Moreover, the variable-centered theme analysis revealed that

victims in this homicide subtype were most often being murdered in their own homes

(63.1%).

Table 6.8 Indoor Location Themes for Far-Right Homicide
Non-[M [M

N (81) Percentgge N (39) Percentage_

Occurred at Victim Residences 37 45,7 3 20.5
Movement-Related at Victim '3 Home I] 29.8 --- «-
Petty Arguments with Acquaintances 10 2 7. 0 --- «-
Rampage Killings Beginning with Loved Ones 6 16.2 m m
Non-Movement Murder of Other F amily/F riends 5 l 3 .5 m «-
Robbery-Murders at Victim 's Home 5 13.5 --- ---

Prison Murders 14 17.4 7 17.9

Restaurant/Bar 5 6.2 3 7.7

Suspect Residences 10 12.3 2 5.1

Motel 4 4.9 3 7.7

Convenience Store 4 4.9 4 10.3

Other Residences 3 3.7 2 5.1

School 3 3.7 3 7.7

Medical Facility --- -- 3 7.7

Government Facility -- -- 3 7.7

Bank 1 1.2 m «-

Religious Institution m m 1 2.6

 

Such homicides involved a number of home invasions, but also included a number of
homicides that involved suspects and victims who lived together. Often suspects felt

victims had wronged them in some way. Indeed, an underlying storyline throughout

156

homicides meeting these criteria was the “settling of scores” by attacking the victims in
their homes.

Approximately the same proportion of non-[M attacks occurring in the home were
movement related (29.8%) as not related to the far-right movement (27.0%). Homicide
events associated with furthering the movement mostly consisted of “internal discipline”
cases that occurred within the home of the victim. As discussed above, these cases often
involved disrespect of group norms by victims, and suspects used homicide as a way to
seek revenge and send a message to others.

Those non-[M residence homicides not related to the far-right movement, in
contrast, consisted of circumstances such as retaliation for taking a bicycle, an insult to a
suspect’s daughter, the alleged theft of marijuana, as well as a list of other more mundane
catalysts of homicide events.

Phillip’s (2009) research on [M crimes in one large Eastern city also found that a
substantial amount of events involved, for example, petty arguments with neighbors and
other ordinary or everyday disagreements that Spiraled out of control and turned fatal.
Table 6.7 reveals, however, that while far-rightists have also been involved in this type of
[M homicide, these events have not occurred at victims’ residences.

F ar-right homicide events occurring at the home of victims is particularly
frightening as they often occurred unexpectedly as victims slept or when simply
answering their doors. Moreover, suspects often had intimate knowledge of victims (e. g.
addresses, work schedules) that made victims relatively easy targets. Also disturbing is
that a number of these events involved within-family homicides that served as only the

beginning of long deadly crime sprees for many far-rightists. Such homicides illustrate

157

the inherent danger of association with violent far-right extremists for those victims
affiliated or not affiliated with the movement themselves.

It is also evident fi'om Table 6.8 that homicides occurring in suspects’ residences
are proportionately more common for far-right non—[M homicide events. These events
disproportionately involved the murder of a police officer, and as discussed above
regarding the findings found in Table 6.7, many of these homicides involved suspects
known to the police officers who fell victim to far-rightists.

Finally, it appears that far-r1 ghtists committed proportionately similar proportions
of IM and non-[M homicide events in prison. Other than revealing that prison is a salient
indoor location for far-right homicide, this finding highlights the dual threat posed by far-
rightists in prison. That is, prison officials must be wary of allowing far-rightists to
associate with Black inmates as all but one of these events targeted Black victims. Prison
officials, though, also have to be cognizant of allowing far-rightists to associate with one
another clue to the within-group fighting that occurs behind prison doors.

Discussion

This chapter showed that it is inadequate to describe far-right extremists as strictly
[M offenders. Focusing solely on their ideological offending is capturing only one aspect
of the harm far-rightists inflict on society and their communities. The threat of far-right
extremist violence extends beyond social minority groups and governmental targets.

This study found that, in comparison to far-right IM homicide events, non-1M far-
right homicides tended to involve profit-seeking circumstances, to take place in private
settings, to target known victims, and tended to be perpetrated by offenders with more

lengthy criminal records. There were a number of similarities and differences across

158

motivational subtypes identified by quantitative analyses, and a number of more nuanced
findings were uncovered using variable-centered theme analyses. For instance, both far-
right non-[M and [M often involved robbery, though the purposes of robbery varied.

This study also built on the findings of Messner et a1. (2004) who found that [M
offenders were generally not crime specialists, but instead were versatile offenders.
Messner et al. (2004) explored racial assaults rather than [M homicides, so while the
findings of this research can extend their findings, they are not directly comparable. As
discussed more fully in the final chapter, further research is needed to comparatively
examine characteristics of offenders (and incidents and victims) involved in less extreme
forms of far-right violence (i.e. assaults, attempted homicides).

It is also clear from the findings that association and interaction with violent far-
right extremists is extremely dangerous, even when victims and offenders are known to
each other. Indeed, it was shown that far-rightists’ violence can often be directed at
acquaintances, as well as loved ones and other family members to a less extent. In
addition, these findings highlight the consequences of incarceration on inmate recidivism.
One reoccurring theme for non-[M far-right homicides was within-group violence, often
involving organized prison gangs. Offenders often stated in the open-source materials
that membership in White supremacist gangs was required for survival in prison. In
groups such as the Aryan Brotherhood the motto remains “blood in, blood out.” These
findings reveal the need for policymakers and criminal justice officials to find ways to
reduce the influence of prison gangs in state prisons and reduce the need for inmates to

join them.

159

Importantly, the findings presented in this chapter illustrated the complex links
between terrorism and crime. Past research has indicated that most terrorists, including
domestic far-right extremists, are charged with criminal acts rather than acts of terrorism
(Smith, 1994). Others have also discussed the ways that frameworks for studying crime
can be translated to the study of terrorism (Hamm, 2007; LaFree and Dugan, 2004). This
research illustrated the extent to which terrorism and crime can overlap. F ar-right
extremists have been responsible for nearly proportional amounts of criminal and terrorist
homicides in the United States between 1990 and 2008.

In addition, this chapter showed that it is possible to distinguish between far-right
homicide events based on motivational subtypes. F ar-right homicide offenders pose a
criminal threat to the public that is unique from that of the threat of [M violence.
Remaining is the question of the uniqueness of the far-right [B homicides, or those
targeting social minority groups. Addressing this set of research questions is important to
confirm or expand past classifications of [B violence. In response, the comparative
analyses in the next chapter will reveal the potential effects of affiliation with the far-
right movement on the relative nature of [B homicide in the United States. Understanding
similarities and differences in [B violence by suspect types can aid in the training and
investigating strategies of law enforcement faced with this type of serious crime. Chapter
7 also addresses how [B violent events are unique from domestic terrorism events, as
these types of violence are often considered separately by different scholars and are
associated with separate scholarly literatures, despite obvious similarities in the nature of

these types of extremist violence (Gruenewald et al., 2009; Hamm, 1998).

160

Chapter 7: A Comparison of Far-Right and Non Far-Right Identity-Based (1B)

Homicides

While Chapter 6 examined the similarities and differences in far-right homicide
motivational subtypes, the focus of the current chapter is on the nature of identity-based
([B) homicide events committed by far-right extremists and their comparison to [B
homicides not perpetrated by far-right extremists. While the last chapter concluded that
disaggregating homicide events by motivational subtypes was useful for understanding
the complexity of far-right homicides, such comparisons reveal little about the nature of
IE, in particular, and how these events may vary by what types of offenders are involved.
Although past homicide research has found that the nature of homicide events vary in
pattern depending on suspects’ affiliations with gangs (e. g. Maxson et al., 1985; Pizarro
& McGloin, 2006), past studies have neglected to consider the influence of far-right
affiliations on [B homicide events.

The findings presented in this chapter are unique from those presented in Chapter
6 in that only [B homicides are considered and the subtypes of homicide compared are
defined by the type of suspect rather than the primary motivation for the event. The first
subtype is far-right [B homicide. These events were perpetrated by far-right extremists
and only those homicides targeting social minorities, including groups defined by
homelessness, racial/ethnic, sexual orientation, religion, and nationality statuses, were
considered for the purpose of meaningful comparison. Therefore, far-right homicides
targeting victims such as government authorities and abortion providers were not
included in this particular analysis. The second subtype of homicide included in this
analysis is [B homicides not perpetrated by far-right extremists. Thus, none of the

suspects in these events had known ties to the far-right extremist movement.

161

Focusing on [B homicide allows the current study to address the third research
question posited regarding the similarities and differences in [B homicide events across
suspect subtypes. This chapter’s exclusive focus on [B homicide is a substantial
contribution to the [B crime (or hate crime) literature. In the past, scholars have suggested
that far-right extremists are responsible for only a minor portion of [B crimes in the
United States. For instance, Levin and McDevitt (1993) have estimated that organized
hate groups, typically far-right extremist groups in the United States, are responsible for
10 to 15 percent of all [B crimes (see also Bradley, 2007; McDevitt et al., 2002).
Although accurately estimating the proportion of [B offenders involved in the far-right
movement may be nearly impossible, it is safe to assume that only a minority of these
crimes involve far-right extremists. Nonetheless, researchers of [B crimes have deemed
far-right extremist [B violence an important topic of research (Levin and McDevitt, 1993;
Perry, 2001), as, in some instances, far-right extremism has fueled high-profile [B
homicides, such as the spree killing case of Benjamin Smith mentioned in the last
chapter.

How the nature of far-right [B homicides is similar to and different from [B
homicide not perpetrated by far-right extremists has not been addressed systematically,
though the limited research that has been conducted suggests that there may be
differences in these forms of violent offending across suspect subtypes. For instance,
Levin and McDevitt’s (1993) important work on classifying [B criminal offenders has
provided a typology that suggests far-right offenders tend to be “mission offenders.” That

is, far-right extremists are more likely to commit violence by seeking out and targeting

162

random offenders based on their minority statuses, while other IB offenders tend to
commit violence for “thrills,” as well as for other reasons (see also McDevitt, Levin, &
Bennett, 2003). Other researchers have examined the violence committed by certain types
of far-right extremists, namely skinheads (Hamm, 1994; see also Heitrneyer, 2003). This
research has suggested the far-right IB violence may be unique from other IB violent
offending in the type of weaponry used. For instance, skinheads committing IB violence
have been associated with the use of “billy clubs,” baseball bats, Doc Martin steel-toed
boots, and other blunt objects that reflect the working-class cultural influences on this
group (Hamm, 1993). It is also plausible to expect that far-right [B violence is more
likely to involve multiple suspects targeting a single or a small number of victims, as
Skinheads and other far-right extremists are bound together by their similarly held beliefs
and organizational structure, however loosely bound either might be. Finally, Hamm
(1993) and others have suggested that far-right extremists tend not to be the criminal-type
or have extensive prior criminal histories. In this way, skinheads and other groups of far-
right extremists may be unique in that the they do not necessarily have lengthy criminal
pasts, but instead are average, law-abiding, working-class young males. Therefore, it is
possible that far-r1 ght extremists who commit [B homicides will be less likely to have
prior criminal histories than those IB homicide offenders not affiliated with the far-right
movement. While the previous chapter found that far-right non-[M offenders were more
likely to have prior criminal histories than far-right [M offenders, the comparison of [B
homicide events in terms of prior criminal records across suspect subtypes remains

unclear.

163

To address these gaps in past research, the current chapter seeks answers to a
number of more focused research questions, including 1) Are far-right [B homicides
more likely to involve random strangers as targets? 2) Are far-right [B homicides more
likely to involve non-firearm weaponry, such as blunt objects? 3) Are far-right [B
homicides more likely to involve multiple offenders? 4) Are far-right [B homicides less
likely to involve profit-related circumstances? and 5) Are far-r1 ght offenders involved in
[B homicides less likely to have prior criminal histories compared to those not associated
with the far-right?

As discussed in the presentation of the methodology in Chapter 4, both far-right
and non far-right IB events were identified from open-sources. The primary difference in
these events is that far-right [B homicides involved at least one suspect associated with
the far-right movement. Open-sources included, for instance, existing databases (e. g.
RAND-MIPT, GTD), official sources (e. g. FBI’s Terrorism in the United States annual
report), scholarly reports and articles, watchdog reports (e. g. SPLC-Intelligence Report),
and systematic media searches. Additional open-sources were also searched to identify 13
homicides not involving far-right extremists for which there were clear indicators that
victims were selected based on their social minority status (e. g. Human Rights Watch,
National Asian Pacific American Legal Consortium, Anti-Violence Project).29 The
collection of a set of [B homicides not involving far-right extremists required the use of
additional resources, separate from those extracted from the Extremist Crime Database

Project (ECDB Project) run by Drs. Freilich and Chermak. As discussed briefly in

 

29 To access these original open-sources, go to hitp:Nwwwavporg/gublicalions.him and
http://www.hrc.org/documents/AfChronology_of Hate__Crimes.ndf.

164

Chapter 1, the primary aim of the F ar-Right Extremist Homicide Project led by the author
and funded by the National Consortium for the Study of Terrorism and Responses to
Terrorism (START) was to supplement the homicide data contained in the ECDB with
non far—right [B homicide data for the purposes of comparison. The resources provided
by START were used partially to fund the author and a number of undergraduate research
assistants who worked to identify these homicide events. Once the set of non far-right IB
homicide events were identified, the author and research assistants collected open-source
information on each homicide. Data were then inputted and coded into three separate
codebooks, similar to far-right homicide events from the ECDB, including variables for
homicide incident, suspect, and the quality/quantity of open-source information for each
event.30 The data preparation process for the non far-right [B homicide events took
approximately one year to complete. Once data preparation was completed for the time
period (1990 to 2008), data were comparatively analyzed using bivariate and multivariate
statistical techniques.

This chapter is presented in a number of sections. First, descriptive statistics on
the nature of far-right [B homicides and non far-right [B homicides are compared and
discussed. Second, multivariate findings are presented and a discussion of statistically
significant predictors of homicide suspect subtypes is provided. Third, this chapter
presents descriptive findings on the multiple types of victims targeted in [B homicide
offending, and then discusses the similarities and differences in those victims, as well as

the implications of these findings.

 

30 Data for non far-right [B homicide victims were also collected but were not used in the current study.

165

Descriptive Findings: Comparing Far-Right [B and Non Far-Right [B Homicide

Based on the descriptive statistics shown in Table 7 .1, it appears that there may be
a number of differences in far-right and non far-right [B homicides, many of which
support the expectations discussed above that are based on past research findings. First, a
relatively high proportion of far-right [B homicides involved multiple offenders and
involved victims unknown to the perpetrators.

These events were also proportionately less likely to involve profit-seeking
behaviors, as expected. As this is the first time that research has moved past anecdotal
evidence to provide an empirically—driven descriptive account of these homicide events,

and how they compare, it is encouraging that findings confirm past assertions made about

far-right [B violence.

Table 7.1 Descriptive Findings on Far-Right and Non Far-Right Identity-Based (IB) Homicide
1N =232) a

Far-Right Ideological Non Far-Right Ideological

 

 

(N891) (IV-141)
N Percent N Percent
Incident Characteristics
Number of Offenders (N=85, 129)
Multiple 37 43.5 55 42.6
Single 48 56.5 74 57.4
Weapon (N=87, 130)
Gun 39 44.8 53 40.8
Knife 19 21.8 27 20.8
Blunt 14 16.1 10 14.6
Bodily 12 13.8 19 14.6
Other 3 3.4 21 16.2
Circumstance (N=91, 141)
Not for profit 74 81.3 110 78.0
Profit-related 17 18.7 31 22.0

Victim-Offender Relationship (N=81, 130)

166

Table 7.1 continued

 

68 84.0 80 61.5
Strangers
Friends/Acquaintances 13 16.0 48 36.9
Domestic Partners 0 0.0 l 0.8
Other family 0 0.0 1 0.8
Total Death (N=91, 129)
Single Victim 78 85.7 121 93.8
Multiple Victim 13 14.3 8 6.2
Location (N=87, 115)
Outdoor 55 63.2 65 56.5
Indoor 32 36.8 50 43.5
Primary Suspect Characteristics
Age (N=90, 131)
1-18 10 11.1 23 17.6
19-28 45 50.1 70 53.4
29-38 28 31.1 20 15.3
39-48 4 4.4 14 10.7
49-58 3 3.3 2 1.5
59 to highest 0 0.0 2 1.5
Criminal History (N=91, 141)
Priors 38 41.8 43 30.5
No Priors 53 58.2 98 69.5
Drug/Alcohol Use During Incident (N=91,
141)
Yes 23 25.3 36 25.5
No 68 74.7 105 74.5

 

:l:Only one far-right homicide suspect for each homicide incident is considered.

What remains unknown, however, is whether these homicide characteristics can
statistically distinguish between homicide suspect subtypes. That is, can homicide
variables significantly predict homicide suspect subtypes? In order to answer this
question, a multivariate binary-logistic regression analysis was conducted.
Multivariate Findings

The regression model shown in Table 7 .2 is similar to the multivariate model
presented in Chapter 6, except the binary-coded dependent variable is suspect subtype

(l=far-right, 2=non far-right) rather than motivational subtype.

167

Table 7.2 Binary-Logistic Regression Findings for Predicting Homicide Suspect Subtype
(N=232):

 

B (SE) Odds Ratio (019

Incident Characteristics

Multiple Offenders -.519(.375) 0.595

Gun-Related -.133(.37l) 0.875

Profit-Motivated .042(448) 0.959

Known Victim 1.248(.455)*" 3.485

Multiple Deaths -.696 (.606) 0.498

Occurred Indoors -.208(.407) 0.812
Primary Suspect Characteristics

Juvenile .452(.504) 1.878

Had Priors -.121(.365) 0.886

Under the Influence During .279(.398) 1.322
Constant .313(.410) ---
Pseudo R2 .120 ---
—2 Log Likelihood 197.310 --

 

“psi"p5.05$'"ps.01""pSDOl
:l:On1y one far-right homicide suspect for each homicide incident is considered.

AS shown in Table 7.2, the only statistically significant variable that distinguishes
between [B homicide suspect subtypes is victim-offender relationship. Specifically,
known victim-offender relationships are statistically associated with [B homicides not
perpetrated by far-right extremists. This supports how far-right extremist homicides have
been characterized in past research (i.e. as mission offenders targeting unknown victims).

In addition, Chapter 5 discussed how some far-rightists have been more or less
intensely affiliated (scale of 1-4) with the far-right extremist movement. Chapter 6 found
that when those cases involving less intensely affiliated suspects (suspects who scored 1
or 2 on the intensity scale) were removed fi'om the multivariate analysis, the variables
that significantly predicted motivational subtypes changed Slightly. Removing the less

intensely affiliated suspects in the model presented in Table 7.2 is also important, as the

168

model is used to predict suspect subtypes (far-right or non far-right). Indeed, it is
plausible that when those cases involving less intensely affiliated suspect are removed,
the suspect subtypes would become more clearly distinguishable, and those predictors
that are most important for distinguishing these subtypes would become more prominent.
Therefore, as done in the last chapter, the multivariate model was re-run with homicides
involving suspects who did not hold strong ties to the far-right movement removed. In
this model, victim-offender relationship remained a strong predictor of suspect subtype
(p305, OR=2.98). Another variable, the number of offenders involved in the homicide
event (p305, OR=O.36), also emerged as a significant predictor of suspect subtype for [B
homicide events. Specifically, homicide events involving multiple offenders were
significantly associated with far-right [B suspects compared to those events involving
only single offenders. These findings align with those that emerged from the affiliation
analysis found in Chapter 5. That is, those suspects less intensely affiliated with the
movement usually offend alone. It is possible to argue, then, that those homicides
involving less intensely affiliated suspects have much in common, in terms of their nature
with non far-right homicide events.

Although [B homicide suspect subtypes can be distinguished by victim-offender
relationships and the number of suspects involved, these subtypes are statistically
indistinguishable across a number of incident and suspect-level characteristics. The
purpose of this chapter is to not only identify differences in [B suspect subtypes, but also
to recognize the similarities and attempt to explain them. The similarities indicate that the
involvement of suspects in the far-right extremism movement does not necessarily

translate into distinct behaviors patterns detectable at the homicide event level as might

169

be expected. For instance, both far-rightists and non far-rightists largely relied on non-
firearm weapons and often offended in public or semi-public locations. One explanation
is that [B violence, such as beatings of homeless persons and gay bashings, and the belief
systems behind this violence, are not necessarily subcultural phenomena (Perry, 2001).
Perry (2001), for instance, has argued that homophobia and animus toward homeless are
mainstream values. Therefore, it should not be surprising that how these brutal attacks
tend to be more similar than different across suspect subtypes.

The similarities in suspects’ prior arrests and use of alcohol and drugs during the
homicide event suggest that far-rightists are not more or less likely to engage in deviant
lifestyles than non far-rightists. Based on these findings, one conclusion is that the far-
right movement does not appear to insulate or promote deviant behaviors among its
affiliates. What is more, the important differences in suspect subtypes may be in the
precursor attributes of homicide events. That is, in how suspects behavior prior to the
attack and how targets are selected. In particular, far-rightists are more likely to select
victims at random or who are otherwise unknown to them. Though these quantitative
analyses provide information on the most important predictors of homicide suspect
subtype, it remains unclear as to the most common situations or homicide storylines that
were most associated with homicides involving known victim-offender relationships
across far-right and non far-right [B homicide events. A quick descriptive analysis
reveals that by far the most prevalent type of known victim targeted for [B homicides not
perpetrated by far-right extremists were members of GLBT communities, while on the
other hand, the most common subtype of homicide committed by far-rightists was anti-

Black homicide.

170

Figure 7.1 Identity—Based (IB) Homicide Victim Types by 5
Suspect Affiliation: it 1

80 T............ _..o_.....-.._ -.... _....._... _ a l

 

Far-Right Identity-
Based

 

 

 

qtact-3:2

W‘ I Non Far—Right Identity—
Based

 

Anti-Black Anti-GLBT

 

:EOnly anti-Black and anti-GLBT honricides are considered in this figure.
:lztOnly one far-right homicide suspect for each homicide incident is considered.

In order to further examine the different types of anti-GLBT, and other [B
situations in which suspects murder victims who were known to them, a variable-
centered theme analysis was conducted for victim-offender relationship. The theme
analysis was conducted in order to identify how the nature of the relationships between
victims and offenders interacted with other situational variables to shape these fatal
outcomes. A number of themes shown in Table 7.3 were identified that shed light on the
influence of victim-offender relationship on this homicide subtype.

Variable-Centered Theme Analysis: Known Victim-Offender Relationships

There were two general types of variable-centered themes of known victims, anti-

GLBT and anti-race/ethnicity homicides. Within the most common subtype of anti-

171

GLBT homicides, a number of unique scenarios emerged from the open-source
information on each event. The most common anti-GLBT theme identified for non far-
right homicide events was “sexual advance” homicides, or homicides that were carried

out following a real or perceived sexual advance by a victim perceived to be gay.

Table 7.3 Known Victim-Offender Relationships by Non Far-Right Suspects

 

 

 

Non Far-Right Far-Right
N (43) Percentage N (l 3) Percentafi
Anti-GLBT Victims 32 74.4 4 20.0
Sexual Advance 12 37.5 --- ---
Anti-Homosexual 10 31.3 --- ---
Anti- Transgender 5 15.6 --- ---
Stranger “Pick-ups ” 4 12.5 --- ---
Anti—Bi Sexual l 3.1 --- ---
Anti-Race/Ethnlcity Victims 11 25.6 7 35.0
Anti-Black 5 45.4 7 100.0
Anti- White 2 I 8.2 --- ---
Anti-Asian 2 18.2 m ---
Anti-Arab 2 18.2 --- ---
Anti-Police -- — 6 30.0
Anti-Abortion —- —- 1 5,0
Anti-Sex Offender -- .— 1 5,0
School Shooting -— - 1 5.0

 

Sexual advance homicides were included in the current study if another indicator of bias
was present (e. g. homophobic slurs were reported). For these homicides, victims were
known to offenders, but often only for a short period of time. For instance, offenders and
victims sometimes met at bars or parties and then returned home together for the first
time. After a real or perceived sexual advance, the victim was killed, often brutally, by

the suspect. Such homicides often went on to be described as situations in which suspects

172

experience “gay panic.” What is important to note about these homicides is how they
have been captured in this study as “known victims” as presented in the quantitative
findings. Prior research on hate crimes, however, has found that events codified as
involving known victim-offender relationships may be misleading. In particular, Mason
(2005), examined racial and homophobic harassment reports recorded to the Metropolitan
Police Service in 2001 and found that while victims were aware of the offenders, they
often did not have close relationships with them (e.g. coworkers and neighbors). As is
also the case in this study, Mason (2005) suggested that studying the contexts in which
victims and offenders interact can advance a more holistic understanding of crime events.

The second most common theme for non far-right homicides is simply referred to
as anti-homosexual homicides, including murders of gays and lesbians. These homicides
involved a mixed bag of circumstances and motives, including what was often described
as “hatred of homosexuals” in open-source materials. Gay males were also targeted in
some instances because they were perceived as “easy targets”, for instance, in robbery-
homicide events. Because the current study defines anti-GLBT events as [B if the victim
was targeted because of their sexual orientation status, regardless of suspects’ thoughts or
feelings concerning homosexuality, it was possible to uncover the heterogeneity within
this form of [B homicide.

Though less common, another anti-GLBT theme identified, anti-transgender
victims, is intriguing in light of a recent court ruling. The majority of states in the United
States do not consider GLBT victims as protected under hate crime laws (Jenness, 2003).
Moreover, the issue of transgender victims being considered a protected group under

such laws is especially controversial. However, a recent decision by a Colorado jury in a

173

landmark case found a man guilty of a hate crime for beating a transgender victim to
death with a fire extinguisher after he discovered that the victim (with whom he had
sexual relations) was a biological male (F rosch, 2009). Likewise, in a number of cases
included in the current study, suspects murdered transgendered victims after acting out in
what has been referred to as violence committed in the “heat of the moment” or was a “
crime of passion.” Other anti-GLBT homicides included “pick-up” events in which
suspects specifically targeted gay men in order to rob and murder them. Interestingly, the
few cases belonging in the category were all instances of serial homicide offending, in
which suspects moved from location to location preying on homosexual men. Again, in
these events, the relationship between victims and offenders was, most often, a day, two
days, or a week, though they were codified as known victims to offenders.

Although anti-GLBT homicides were proportionately more associated with non
far-right homicides involving known victims, there were also homicide events that did
not fit into this pattern. Indeed, 20 percent of far-right homicides involving known
victims targeted members of the GLBT community. For example, in one case two
brothers belonging to a far-r1 ght extremist group broke into the home of two homosexual
men as they lay asleep in their bedroom. The couple was openly gay and well-known in
the community. During the trial, one brother stated that his older brother had informed
him that it was necessary to murder the gay men if they really believed that
homosexuality was evil, as the group they belonged to preached. Following the murder,
the younger brother was imprisoned for life and the older brother committed suicide in

prison.

174

Anti-GLBT Homicides as Crime Events

The components of the CEP facilitates a more comprehensive understanding of
anti-GLBT homicides committed by both far-rightists and non far-rightists. This research
approach emphasizes that there are multiple stages of a criminal event, each affecting the
other, including precursor factors, transaction factors, and aftermath factors (Sacco &
Kennedy, 2002). There were a number of unique precursor factors that fireled these
particular anti-GLBT homicide events. For example, though the exact perceptions of the
situation or the precise motives of the crime participants prior to the event cannot be
known, sexual advance homicide precursor factors often involved brief meetings between
victims and offenders at establishments or parties where alcohol was served. It then
followed that victims and offenders would go to another location, such as the victim’s
residence or motel. Transgender homicides often involved the same precursor attributes.
Though various types of anti-GLBT homicides, such as these, are similar in more
intimate interaction between participants, the transactions of sexual advance and
transgender homicides usually involved slightly different elements. In particular, sexual
advance homicides involved a real or perceived sexual threat from the victim. The violent
act of murder occurred both immediately and shortly thereafter when suspects planned
the homicide (and often recruited accomplices). On the other hand, homicides of
transgender victims often involved a sexual encounter between the victim and offender in
which the suspect was apparently unaware of the “true” gender of the victim. Violence
occurred before the encounter could be completed upon discovery of the victim’s gender,
or much later in cases where the victim’s gender was discovered by the suspect following

the sexual encounter. One common element of anti-GLBT homicide events was the

175

apparent overkill involved, including acts of mutilation of body parts, decapitation and
the beating of victims beyond recognition. Interestingly, a common characteristic of the
aftermath of anti-GLBT homicides was the immediate disclosure (or bragging) of the
violent transaction to friends, family, and even the police in some instances. A number of
explanations for this subtype of [B homicide events can be applied to the behaviors of
those who perpetrate them.
Explaining Anti-GLBT Homicide

One theoretical perspective that works toward explaining the various forms of
anti-GLBT homicide described in this chapter is the “doing gender” perspective (West &
Zimmerman, 1987). This perspective supposes that criminal activity is used as one tool or
resource for reaffirming one’s sexual identity once it has been threatened or compromised
in some way. Indeed, real or perceived sexual advances by victims and the feelings of
being deceived by transgender victims following sexual encounters were viewed by
suspects as threats to their masculinity. As immediate or delayed responses, the violent
and deadly acts of violence can be conceptualized as resources for re-establishing
suspects’ heterosexual sexual identity.

The concept of “saving face” can also be used to explain the often brutal anti-
GLBT homicide events. Scholars, like Athens (1980) and Katz (1988), have used the
notion of saving face in their qualitative descriptions of dynamic homicide interactions
(see also Goffrnan, 1967). Indeed, sexual orientation is arguably an integral aspect of
suspects’ identities and experiencing threats to this part of themselves, especially in
public settings, can cause suspects to experience the loss of face. Luckenbill and Doyle’s

(1989) integrated approach to explaining violence sheds light on this form of violence.

176

Their perspective integrated concepts of cultural honor, structural position, symbolic
interaction, and situational context. They hypothesized that the chance of violence was
likely to increase when the “self” was attacked in public settings and when perceived
threats came from those of a different social status. Moreover, the need to inform others
of the murderous transactions following the homicide could be explained by the need of
suspects to regain “face” following such attacks to the self.

In addition to using homicide as a resource for “doing gender,” one possible
situational explanation for the weak relationship between far-right [B homicide and
known victim-offender relationships compared to non far-rightists is that gay victims are
less likely to associate with far-right extremists due to their appearance and their attitudes
toward social minority groups. It is less likely, for instance, that far-right extremists and
gay males would find themselves in the same types of meeting places, such as bars and
parties to which both would be invited. It is also less likely, in general, that there would
be ambiguity in the relationships between far-right extremists and gay companions after
their initial meetings. In this way, it is becomes evident how daily routines, perceptions
of victims and offenders, and location of homicide events could interact to shape these
fatal outcomes.

Moreover, one of the key aspects of the CEP is to firrther understanding of
complex phenomena like [B homicide events, and this study has attempted to do so by
also considering homicide events that do not fit the patterns (i.e. negative cases) indicated
by the quantitative findings. To illustrate, as shown in Table 7.3, there were a limited
number of far-right [B homicides that involved known victim-offender relationships. The

most prominent theme that emerged from this subset of events was anti-race/ethnicity

177

homicide circumstances. Interestingly, most of these events (71%) involved
circumstances in which suspects disapproved of the victims’ bi-racial relationships.
White supremacists and White separatists often view “race-mixing” as the most offensive
type of behavior as it is viewed as tainting the purity of the White race (Ferber, 1998).

In short, multivariate analysis showed that known victim-offender relationships
were significantly related to non far-right [B homicides. Theme analysis further explored
this subtype of homicide and found that sexual advance homicides were the most
common situational theme identified. This qualitative analysis also showed that while
victims were known, these relationships were often short-lived, and commonly involved
real or perceived sexual advances and deceit by victims. Theme analysis was a useful
technique for advancing knowledge of nuances in how victim-offender relationships
shape fatal outcomes in different forms of anti-GLBT homicide. The stages of anti-GLBT
homicides were presented and the precursor attributes contribution to the nature of the
violent transactions was discussed for different subtypes of anti-GLBT homicide. Finally,
the doing gender approach and concepts of “saving face” and the construction of the
“self” or self identity, as well as the consideration of the intertwining of situational
factors and actors’ decisions, were used to provide partial explanations of these complex

and dynamic homicide events.

178

CHAPTER 8: DISCUSSION

This final chapter accomplishes a number of objectives. First, it reviews the major
components of the current project and the major findings stemming from it. Second, this
chapter highlights the implications of the major findings relevant to academic, law
enforcement, and policyrnaking communities. Third, research limitations of this
dissertation research are reviewed. Finally, plans for future research that extend the
current study and advance knowledge of far-right homicide events are outlined.
A Review of the Extremist Homicide Project and Major Findings

Prior to this research, the topic of homicide events perpetrated by far-right
extremists had not been systematically examined for a number of reasons, including the
great obstacle of a lack of available data on these events. This study approached the topic
of far-right extremist homicide from a criminal event perspective (CEP) (Block, 1981;
Sacco & Kennedy, 2002), using a comprehensive research framework to guide the
theoretical, research design, and methodological elements of this study. One advantage of
utilizing the CEP is its facilitation of the positing of original research questions and
nuanced research methods. In particular, this dissertation was a comparative study that
examined questions regarding the comparison of far-right extremist homicides and
“average” American homicides (Chapter 5), the far-right [M and non-[M homicides
(Chapter 6), and the far-right [B homicides and non far-right extremist [B homicides
(Chapter 7).

Addressing these research questions required the creation of an original open-
source comprehensive database on all far-right extremist homicide events, as well as on

all [B homicides not perpetrated by far-right extremists. In addition, a random sample of

179

all homicides that occurred in the study’s time frame was gathered from the F BI’S UCR-
Supplementary Homicide Reports for purposes of comparison in Chapter 5. Far-right
extremist homicides were identified fiom a number of open-sources, and information on
each homicide event was gathered fi'om over twenty web-based search engines. The
creation of the open-source database allowed this study to avoid the many pitfalls of
relying on official data to study “deviant” forms of homicide, including most crucially,
the inability to identify far-right extremist homicide in existing homicide and terrorism
databases.

Specifically, this dissertation conducted quantitative analyses in order to identify
key predictor variables that could distinguish between far-right extremist subtypes of
interest. Then, qualitative variable-centered theme analyses were conducted for each of
these important homicide characteristics in order to further knowledge of the interaction
and operation of these variables with other situational variables in the context of the
‘ entire event to Shape these fatal outcomes.

There were a number of important findings in each of the three analysis chapters
in this dissertation. Chapter 5 provided the context for the comparative study of far-right
extremist homicide. It was recognized that homicide events do not occur in a vacuum and
that it was impossible to capture the “true” nature of far-right homicide events separate
from their social context.

More specifically, it was discussed how many external social factors, such as
moral panics following high-profile events, hate crime, terrorism legislation, and
advocacy efforts by social movement organizations can shape the geospatial and

temporal natures of far-right homicide occurrences. This chapter showed that not all

180

states have experienced far-right extremist homicides since 1990. Moreover, some states
that have been usually associated with far-right extremist activity had a relatively low
number of these occurrences. A number of partial explanations were posited for states
that had experienced disproportionately large and small amounts of far-right homicide in
comparison to their population size, including the presence of state prisons and the
historical presence of White supremacist gangs in some cities (e. g. skinheads, KKK). In
addition, Chapter 5 found that while the rate of the average homicide has generally
declined since 1990, the rate of far-right extremist homicide has been more volatile,
experiencing a number of valleys and peaks. A number of partial explanations for these
fluctuations were discussed, including the influences of hate crime legislation, moral
panics following high-profile events, and the social movement activity of advocacy
groups like the Southern Poverty Law Center. Most recently the number of far-right non-
[M homicide has increased while [M homicide has decreased, suggesting that there is a
need to focus on the former. When quantitatively compared with the average homicide,
far-r1 ght extremist homicides were found to have both similarities and differences in
characteristics. Unexpectedly, far-right suspect age distributions were similar to those of
the average homicide suspect, which conflicts with past research on domestic terrorist
suspects (Smith, 1994). More expectedly, far-r1 ght extremist homicide events were
proportionately more likely to involve multiple offenders, intimate weaponry, unknown
victims, and profit-seeking behaviors. These characteristics of far-right homicide
revealed that far-right extremist homicides were a more complex phenomenon than has

been described anecdotally in past research. Indeed, this chapter suggested that

181

comparatively examining different subtypes of far-right extremist homicide could
illuminate the heterogeneity within this form of violence.

Chapter 6 examined two motivational subtypes of far-right extremist homicide,
[M and non-[M events. Past homicide research has found that disaggregating events by
motivational subtypes can be useful for understanding the offending and situational
patterns associated with distinct categories of homicides (Decker, 1996; F lewelling &
Williams, 1999). However, past research has ignored motivational variations in far-right
homicide. Therefore, this study was interested in the similarities and differences in far-
right homicide [M and non-[M motivational circumstances. Descriptive analysis revealed
that far-r1 ght extremists in the United States have committed, proportionately, slightly
more non-[M homicides than [M events. Thus, this chapter largely focused analytical
attention on this motivational subtype of far-right extremist homicide. Multivariate
quantitative analysis revealed that far-1i ght offenders committing non-[M homicides were
more likely to have prior criminal histories. These offenders often belonged to organized
White supremacist groups. In addition, this analysis showed that known victim-offender
relationships, profit-seeking behaviors, and indoor homicide locations were significantly
related to far-r1 ght non-[M homicide events. Variable-centered theme analyses were
conducted for each of the statistically significant situational variables. These analyses
identified the variety of situational themes in which these particular variables operated,
within the context of the entire homicide event, to shape these fatal outcomes. Themes
were identified for [M and non-[M homicide events in order to compare the similarities
and differences across motivational subtypes. Considering the theme analyses findings as

a whole, it became apparent that far-right non-[M homicide offenders pose threats on

182

multiple fionts. For instance, it was clear that far-right extremists, often White
supremacists, committed robbery-homicide for reasons that had nothing to do with the
far-right movement, as the attainment of money and property was the primary motive.
Although non-[M offenders often belonged to prison-based White supremacist groups,
these events often occurred in locations such as residences (and often in victims’ homes).
It was also clear that far-right extremists often traveled to victims’ homes for the
purposes of enforcing internal discipline through murder when it is thought that victim’s,
usually group affiliates or members, were disloyal to the group in some way. This chapter
revealed for the first time the heterogeneous and brutal nature of non-[M violence
committed by far-right extremists.

Chapter 7 addressed the final research question posited in this study regarding
suspect subtypes of identity-based (IB) homicide events. That is, this chapter shifted the
focus of comparative analysis to far-r1 ght [B events compared to [B homicides not
perpetrated by far-right extremists. One advantage of creating an original open-source
database on homicide events was that it allowed for the examination of multiple subtypes
based on motivational and suspect categories so that it was possible to respond to a
number of related research questions of interest to this dissertation.

Only one variable, homicide victim-offender relationship, was found to
statistically distinguish between homicide suspect subtypes. In particular, known victim-
offender relationships were most associated with those [B homicides not perpetrated by
far-right extremists. Further descriptive analysis revealed that anti-GLBT homicides were
likely driving this finding, as these types of homicides were, proportionately, more likely

to be committed by non far-rightists. Theme analysis then showed that the most common

183

lulu cpl... .7 PD: med aging

type of situation for this homicide subtype were sexual advance homicides, in which
suspects claimed that victims were murdered, often brutally, due to an unwanted sexual
advance. Chapter 7 used the components of the CEP to advance a more in-depth
understanding of anti-GLBT homicides, and put forward theories (e. g. doing gender) on
what influences offenders to commit these offenses and discussed, briefly, possible
explanations for non—far-rightists being more likely not to engage in these types of
situations.

Research Implications

This study has implications for the academic, law enforcement, and policymaking
communities. First, this study illustrated how there are substantial overlaps in types of
domestic terrorism, hate crime, and homicide despite research in these areas often being
alienated from one another. This study showed that violence committed by far-rightists
falls into each of these research areas and it is important to integrate the literatures of
each topic to examine the nature of far-right homicide. In particular, this study illustrated
how it is possible and useful to study domestic terrorism from a criminological
perspective.

Second, this project successfully integrated multiple sources of data to produce an
original and comprehensive database that can be used by scholars to further
comparatively examine far-right and other forms of homicide. Examples of how this
database can be used in future research are provided below.

Third, this study contributed to theoretical explanations of far-right homicide. By
emphasizing the situational contexts of far-right homicide, this study revealed how

suspect and incident-level homicide variables interacted to shape variations of fatal

184

ID: .ID

.I.~.ul...il.r|
rung

afll . :1

outcomes. In this way, this study also emphasized the multi-dimensionality of far-right
homicide events and moved beyond a reliance on suspect-based theoretical descriptions
of extremist crime found in past research. Furthermore, by focusing on combinations of
situational and suspect homicide variables, it was possible to uncover how situational
characteristics of violent transactions affected the decisions of homicide actors and
shaped the human agency or ability of actors to act in a self-determined way. For
instance, this study showed how in some situations far-rightists decided long beforehand
to murder their victims, while in other situations unplanned and tragic interaction with
victims led to the fatal violence. These theoretical concepts were directly linked to the
methods employed in this study, as variable-centered theme analysis was used to identify
variations in how statistically important situational variables interacted with suspect
variables to result in such violent transactions.

Fourth, the findings of this study can be integrated into police training materials
on extremist violence. This project examined that nature of both far-right IM and non-IM
homicide and the findings can aid law enforcement in the recognition and investigation of
these forms of violence. For instance, many of the anti-police far-right homicides were
non-1M events and began as routine traffic stops or calls to service. Identifying cues to
impending escalation of violence by far-rightists is one way this study can contribute to
the prevention of far-right extremist violence.

Fifth, a number of the findings from the current study can contribute to the policy
discussion revolving around extremist crimes of violence. In particular, findings
involving the similarities and differences of far-right and average homicides, as well as

comparisons of far-right and other [3 homicides, can be used to inform the debate on if

185

and how extremist violence is different in nature from traditional violence. In this way,
the policy debate can move beyond a discussion of strictly high-profile extremist
homicides and anecdotal evidence. Findings from this research may also be used to
inform policymakers on how best to distribution particular resources (e. g. funds,
trainings) across their constituencies.
Research Limitations

A discussion of the research limitations is also required. As stated previously,
there were a number of potential biases introduced by the use of open-source data in this
dissertation, including media biases and newsworthiness biases, which could potentially
affect the availability of data for particular homicide characteristics (see Chapter 4).
Repeating this discussion is not necessary here, but it is worth repeating that focusing on
homicide events that have usually caught the attention of media, policymakers, and the
public ensured that lack of information on homicide events was not a major obstacle to
overcome in this study. Nonetheless, one research limitation related to the availability of
information is that it remains nearly impossible to know if all far-right extremist
homicides that occurred in the United States since 1990 were captured in this study. This
study had to rely on open-sources to identify these events. Again, though, because of the
relative severity of these violent crimes, and the newaorthiness of homicides committed
by those with extremist beliefs, it is reasonable to expect that this study captured the
majority of all far-right homicides in which information was available, and this set of
homicides represents the complex phenomenon of far-right homicide in the United States.

A limitation of research that has not been discussed is the somewhat arbitrary

starting time period of 1990. Undoubtedly, far-right extremist homicides occurred prior

186

to 1990, and the nature of these homicides likely differed from those homicides examined
in this study in important ways. In addition to factors involving available resources for
conducting this research, available open-source information may be more difficult to
attain via the Internet on far-right homicides prior to 1990. Therefore, additional
resources would be required to attain information on homicide events occurring before
this time.

Another limitation is the focus on far-right extremist homicide, in general, and the
only brief exploration of a number of important subtopics, such as far-right extremist
police killings and prison gang killings. Moreover, other types of domestic extremist
homicides that have occurred in the United States, including possible lefl-wing group
homicides, were not examined in the current study. Therefore, little is known about how
far-right extremist homicides compare to these events. In addition, the crime of homicide
as the sole focus of this study is another limitation to this research. Studying violent
crimes such as attempted homicides, conspiracies to commit homicides, and aggravated
assaults committed by far-right extremists in combination with far-right extremist
homicides could substantially further understanding of the nature of far-right extremist
homicide.

Finally, this study is also limited in its nearly exclusive focus on the suspect and
incident dimensions of far-right homicide events. As discussed in Chapter 3, the CEP
emphasizes the benefits of considering all homicide dimensions simultaneously. For

practical reasons, however, this was not possible in the current study.

187

Future Research

This dissertation and the original open-source database on far-right extremist
homicide and other IB homicide events will serve as the basis for my immediate research
agenda. This final section outlines the plans for a number of empirical studies that
advance knowledge and an in-depth understanding of far-right extremist homicide, and
how these events compare to other types of IM and non-1M forms of violence. As
discussed throughout this dissertation, empirical research on this topic is extremely
limited, thus this research should make substantial contributions to the literature.

The future studies discussed here will accomplish a number of general objectives,
including: 1) building off of the quantitative analyses presented in Chapters 5, 6, an 7, to
explore the similarities and differences in far-right homicide and the “average” homicide
in the United States, as well as across far-right homicide event subtypes. These studies
will rely on a mixed-methods approach to examine in-depth how particular theoretical
variables of interest (e. g. victim-offender relationship, location, etc.) operate to shape
fatal outcomes 2) exploring homicide subtypes that are especially relevant to law
enforcement, judicial officials, and policymakers who are responsible for the
investigation, prosecution, and legislation against such violence 3) studying how these
homicides included in the current study compare across to other forms of violence,
including far-left and jihadi events, non-fatal IM and IB violence, and finally 4)
supplementing the current data with national news print media data and Census data, thus
branching into areas of media coverage of far-right violence and the influence of macro-
level predictors of far-right 1M and non-1M homicide across the states. Below these

studies are further outlined.

188

The first three future articles on far-right extremist homicide will involve the
reworking of Chapters 5, 6, and 7 of this dissertation. These will be quantitative and
mixed-method comparative analyses. The first study will compare far-right homicide in
the United States to the “average” homicide during the same time period, 1990 to 2006.
Producing this particular paper at the outset will be advantageous, as it will establish that
far-right homicide is a unique phenomenon from traditional homicide and it will provide
the temporal and geospatial context required for understanding these events. The second
paper will comparatively examine two motivational subtypes of far-right homicide, IM
and non-1M homicide events, while the third paper will compare suspect subtypes, far-
right and non far-right suspect homicides. In addition, the second and third paper will use
qualitative variable-centered theme analyses to explore how significant predictors from
multivariate models operate to shape fatal outcomes. This initial series of papers will
describe the extent and heterogeneity of far-right homicide, while also introducing the
far-right homicide project and the open-source database created to study these events.

Future research also needs to examine in more depth the subtypes of far-right
homicide associated with relevant policy concerns in the United States, such as the
prosecution of far-right IB homicide offenders. Indeed, one potentially useful area of
future research relevant to debates involving hate crime legislation is the comparative
examination of those homicide involving suspects who were charged as hate-motivated
offenders and those events that were not. Therefore, the fourth study stemming from the
far-right homicide project will examine the legal and non-legal factors associated with
how far-right [B are charged and prosecuted. Using a comparative case-study approach,

this study will select and compare a number of different types of homicides, including

189

far-right IB homicides charged as hate crimes, non far-right IB homicides charged as hate
crimes, far-right IB homicides not charged as hate crimes, and non far-right [B homicides
not charged as hate crimes. In this way, it will be possible to examine the factors most
associated with how events are charged and prosecuted across suspect subtypes.

Future research will also expand on the data currently available in the far-right
extremist open-source database. Homicides committed by other types of domestic
extremists will be identified and compared with far-right extremist homicides, including
left-wing, single-issue, and jihadi homicides occurring in the United States. The fifth
paper will focus on these comparisons and address how far-right homicides are similar
and different to these occurrences. The sixth paper will then examine how far-right
homicides compare to types of non-fatal violence committed by far-1i ght extremists. The
ECDB contains information on far-right assaults and attempted homicides. This paper
will extract these data on non-fatal events and comparatively examine how far-right and
other [3 non-fatal attacks compare to homicides.

In addition, by integrating data from sources such as the Census Bureau and
particular national news print media sources (e. g. The New York Times) future research
will extend outside the boundaries of events-based violence research and into the areas of
macro-level violence research and media representations of violence in the United States.
This dissertation found that not all states have experienced far-right extremist homicide.
Therefore, the seventh paper stemming from this study will examine how socio-
demographic, socioeconomic, and far-right movement factors affect the rate of far—right
extremist occurrences by region or states. Finally, in one, possibly less obvious area of

future research, the eighth paper will examine the media coverage of far-right extremist

190

homicides. Media is the primary way that the public learns about these homicide events,
and how media choose to cover these events shapes the public understanding of this
phenomenon (Chermak, 2002; Surrette, 1998). It was stated previously in this dissertation
that homicide events tend to be high-profile events that capture the attention of the media.
However, a number of studies have found that not all homicide events, as well as
domestic terrorism events (Chermak & Gruenewald, 2006), have received equal amounts
or types of media coverage (Buckler & Travis, 2005; Gruenewald, Pizarro, & Cherrnak,
2009; Johnstone, Hawkins, & Michener, 1995; Lundman, 2003; Peelo, Francis, Soothill,
Peerson, & Ackerly, 2004; Pritchard, 1985; Pritchard & Hughes, 1997; Sorenson, Manz,
Berk, 1998; Weiss & Chermak, 1998; Wilbanks, 1984). By integrating national news
print media with 1M homicide event-based data, this paper will address how media
coverage compares and how certain events are framed for different types of suspect types

and victim groups.

191

APPENDIX A

Table A.l. Indicators of Far-Right Extremist Affiliation

 

Indicator

Description and/or Examples

 

Self-admission

Established membership to national or
localized far-right extremist group

Found verbal or written evidence of far-
right extremist ideology during police
investigation.

Bodily markings and attire

Found possession of literature and
music

Witness testimony

Non-violent specific involvement in
far-right movement activities

 

Homicide suspect(s) admit to friends, police,
or others their involvement with far-right
extremist movement.

Examples include Ku Klux Klan, Aryan
Nation, Michigan Militia, etc.

Examples include bigoted innuendo or slang,
written manifestos, letters to family, friends,
newspaper editors, or others.

Examples include symbolic tattoos (e. g. “white
power) and items of dress (e.g. Doc Martin
boots).

Examples include extremist books (e. g. Mein
Kampf, Turner Diaries), manuals (e.g.
survivalist manuals, bomb-making manuals),
and particular types of music (white power
rock/punk music).

Examples include court testimony and witness
accounts gathered by media or advocacy

groups.

Evidence that suspect was host or publisher of
media outlet distributing far-right extremist
materials, a far-right event organizer, attended
far-right rallies, or participated in leafleting of
propaganda materials.

 

192

W

Table B.1. Indicators of Ideological Motivation

 

Indicator

Description and/or Examples

 

Lack of known or ulterior motive

Written and verbal statements by the homicide
perpetrator(s) to indicate far-right extremist
beliefs

Written drawings, markings, symbols, or other
graffiti found during investigation.

Location of homicide

Witness perceptions

 

Available evidence shows that bigotry or hatred
toward social minority group was only motive.

Examples include bigoted innuendo or slang,
written manifestos, letters to family, friends,
newspaper editors, or others.

Examples would include racial or anti-Semitic
epithets found at or near homicide location.

Examples include symbolic sites, such as gay
bars, churches, abortion clinics, homeless
establishments, advocacy group offices. Also,
other sites of IM crime.

Examples include testimony by witnesses,
family members, or others with knowledge of
offender(s) motivation.

 

193

 

REFERENCES

Akers, R.L., & Jensen, OF (2006). The empirical status of social learning theory of
crime and deviance: The past, present, and future. In F.T. Cullen, J .P.,
Wright, & KR. Blevins (Eds.), Taking stock: The status of criminological theory
(V 01. 15, pp. 37-76). New Brunswick: Transaction Publishers.

Aho, J .A. (1990). The politics of righteousness: Idaho Christian Patriotism. Seattle, WA:
University of Washington Press.

Athens, L.H. (1980). Violent criminal acts and actors: A symbolic interactionist study.
Boston, MA: Routledge and Kegan Hall. 3.

Babbie, E. (1990). Survey research. Belmont, CA: Wadsworth.

Barkun, M. (1997). Religion and the racist right: The origins of the Christian Identity
movement (2"d ed.). Chapel Hill, NC: University of North Carolina Press.

 

Berk, R.A., Boyd, E.A., & Hamner, K.M. (1994). Thinking more clearly about hate- I,
motivated crimes. In G.M. Herek & K.T. Berrill (Eds.), Hate crimes: Confronting '
violence against lesbians and gay men (pp. 123-138). Newbury Park: Sage
Publications.

Berlet, C., & Lyons, MN. (2000). Right-wing populism in America: T 00 close for
comfort. New York: The Guilford Press.

Berrill, K.T. (1992). Anti-gay violence and victimization in the United States: An
overview. In G.M. Herek, & K.T. Berrill (Eds.), Hate crimes: Confronting
violence against lesbians and gay men (1945). Newbury Park, CA: Sage
Publications.

Best, J. (1995). Typification and social problem construction. In J. Best (Ed.), Images of
issues: T ypijying contemporary social problems (2“ ed.) (1-16). Hawthorne, NY:
Aldine de Gruyter.

Birkbeck, C., & LaFree, G. (1993). The situational analysis of crime and deviance.
Annual Review of Sociology, 19, 113-137.

Blazak, R. (2001). White boys to terrorist men: Target recruitment of Nazi skinheads.
American Behavioral Scientist, 44(6), 982-1000.

Block, R. (1981). Victim-offender dynamics in violent crime. The Journal of Criminal
Law and Criminology, 72(2), 743-761.

194

Blumer, H. (1969). Symbolic interactionism: Perspective and method. Englewood Cliffs,
N.J.: Prentice Hall.

Blumstein, A., Rivara, F.P., & Rosenfeld, R. (2000). The rise and decline of homicide-
and why. Annual Review of Public Health, 21, 505-541.

Boyd, E.A., Berk, R.A., & Hamner, K.M. (1996). "Motivated by hatred or prejudice":
Categorization of hate-motivated crimes in two police divisions. Law and Society
Review, 30(4), 819-850.

Bradley, MS. (2007). Symbolizing hate: The extent and influence of organized hate
group indicators. Journal of Crime and Justice, 30, 1-15.

Brantingham, P.L., & Brantingham, R]. (1993). Environment, routine, and situation:
Toward a pattern theory of crime. In R.V. Clarke, & M. Felson (Eds.), Routine

activity and rational choice. Advances in criminological theory (Vol 5, pp. 259-
294).

Bruce, T. (2001). The new thought police. New York: Three Rivers Press.

Byers, B., Crider, B.W., & Biggers, G.K. (1999). Bias crime motivation: A study of hate
crime and offender neutralization techniques used against the Amish. Journal of
Contemporary Criminal Justice, 15, 78-96.

Buckler, K., & Travis, L. (2005). Assessing the newsworthiness of homicide events: An
analysis of coverage in the Houston Chronicle. Journal of Criminal Justice and
Popular Culture, 12(1), 1-25.

Carlson, J .R. (1995). The future terrorists in America. American Journal of Police,
14(3/4), 71- 91.

Chermak, SM. (2002). Searching for a demon: The media construction of the militia
movement. Boston, MA: Northeastern University Press.

Chermak, S.M., Freilich, J .D., & Shemtob, Z. (2009). Law enforcement training and the
domestic far-right. Forthcoming in Criminal Justice and Behavior.

Chermak, S.M., & Gruenewald, LA. (2006). The media’s coverage of domestic terrorism.
Justice Quarterly, 23(4), 428-461.

Cohen, L.E., & Felson, M. (1979). Social change and crime rate trends. American
Sociological Review, 44, 588-607.

Colomb, W., & Damphousse, K. (2004). Examination of newspaper coverage of hate
crimes: A moral panic perspective. American Journal of Criminal Justice, 28(2),
147-163.

195

Comstock, GD. (1991). Violence against lesbians and gay men. New York: Columbia
University Press.

Cook, W., Jr., & Kelly, R.J. (1999). The dispossessed: Domestic terror and political
extremism in the American heartland. International Journal of Comparative
and Applied Criminal Justice, 23(2), 241-256.

Clarke, R.V, & M. Felson (Eds.), (1993). Routine activity and rational choice. Advances
in criminological theory (V 01 5). New Brunswick, NJ: Transaction Publishers.

Cotter, J. M. (1999). Sounds of hate: White power rock and roll and the neo-Nazi
skinhead subculture. Terrorism and Political Violence, 11(2), 11 1-140.

Coverson, L. (2006). Aryan Brotherhood tried for 40 years of prison Mayhem. ABC
News. Retrieved on May 15, 2009, from
http://abcnewsggo.com/US/story?id= l 7262 15&page= l .

Cresswell, J .W., & Clark, V.L. Plano. (2007). Designing and conducting mixed methods
research. Thousand Oaks, CA: Sage Publications.

Decker, S. H. (1993). Exploring victim—offender relationships in homicide: The role of
individual and event characteristics. Justice Quarterly, 10(4), 585—612.

Decker, S. H. (1996). Deviant homicide: A new look at the role of motives and victim-
offender relationships. Journal of Research on Crime and Delinquency, 33(4),
427—449.

Decker, S.H., & Curry, GD. (2002). Gangs, gang homicides, and gang loyalty:
Organized crimes or disorganized criminals. Journal of Criminal Justice, 30, 343-
352.

Dillman, DA. (1978). Mail and telephone surveys: The total design method. New York:
Wiley-Intersciences.

Drass, K.A., & Ragin, CC. (1989). QCA: Qualitative comparative analysis. Evanston,
IL: Center for Urban Affairs and Policy Research, Northwestern University.

Falkenrath, R. (2001). Analytical models and policy prescription: Understanding recent
innovation in US. counterterrorism. Studies in Conflict and Terrorism, 24, 159-
181.

Federal Bureau of Investigation. (2004). Crime in the United States, 2004. United States
Department of Justice: Federal Bureau of Investigation.

196

Federal Law Enforcement Training Center. (2005). Hate and bias crimes train-the-
trainer program. Glynco, GA: Author.

Finz, S. (1999, July 20). Brothers charged in slaying of gay couple; Suspects also in fires
at Sacramento synagogues. The San Francisco Chronicle, p. A15.

Flewelling, R.L., & Williams, KR. (1999). Categorizing homicides. In M.D. Smith, &
M.A. Zahn (Eds.), Homicide, a sourcebook of social research (pp. 96-106).
Thousand Oaks, CA: Sage Publications.

Franklin, K. (2000). Anti-gay behaviors among young adults: Prevalence, patterns, and
motivators in a noncriminal population. Journal of Interpersonal Violence, 15,
339-362.

Freilich, J .D., & Chermak, S.M. (2006). Developing afar-right crime database.
Unpublished grant proposal to National Consortium for the Study of Terrorism
and Responses to Terrorism.

Freilich, J .D. & S.M. Chermak. (2007). Final DHS Summer Faculty and Student
Research Team Grant Report: Creation of a database of (IS. extremist crime,
1995- 2005. Department of Homeland Security, Science and Technology
Directorate.

F reilich, J .D., & Chermak, S.M. (2009). Preventing deadly encounters between law
enforcement and American far-rightists. In J .D. Freilich, & G.R. Newman (Guest
Eds.), Reducing terrorism through situational crime prevention. Forthcoming in
Crime Prevention Studies.

Freilich, J.D., & Chermak, S.M. (2009). United States Crime Database (ECDB), 1990-
2008: Preliminary results. Paper presented at the annual Department of
Homeland Security University Network Research and Education Summit.
Washington DC.

Freilich, J.D., S.M. Chermak, and J. Simone. (2009). Surveying American state police
agencies about terrorism threats, terrorism sources, and terrorism definitions.
Forthcoming in Terrorism and Political Violence.

Freilich, J.D., & Pridemore, W.A. (2005). A reassessment of state-level covariates of
militia groups. Behavioral Sciences & the Law, 23(4), 527-546.

Freilich, J .D., & Pridemore, W.A. (2006). Mismeasuring militias: Limitations of

advocacy group data and of state-level studies of paramilitary groups. Justice
Quarterly, 23(1), 147-162.

197

Freilich, J .D., & Pridemore, W.A. (2007). Politics, culture, and political crime:
Covariates of abortion clinic attacks in the United States. Journal of Criminal
Justice, 31(3), 323-336.

Frosch, D. (2009, April 16). Murder trial tests Colorado hate-crime statute. The New York
Times. Retrieved May 6, 2009 fi'om
hip://www.nytimes.com/2009/04/17/us/l 7tr2_rnsgen.html.

Garofalo, J ., & Martin, SE. (1993). Bias-motivated crimes: The law enforcement
response. Carbondale, 111: Center for the Study of Crime, Delinquency, and
Corrections.

Goffman, E. (1967). Interaction ritual: Essays on face-to-face behavior. New York:
Anchor Books.

Green, D. P. & Strolovitch, D. Z. & Wong, J .S. (1998). Defended neighborhoods,
integration, and racially motivated crime. American Journal of Sociology, 104(2),
372-403.

 

Grattet, R., & Jenness, V. (2001). The birth and maturation of hate crime policy in the i.
United States. The American Behavioral Scientist, 45(4), 668-696.

Gottfredson, M., & Hirschi, T. (1990). A general theory of crime. Stanford, CA: Stanford
University Press.

Gruenewald, J .A. (2004). Timothy McVeigh and the Oklahoma City Bombing. In F.
Bailey & S.M. Chermak. (Eds.), Famous American crimes and trials (Vol.5, pp.
249-272). Westport, CT: Praeger.

Gruenewald, J .A., Pizarro, J ., & Chermak, S.M. (2009). Evaluating the newsworthiness
of homicide incidents. Forthcoming in Journal of Criminal Justice.

Gruenewald, J .A, Freilich, J ., Cherrnak, S. (2009). A review of the far-right extremism
literature. In R. Blazak & B. Perry (Eds.), Hate crime issues and perspectives (1-
21). Westport, CT: Praeger.

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

Hamm, MS. (1993). American skinheads: The criminology and control of hate crime.
Westport, CT: Praeger.

Hamm, MS. (1994). A modified-social control theory of terrorism: An empirical and
ethnographic assessment of the American neo-nazi skinhead. In M.S. Hamm
(Ed.). Hate crime: International perspective on causes and control (105-150).

Cincinnati, OH: Anderson Publishing Co.

198

Hamm, MS. (1998). Terrorism, hate crime, and antigovemment violence: A review of
the research. In H.W. Kushner (Ed.), The future of terrorism: Violence in the new
millennium (59-96). Thousand Oaks, CA: Sage Publications.

Hamm, MS. (2007). Terrorism as crime: From Oklahoma City to Al Qaeda and beyond.
New York: New York University Press.

Heitrneyer, W. (2003). Right-wing extremist violence. In W. Heitrneyer, & J. Hagan
(Eds.), International handbook of violence research (pp. 399-436). Boston, MA:
Kluwer Academic Publishers.

Herek, G.M. (2007). Hate crimes and stigma-related experiences among sexual minority
adults in the United States: Prevalence estimates from a national probability
sample. Journal of Interpersonal Violence, 24, 54-74.

Hewitt, C. (2003). Understanding terrorism in America. New York: Routledge.

Hindelang, M.J., Gottfredson, M.R., & Garofalo, J. (1978). Victims of personal crime: An
empirical foundation for a theory of personal victimization. Cambridge, MA:
Ballinger.

Holmes, W. (1992). Hate crime reporting: Obstacles, facilitators, and strategies. Boston,
MA: Statistical Analysis Center, Massachusetts Committee on Criminal Justice.

Holthouse, D. (2009, Spring). Intelligence Report: Number of hate groups tops 900.
Southern Poverty Law Center. Retrieved May 1, 2009, from
http://www.spleenter.org/intel/intelreport/article.isp? aid=1027.

Homicide Studies. (2004). Special issue on missing data in homicide research. Homicide
Studies, 8, 163-321.

Hudson, A. (2009, May 14). Report citing veteran extremism is pulled. Washington
Times. Retrieved on May 16, fi'om
http://www.washingtontimes.com/news/2009/may/ 14/report-citing-vet-
extremism-is-pulled/.

Jackson, S.L., Zahn, M.A., & Brownstein, H.H. (2004). The need for a theory of
violence. In M.A. Zahn, H.H. Brownsteim, & S.L. Jackson (Eds.), Violence:
From theory to research (pp.227-236). LexisNexis Group.

Jacobs, J .B., & Henry, J .S. (1996). The social construction of a hate crime epidemic. The
Journal of Criminal Law and Criminology, 86(2), 366-391.

J enness, V. (2003). Engendering hate crime policy: Gender, the “dilemma of difference,”
and the creation of legal subjects. Journal of Hate Studies, 2, 73-97.

199

Jenness, V., & Broad, K. (1997). Hate crimes: New social movements and the politics of
violence. New York: Aldine De Gruyter.

J enness, V. 8: Grattet, R. (2005). The law-in-between: The effects of organizational
perviousness on the policing of hate crime. Social Problems, 52(3), 337-359.

Jensen, G.F., & Brownfield, D. (1986). Gender, lifestyles, and victimization. Beyond
routine activity. Violence and Victims, 1 (2), 85-99.

Johnstone, J .W.C., Hawkins, D.F., & Michener, A. (1995). Homicide reporting in
Chicago dailies. Journalism Quarterly, 71, 860-872.

Kaplan, J. (1995). Right wing violence in North America. Terrorism and Political
Violence, 7(1), 44-95.

Katz, J. (1988). Seductions of crime: Moral and sensual attractions in doing evil. New
York: Basic Books.

LaFree, G., & Birkbeck, C. (1991). The neglected situation: A cross-national study of the
situational characteristics of crime. Criminology, 29, 73-98.

LaFree, G., & Dugan, L. (2004). How does studying terrorism compare to studying
crime? Sociology of Crime, Law and Deviance, 5, 53-74.

LaFree, G., & Dugan, L. (in press). Terrorism and counter-terrorism. In M. Tonry (Ed),
Crime and Justice: A Review of Research (V 01. 38). Chicago, IL: University of
Chicago Press.

LaFree, G., Dugan, L., Fogg, H., & Scott, J. (2006). Building a global terrorism
database. Final report to the National Institute of Justice.

Langer, E. (2003). A hundred little Hitlers: The death of a black man, the trial of a white
racist, and the rise of the neo-Nazi movement in America. New York:
Metropolitan Books.

Lawrence, RM. (1999). Punishing hate: Bias crimes under American Law. Cambridge:
MA: Harvard University Press.

Lee, M.R., & Bartkowski, J .P. (2004). Civic participation, regional subculture, and
violence: The differential effects of secular and religious participation on adult
and juvenile homicide. Homicide Studies, 8, 5-39.

Levin, B. (1999). Hate crimes: Worse by definition. Journal of Contemporary Criminal
Justice, 15(1), 6-21.

200

 

Levin, J. 8L McDevitt, J. (1993). Hate crimes: The rising tide of bigotry and bloodshed.
New York: Plenum.

Liao, T.F. (1994). Interpreting probability models: Logit, probit, and other generalized
models. Thousand Oaks, CA: Sage Publications.

Lofland, J. (1969). Deviance and identity. Englewood Cliffs, NJ: Prentice Hall.
Loftin, C., Kindley, K., Norris, S., & Wiersema, B. (1987). An attribute approach to
relationships between offenders and victims in homicide. Journal of Criminal

Law and Criminology, 78, 259-271.

Long, J .S. (1997). Regression models for categorical and limited dependent variables.
Thousand Oaks, CA: Sage Publications.

Loseke, DR. (2003). Thinking about social problems. New Brunswick, NJ: Transaction
Publishers.

Luckenbill, DP. (1977). Criminal homicide as a situated transaction. Social Problems,
25, 176-86.

Luckenbill, D.F., & Doyle, DP. (1989). Structural position and violence: Developing a
cultural explanation. Criminology, 27(3), 419-36.

Lundman, R.J. (2003). The newsworthiness and selection bias in news about murder:
Comparative and relative effects of novelty and race and gender typifications on
newspaper coverage of homicide. Sociological Forum, 18, 357-3 86.

Maguire, K. (1995). Crime and the Ku Klux Klan: A study of political criminality in the
USA. Criminologist, 19(3), 131- 137.

Mariani, M. (1998). The Michigan militia: Political engagement or political alienation?
Terrorism and Political Violence, 10(4), 122-148.

Martin, SE. (1996). Investigating hate crimes: Case characteristics and law enforcement
responses. Justice Quarterly, 13(3), 455-480.

Mason, G. (2005). Hate crime and the image of the stranger. British Journal of
Criminology, 45, 837-859.

Maxfield, MG. (1989). Circumstances in Supplementary Homicide Reports: Variety and
validity. Criminology, 27(4), 671-696.

Maxson, C.L., Gordon, M.A., & Klein, M.W. (1985). Differences between gang and
nongang homicides. Criminology, 23, 209-222.

201

Maxwell, C., St Maxwell, SR. (1995). Youth participation in hate-motivated crimes:
Research and policy implications. Boulder, CO: Center for the Study and
Prevention of Violence.

McVeigh, R., Welch, M.R., & Bjamason, T. (2003). Hate crime reporting as a successful
social movement outcome. American Sociological Association, 68(6), 843-867.

Meier, R.F., Kennedy, L.W., & Sacco, V.F. (Eds). (2001). The process and structure of

crime: Criminal events and crime analysis. Advances in criminological theory
(V 01. 9). New Brunswick, NJ: Transaction.

Miethe, T.D., & Drass, K.A. (1999). Exploring the social context of instrumental and
expressive homicide: An application of qualitative comparative analysis.

Journal of Quantitative Criminology, 15(1), 1-21.

Meithe, T.D., & Regoeczi, W. (2004). Rethinking homicide: Exploring the structure and
process underlying deadly situations. New York: Cambridge University Press.

Miethe, T.D., & Meier, R.F. (1990). Criminal opportunity and victimization rates: A
structural-choice theory of criminal victimization. Journal of Research in Crime
and Delinquency, 27, 243-266.

Meithe, T.D. & Meier, R.F. (1994). Crime and its social context: Toward and integrated
theory of offenders, victims, and situations. Albany, NY: State University of New
York Press.

Merriam, SB. (1998). Qualitative research and case study applications in education
(Rev. ed.). San Francisco, CA: Jossey-Bass.

Messner, S., McHugh, S., & Felson, R. (2004). Distinctive characteristics of assaults
motivated by bias. Criminology, 42(3), 585-618.

Michel, L., & Herbeck, D. (2001). American terrorist: Timothy McVeigh and the
Oklahoma City bombing. New York: HarperCollins.

Mickolus, E.F., Sandler, T., Murdock, J .M., & Fleming, P. (1993). International
Terrorism: Attributes of Terrorist Events [988-91 (IT ERA T E 4). Dunn Loring,
VA: Vineyard Software.

Morash, M. (2006). Understanding gender, crime, and justice. Thousand Oaks, CA:
Sage Publications.

McDevitt, J ., Balboni, J .M, Bennett, S., Weiss, J .C., Orchowsky, S., & Walbolt, L.
(2000). Improving the quality and accuracy of bias crime statistics nationally: An
assessment of the first ten years of bias crime data collection (Executive

202

Summary). Boston, MA: Center for Criminal Justice Policy Research,
Northeastern University.

McDevitt, J ., Levin, J ., & Bennett, S. (2002). Hate crime offenders: An expanded
typology. Journal of Social Issues, 58(2), 303-317.

National Consortium for the Study of Terrorism and Responses to Terrorism. (2009).
Press release: Background information: Far-right attacks on US. law
enforcement. National Consortium for the Study of Terrorism and Responses to
Terrorism (START). Retrieved on May 15, 2009 from
http://www.start.umd.edu/start/media/Far-
_R_ight_Attacks_on_US_Law_Enforcement_PressReleasepdf.

National Counterterrorism Center. (2009). 2008 report on terrorism. Retrieved on May 1,
2009, from http://www.nctc.gov/.

Neiwert, D. A. (1999). In God '3 country: The Patriotic Movement and the Pacific
Northwest. Pullman, WA: Washington State University Press.

Nice, David C. (1988). Abortion clinic bombings as political violence. American Journal
of Political Science, 32(11), 178- 195.

Noble, DR (2004). Assessing the reliability of open source information. Vienna, VA:
Evidence Based Research, Inc.

Nolan, J .J ., III, Akiyama, Y., & Berhanu, S. (2002). The Hate Crime Statistics Act of
1990. The American Behavioral Scientist, 46(1), 136-153.

O’Brien, S. P., & D. P. Haider-Markel. (1998). Fueling the fire: Social and political
correlates of citizen militia activity. Social Science Quarterly, 79(2), 456-465.

Parker, KR (2001). A move toward specificity: Examining urban disadvantage and race-
and relationship-specific homicide rates. Journal of Quantitative Criminology,
1 7, 89-110.

Parker, R. (2009). Police training. In B. Perry, & F .M. Lawrence (Eds.), Hate Crimes:
Responding to hate crime (pp. 51-69). Westport, CT: Praeger.

Peelo, M., Francis, B., Soothill, K., Pearson, J ., & Ackerly, E. (2004). Newspaper
reporting and the public construction of homicide. British Journal of Sociology,
33, 256-275.

Pelz, M.E., Marqaurt, J .W. & Pelz, CT. (1991). Right-wing extremism in the Texas

Prisons: The rise and fall of the Aryan Brotherhood of Texas. Prison Journal,
71(2), 23-37.

203

 

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

Perry, B. (2003). Where do we go from here? Researching hate crime. Internet Journal of
Criminology.

Perry, B. & Iganski, P. (Eds.). (2009). Hate crimes: The consequences of hate crimes
(V 01. 2). Westport, CT: Praeger.

Phillips, ND. (2009). The prosecution of hate crimes: The limitations of the hate crime
typology. Journal of Interpersonal Violence, 25(5), 883-905.

Pitcavage, M. (1998). Flashpoint America: Surviving a traffic stop confrontation with an
anti-government extremist. Anti-Defamation League-T he Militia Watchdog.
Retrieved, May 1, 2009, http://www.adl.org/MWD/trafstop.a_sp.

Pizarro, J .M. (2008). Reassessing the situational covariates of homicides: Is there a need
to disaggregate? Homicide Studies, 12(4), 323-349.

Pizarro, J .M., & McGloin, J .M. (2006). Explaining gang homicides in Newark, New

Jersey: Collective behavior or social disorganization? Journal of Criminal Justice,
34, 195-207.

Polk, K. (1994). When men kill: Scenarios of masculine violence. New York: Cambridge
University Press.

Pritchard, D. (1985). Race, homicide, and newspapers. Journalism Quarterly, 62, 500-
507.

Pritchard, D., & Hughes, K.D. (1997). Patterns of deviance in crime news. Journal of
Communication, 47, 49-67.

Pridemore, W.A., & Freilich, J .D. (2005). A test of recent subcultural explanations of
White violence in the United States. Journal of Criminal Justice, 34(1), 1-16.

Riedel, M., & Welsh, W.N. (2001). Criminal violence: Patterns, causes, and prevention.
Beverly Hills, CA: Roxbury Park.

Riley, K.J. & Hoffman, B. (1995). Domestic Terrorism: A National Assessment of State
and Local Preparedness. Santa Monica: Rand (supported by the National Institute
of Justice and the US. Department of Justice).

Riley, K. J ., Treverton, G. R, Wilson, J. M., & Davis, L. M. (2005). State and local
intelligence in the war on terrorism. Santa Monica, CA: Rand.

Sacco, V.F., & Kennedy, L. (2002). The criminal event: Perspectives in space and time
(2“d ed.). Belmont, CA: Wadsworth/Thompson Learning.

204

 

Sageman, M. (2004). Understanding terror networks. Philadelphia, PA: University of
Pennsylvania Press.

Salfati, CG. (2000). The nature of expressiveness and instrumentality in homicide:
Implications for offender profiling. Homicide Studies, 4(3), 265-293.

Schmid, AP. (2004). Frameworks for conceptualizing terrorism. Terrorism and Political
Violence, 16(2), 197-221.

Schmid, A.P., & Jongrnan, A.J. (1988). Political terrorism: A new guide to actors,
authors, concepts, databases, theories and literature. Amsterdam: North-Holland
Publishing Company.

Senate Republican Policy Committee. (2000). Crimes and hate crimes: What the
numbers tell us? Retreived May 1, 2009, fi'om
http://rpcsenategov/releases/ l999/cr061500.htm.

Sherman, L.W., Gartin, P.R., & Buerger, ME. (1989). Hot spots of predatory crime:
Routine activities and the criminology of place. Criminology, 27, 27-56.

Short, J .F., Jr. (1997). Poverty, ethnicity, and violent crime. Boulder, CO: Westview.

Silke, A. (1996). Terrorism and the blind men’s elephant. Terrorism and Political
Violence, 8(3), 12-28.

Silke, A. (2001). The devil you know: Continuing problems with research on terrorism.
Terrorism and Political Violence, 13(4), 1-14.

Smith, B. (1994). Terrorism in America: Pipe bombs and pipe dreams. Albany, NY:
State University of New York Press.

Smith, M.D., & Parker, RN. (1980). Type of homicide and variation in regional rates.
Social forces, 59(1), 136-147.

Sorenson, S.B., Manz, J .G.P., 8L Berk, R.A. (1998). News media coverage and the
epidemiology of homicide. American Journal of Public Health, 88, 1510-1514.

Southern Poverty Law Center. (1999, Winter). Intelligence Report: Hate group count
tops 500, internet sites soar. Retrieved May 1, 2009, from
httpzl/www.spleenter.org/intel/intelreport/artic1e.jsp? aid=356.

Spector, M., & Kitsuse, J .I. (1977). Constructing social problems. Menlo Park, CA:
Cummings.

205

Spohn, C., & Holleran, D. (2001). Prosecuting sexual assault: A comparison of charging
decisions in sexual assault cases involving strangers, acquaintances, and intimate
partners. Justice Quarterly, 8(3), 651-688.

Strauss, A., & Corbin, J. (1998). Qualitative research techniques and procedures for
developing grounded theory (2"d ed.). London: Sage Publications.

Strom, KJ. (2001). Hate crimes reported in NIBRS, 1997-1999. Bureau of Justice
Statistics Special Report. Washington DC: US. Department of Justice.

Suprynowicz, V. (2004). The ballad of Carl Drega. Saratoga Springs: NY: Mountain
Media.

Surrette, R. (1998). Media, crime, and criminal justice: Images, realities, and policies.
New York: Wadsworth.

Sutherland, E. H. (1947). Principles of criminology (4th ed.). Philadelphia: Lippincott.

Sykes, G.M., & Matza, D. (1957). Techniques of neutralization: A theory of delinquency.
American Sociological Review, 22, 664-670.

Tedeschi, J .T., & Felson, RB. (1993). Aggression and violence: Social interactionist
perspectives. Washington, DC: American Psychological Association.

Walchholz, S. (2009). Pathways through hate: Exploring the victimization of the
homeless. In B. Perry (Ed.), Hate crimes: The victims of hate crime (Vol. 3, pp.
173-197). Westport, CT: Praeger.

Washington Post. (1999, July 11). Shooter left journal; Police seek rampage motive in
entries. The Washington Post, p. A02.

Weinberg, L., Pedahzur, A., & Hirsch-Hoefler, S. (2004). The challenges of
conceptualizing terrorism. Terrorism and Political Violence, 16(4), 777-794.

Weiss, A., & Chermak, S.M. (1998). The news value of Afiican-American victims: An
examination of the media’s presentation of homicide. Journal of Crime and
Justice, 21, 71-88.

Wilbanks, W. (1984). Murder in Miami: An analysis of homicide patterns and trends in
Dade County (Miami) Florida, 1917-1988. Lanham, MD: University Press of
American.

West, C., 8: Zimmerman, DH. (1987). Doing gender. Gender and Society, I (2), 125-151.

Wilkinson, D.L., & F agan, J. (2001). A theory of violent events. In R.F. Meier, L.W.
Kennedy, & V.F. Sacco (Eds.), The process and structure of crime: Criminal

206

events and crime analysis. Advances in criminological theory (Vol. 9). New
Brunswick, NJ: Transaction.

Wolfe, L., & Copeland, L. (1994). Violence against women as bias-motivated hate crime:
Defining the issues in the USA. In M. Davies (Ed.), Women and violence:
Realities and responses worldwide (pp. 200-213). Atlantic Highlands, NJ: Zed
Books.

207