Posted: August 6th, 2022

Discussion Question

 Discussion: Does Hate Crime Legislation Reduce Hate Crimes and Contribute to Social Change?

Consider a crime like murder that might have different legal consequences based on whether it is considered to be a hate crime. When someone murders another person, the offender can be prosecuted and sentenced for this offense, given sufficient evidence to prove that the individual committed this act. Let’s say that the offender left placards around the dead body that stated: “All Muslims must leave the country or die.” In this case, the prosecutor might be able to prove that the murder was a hate crime. What would be the advantages and disadvantages of prosecuting murder as a hate crime?

For this Discussion, you will research and analyze hate crime legislation. In your post, you will discuss whether hate crime legislation affects the occurrence of certain types of criminal acts, victim impact, or social change.

 Review the Learning Resources.

  • Consider your empathy-bias regarding hate crime legislation.
  • Set your empathy-bias aside and consider the effects of having hate crime legislation and prosecution on the occurrence of hate crimes, victim impact, and social change.
  • Consider the differences in sentencing for hate crimes.
  • Gather scholarly literature, which may include Learning Resources to support the points you plan to make in your discussion post.

Crime & Delinquency
2017, Vol. 63(10) 1191 –1223

© The Author(s) 2015
Reprints and permissions:
DOI: 10.1177/0011128715620626


Extreme Hatred:
Revisiting the Hate
Crime and Terrorism
Relationship to
Determine Whether
They Are “Close
Cousins” or “Distant

Colleen E. Mills1, Joshua D. Freilich1,
and Steven M. Chermak2

Existing literature demonstrates disagreement over the relationship between
hate crime and terrorism with some calling them “close cousins,” whereas
others declare them “distant relatives.” We extend previous research by
capturing a middle ground between hate crime and terrorism: extremist hate
crime. We conduct negative binomial regressions to examine hate crime
by non-extremists, fatal hate crime by far-rightists, and terrorism in U.S.
counties (1992-2012). Results show that counties experiencing increases in
general hate crime, far-right hate crime, and non-right-wing terrorism see
associated increases in far-right hate crime, far-right terrorism, and far-right
hate crime, respectively. We conclude that hate crime and terrorism may be
more akin to close cousins than distant relatives.

1John Jay College of Criminal Justice; City University of New York, The Graduate Center
New York City, NY, USA
2Michigan State University, East Lansing, MI, USA

Corresponding Author:
Colleen E. Mills, John Jay College of Criminal Justice; City University of New York, The
Graduate Center, 524 West 59th St., 2103 North Hall, New York, NY 10019, USA.

620626CADXXX10.1177/0011128715620626Crime & DelinquencyMills et al.

1192 Crime & Delinquency 63(10)

terrorism, violence, minorities, hate crime


In June 2015, Dylann Storm Roof opened fire on Black congregants in the
Emmanuel AME church in Charleston, South Carolina. Roof previously
posted a manifesto, detailing his hatred for non-White races and confirming
the racial motivation behind the shooting (Robles, 2015). Three years prior,
Wade Michael Page stormed a Sikh Temple in Wisconsin, executing a mass
shooting that was widely recognized as bias-motivated against the Sikh con-
gregants. The authorities later revealed that Page was a White supremacist,
active in the neo-Nazi music scene, and often spoke of the racial holy war
(Elias, 2012). After both of these attacks, many designated the attack as
domestic terrorism, lone wolf terrorism, as well as a hate crime (Elias, 2012;
Gladstone & Zraick, 2015; Goode & Kovaleski, 2012; B. Levin, 2012;
Murphy, 2012; Robles, 2015; “Unprosecuted Hate Crimes,” 2012). Incidents
such as Page’s and Storm’s rampages blur the line between certain hate
crimes and terrorism. Such confusion extends beyond the media to the schol-
arly community. Existing literature recognizes the parallels between hate
crime and terrorism (Deloughery, King, & Asal, 2012; Green, McFalls, &
Smith, 2001; Hamm, 1993; Herek, Cogan, & Gillis, 2002; Krueger &
Malečková, 2002, 2003), but some scholars disagree over the nature of the
relationship. Krueger and Malečková (2002, 2003) deem them “close cous-
ins” with their similarities outweighing their differences, whereas Deloughery
et al. (2012) characterize them as “distant relatives,” finding their differences
set them apart.

Although debate exists over the hate crime–terrorism relationship, only a
limited body of research has empirically examined this relationship (Byers &
Jones, 2007; Deloughery et al. 2012; Disha, Cavendish, & King, 2011; R. D.
King & Sutton, 2013). Much of this work centers on the impact of the
September 11 attack on hate crime offending (Disha et al., 2011; R. D. King
& Sutton, 2013). To date, Deloughery et al.’s (2012) temporal analysis is the
only study that examines the effects of the full range of anti-U.S. terrorist
attacks on hate crimes. In addition, it is the only known study that tests for
escalation from hate crimes to right-wing terrorism.

The current study extends Deloughery et al.’s (2012) important work by
testing the spatial relationship between hate crime and terrorism on the
county level. We unpack the relationships among (a) non-fatal hate crimes
committed by non-extremists, (b) fatal far-right hate crime, and (c) terrorist

Mills et al. 1193

attacks from 1992 to 2012. The current study fills another gap by utilizing
bias-motivated homicides by the far-right catalogued in the Extremist Crime
Database (ECDB). The use of the ECDB addresses a limitation acknowl-
edged by Deloughery et al. in their study as they used Hate Crime Statistics
Act (HCSA) data, which fails to account for perpetrator ideological strength
or affiliation. The current study seeks to answer the following research

Research Question 1: Are hate crimes and terrorism more interrelated
than prior research has demonstrated?

Literature Review

The Similarities and Differences Between Terrorism and Hate

Some past research has highlighted the similarities between terrorism and
hate crime. For example, one of the earliest forms of terrorism in the United
States was racially and politically motivated violence of the postbellum Ku
Klux Klan, which ushered in early legislative attempts to address the com-
mon phenomenon of racially motivated terrorism. Examining the Klan’s use
of violence to block African Americans’ political involvement, Law (2009)
calls the Klan the “terrorist wing of the Democratic Party” (p. 132), high-
lighting political motivations of the Klan’s reign of terror. Arguing that early
efforts to combat political violence coincided with tackling racially moti-
vated violence, Shimamoto (2004) remarks that the Enforcement Act of 1870
and the Ku Klux Klan Act of 1871 were the first measures taken by the United
States to handle terroristic racially motivated violence, so as to preserve the
rights of targeted citizens much like hate crime legislation. Thus, the line
between hate crime and terrorism proves blurred historically as early
American terrorism was both politically and racially motivated.

Hamm (1993) notes the similarities between the language of terrorism
and hate crime definitions as stated by the U.S. government. Hamm (1993)
cites the Federal Bureau of Investigation’s (FBI) definition of a terrorist
incident as a “violent act or an act dangerous to human life in violation of
the criminal laws . . . to intimidate or coerce a government, the civilian
population, or any segment thereof, in the furtherance of political or social
objectives” (pp. 106-107). The most recent language of federal hate crime
legislation defines hate crimes as offenses motivated by “prejudice based
on race, gender and gender identity, religion, disability, sexual orientation,
or ethnicity” (U.S. Department of Justice, FBI, 2011b). Given the statutory

1194 Crime & Delinquency 63(10)

language, Hamm (1993) argues that skinhead violence can be classified as
hate crimes or terrorist acts, evidencing the similar nature of such acts.

Definitions of hate crime and terrorist acts reveal a number of shared
traits. Both involve acts of violence against persons and property. Hate crime
and terrorism definitions both focus on classifying civilian populations, or
subgroups thereof, as victims (Shimamoto, 2004). Many definitions of ter-
rorism rely on the political, social, and/or religious nature of the goals of
terrorist perpetrators (Hoffman, 1998). Like terrorism, hate crimes express a
number of socio-political objectives by targeting individuals based on their
perceived group membership. Biases often prove intricately related to socio-
political and/or religious views. Both acts serve as tactics in the arsenal of
hate groups, a number of which are also labeled as terrorist organizations
such as the Ku Klux Klan (Atkins, 2006; J. Levin, 2013). Similarly,
McDevitt, Levin, and Bennet’s (2002) typology of hate crime offenders
includes the “mission” category made up of members or supporters of orga-
nized hate groups. “Mission” offenders are often racist White supremacist
extremists who believe that they must purge the world of evil by eliminating
the “other” group that threatens their group. B. Levin (2012) notes that these
“hard core hatemongers are believed to be responsible for about 33%-40%
of hate motivated homicides” (para. 7).

Hate crimes and terrorist incidents act as message crimes, instilling fear
and psychological harm, as well as behavioral modification. Noting the
close relationship between hate crime and terrorism, Krueger and Malečková
(2002, 2003) describe the goal of hate crimes to terrorize a larger group
beyond the immediate victim, who is selected on the basis of her or his
group identity. Hate crimes constitute not only an attack on a single person,
but also they send an anti-“other” message to the target’s larger community.
Hate crimes thus present unique harms that distinguish them from ordinary
crimes as they align more closely with terrorism. Several studies (Barnes &
Ephross, 1994; Iganski & Lagou, 2009; Lim, 2009; McDevitt, Balboni,
Garcia, & Gu, 2001) also show that victims of hate crime suffer greater
psychological and emotional harms, including depression, increased fear of
victimization, anger, and stress. For example, Iganski and Lagou (2009)
find that both racial minority (and the larger minority communities) and
White victims of racially motivated crimes avoid certain places and are
more likely to have moved (i.e., changed residences). Increased avoidance
behaviors and other behavioral changes also follow in the aftermath of ter-
rorist attacks, such as those of 9/11 and the 2005 London bombings
(Gigerenzer, 2004, 2006; McArdle, Rosoff, & John, 2012; Prager, Beeler
Assay, Lee, & von Winterfeldt, 2011; Rubin, Brewin, Greenberg, Simpson,
& Wessely, 2005).

Mills et al. 1195

Hate crimes and terrorist acts can be defensive or retaliatory. Defensive
hate crimes are those in which offenders “defend their turf” and send a
message to the larger community to which the victim belongs (Green,
Glaser, & Rich, 1998; McDevitt et al., 2002). Interviewing White youth
in Brooklyn, Pinderhughes (1993) finds that youth committed racially
motivated attacks to defend their turf as they believed that the govern-
ment was taking their jobs and giving them to racial minorities while
Whites suffered unemployment and homelessness. Retaliatory hate crimes
occur in response to some precipitating event, specifically a perceived or
actual hate crime against a member of the offender’s ingroup (McDevitt
et al., 2002). One study of hate crimes in New York City, for example,
found that “cross-sectionally, antiwhite incidents correlate with the num-
ber of antiblack incidents, and temporally these two monthly time series
seem to follow a tit-for-tat pattern” (Green, Glaser, & Rich, 1998 in
Green, Strolovitch, & Wong, 1998, p. 399). Terrorist acts can also be con-
ceptualized as defensive or retaliatory. For example, the Troubles in
Ireland exemplify both models with republican dissidents “defending”
Ireland from the British or acting in retaliation with tit-for-tat attacks by
republican dissidents and loyalists or British forces (LaFree, Dugan, &
Korte, 2009).

Although many scholars argue for the similarities between the two, oth-
ers note how each are unique. In investigating the association between hate
crimes and terrorism, Deloughery et al. (2012) address the claim that hate
crime acts as a “poor man’s terrorist attack” that eventually escalates to
more serious acts of terrorism (p. 665). Unlike most terrorist attacks that
require some level of planning and resources, hate crimes are usually com-
mitted on the spur of the moment.1 Therefore, hate crimes present an ave-
nue for extremists to pursue their socio-political objectives without the
necessity of planning. Hate crimes also pose less danger of arrest. Hate
crimes are underreported and under-investigated and prosecuted (Freilich
& Chermak, 2013; R. D. King, 2007; R. D. King, Messner, & Baller, 2009).
Terrorist attacks garner media, government, and law enforcement attention
and pose a greater threat of apprehension. Hate crimes thus present an
effective route for upholding ideological beliefs while minimizing the costs
of resources and risks.

Hate crime and terrorism further differ in certain ways. Hamm (1993)
argues that the distinction between hate crime and terrorism is nuanced,
remarking that only extreme hate crimes driven by socio-political goals
should be considered terrorism. Deloughery et al. (2012) find that hate
crimes constitute more of a “downward” offense with a majority party
attacking a member of a minority, whereas terrorism proves to be an

1196 Crime & Delinquency 63(10)

“upward” crime with a less powerful group attacking a more powerful one.
Although this may be a valid distinction, it fails to acknowledge the nature
of a terrorist group’s social, political, or religious objectives. Michael (2003)
comments that “terrorism is almost always linked to a wider social move-
ment. . . . Klan terrorism in the South was part of a broader pattern of white
resistance to the civil rights struggle” (p. 105). Shimamoto (2004) also
argues that both terrorism and hate crime attack fundamental notions of
democracy and the state. Therefore, the argument remains that hate crimes
attack society at large by attacking its norms, targeting dearly held values of
equality, liberty, and basic human rights. Such a conception of hate crimes
aligns them with the “upward” nature of terrorism, refuting that hate crimes
are only a downward crime.

Other differences between hate crime and terrorism pertain to offender
and incident characteristics. J. Levin and McDevitt (2002; see also Phillips,
2009) find that the majority of hate crimes are actually committed by
groups of thrill-seeking youth who lack firm ideological beliefs or hate
group affiliation.2 In addition to this thrill nature and the peer dynamics,
these incidents usually involve alcohol or drug consumption and are
unplanned (J. Levin & McDevitt, 2002; McDevitt et al., 2002; Messner,
McHugh, & Felson, 2004). It must be noted though, that thrill-seeking hate
crimes still send a message to the targeted group and often are an out-
growth of societal cultural norms. Byers, Crider, & Biggers (1999) shows
that many in their sample of thrill offenders expressed negative views of
their Amish victims. These thrill hate crime offenders also thought that the
larger community agreed with them that Amish persons were inferior and
not a part of society. Although thrill-seeking hate crime offenders are not
political extremists and are far from being firmly committed terrorists,
they still may be motivated by quasi-political motives. Thrill-seeking
offenders, in other words, often commit these attacks to send a message
that reflects both their personal biases and what they believe to be their
society’s cultural norms.

Another difference is that offenders typically do not claim responsibil-
ity for the attack or publicize it as terrorists often do, but they do not
necessarily need to publicize their crimes themselves (LaFree & Dugan,
2004). As message crimes, hate crimes themselves issue a warning to the
victim’s larger group. Such crimes often garner enough media attention to
get their objective publicized. Despite important differences between hate
crime and terrorism, their similarities provide the groundwork for further
investigation of the relationship between the two phenomena. See Figure
1 for summary of similarities and differences between hate crime and

Mills et al. 1197

The Theoretical Context of the Hate Crime and
Terrorism Relationship

General Hate Crime Escalating to Extremist Hate Crime

The theoretical basis for investigating the relationships between hate crime,
extremist hate crimes, and terrorism relies on intergroup conflict and related
theories, including normative support and social identity theory. Several
studies (Grattet, 2009; Green, Glaser, & Rich, 1998; Green, Strolovitch, &
Wong, 1998; Jacobs & Wood, 1999; R. D. King & Brustein, 2006; R. D. King
& Sutton, 2013; C. J. Lyons, 2007) investigate the role of intergroup conflict
and hate crime. Green, Strolovitch, and Wong’s (1998) “defended neighbor-
hoods thesis” draws on realist group conflicts theories. In brief, realist group
conflict theories posit that White intolerance manifests when racial and

Similarities Differences

•• Early U.S. terrorism linked
to KKK, a notorious hate

•• Hate crimes are often committed on
the spur of the moment; and usually
require less planning and resources

•• Similar language in statutes
(violence, civilian populations,
socio-political objectives)

•• Hate crimes are less likely to result
in an arrest; are under-reported,
-investigated, -prosecuted

•• Biases linked to socio-political
and religious ideologies

•• Hate crimes can be downward
(powerful subgroup attacking a
minority subgroup)

•• Overlap between hate groups
and terrorist groups

•• Many hate crimes are committed by
offenders fueled by alcohol, and drugs.

•• Communicative nature •• Many hate crimes are committed by
non-extremist youths, acting with
others, for the “thrill” of it

•• Instill psychological harms, fear,
and behavior modification

•• Hate crimes can lack a publicity aspect

•• Both can be upward (terrorism
and hate crime attack notions
of democracy, equality, human

Figure 1. The Similiarities and Differences Between Hate Crime and Terrorism
(Barnes & Ephross, 1994; Deloughery, King, & Asal, 2012; Hamm, 1993; Iganski &
Lagou, 2009; LaFree & Dugan, 2004; J. Levin & McDevitt, 2002; Law, 2009; Lim,
2009; Messner et al., 2004; McDevitt et al., 2001; McDevitt et al., 2002; Michael,
2003; Phillips, 2009; Shimamoto, 2004).

1198 Crime & Delinquency 63(10)

ethnic minority groups move into their areas, thereby representing a threat to
their political and economic interests as a competing force for resources
(Green, Strolovitch, & Wong, 1998). Several studies (Grattet, 2009; Green,
Strolovitch, & Wong, 1998; C. J. Lyons, 2007) support the “defended neigh-
borhoods” thesis, evidencing that racially/ethnically based hate crimes are
highly correlated with the influx of minorities into former almost all-White

The following section examines how the presence of general anti-“other”
hate crimes can result in fatal hate crimes by far-rightists. Normative support
proves to be a very significant factor in influencing the use of violence in
intergroup conflict. Louis and Taylor (2002) explain how group norms shape
individual members’ perceptions, specifically perceptions of intergroup con-
flict. Senechal de la Roche (1996) explains that solidarity is integral for
collective violence, permitting violent expression of group grievance. As
such, lynching varied with local solidarity in the American South with close-
knit communities seeing increased lynchings. Gurr (1968) explains that
experimental evidence demonstrates that highly cohesive groups are much
more likely to express hostility against “outsiders” (p. 272). Regarding the
importance of normative support for violence in enabling terrorists, M.
King, Noor, and Taylor (2011) note how Milgram’s experiments demon-
strated that individuals are susceptible to accepting violence when sur-
rounded by others who were compliant with engaging in violence. M. King
et al. (2011) find that jihadi terrorists receive normative support from their
families, as well as the larger community. Ingroup identification provides a
mechanism for individuals to positively see themselves. Social identity the-
ory dictates that people derive self-esteem through their group membership
and by viewing their group positively compared with other groups; further-
more, such group identification strengthens individual conformity to group
norms (Cohrs & Kessler, 2013; Federico, 2013; Louis & Taylor, 2002; P. A.
Lyons, Kenworthy, & Popan, 2010). P. A. Lyons et al. (2010) find that the
interaction of ingroup identification and mean and high-level group narcis-
sism among U.S. citizens was associated with negative attitudes and behav-
iors toward Arab immigrants.

The research demonstrates that extremists are more likely to resort to
violence against a perceived threat when they receive normative support
from their ingroup. Hamm (1993) finds that skinheads are synanomic,
which he defines as “hyperactively bonded to the dominant social order
and to one another” (p. 212). As a result, far-right extremists should be
more likely to not only be more aggressively bonded to their goals of car-
rying out their socio-political objectives in sustaining “traditional” values,
but also their ingroup (White, heterosexual, working-class men).

Mills et al. 1199

Furthermore, far-right extremists should prove to feed off of the normative
support of their ingroup in exercising violence against outgroups.

In sum, there are two causal mechanisms that could explain how regu-
lar hate crimes committed by non-extremists lead to fatal ideologically
motivated attacks committed by far-right extremists. First, regular hate
crimes often attract attention from the media, the larger community, and
committed far-rightists. These regular hate crimes may encourage far-
rightists to conclude that “regular” persons in the general community
share their racial and extremist grievances. For example, Green and Rich
(1998) investigated the association between White supremacist rallies and
demonstrations and cross burnings on the county level in North Carolina.
They found that in counties where White supremacist rallies occurred, the
likelihood of subsequent cross burnings increased. The authors concluded
that White supremacist rallies could encourage individuals traveling to
the event by drawing attention to racial grievances, and therefore facili-
tating action in the form of racial intimidation. Deloughery et al. (2012)
similarly explain that anti-minority hate crimes can highlight growing
anti-minority sentiment to which extremists may respond with more seri-
ous violence. Our argument is that regular hate crimes committed by non-
extremists could (perhaps unintentionally) highlight these same racial
grievances that then encourage far-right extremists to commit fatal acts of
ideologically motivated hate crimes.

Second, these regular hate crimes are often fiercely denounced by gov-
ernment officials, minority communities, and advocacy groups (Jenness &
Grattet, 2004; J. Levin & McDevitt, 2002). Simi and Futrell (2010) have
explained that far-rightists commonly feel stigmatized by mainstream soci-
ety. Many therefore retreat to “free places” where they are better able to
subscribe to and act upon their extremist beliefs, and interact with others
who think like them. It is possible that these denunciations of regular hate
crimes aggravate the feelings of persecution held by many far-rightists that
is reinforced by others who share their views. This in turn could create a
backlash effect that results in some far-rightists committing fatal hate
crimes. As the far-right movement often attracts violent individuals (see,
for example, Ezekiel, 1995; Freilich, Adamczyk, Chermak, Boyd, & Parkin,
2015), we wonder, in other words, whether some far-rightists engage in
fatal bias-motivated violence in response to the condemnation of regular
anti-minority hate crime, which presents an attack on their grievances and

Based upon both of these possible causal mechanisms, we hypothesize
that places experiencing hate crimes in general are more likely to experience
fatal hate crimes by far-right extremists.

1200 Crime & Delinquency 63(10)

Extremist Hate Crime Escalating to Extremist Terrorism

In terms of extremist hate crime escalating to terrorism in the intergroup context,
Michael (2003) discusses Sprinzak’s theory of “split-delegimitization” as
applied to right-wing terrorism, which asserts that “outsiders” as well as the
state simultaneously come under attack (p. 95). Michael (2003) contends that
this theoretically supports the evolution of right-wing terrorism with attacks
escalating from those against the “outsiders” to the state due to the state’s per-
ceived alliance with the “outsiders” (p. 95). Supporting the theoretical escala-
tion, Michael (2003) looks to Hewitt’s (2000) descriptive data on American
domestic right-wing terrorism, which evidences a demonstrable escalation in
violence against the state. Hewitt’s (2000) data show that the majority of the first
wave of far-right attacks from the 1950s to the 1970s was against people based
on their race or ethnicity, followed by civil rights workers. The second wave
from the 1970s to the present shows that the far-right has increasingly targeted
the government, including attacks against law enforcement, politicians, and
government facilities. Examining the life course of American far-right groups,
Kerodal, Freilich, Chermak, and Suttmoeller (2015) empirically test and find
support for Sprinzak’s theory, uncovering that the far-right initially attacked
non-government targets but began to equally strike both non-government and
government targets after becoming disillusioned with the government. Such
findings support the idea that the far-right may move from simply engaging in
hate crimes against minorities to anti-government attacks as well, signaling an
escalation in their activities. Deloughery et al.’s (2012) case study of Timothy
McVeigh’s horrific anti-government bombing attack of the Federal building in
Oklahoma City in 1995 demonstrated that increases in anti-minority hate crimes
were a way to express right-wing grievances and can act as a warning or a signal
that some extremists will subsequently potentially “upgrade” to (anti-govern-
ment or American society at large) terrorism (p. 668). As a result, we hypothe-
size that counties experiencing fatal hate crimes by far-rightists would also see
far-right terrorism with these extremists employing violence against both minor-
ity and government targets and American society at large.

Extremist Hate Crime as Response to Terrorism

Regarding extremist hate as a response to anti-American terrorism, a review of
the literature on group grievance, social control, and retaliation is useful. Black
(1983) posits a theory of crime as social control, in which individuals use crime
as “self-help” to express their group’s grievance against a particular subgroup to
maintain social control. McCauley and Moskalenko (2011) define group (or
political) grievance as a mechanism for radicalization and as the “threat or harm

Mills et al. 1201

to a group or cause the individual cares about can move the individual to hostil-
ity and violence toward perpetrators” (p. 21). Terrorist attacks perceived to
attack “traditional” or “American” values thus present extremists with a group
grievance that manifests in violent retribution. Vicarious retribution occurs
when an ingroup member views an entire outgroup responsible for a harm
against a fellow ingroup member and thus attacks an outgroup member for ret-
ribution (Lickel, Miller, Stenstrom, Denson, & Schmader, 2006). Several schol-
ars (Lickel et al., 2006; McCauley & Moskalenko, 2008, 2011) explain that a
popular mechanism for both radicalization and vicarious retribution is dehu-
manization of the “enemy.” Lickel et al. (2006) comment that intergroup con-
flict sees dehumanization of the outgroup, which facilitates vicarious retribution
as outgroup members are seen “as being interchangeable and therefore equally
deserving of retaliation” (p. 378). Retaliatory hate crimes involve individuals
who seek revenge by targeting innocent bystanders whom they perceive as rep-
resentative of a larger enemy. Several studies (Byers & Jones, 2007; Deloughery
et al., 2012; Disha et al., 2011; R. D. King & Sutton, 2013; McDevitt et al.,
2002) demonstrate the prevalence of hate crimes following terrorist attacks.
Hate crimes targeting perceived Middle Eastern victims occurred not only after
9/11 and the Boston Marathon, but also immediately at the start of the Iran hos-
tage crisis in 1979 (J. Levin & McDevitt, 2002). Retaliatory hate crimes thus act
as micro-level manifestation of broader conflicts on the international scale.

The synanomic nature of far-right extremists thus explains why they are
likely to respond to terrorist attacks against “traditional” or “American” val-
ues with hate crimes against outgroups they perceive as a threat or as respon-
sible for precipitating terrorist attacks. As a result, extremists prove more
likely to exercise hate crime as a form of social control. Furthermore, norma-
tive support exists for retributive violence in the course of intergroup conflict
(Lickel et al., 2006). Therefore, extremists feed off of normative support to
not only engage in hate crimes in general, but also specifically as a form of
vicarious retribution. Retaliatory hate crimes following terrorist attacks thus
express group grievance, as well as social control, by those ultra-committed
to upholding the dominant social order. We hypothesize that counties that
experience terrorist attacks by non-right-wing groups would be more likely
to see an increase in fatal hate crimes by far-right extremists.

Revisiting Deloughery et al. (2012): Are Hate Crimes Only
Distant Relatives?

Using HCSA and Global Terrorism Database (GTD) data, Deloughery et al.
(2012) examine the temporal proximity of hate crimes and terrorism and find
that (a) hate crimes do not necessarily lead to future right-wing terrorism, (b)

1202 Crime & Delinquency 63(10)

hate crimes are more often a response to terrorism, and (c) anti-minority hate
crimes prove especially prevalent after non-right-wing terrorist attacks that
seem to attack traditional “American” values. As a result, they conclude that
hate crimes and terrorism are more akin to “distant relatives” as hate crimes
are not indicative of future terrorist attacks. One limitation acknowledged by
the authors is the absence of a measure for ideological strength or offender
affiliation. The majority of hate crime offenders consist of thrill offenders,
who lack firmly committed extremist ideological beliefs (J. Levin & McDevitt,
2002; McDevitt et al., 2002). Therefore, HCSA data does not allow research-
ers to identify which offenders subscribe to extremist views which undermines
our ability to study this phenomenon. Offenders who subscribe to extremist
right-wing ideology, however, prove more likely to resort to both ideologi-
cally motivated hate crimes and terrorist acts than non-ideological offenders.
This study extends Deloughery’s important work in two ways: by utilizing
fatal bias-motivated homicides committed by far-rightists as contained in the
ECDB and by examining the county-level association of hate crime and

Finally, based upon the prior literature we also examine four additional
hypotheses (for a total of seven). Research establishes that intergroup conflict
occurs when minorities pose a threat to the interests of the dominant group
(Blalock, 1967; Green, 1998, Green, Strolovitch, & Wong, 1998, R. D. King
& Brustein, 2006). Greater rates of minority presence are said to lead to
White intolerance, and in turn violence against minorities, as their presence
poses a threat to White economic interests (Green, Strolovitch, & Wong,
1998). Previous studies (Disha et al., 2011; C. J. Lyons, 2007) find that
greater racial/ethnic minority presence explains interracial violence.
Intergroup conflict theories also posit that ethnic heterogeneity can also lead
to greater conflict (Olzak, Shanahan, & McEneaney, 1996; Shanahan &
Olzak, 1999). Another important predictor in studies examining intergroup
conflict is demographic change. Green, Strolovitch, and Wong’s (1998)
defended neighborhoods thesis holds that demographic change over time
with minority growth in areas contributes to White violence against “invad-
ing” minorities. We hypothesize that those counties with greater minority
presence as well as greater ethnic heterogeneity will be more likely to see
far-right activity. We further hypothesize that demographic change (i.e.,
minority presence increasing over time) will be associated with far-right
activity as well.

The literature on intergroup conflict often relies on measures of eco-
nomic competition. Theoretically, poor economic conditions foster
increased racial competition for resources, which, in turn, fosters increased
intergroup conflict leading to violent outcomes such as hate crimes and

Mills et al. 1203

terrorist incidents (Corzine, Huff-Corzine, & Creech, 1988; C. J. Lyons,
2007; Olzak, 1989, 1990; Soule, 1992; Tolnay & Beck, 1995; Tolnay,
Deane, & Beck, 1996). Relatedly, deprivation frameworks in criminology
such as the classic strain theories maintain that poorer locations usually
provide fewer opportunities for success. Some persons use crime as an
alternative way to achieve financial success as the legal opportunities are
closed to them (Merton, 1938). Often, persons in these areas are socially
isolated from mainstream society. These areas may also attract offenders
from other locations who exacerbate this locale’s crime problem (Messner
& Rosenfeld, 2007). Significantly, economic deprivation has also been
seen as linked to far-right extremism (Lipset & Raab, 1977). Freilich et al.
(2015; see also Pridemore & Freilich, 2006), for example, discuss how far-
right extremists residing in poorer locations might conclude that their ide-
ological opponents are responsible for their economic deprivation. These
far-rightists may therefore then attack these opponents. We hypothesize
that counties experiencing both higher rates of unemployment and poverty,
as well as increased rates in both of these domains over time, will see
higher numbers of far-right activity.

Data and Methods

This study investigates the research question, “Are hate crimes and terrorism
more interrelated than prior research has demonstrated?” Using incident data
aggregated to the county level, we seek to address whether hate crime and
terrorism prove more similar to each other than not by studying the spatial
association between the two phenomena. We test the following seven

Hypothesis 1: An increase in counties’ non-fatal anti-minority/anti-
“other” hate crimes committed by all type of perpetrators is associated
with an increase in counties’ fatal hate crimes committed by far-rightists.
Hypothesis 2: An increase in counties’ fatal hate crimes committed by
far-rightists is associated with an increase in counties’ far-right terrorist
Hypothesis 3: An increase in counties’ terrorist attacks by non-right-wing
groups that attack “traditional/American values” is associated with an
increase in counties’ fatal hate crimes committed by far-rightists.
Hypothesis 4: An increase in counties’ levels of minority presence and
diversity is associated with increases in counties’ far-right activity.
Hypothesis 5: Growing minority presence and diversity over time is asso-
ciated with increases in counties’ far-right activity.

1204 Crime & Delinquency 63(10)

Hypothesis 6: An increase in counties’ levels of poor economic condi-
tions (poverty and unemployment) is associated with increases in coun-
ties’ far-right activity.
Hypothesis 7: Worsening economic conditions (poverty and unemploy-
ment) over time is associated with increases in counties’ far-right

To address these hypotheses, this study conducts a county analysis using
pooled event counts by county over a 20-year period (1992-2012) from three
different databases: the U.S. ECDB, the HCSA from the Uniform Crime
Report (UCR), and U.S. cases from the GTD.

The GTD is a terrorist event database that includes all terrorist attacks that
occur around the globe using open-source data (see LaFree & Dugan, 2007,
for information on incident inclusion criteria). This study uses incident-level
data from the GTD to examine pooled counts of 223 right-wing3 and 225
non-right-wing/anti-“American” (primarily far-left animal/environmental
terrorists and radical Islamists) terrorist attacks in the United States over the
20-year period from 1992 to 2012 (excluding 1993, which is missing from
the GTD) (data from National Consortium for the Study of Terrorism &
Responses to Terrorism, 2014). As Deloughery et al. (2012) do in their study,
this study identifies the perpetrator type (i.e., far-right vs. far-left) by using
data from the Terrorist Organization Profiles (TOPS), which codes and orga-
nizes groups by ideology (Deloughery et al., 2012). The current study also
evaluates each potential individual or unknown perpetrator attack to discern
and classify terrorist attacks according to evidence indicating far-right or
non-right-wing/anti-“American” perpetrators or motivations.4

The HCSA of 1990 provides for the collection of data on hate crime inci-
dents in the United States with law enforcement agencies recording and sub-
mitting counts and other possible descriptive information of hate crime
incidents in their jurisdictions to the FBI for inclusion in the UCR (U.S.
Department of Justice, FBI, 2011a). Currently, the federal hate crimes act
charges the Attorney General with collecting data on designated crimes moti-
vated by “prejudice based on race, gender and gender identity, religion, dis-
ability, sexual orientation, or ethnicity” (U.S. Department of Justice, FBI,
2011b). Similar to other official crime databases, limitations exist with the
HCSA data as hate crimes suffer from underreporting as well as differential
compliance with recording and reporting hate crimes by location (R. D. King,
2007; R. D. King et al., 2009).

Importantly though, despite its limitations the FBI’s HCSA is recognized
as one of the most reliable sources available for county-level hate crime data.
The HCSA includes more participating police agencies and covers more of

Mills et al. 1205

the nation’s population than the FBI’s National Incident Based Recording
System’s (NIBRS) bias crime data. In addition, watch-groups do not publish
annual listings of all hate crimes for the entire nation in any systematic fash-
ion (Freilich & Chermak, 2013). Regarding the National Crime Victimization
Survey, there could be significant variation in the respondents’ understanding
of hate crime victimization. In sum, the HCSA is one of the more reliable
sources for hate crimes data. Indeed, a series of studies have used HCSA data
to investigate a variety of important issues (see, for example, Byers & Jones,
2007; Deloughery et al., 2012; Disha et al., 2011; R. D. King et al., 2009; R.
D. King & Sutton, 2013).

The current study uses a pooled count of 130,289 non-fatal anti-“other” or
anti-minority (including all minority groups protected by the federal legisla-
tion listed above) hate crimes from 1992 to 2012 (excluding 1993). Data on
only non-fatal acts ensure there is no overlap with any of the fatal far-right
bias crimes obtained by the ECDB (data obtained from U.S. Department of
Justice, Federal Bureau of Investigation, 2014).

The U.S. ECDB provides a rich source of data on violent and financial
crimes committed by extremists, specifically far-rightists, Al-Qaeda inspired,
as well as extremist animal or environmental rights advocates (Freilich,
Chermak, Belli, Gruenewald, & Parkin, 2014). Unlike other databases, the
ECDB includes only incidents, plots, or schemes in which at least one extrem-
ist was involved. In addition to non-ideological violence, fatal incidents cap-
tured by the ECDB include ideologically motivated homicides against
government targets as well as other ideological targets based on biases (i.e.,
against racial groups). In addition to ideologically motivated violence most
closely matching common definitions of terrorism, the specification of bias-
motivated violence by extremists most closely approaches the phenomenon
of interest for this study. The ECDB has proved to be a valid source of data
on fatal far-right ideologically motivated attacks (Chermak, Freilich, Parkin,
& Lynch, 2012). Recent studies have relied on the ECDB to examine the
evolution of domestic extremist groups (Freilich, Chermak, & Caspi, 2009),
differences between violent and non-violent extremist groups (Chermak,
Freilich, & Suttmoeller, 2013; Suttmoeller, Chermak, & Freilich, 2015),
comparisons between far-right homicides and “regular” non-extremist homi-
cides (Gruenewald & Pridemore, 2012), fatal far-right attacks against the
police (Freilich & Chermak, 2009; Suttmoeller, Gruenewald, Chermak, &
Freilich, 2013), lone wolf attacks (J. Gruenewald, Chermak, & Freilich,
2013a, 2013b), and county-level variation in extremist violence (Chermak &
Gruenewald, 2015; Freilich et al., 2015).

The use of the ECDB improves upon the use of HCSA data in Deloughery
et al.’s (2012) study because it provides data on bias-motivated violent

1206 Crime & Delinquency 63(10)

incidents in which at least one perpetrator is a far-rightist who committed
a fatal attack to further their extremist ideology. This study uses 118 bias-
motivated fatal attacks committed by far-rightists pooled over a 20-year
period (1992-2012, excluding 1993).5 The attacks include those commit-
ted because of the suspects’ bias against persons based on sexual orienta-
tion, homelessness, or membership in minority racial/ethnic or religious

Importantly, although we are examining two measures of far-right extrem-
ist criminal activity, they are distinct universes and no case was double-
counted. The ECDB’s data on bias-motivated homicides include far-right
fatal attacks that targeted racial/ethnic and religious minorities, gay, bisexual,
and homeless persons. All of these attacks are ideologically motivated but
many (though by no means all) are also the outcome of “presented opportuni-
ties” (Freilich et al., 2015). In these cases, the offenders’ paths crossed with
the victim at which time the perpetrators seized the opportunity to attack.
These incidents are also often labeled as “hate crimes” and not “terrorism.”
All ECDB anti-government and anti-abortion attacks were excluded as they
already appeared in the GTD, and we thus insured that they were not double-
counted. However, the GTD far-right cases include mostly planned attacks
against the government or American society at large, abortion-related targets
as well as anti-minority (i.e., bias/hate) cases. Importantly though, the 11
fatal far-right attacks in the GTD that targeted a minority, gay, or homeless
person were removed as they were already included in the ECDB universe
just discussed. This allowed us to better capture this middle ground of extrem-
ist hate crime and insured that no case was double-counted.

Demographic indicators come from the Decennial Census from the
years 1990, 2000, and 2010.6 To account for the county racial/ethnic
minority presence, we use the average percentage of the non-White/non-
Hispanic population from 1990 to 2010. Given far-right’s general preju-
dice against all non-White racial and ethnic groups, this analysis considers
the entire non-(non-Hispanic) White population that we label as minority
presence. The average diversity index, as well as the accompanying change
in the index over the 20-year period, is another predictor of interest,
accounting for ethnic heterogeneity.7 As the average minority presence and
change predictors are both highly correlated with the average diversity
index and change variable, respectively, we use them in separate models.
We also include a measure of demographic change, specifically account-
ing for the absolute change in the non-White population from 1990 to
2010. Using data from the U.S. Bureau of Labor Statistics for 1990, 2000,
2010 (U.S. Bureau of Labor Statistics, 1990; 2000; 2010 and the Census
Bureau for1989, 1999, and 2009 (U.S. Census Bureau, 1990; 2000; 2010),

Mills et al. 1207

we use two oft-used economic indicators: unemployment and percent
below the poverty level. We calculate the average unemployment and pov-
erty and their absolute changes in these rates, respectively, over the 20-year

Analytical Approach

The current study investigates “spaces of hate” (Disha et al., 2011, p. 40),
testing the association between hate crimes and terrorist acts at the county
level. We focus on the dependent variables of fatal hate crimes committed by
far-rightists (ECDB) and far-right terrorist acts (GTD). We use the total
counts at the county level over the 20-year period (1992-2012) for these two
types of events in 3,137 U.S. counties. Given the rare event nature of extrem-
ist activity (specifically homicides and terrorist acts by far-rightists), both
dependent variables are skewed with many counties failing to experience
either type of extremist activity. With the skewed distribution and overdisper-
sion, we conduct a series of negative binomial regressions to test the associa-
tions between hate crime and terrorism, as well as county-level demographic
and economic characteristics.8


Descriptive statistics are presented in Table 1. They show that extremist
activity is very rare with county-level means close to zero for fatal far-right
hate crime, far-right and non-right-wing terrorist acts. General hate crimes
average about 42 in 3,132 U.S. counties9 over 1992 to 2012.

The first set of analyses investigates the association between general
hate crime and fatal hate crime by far-rightists with the results presented
in Table 2. The results show a significant, yet weak, positive relationship
between general hate crime and bias-motivated homicides by extremists.
Model 1 presents the baseline model regressing general hate crimes on
fatal hate crimes by far-rightists. Regarding the economic and demo-
graphic predictors relevant to intergroup conflict, the full models (Models
2 and 3) present a number of interesting findings. Model 2 includes the
average minority presence, and Model 3 includes the average diversity
index. Inspecting demographic predictors, both minority presence and
ethnic heterogeneity explain increases in far-right hate crime. Minority
presence is a much weaker predictor accounting for only a 2% increase in
such events. The diversity index shows that increased ethnic heterogene-
ity is associated with an increase in far-right hate crime at a rate of
approximately 93 times greater. This is most likely due to the

1208 Crime & Delinquency 63(10)

ratio measure of the diversity index. Similarly, demographic change is
associated with increases in far-right hate crime. Although changes in the
unemployment rate over time fails to achieve significance, increases in
average unemployment see a 10% (p < .1) and 15% increase in extremist hate crime in Models 2 and 3, respectively. Contrary to expectations, pov- erty is negatively associated with far-right hate crime with approximately an 8% decrease in such events. Growth in poverty over time, however, accounts for about an 11% increase in far-right hate crime. Counties expe- riencing a growth in poverty see greater numbers of far-right hate crime, which is in line with the predicted relationship. All in all, there is weak evidence to support the first hypothesis.

The second hypothesis, however, receives much more support. Table 3
shows strong, positive associations between fatal far-right hate crimes
and far-right terrorist attacks. Model 1 presents the baseline model with
an increase in fatal far-right hate crimes seeing 9 times more far-right ter-
rorist acts. Models 2 and 3 show that counties seeing increased far-right
hate crime are about 4 times more likely to see far-right terrorist acts,
respectively. As in Table 2, minority presence and ethnic heterogeneity
significantly increase far-right terrorism while the average poverty pres-
ence observes significant declines. In Models 2 and 3, the change in pov-
erty is significant with increased poverty over time seeing an 11% and 9%
increase in far-right terrorism. In addition to change in minority presence

Table 1. Descriptive Statistics.


No. of general hate crimes (HCSA) 41.59 264.64
No. of fatal FR hate crimes (ECDB) 0.04 0.26
No. of FR terrorist acts (GTD) 0.07 0.44
No. of non-right-wing/anti-“American” terrorist acts (GTD) 0.07 0.54
Average % minority presence 18.59 18.86
Change in minority presence (1990-2010) 6.28 6.13
Average unemployment rate 6.58 2.28
Change in unemployment rate (1990-2010) 3.07 2.66
Average % below poverty level 15.72 6.74
Change in % below poverty level (1989-2009) −0.36 4.05
Average diversity index 0.25 0.18
Change in diversity index (1990-2010) 0.08 0.07

Note. N = 3,137 U.S. counties from 1992 to 2012 (excluding 1993). HCSA = Hate Crime
Statistics Act; FR = far-right; ECDB = Extremist Crime Database; GTD = Global Terrorism

Mills et al. 1209

and ethnic heterogeneity, the average unemployment rate and its change
over time, however, fail to achieve significance.

Finally, the third set of analyses tests the association between non-right-
wing/anti-“American” terrorism and extremist hate crime. Once again, the
results in Table 4 demonstrate a significant, positive association between
such terrorist acts and far-right hate crime. The results in the full Models 2
and 3 show that increases in non-right-wing/anti-“American” terrorist acts
see a 78% and a 64% increase in far-right hate crime, respectively.
Regarding the demographic and economic predictors, the results remain
largely the same as Table 2 as the main predictors are just alternated in
Table 4. The three sets of analyses confirm our original hypotheses regard-
ing the associations between hate crime and extremist activities.

Table 2. Negative Binomial Regression Models: Fatal Hate Crimes by Far-

Independent variables

Model 1 Model 2 Model 3

I.R.R. (SE) I.R.R. (SE) I.R.R. (SE)

General hate crimes 1.00*** 1.00* 1.00*
(0.00) (0.00) (0.00)
Average % minority presence 1.02**
Change in minority presence


Average diversity index 93.03***
Change in diversity index


Average unemployment rate 1.10† 1.15**
(0.05) (0.06)
Change in unemployment rate

0.97 0.95

(0.05) (0.05)
Average % below poverty level 0.92*** 0.91***
(0.02) (0.02)
Change in % below poverty level

1.11*** 1.11***

(0.03) (0.03)
Wald χ2 26.50*** 156.42*** 160.27***
Log pseudolikelihood −415.39 −390.55 −387.91

Note. N = 3,132 U.S. counties from 1992 to 2012 (excluding 1993).
†p <≤.1. *p ≤ .05. **p ≤ .01. ***p ≤ .001. I.R.R. = Incident Rate Ratio.

1210 Crime & Delinquency 63(10)


Much of the prior literature disagrees over the nature of the hate crime–ter-
rorism relationship with some calling the two phenomena “close cousins,”
whereas others call them “distant relatives.” As previous studies focus on
either general hate crime and terrorism data sources, they miss an important
middle ground between the two phenomena, specifically fatal hate crimes
committed by far-right extremists that are not included in the data on terror-
ism. Data on extremist hate crime can provide a promising avenue for future
examination of the hate crime–terrorism relationship. Our analysis runs
counter to Deloughery et al.’s (2012) findings with positive associations
between hate crime and terrorism at the county level. Whereas there is only a
small positive association between general hate crime offending and fatal

Table 3. Negative Binomial Regression Models: Far-Right Terrorist Acts.

Independent variables
Model 1 Model 2 Model 3
I.R.R. (SE) I.R.R. (SE) I.R.R. (SE)

Fatal FR hate crimes 9.33***



Average % minority presence 1.05***
Change in minority presence


Average diversity index 258.01***
Change in diversity index (1990-2010) 10.20
Average unemployment rate 0.93 0.97
(0.04) (0.05)
Change in unemployment rate

1.08 1.06

(0.05) (0.05)
Average % below poverty level 0.89*** 0.91***
(0.02) (0.02)
Change in % below poverty level

1.11*** 1.09**

(0.03) (0.03)
Wald χ2 75.26*** 167.53*** 184.42***
Log pseudolikelihood −651.50 −596.42 −589.16

Note. N = 3,132 U.S. counties from 1992 to 2012 (excluding 1993). FR = far-right; I.R.R. =
Incident Rate Ratio.

Mills et al. 1211

far-right hate crime, there are much stronger positive associations between
fatal far-right hate crime and far-right terrorism, as well as fatal hate crime
and non-right-wing terrorism that targets “traditional” American values. The
results show that counties experiencing increases in general hate crime, far-
right hate crime, and non-right-wing terrorism see associated increases in
far-right hate crime, far-right terrorism, and far-right hate crime, respectively.
In summary, counties undergoing increases in one type of extremist activity
are likely to see increases in other types of extremist activity.

In addition to supporting our main hypotheses, the results also corroborate
hypotheses stemming from intergroup conflict theories. Regarding minority
group threat, the analyses consistently show significant, positive associations
between both measures of minority presence and ethnic heterogeneity and

Table 4. Negative Binomial Regression Models: Fatal Hate Crimes by Far-

Independent variables
Model 1 Model 2 Model 3
I.R.R. (SE) I.R.R. (SE) I.R.R. (SE)

Anti-U.S. terror acts 2.55*** 1.78*** 1.64***
(0.44) (0.19) (0.15)
Average % minority presence 1.03***
Change in minority presence


Average diversity index 213.73***
Change in diversity index (1990-2010) 34.15*

Average unemployment rate 1.10† 1.16**
(0.05) (0.06)
Change in unemployment rate

0.96 0.94

(0.05) (0.05)
Average % below poverty level 0.90*** 0.89***
(0.02) (0.02)
Change in % below poverty level

1.13*** 1.12***

(0.04) (0.04)
Wald χ2 29.21*** 153.78*** 175.97***
Log pseudolikelihood −436.58 −388.62 −386.76

Note. N = 3,132 U.S. counties from 1992 to 2012 (excluding 1993).
†p ≤ .1. *p ≤ .05. **p ≤ .01. ***p ≤ .001.
I.R.R. = Incident Rate Ratio.

1212 Crime & Delinquency 63(10)

far-right hate crimes and terrorist acts. Furthermore, demographic change
measures of increased minority presence and ethnic heterogeneity over time
consistently prove positively associated with far-right bias-motivated homi-
cides. Contrary to the theoretical framework, increased poverty proves sig-
nificantly associated with fewer far-right hate crimes and terrorist acts. This,
however, corresponds with previous research finding a negative relationship
between poverty and extremist presence and activities (Freilich et al., 2015;
LaFree & Bersani, 2014). As more far-right acts are committed in counties
with less poverty, it could be that those with “more to lose” are committing
these attacks (Freilich et al., 2015). In this sense, this finding would be con-
sistent with backlash models that view hate crime offenders as reacting to
perceived threats. However, increased poverty over time corresponds with
the predicted relationships in all of the models. This finding may illustrate
that the far-right is reacting to worsening poverty as more of perceived threat
than the general level of poverty. The current findings corroborate the previ-
ous research showing that counties experiencing higher levels of poverty are
not at risk of extremist activities; however, the findings do demonstrate the
importance of investigating change in poverty over time. Counties coping
with higher unemployment rates also see increased far-right hate crimes. As
a result, poor or worsening economic conditions over time are more strongly
associated with far-right activities. For the most part, the analyses support the
major tenets of minority group threat with growing minority presence and
poor or worsening economic conditions being linked to intergroup violence,
with White intolerance manifesting in far-right extremist acts. As a result, all
levels of government should be concerned with the effects of worsening eco-
nomic conditions and work to improve such conditions. Addressing such
macro-level economic conditions can potentially reduce the appeal of the
far-right and its ideology and thus reduce the threat of its violent activities.

Given the results, it appears that counties experiencing any type of extrem-
ist activity are likely to be targets for other extremist activities. Such results
have potential implications for law enforcement. Since the passage of hate
crime legislation, law enforcement agencies across the country established
specialized bias crime units to handle the unique threat of bias-motivated
crimes. Due to the unique harms caused by hate crimes and terrorism, both
require specialized attention by law enforcement. Such extremist acts inflict
injury and death to both law enforcement as well as citizens, especially those
targeted for their inherent characteristics (race, ethnicity, religion, etc.).
Freilich, Chermak, and Simone (2009) present survey data that show that 85%
of state police agencies reported the presence of right-wing groups. They also
find that these state police agencies consider Islamic terrorism a greater threat
on the national and state level than that of far-right terrorism. Given the

Mills et al. 1213

variation in far-right groups and actions, Chermak, Freilich, and Shemtob
(2009) emphasize the importance of understanding the distinctions between
different groups, their beliefs, and how they inform their extremist activities.
Greater attention to such details should figure into training for law enforce-
ment in dealing with the far-right. The current study shows that law enforce-
ment should be attentive to the gradations in far-right extremist crimes.

The results reinforce the need for government policy makers and practitio-
ners, especially law enforcement to defuse tensions and strengthen community
relations in counties seeing such extremist activities. The presence of far-right
activities can reveal the underlying issues at work in the county. Demographic
change and worsening economic conditions can exacerbate group tensions, dam-
aging community relations; intervention, however, can defuse such tensions. For
example, this study’s results would be useful for the U.S. government’s
Community Relations Service (CRS). The CRS exists as a mediating agency,
working with various types of institutions at the government and organizational
levels, including community and civil rights groups. The CRS endeavors to
address “community conflicts and tensions arising from differences of race,
color, national origin, gender, gender identity, sexual orientation, religion, and
disability” (U.S. Department of Justice, n.d.). Agencies such as the CRS would
benefit from this study’s results by addressing what states, and more specifically
what counties, need their services to address their local-level conflicts evidenced
by the higher rates of extremist activities. Turning to specifically addressing
counties’ extremist presence, past research uncovering the county-level processes
facilitating extremist activities emphasizes the need for law enforcement to estab-
lish communication with far-right groups (Adamczyk, Gruenewald, Chermak, &
Freilich, 2014). In addition to recommending that police monitor hate groups and
track bias-motivated incidents, Freilich and Chermak (2013) stress the impor-
tance of law enforcement reaching out to the various community stakeholders
invested in the problem of bias-motivated violence, including schools, academ-
ics, victims services, as well as other community organizations. Further explora-
tion of the relationship between hate crime and terrorism will contribute to the
production of policies aimed at preventing the escalation of violence by far-right
extremists, thus preventing harm to citizens and law enforcement.

There exist several limitations with the current analysis. The first limita-
tion lies with the HCSA data from the UCR. Some counties have zero hate
crimes due to either a lack of compliance with reporting requirements or the
inability of their law enforcement agencies to recognize and investigate hate
crimes as such. This shortcoming, however, is limited to only one predictor
(general hate crime) in the analyses testing our first hypothesis. Second, the
analysis is one that is concerned with what Disha et al. (2011) term “spaces
of hate” (p. 40); as such, the data include the cumulative totals for each

1214 Crime & Delinquency 63(10)

county over a 20-year period. As a result, the issue of time arises in our analy-
ses. Tita and Cohen (2004) address how research often examines the effects
of space and time on crime separately; however, they note it is important to
consider space and time simultaneously in their analysis as phenomena such
as the “mechanisms of crime . . . are interdependent both over time and across
geographic space” (p. 171). Future analysis may want to control for the time
period as well. It may prove necessary to further unpack these associations as
it may be possible that these relationships are working in reverse. Regardless,
the positive associations at the county level evidence that those counties that
experience higher levels of various types of bias-motivated or extremist vio-
lence are more likely to witness higher levels of other types of bias-motivated
or extremist violence. In summary, the current study provides the ground-
work for further analysis to more deeply investigate these interesting associa-
tions between different types of extremist activities at the county level.

Conclusion: Extremist Hate Crime as Common

Although hate crime and terrorism differ in important ways, their similarities
warrant further investigation into the relationship between the two phenom-
ena. In addition to both serving as tactics by hate and terrorist groups, hate
crime and terrorism share common characteristics, including their socio-
political, communicative aspects, as well as their use as defensive or retalia-
tory tactics. Through the use of multiple data sources, this study uncovers the
positive associations between hate crime and terrorism. In the context of
intergroup conflict, there appears to be a continuum between the bias-moti-
vated actions of non-extremists to the hate crimes and terrorist acts commit-
ted by far-rightists, with the presence of one type of activity seeing an
escalation in the next type. As a result, it appears that hate crime and terror-
ism may be more akin to close cousins than distant relatives.


We thank Dr. Mike Maxfield for his invaluable feedback on earlier drafts of this
article, Dr. Jeremy Porter for his helpful advice on methods, Dr. Ashmini Kerodal for
all of her feedback on this project, as well as Maggie Schmuhl, M.A. for her help and
for providing the Diversity Index for this project.

Authors’ Note

The views and conclusions contained in this document are those of the authors and
should not be interpreted as necessarily representing the official policies, either
expressed or implied, of the U.S. Department of Homeland Security, or START.

Mills et al. 1215

Declaration of Conflicting Interests

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


The authors disclosed receipt of the following financial support for the research,
authorship and/or publication of this article: This research was supported by the
Office of University Programs Science and Technology Directorate of the U.S.
Department of Homeland Security through the Center for the Study of Terrorism and
Behavior (CSTAB–Center Lead) Grant made to the START Consortium (Grant


1. There are exceptions to both of these claims. A few hate crimes are planned (e.g.,
a group of youths plan to go “gay-bashing” later that night), and certain terror-
ist attacks arise from “presented opportunities” (e.g., an anti-government patriot
who is pulled over by the police and then spontaneously kills the officers due to
his anti-government ideology).

2. Although Phillips’ (2009) research on a sample of hate crime offenders from a
New Jersey county also found that thrill seekers were the most common type
of hate crime offenders, it was only a plurality (43%) of the total. Significantly
though, Phillips also found that ideologically motivated extremist mission
offenders comprised a larger share of all hate crime offenders compared with
Levin and McDevitt’s sample.

3. Right-wing traits include

“fiercely nationalistic, anti-global, suspicious of federal authority and
reverent of individual liberties . . . believe in conspiracy theories involving
imminent threats to national sovereignty or personal liberty and beliefs that
their personal or national “way of life” is under attack . . . for some the threat
also originates from specific racial or religious groups. They believe that
they must be prepared to defend against this attack by participating in
paramilitary training or survivalism.” (Freilich, Chermak, Belli, Gruenewald,
& Parkin, 2014, p. 380).
Anti-abortion attacks included.

4. For such individual/unknown cases in the GTD, we used the incident’s GTD
incident description and follow-up open-source searching to evaluate and deter-
mine whether the cases evidenced right-wing or non-right-wing perpetrators or
motivations (especially in cases where there were no sources for the incident) for
inclusion in the study.

5. The 118 incidents from the ECDB do not include all ideologically motivated homi-
cides contained in the database. This analysis excludes attacks prior to 1992, as
well as those that occurred in 1993 (anti-government and anti-abortion ideological

1216 Crime & Delinquency 63(10)

homicides also excluded). The analysis, however, was more inclusive than the
ECDB when classifying cases as bias-motivated when ECDB guidelines found
incidents non-ideological as long as the attack was motivated as least “in part” by
bias (based on statutory membership categories) as hate crime legislation dictates.

6. Census Bureau Data on Race/Hispanic Origin for 1990, 2000, and 2010
obtained from the Minnesota Population Center. National Historical Geographic
Information System: Version 2.0. Minneapolis, MN: University of Minnesota
2011 (2011a;2011b;2011c).

7. The diversity index is a ratio measure, showing the likelihood that two randomly
selected people would differ by race/ethnicity with the formula, “Square the per-
cent for each group B. Sum the squares, and subtract the sum from 1” (U.S.
Census Bureau, 2001). We calculate the average diversity index using the com-
puted indices from the Census in 1990, 2000, and 2010.

8. Poisson regressions for our baseline models returned significant Pearson good-
ness-of-fit statistics, so we proceeded with negative binomial regressions. We
also use robust standard errors as they slightly reduced the standard errors in
most cases in our models.

9. Due to Hate Crime Statistics Act (HCSA) reporting, this analysis used the bor-
ough/county of New York for all New York City-based acts and dropped the four
remaining boroughs; acts based in St. Louis were also consolidated into St. Louis
County and St. Louis City was dropped from the analysis.


Adamczyk, A., Gruenewald, J., Chermak, S. M., & Freilich, J. D. (2014). The rela-
tionship between hate groups and far-right ideological violence. Journal of
Contemporary Criminal Justice, 30, 310-322.

Atkins, J. K. (2006). The Ku Klux Klan: America’s forgotten terrorists. Law
Enforcement Executive Forum, 5(7), 127-144.

Barnes, A., & Ephross, P. H. (1994). The impact of hate violence on victims:
Emotional and behavioral responses to attacks. Social Work, 39(3), 247-251.

Black, D. (1983). Crime as social control. American Sociological Review, 48, 34-45.
Blalock, H. M. (1967). Toward a theory of minority-group relations. New York, NY:

John Wiley.
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 Criminology, 15, 78-96.

Byers, B. D., & Jones, J. A. (2007). The impact of the terrorist attacks of 9/11 on anti-
Islamic hate crime. Journal of Ethnicity in Criminal Justice, 5(1), 43-56.

Chermak, S. M., Freilich, J. D., Parkin, W., & Lynch, J. P. (2012). American terrorism
and extremist crime data sources and selectivity bias: An investigation focusing
on homicide events committed by far-right extremists. Journal of Quantitative
Criminology, 28, 191-218.

Chermak, S. M., Freilich, J. D., & Shemtob, Z. (2009). Law enforcement training and
the domestic far right. Criminal Justice and Behavior, 36, 1305-1322.

Mills et al. 1217

Chermak, S. M., Freilich, J. D., & Suttmoeller, M. (2013). The organizational dynam-
ics of far-right hate groups in the United States: Comparing violent to non-violent
organizations. Studies in Conflict & Terrorism, 36, 193-218.

Chermak, S. M., & Gruenewald, J. (2015). Laying the foundation for the criminologi-
cal examination of right-wing, left-wing, and Al Qaeda-inspired extremism in the
United States. Terrorism and Political Violence, 27, 133-159.

Cohrs, J. C., & Kessler, T. (2013). Negative stereotypes, prejudice and discrimi-
nation. In A. Golec de Zavala & A. Cichocka (Eds.), Social psychology of
social problems: The intergroup context (pp. 3-29). Chippenham, UK: Palgrave

Corzine, J., Huff-Corzine, L., & Creech, J. C. (1988). The tenant labor market and
lynching in the South: A test of split labor market theory. Sociological Inquiry,
58, 261-278.

Deloughery, K., King, R. D., & Asal, V. (2012). Close cousins or distant relatives?
The relationship between terrorism and hate crimes. Crime & Delinquency, 58,

Disha, I., Cavendish, J. C., & King, R. D. (2011). Historical events and spaces of hate:
Hate crimes against Arabs and Muslims in post-9/11 America. Social Problems,
58, 21-46.

Elias, M. (2012). Sikh temple killer Wade Michael Page radicalized in army (Intelligence
Report, 148). Retrieved from

Ezekiel, R. S. (1995). The racist mind: Portraits of American Neo-Nazis and
Klansmen. New York, NY: Penguin Books.

Federico, C. M. (2013). The social context of racism. In A. Golec de Zavala & A.
Cichocka (Eds.), Social psychology of social problems: The intergroup context
(pp. 30-56). Chippenham, UK: Palgrave Macmillan.

Freilich, J. D., Adamczyk, A., Chermak, S. M., Boyd, K., & Parkin, W. S. (2015).
Investigating the applicability of macro-level criminology theory to terrorism: A
county-level analysis. Journal of Quantitative Criminology, 31, 383-411.

Freilich, J. D., & Chermak, S. M. (2009). Preventing deadly encounters between law
enforcement and American far-rightists. Crime Prevention Studies, 25, 141-172.

Freilich, J. D., & Chermak, S. M. (2013). Hate crimes. Problem-oriented guides for
police, problem-specific guides (Series, No. 72). Office of Community Oriented
Policing Service. Washington, DC: U.S. Department of Justice. Retrieved from

Freilich, J. D., Chermak, S. M., Belli, R., Gruenewald, J., & Parkin, W. S. (2014).
Introducing the United States Extremist Crime Database (ECDB). Terrorism and
Political Violence, 26, 372-384.

Freilich, J. D., Chermak, S. M., & Caspi, D. (2009). Critical events in the life trajec-
tories of domestic extremist white supremacist groups: A case study analysis of
four violent organizations. Criminology & Public Policy, 8, 497-530.

Freilich, J. D., Chermak, S. M., & Simone, J., Jr. (2009). Surveying American state
police agencies about terrorism threats, terrorism sources, and terrorism defini-
tions. Terrorism and Political Violence, 21, 450-475.

1218 Crime & Delinquency 63(10)

Gigerenzer, G. (2004). Dread risk, September 11, and fatal traffic accidents.
Psychological Science, 15, 286-287.

Gigerenzer, G. (2006). Out of the frying pan into the fire: Behavioral reactions to ter-
rorist attacks. Risk Analysis, 26, 347-351.

Gladstone, R., & Zraick, K. (2015, June 18). Charleston, S.C., church shooting: Live
updates. The New York Times. Retrieved from

Goode, E., & Kovaleski, S.F. (2012, August 6). Wisconsin killer fed and was fueled
by hate-driven music. The New York Times. Retrieved from http://www.nytimes.

Grattet, R. (2009). The urban ecology of bias crime: A study of disorganized and
defended neighborhoods. Social Problems, 56(1), 132-150.

Green, D. P., Glaser, J., & Rich, A. (1998). From lynching to gay bashing: The elusive
connection between economic conditions and hate crime. Journal of Personality
and Social Psychology, 75, 82-92.

Green, D. P., McFalls, L. H., & Smith, J. K. (2001). Hate crime: An emergent agenda.
Annual Review of Sociology, 27, 479-504.

Green, D. P., & Rich, A. (1998). White supremacist activity and crossburnings in
North Carolina. Journal of Quantitative Criminology, 14, 263-280.

Green, D. P., Strolovitch, D. Z., & Wong, J. S. (1998). Defended neighborhoods,
integration, and racially motivated crime. American Journal of Sociology, 104,

Gruenewald, J., Chermak, S. M., & Freilich, J. D. (2013a). Distinguishing “loner”
attacks from other domestic extremist violence. Criminology & Public Policy,
12, 1-27.

Gruenewald, J., Chermak, S. M., & Freilich, J. D. (2013b). Far-right lone wolf terror-
ism in the United States. Studies in Conflict & Terrorism, 36, 1005-1024.

Gruenewald, J. A., & Pridemore, W. A. (2012). A comparison of ideologically-
motivated homicides from the new Extremist Crime Database and homicides from
the Supplementary Homicides Reports using multiple imputation by chained equa-
tions to handle missing values. Journal of Quantitative Criminology, 28, 141-162.

Gurr, T. R. (1968). Psychological factors in civil violence. World Politics, 20,

Hamm, M. (1993). American skinheads: The criminology and control of hate crime.
Westport, CT: Greenwood.

Herek, G. M., Cogan, J. C., & Gillis, J. R. (2002). Victim experiences in hate crimes
based on sexual orientation. Journal of Social Issues, 58, 319-339.

Hewitt, C. (2000). Patterns of American terrorism 1955-1998: An historical perspec-
tive on terrorism-related fatalities. Terrorism and Political Violence, 12, 1-14.

Hoffman, B. (1998). Inside terrorism. New York, NY: Columbia University Press.
Iganski, P., & Lagou, S. (2009). How hate crimes hurt more: Evidence from the

British crime survey. In B. Perry & P. Iganski (Eds.), Hate crimes, volume 2: The
consequences of hate crime (pp. 107-122). Westport. CT: Praeger.

Mills et al. 1219

Jacobs, D., & Wood, K. (1999). Interracial conflict and interracial homicide: Do polit-
ical and economic rivalries explain white killings of blacks or black killings of
whites? American Journal of Sociology, 105, 157-190.

Jenness, V., & Grattet, R. (2004). Making hate a crime: From social movement to law
enforcement. New York, NY: Russell Sage Foundation.

Kerodal, A., Freilich, J. D., Chermak, S. M., & Suttmoeller, M. J. (2015). A test of
Sprinzak’s split delegitimization’s theory of the life course of far-right organi-
zational behavior. International Journal of Comparative and Applied Criminal
Justice, 39, 307-329.

King, M., Noor, H., & Taylor, D. M. (2011). Normative support for terrorism: The
attitudes and beliefs of immediate relatives of Jema’ah Islamiyah members.
Studies in Conflict & Terrorism, 34, 402-417.

King, R. D. (2007). The context of minority group threat: Race, institutions, and com-
plying with hate crime law. Law & Society Review, 41, 189-224.

King, R. D., & Brustein, W. I. (2006). A political threat model of intergroup violence:
Jews in pre-World War II Germany. Criminology, 44, 867-891.

King, R. D., Messner, S. F., & Baller, R. D. (2009). Contemporary hate crimes, law
enforcement, and the legacy of racial violence. American Sociological Review,
74, 291-315.

King, R. D., & Sutton, G. M. (2013). High times for hate crimes: Explaining the tem-
poral clustering of hate-motivated offending. Criminology, 51, 871-894.

Krueger, A. B., & Malečková, J. (2002, June 24). Does poverty cause crime? The New
Republic, 226(24), 27-33.

Krueger, A. B., & Malečková, J. (2003). Education, poverty and terrorism: Is there a
causal connection? Journal of Economic Perspectives, 17(4), 27-33.

LaFree, G., & Bersani, B. E. (2014). County-level correlates of terrorist attacks in the
United States. Criminology & Public Policy, 13, 455-481.

LaFree, G., & Dugan, L. (2004). How does studying terrorism compare to studying
crime? In M. DeFlem (Ed.), Terrorism and counterterrorism: Criminological
perspectives (pp. 53-74). Amsterdam, The Netherlands: Elsevier.

LaFree, G., & Dugan, L. (2007). Introducing the global terrorism database. Political
Violence and Terrorism, 19, 181-204.

LaFree, G., Dugan, L., & Korte, R. (2009). The impact of British counter terrorist
strategies on political violence in Northern Ireland: Comparing deterrence and
backlash models. Criminology, 47, 501-530.

Law, R. (2009). Terrorism: A history. Malden, MA: Polity Press.
Levin, B. (2012, August 6). Lone wolf killers are often a combination of hatreds and

frustrations. The Huffington Post. Retrieved from http://www.huffingtonpost.

Levin, J. (2013, May 1). Domestic terrorism: Myths and realities. The Huffington
Post. Retrieved from

Levin, J., & McDevitt, J. (2002). Hate crimes revisited: America’s war on those who
are different. Boulder, CO: Westview Press.

1220 Crime & Delinquency 63(10)

Lickel, B., Miller, N., Stenstrom, D. M., Denson, T. F., & Schmader, T. (2006).
Vicarious retribution: The role of collective blame in intergroup aggression.
Personality and Social Psychology Review, 10, 372-390.

Lim, H. A. (2009). Beyond the immediate victim: Understanding hate crimes as mes-
sage crimes. In B. Perry & P. Iganski (Eds.), Hate crimes, volume 2: The conse-
quences of hate crime (pp. 107-122). Westport. CT: Praeger.

Lipset, S. M., & Raab, E. (1977). The politics of unreason: Right-wing extremism in
America. New York, NY: Harper Collins.

Louis, W. R., & Taylor, D. M. (2002). Understanding the September 11 terrorist attack
on America: The role of intergroup theories of normative influence. Analyses of
Social Issues and Public Policy, 2, 87-100.

Lyons, C. J. (2007). Community (dis)organization and racially motivated crime.
American Journal of Sociology, 113, 815-863.

Lyons, P. A., Kenworthy, J. B., & Popan, J. R. (2010). Ingroup identification and
group-level narcissism as predictors of U.S. citizens’ attitudes and behavior
toward Arab immigrants. Personality and Social Psychology Bulletin, 36,

McArdle, S. C., Rosoff, H., & John, R. S. (2012). The dynamics of evolving beliefs,
concerns emotions, and behavioral avoidance following 9/11: A longitudinal
analysis of representative archival samples. Risk Analysis, 32, 744-761.

McCauley, C., & Moskalenko, S. (2008). Mechanisms of political radicalization:
Pathways toward terrorism. Terrorism and Political Violence, 20, 415-433.

McCauley, C., & Moskalenko, S. (2011). Friction: How radicalization happens to
them and us. New York, NY: Oxford University Press.

McDevitt, J., Balboni, J., Garcia, L., & Gu, J. (2001). Consequences for victims:
A comparison of bias- and non-bias-motivated assaults. American Behavioral
Scientist, 45, 697-713.

McDevitt, J., Levin, J., & Bennet, S. (2002). Hate crime offenders: An expanded
typology. Journal of Social Issues, 58, 303-317.

Merton, R. K. (1938). Social structure and anomie. American Sociological Review,
3(5), 672-682.

Messner, S. F., McHugh, S., & Felson, R. B. (2004). The distinctive characteristics of
assaults motivated by bias. Criminology, 42, 585-615.

Messner, S. F., & Rosenfeld, R. (2007). Crime and the American dream. New York,
NY: Wadsworth Publishing Co.

Michael, G. (2003). The far right and terrorism. In G. Michael (Ed.), Confronting right-
wing extremism & terrorism in the USA (pp. 93-128). New York, NY: Routledge.

Minnesota Population Center. (2011a). U.S. Census Bureau: Persons by Hispanic or
Latino origin by race 1990 (National Historical Geographic Information System:
Version 2.0). Minneapolis: University of Minnesota. Available from http://www.

Minnesota Population Center. (2011b). U.S. Census Bureau: Persons by Hispanic or
Latino origin by race 2000 (National Historical Geographic Information System:
Version 2.0). Minneapolis: University of Minnesota. Retrieved from http://www.

Mills et al. 1221

Minnesota Population Center. (2011c). U.S. Census Bureau: Persons by Hispanic or
Latino origin by race 2010 (National Historical Geographic Information System:
Version 2.0). Minneapolis: University of Minnesota. Retrieved from http://www.

Murphy, K. (2012, August 9). Sikh temple shooter seems to have followed “lone wolf”
path. The Los Angeles Times. Retrieved from

National Consortium for the Study of Terrorism and Responses to Terrorism. (2014).
Global terrorism database. Retrieved from

Olzak, S. (1989). Labor unrest, immigration, and ethnic conflict in urban America,
1880-1914. American Journal of Sociology, 94, 1303-1333.

Olzak, S. (1990). The political context of competition: Lynching and urban racial
violence, 1882-1914. Social Forces, 69, 395-421.

Olzak, S., Shanahan, S., & McEneaney, E. H. (1996). Poverty, segregation, and race
riots: 1960–1993. American Sociological Review, 1996, 590-613.

Phillips, N. D. (2009). The prosecution of hate crimes: The limitations of the hate
crime typology. Journal of Interpersonal Violence, 24, 883-905.

Pinderhughes, H. (1993). The anatomy of racially motivated violence in New
York City: A case study of youth in southern Brooklyn. Social Problems, 40,

Prager, F., Beeler Assay, G. R., Lee, B., & von Winterfeldt, D. (2011). Exploring
reductions in London underground passenger journeys following the July 2005
bombings. Risk Analysis, 31, 773-786.

Pridemore, W. A., & Freilich, J. D. (2006). A test of recent subcultural expla-
nations of White violence in the United States. Journal of Criminal Justice,
34(1), 1-16.

Robles, F. (2015, June 20). Dylann Roof photos and a manifesto are posted on web-
site. The New York Times. Retrieved from

Rubin, G. J., Brewin, C. R., Greenberg, N., Simpson, J., & Wessely, S. (2005).
Psychological and behavioural reactions to the bombings in London on 7 July
2005: Cross sectional survey of a representative sample of Londoners. British
Medical Journal, 331, 606-611.

Senechal de la Roche, R. (1996). Collective violence as social control. Sociological
Forum, 11, 97-128.

Shanahan, S., & Olzak, S. (1999). The effects of immigrant diversity and ethnic com-
petition on collective conflict in urban America: An assessment of two moments
of mass migration, 1869-1924 and 1965-1993. Journal of American Ethnic
History, 18(3), 40-64.

Shimamoto, E. (2004). Rethinking hate crime in the age of terror. University of
Missouri-Kansas City Law Review, 72 (2003-2004), 829-844.

Simi, P., & Futrell, R. (2010). American Swastika: Inside the White power move-
ment’s hidden spaces of hate. New York, NY: Rowman & Littlefield.

Soule, S. (1992). Populism and black lynching in Georgia, 1890-1900. Social Forces,
71, 431-449.

1222 Crime & Delinquency 63(10)

Suttmoeller, M., Chermak, S. M., & Freilich, J. D. (2015). The influence of external
and internal correlates on the organizational death of domestic far-right extremist
groups. Studies in Conflict & Terrorism, 38, 734-758.

Suttmoeller, M., Gruenewald, J., Chermak, S. M., & Freilich, J. D. (2013). Killed in
the line of duty: Comparing police homicides committed by far-right extremists
to all police homicides. Law Enforcement Executive Forum, 13(1), 45-64.

Tita, G., & Cohen, J. (2004). Measuring spatial diffusion of shots fired activity across
city neighborhoods. In M. F. Goodchild & D. G. Jannelle (Eds.), Spatially inte-
grated social science (pp. 171-204). New York, NY: Oxford University Press.

Tolnay, S. E., & Beck, E. M. (1995). A festival of violence: An analysis of southern
lynchings, 1882-1930. Urbana: University of Illinois Press.

Tolnay, S. E., Deane, G., & Beck, E. M. (1996). Vicarious violence: Spatial effects
in southern lynchings, 1890-1919. American Journal of Sociology, 102, 788-815.

Unprosecuted hate crimes. (2012, August 14). The New York Times. Retrieved from


U.S. Bureau of Labor Statistics. (1990). Labor force data by county, 1990 annual
averages. Retrieved from

U.S. Bureau of Labor Statistics. (2000). Labor force data by county, 2000 annual
averages. Retrieved from

U.S. Bureau of Labor Statistics. (2010). Labor force data by county, 2010 annual
averages. Retrieved from

U.S. Census Bureau. (1990). Social, economic, and housing statistics division: Poverty.
Retrieved from

U.S. Census Bureau. (2000). Social, economic, and housing statistics division: Poverty.
Retrieved from

U.S. Census Bureau. (2001, December 7). Mapping the census: The geography of
U.S. diversity. Retrieved from

U.S. Census Bureau. (2010). Small area income and poverty estimates: State and
county. Retrieved from

U.S. Department of Justice. (n.d.). Community Relations Service. Available from

U.S. Department of Justice, Federal Bureau of Investigation. (2011a). Hate crime
statistics act, 2010. Retrieved from

U.S. Department of Justice, Federal Bureau of Investigation. (2011b). Hate crime
statistics act, 2010. Retrieved from

U.S. Department of Justice, Federal Bureau of Investigation. (2014). Uniform Crime
Reporting Program Data [United States]: Hate Crime Data, 1992-2012 (exclud-
ing 1993) [Record-Type Files]. ICPSR35086-v1. Ann Arbor, MI: Inter-university
Consortium for Political and Social Research [distributor]. Retrieved from http://

Mills et al. 1223

Author Biographies

Colleen E. Mills is a doctoral student at John Jay College of Criminal Justice/The
City University of New York, The Graduate Center. She is a Project Manager for the
Extremist Crime Database (ECDB) at John Jay College. Her research focuses on hate
crime, far-right extremism, racism, and group conflict.

Joshua D. Freilich is a member of the Criminal Justice Department and the Criminal
Justice PhD Program at John Jay College. He is the Creator and co-Director of the
United States Extremist Crime Database (ECDB), an open source relational database
of violent and financial crimes committed by political extremists in the U.S. Professor
Freilich’s research has been funded by the Department of Homeland Security (DHS)
and the National Institute of Justice (NIJ). His research focuses on the causes of and
responses to terrorism, measurement issues, and criminology theory, especially envi-
ronmental criminology and crime prevention.

Steven M. Chermak is a professor in the School of Criminal Justice at the Michigan
State University. He studies domestic terrorism, media coverage of crime and justice
issues, and the effectiveness of specific policing strategies. Recent publications have
appeared in Terrorism and Political Violence, Crime and Delinquency, and the Journal
of Quantitative Criminology.


Richard Delgado*
Jean Stefancic*’


Hate speech on the Internet has become a source of concern
among many in the civil rights community, i African Americans,
Latinos, women, Jews, gays, and Muslims report a rise in
vituperative messages in this medium—some aimed at them, others
about them.2 At the same time, organizations like the Southern
Poverty Law Center and the FBI t h a t compile figures on hate crime,
which is more easily quantified, find t h a t it too is on the rise.^ The
subjective experiences of those on the receiving end of both forms of
hate are, thus, probably valid and worthy of attention.

* John J. Sparkman Chair of Law, University of Alabama School of Law.
** Professor of Law and Clement Research Affiliate, University of

Alabama School of Law.
1. See, e.g., JEREMY WALDRON, THE HARM IN HATE SPEECH 37, 89, 149, 176

(2012); Shannon Gilreath, The Internet and Inequality: A Comment on the NSA
Spying Scandal, 49 WAKE FOREST L. REV. (forthcoming 2014); Alexander Tsesis,
Hate in Cyberspace: Regulating Hate Speech on the Internet, 38 SAN DIEGO L .
REV. 817, 818-20 (2001) [hereinafter Tsesis, Hate in Cyberspace]; Alexander
Tsesis, Inflammatory Speech: Offense Versus Incitement, 97 MiNN. L. REV. 1145,
1167-73 (2013) [hereinafter Tsesis, Inflammatory Speech].

2. See About Us, TEACHING TOLERANCE, (last
visited Jan. 25, 2014) (describing a task force to counter Internet-based hate
and intolerance); Intelligence Project, S. POVERTY L. CENTER, (last visited Jan. 25, 2014);
see also Grace Gedye, Op-Ed., The Rise of Sexism and Misogyny in a Facebook
Era, SEATTLE TIMES (Apr. 17, 2013),
/2020793295_gracegedyeopedxml.html (noting that “one person can start a
misogynist Facebook group, type an insulting comment or make a sexist status
update and amass support from the Facebook community in the form of ‘likes'”
and noting a subtle change in the portrayal of women in the media); 163 and
Counting. . . Hate Groups Find a Home on the Net, S. POVERTY L. CENTER

3. See sources cited supra note 2; Mark Potok, DOJ Study: More than
250,000 Hate Crimes a Year, Most Unreported, S. POVERTY L. CENTER (Mar. 26,
hate-crimes-a-year-a-third-never-reported/ (analyzing the rise in hate crime and
why FBI statistics are lower than they should be due to unreported hate
crimes); see also Hate Crime Statistics, FBI (2011),
us/cjis/ucr/hate-crime/2011/tables/table-l (providing statistics for a recent year).



Although we have written a good deal on hate speech,” we have
until now confined our attention mainly to its written or printed
form. 5 The rise in the Internet version, however, calls for an
examination focused on it alone.^

Hate in cyberspace is part of a broader pattern of antisocial
behavior that finds a home in that realm.” Most observers believe
that the advent of the Internet has witnessed an increase in certain
types of criminal behavior, including identity theft,^ threats,^
financial and consumer fraud and scams,lo hoaxes and pranks,ii and
hackingi2—and others that are merely despicable. By this we mean
behavior that decreases trust, weakens social bonds, or erodes
quality of Under this heading, we would place not only hate

THAT WOUND 1-2 (2004); Richard Delgado & Jean Stefancic, Four Observations
About Hate Speech, 44 WAKE FOREST L. REV. 353, 354-55 (2009); Richard
Delgado, Words That Wound: A Tort Action for Racial Insults, Epithets, and
Name-Calling, 17 HARV. C . R . – C . L . L . REV. 133, 134-36 (1982).

Levmore & Martha C. Nussbaum eds., 2010) is a notable exception. See
DELGADO & STEFANCIC, supra note 4, at 123-38 (devoting a chapter to “Hate in

6. See Gilreath, supra note 1; Tsesis, Hate in Cyberspace, supra note 1, at
818-19; Tsesis, Inflammatory Speech, supra note 1, at 1166-68 (noting that
Internet communications can rise to the level of imminent harm).

7. See DELGADO & STEFANCIC, supra note 4, at 123 (describing cyberspace
as “a worldwide connection of computers over telephone, cable, and fiber optics
lines . . . [that] allow[s] anyone with access and a computer to send and receive
data” and enumerating some of the most common modes of communication,
including e-mail, chat rooms or newsgroups, and the World Wide Web); Danielle
Keats Citron, Cyber Civil Rights, 89 B.U. L. REV. 61, 62-63 (2009). The present
Article considers a few additional cyberavenues, including social media, such as
Twitter and YouTube, and instant messaging. See infra Part II.

8. For information on identity theft by means of cyberhacking and scams,
see infra note 12 and accompanying text, and Nathaniel Popper & Somini
Sengupta, U.S. Says Ring Stole 160 Million Credit Card Numbers, N.Y. TIMES,
July 26, 2013, at B7.

9. See, e.g., Robbie Brown, 140 Characters Spell Charges and Jail, N.Y.
TIMES, July 3, 2013, at A15 (noting that “Twitter makes it easier for people to
say things they don’t mean seriously. . . . If I say online that I want to kill
Obama, it’s far harder to assess how serious I am than if I’m standing across
the street from the White House and I have a gun”).

10. See infra note 12 and accompanying text (discussing a common form of
computer-assisted fraud).

11. See Amy Chozick & Nicole Perlroth, Twitter Speaks, Markets Listen,
and Fears Rise, N.Y. TIMES, Apr. 29, 2013, at Al (noting that Twitter is rife
with flippant or mean-spirited messages, including hoax material that can
cause serious harm).

12. On hacking, see Danielle Keats Citron, Civil Rights in Our Information
5, at 31, 31-33, and Nathaniel Popper, Wall Street’s Exposure to Hacking Laid
Bare, N.Y. TIMES, July 26, 2013, at Bl.

13. That is, behavior that is not clearly illegal, like sending someone an
anonymous message saying that the person’s boyfriend is fiirting with another


but student cheating,!^ plagiarism,!^ “cyberbuUjdng,”!”
“cyberstalking,”!^ snooping,i9 and circulation of damaging rumors or

person, or posting a negative comment in order to bias other readers. See, e.g.,
Dominique Brossard & Dietram A. Scheufele, This Story Stinks, N.Y. TIMES,
Mar. 3, 2013, at SR5 (describing an experiment that found a “nasty effect,” in
which a single prearranged negative comment “was enough to make study
participants think the downside” of a passage they had been reading).

14. See, e.g., Brian Leiter, Penn Law Student, Anthony Ciolli, Admits to
Running Prelaw Discussion Board Awash in Racist, Anti-Semitic, Sexist Abuse,
LEITER REP.: PHIL. BLOG (Mar. 11, 2005, 6:12 PM), http://leiterreports (noting hundreds of entries
on a law student’s blog containing grossly offensive language about African
Americans and women).

15. For an example of cheating carried out via internet connections, see
Richard Pérez-Peña, Studies Find More Students Cheating, with High Achievers
No Exception, N.Y. TIMES, Sept. 8, 2012, at A14 (noting that “Internet access
has made cheating easier, enabling students to connect instantly with answers,
friends to consult and works to plagiarize. And generations of research has
shown that a major factor in unethical behavior is simply how easy or hard it

6 (2011), available at
Online-Learning (discussing the rise of plagiarism coinciding with the
advent of technology).

17. See, e.g.. Editorial Board, Hate Speech on Facebook, N.Y. TIMES, May
31, 2013, at A20 (noting the proliferation of material glorifying violence against
women and observing that “[t]he company’s slow response may be indicative of
a deeper problem in technology and Internet-based companies—most of them
are primarily run by men”); Tanzina Vega, Facebook Promises to Address Hate
Speech, N.Y. TIMES, May 29, 2013, at BI; Tamara Lush, Suicide Victim Had
Been Bullied for Nearly a Year, TUSCALOOSA NEWS (Sept. 13, 2013), (noting how
prolonged bull5dng can prompt a sensitive child to commit suicide); see also
Citron, supra note 12, at 31, 33 (discussing “cybermobs” and bulljdng on the
Internet); Martha C. Nussbaum, Objectification and Internet Misogyny, in THE
T^ll (noting that that medium is rife with messages and images objectif5ring
women, even historical figures); Daniel J. Solove, Speech, Privacy, and
REPUTATION, supra note 5, at 15, 15, 17 (“Now, people are judged out of context
based on information fragments found online. The amount of these fragments
is vastly increasing a n d . . . becoming more personal and potentially
discreditable.”); Jessica S. Henry, Bias-Based Cyberbullying; The Next Hate
Crime Frontier?, 49 CRIM. L . BULL. 481, 481 (2013).

18. For more information on cyberstalking, see, for example, James
Lasdun, “/ Will Ruin Him”; How It Feels to Be Stalked, CHRON. HIGHER EDUC.
(Jan. 21, 2013), (describing
how the writer “found [him]self, to [his] surprise, the victim of a campaign of
malicious e-mail stalking and online defamation by a former . . . student” who
was disappointed when the instructor did not back publication of her novel and
resisted her amorous overtures and how “a deluge of emails” accusing him of
fabricated acts nearly destroyed his life and career).

19. See Somini Sengupta, Digital Tools to Curb Snooping, N.Y. TIMES, July
18, 2013, at BI.


innuendo.2o We would also place in this category “swarming”—
posting negative reviews of a product or book with the aim of
decreasing its sales^i—and “revenge porn,” in which rejected lovers
post degrading photos of their former boyfriends or girlfriends
online.22 Many such deeds, the criminal and the noncriminal alike,
proceed in secrecy or stealth. Because the Internet makes it easy to
act and speak without self-identification, these acts are easy to carry
out without fear of discovery.23 This ease of commission probably
contributes to their increased incidence and virulence, as weU.24

20. See Saul Levmore & Martha C. Nussbaum, Introduction, in THE
(discussing slurs, negative reviews, and slander on the Internet); Brossard &
Scheufele, supra note 13 (noting that our “emerging online media landscape has
created a new public forum without the traditional social norms and self-
regulation that typically govern our in-person exchanges—and that medium,
increasingly, shapes both what we know and what we think we know”); see also
Cass R. Sunstein, Believing False Rumors, in THE OFFENSIVE INTERNET: SPEECH,
PRIVACY, AND REPUTATION, supra note 5, at 91, 91.

21. See David Streitfeld, Swarming a Book Online: Amazon Becomes a
Battlefield as Reviewers Go on the Attack, N.Y. TIMES, Jan. 21, 2013, at Bl;
David Streitfeld, Why Web Reviewers Make Up Bad Things, BITS (July 15, 2013,
8:30 AM),
up-bad-things/?_r=O (noting that “the problem is much bigger than a few
malicious operators”).

22. See Jill Filipovic, “Revenge Porn” Is About Degrading Women Sexually
and Professionally, What Does It Say About Society that Websites Where Angry
Men Shame Their Ex-lovers Are Thriving?, THEGUARDIAN (Jan. 28, 2013, 5:23
degrades-women; see also Anupam Chander, Youthful Indiscretion in an
supra note 5, at 124, 124-26; Lorelei Laird, Striking Back at Revenge Porn:
Victims Are Taking On Websites for Posting Photos They Didn’t Consent to, 99
A.B.A. J. 45, 46 (2013); Mary Elizabeth Williams, The War on 12-Year-Old
Girls: An Epidemic of High-Profile Trolling Is a Testament to How Pathological
Misogyny Is—and How Early It Begins, SALON (Oct. 19, 2012, 3:05 PM),
/2012/10/19/the_war_on_12_year_old_girls/ (describing aggressive males who
importune young girls to supply a picture of themselves, partially disrobed, and
then place it on the Internet).

23. See Helen Norton, Setting the Tipping Point for Disclosing the Identity
of Anonymous On-Line Speakers, 49 WAKE FOREST. L. REV. (forthcoming 2014)
(discussing the problem of anonymous Internet speech). The Internet’s secrecy,
coupled with a herd instinct, sets the stage for conspiracies and conspiratorial
thinking. See Marc Fisher, The Term They Can’t Come to Terms with:
Reelection Hasn’t Dimmed the Counterfactual Theories of Obama’s Most Fervent
Foes, N.Y. TIMES, Jan. 15, 2013, at Cl (noting that the “Internet and cable TV
have made it far easier to connect with like-minded souls”).

24. See Citron, supra note 12, at 36-37; Micah S. Thompson et al.. The
Consequences of Communicating Social Stereotypes, 36 J. EXPERIMENTAL SOC.
PYSCHOL. 567, 593-96 (2000) (noting that repeating stereotypes, as well as
hearing them often, reinforces them, especially if they appear without


None of these behaviors, of course, is entirely new; indeed, some
of them were common even before the Internet came into wide use
in the mid-1990s.25 But these behaviors have become easier, more
cost free, and more ubiquitous since the medium became popular.^e
More people now misbehave online, and others, seeing how easy it
is, follow suit. Unlike an oral remark, which disappears as soon as
it is spoken, or a graffito, which will be erased or painted over
sooner or later, much material posted on the Internet will remain
there indefinitely, becoming “a permanent or semi-permanent part
of the visible environment in which our lives, and the lives of
vulnerable minorities, have to be lived.” ‘̂? If the hate message “goes
viral,” it may attract millions of viewers and remain in cyberspace,
perhaps forever.^s

The two of us are not Luddites who resist technological
innovation. We use e-mail daily to correspond with colleagues and
friends and have employed computerized research tools since their
very beginning. One of us, in a former life, was a law librarian who
used electronic search tools every day.̂ s We recognize that the
Internet has done a great deal of good, perhaps enough to
counterbalance the drawbacks just In addition to
making communication and research easier, the Internet enahles
millions of people to shop, bank, market products, keep records, find
information, and amuse themselves cheaply and efficiently.^i The
Internet enables the police to notify citizens of neighborhood

25. See Solove, supra note 17, at 16-17.
26. See supra notes 1-3 and accompanying text (noting that exact

quantification is difficult).
27. WALDRON, supra note 1, at 37.
28. Not only will the remark remain potentially forever, injuring the

feelings of the target, stereotypes engrain themselves in the minds of listeners
and readers who hear them often, so that they begin to seem true. Tsesis, Hate
in Cyberspace, supra note 1, at 847-53.

29. In a former life, Jean Stefancic served as a librarian for the Sierra Club
and the University of San Francisco School of Law.

30. See Levmore & Nussbaum, supra note 20, at 1-3 (positing that the
overall balance may well be positive); see also Alexander Tsesis, Challenges to
Privacy in Cyber Transactions, 49 WAKE FOREST L. REV. (forthcoming 2014)
(noting the many positive features of the Internet).

31. See Tsesis, Hate in Cyberspace, supra note 1 (noting that the Internet
“has made available . . . educational opportunities, increased citizens’ role in
government, given greater access to health related resources, made available
library catalogues, and allowed people to find employment far from their homes.
In those ways, it has been an invaluable tool for thriving democracies,” but
noting that it has also served as a tool for division and hate); Natasha Singer,
Under Code, Apps Would Disclose Collection of Data, N.Y. TIMES, July 26, 2013,
at B7 (describing a proposal to alert consumers “to decide at a glance whether
[certain] apps are good for them”).

32. See, e.g., Winnie Hu & J. David Goodman, Wake-Up Call for New
Yorkers as Police Seek Abducted Boy, N.Y. TIMES, July 18, 2013, at Al


and people seeking answers to difficult questions to secure them
through crowdsourcing—combining the knowledge of a host of
readers.33 Social media allow people to follow developments in their
friends’ lives and find companions with whom they share common
interests.34 The Internet, in short, has brought both good and bad.
Although we focus on one problematic use of that medium, we do not
suggest that it lacks impressive redeeming features.35

With those provisos in mind, let us consider the rise of hate
speech on the Internet.


A Naming and Countering a Form of Socially Pernicious Behavior
As mentioned, agencies that monitor hate speech and crime

report an upsurge in hate messages and sites on the Internet and
warn that this development may erode public discourse while
exposing minorities, gays, women, and other disempowered groups
to ridicule and contempt.36 These observers are not crying wolf
The Internet is rife with hate speech, including e-mails, chat groups,
blogs, and websites touting white supremacy and asserting the

(describing computerized robocalls that arrived too early for many sleeping New

33. For some background on crowdsourcing as a means of solving problems,
DRIVING THE FUTURE OF BUSINESS (2009). Crowdsourcing may also be a force for
evil. See Jay Caspian Kang, Crowd-Sourcing a Smear: When an Insidious
Rumor Went Viral in the Aftermath of the Boston Marathon Bombing, It Laid
Bare the Dysfunctional Codependence Between New and Old Media, N.Y. TIMES
MAG., July 28, 2013, at 36.

34. See, e.g.. Online Dating, Life & Style, THEGUARDIAN, (last visited Jan. 26, 2014)
(featuring a blog in the Guardian’s “Life & Style” section devoted to the subject
of online dating).

REPUTATION, supra note 5 (noting, in practically every essay, that the Internet
has brought both good and bad).

36. See, for example, any issue of the Southern Poverty Law Center’s
journal Intelligence Report for coverage of the center’s efforts in this area. See
also Abraham H. Foxman & Christopher Wolf, Letter to the Editor, Internet
Hate Speech, N.Y. TIMES (June 3, 2013),
/06/04/opinion/internet-hate-speech.html?_r=0 (noting that the “writers,
national director and civil rights chairman, respectively, of the Anti-Defamation
League” are writing a book on this issue). For information on a recent meeting
of the group, see Jeffrey Rosen, The Delete Squad: Google, Twitter, Facebook,
and the New Global Battle over the Future of Free Speech, NEW REPUBLIC (Apr.
29, 2013), (noting that “the Internet
giants are grappling with the challenge of enforcing their community guidelines
for free speech” in the face of demands, especially from Europe, for what the
author considers “pohtical censorship” endangering free speech). But see
Citron, supra note 12, at 31 (noting that the Internet is rife with sexist and
racist material and urging an application of civil rights law to this area).


inferiority of gays, blacks, Jews, Latinos, Muslims, foreigners, and

1. An Early, Hard-Fought Victory
The development of hate messages and sites on the Internet

compels attention, for the battle against hate speech—which sprang
up in the 1980s (i.e., during the pre-Internet era) with the
publication of a number of key books and articles^» followed by the
enactment of many university conduct codes and state and federal
laws penalizing it—had seemed largely won.̂ ^ Although the
American Civil Liberties Union (“ACLU”) and a few other First
Amendment absolutists defended hate speakers in the early years,
their advocacy in recent times has been muted, perhaps because the
need for it has abated.^o Hardly anyone today maintains that hate
speech is permissible or a minor inconvenience that minorities
should be prepared to tolerate as one of life’s ordinary burdens.*i

The norm against hate speech, in other words, is now firmly
established. This of course does not mean that the occasional
incident of gay bashing or derogation of minorities never appears.42
Rather, when it does, it receives little public support. In short, the
battle against hate speech had seemingly been won.

37. See infra Part II.
38. See, e.g.. Delgado, supra note 4; Mari Matsuda, Public Response to

Racist Speech: Considering the Victim’s Story, 87 MiCH. L. REV. 2320, 2320
(1989); see also WALDRON, supra note 1 (drawing on these earlier works).

REGULATION 5-7 (2005) (describing the effort by many universities and colleges
to forbid hate speech).

40. Compare Nadine Strossen, Regulating Racist Speech on Campus: A
Modest Proposal?, 1990 DUKE L.J. 484, 523-61 (defending the free speech
position), with Nadine Strossen, Professor, N.Y. Law Sch., & Immediate Past
President, Am. Civil Liberties Union, Address at the Seattle University Law
Review Symposium: Obscenity & Indecency Law: Why Howl Is Still Silenced
(Apr. 5, 2013) (reminding her audience that eternal vigilance is the price of
liberty). For examples of free speech originalism, see Levmore & Nussbaum,
supra note 20, at 7 (discussing authors who urge broad protection for Internet
speech) and Alexander Meiklejohn, The First Amendment Is an Absolute, 1961
SUP. CT. REV. 245, 245-66.

41. See WALDRON, supra note 1, at 2-10, 33 (arguing that this would be an
unfair burden).

42. See id. (noting that the former, at least, is all too common).


2. Two Exceptions to This Development: Courts and

a. The First Exception: The Lower Judiciary

The first is the judiciary, especially the lower tier. Although the
U.S. Supreme Court upheld a cross-burning statute (forbidding one
spectacular and historically rooted form of hate speech),44 lower
courts still regularly strike down campus speech codes, especially
ones that strike them as vague and overbroad.^s On the level of
social (not legal) norms—what one can say without incurring
disapproval—the battle is over. Hate speech is decisively out of
favor; any instance of it, whether in a classroom, speech, television
show, or newspaper column, brings immediate condemnation.^e
(Indeed, the speaker or writer often apologizes and promises not to
do it again.47) As recently as thirty years ago, this was not the
case.48 Now, it is. In short, this is an area, like several others,49
where the legal system lags behind the popular consensus.

b. The Second Exception: The Internet

The second area, the Internet, is our subject in this Article.
Here we find an even more striking exception to the dominant social
norm. One encounters a host of websites with users inveighing
against racial and sexual minorities, women, liberals, foreigners,
Jews, Muslims, and recent immigrants as well as anonymous

43. Namely, that the war against hate speech has been largely won and
that a new norm has entered the public arena.

44. See Virginia v. Black, 538 U.S. 343, 347-48 (2003).
45. See UWM Post, Inc. v. Bd. of Regents of Univ. of Wis. Sys., 774 F. Supp.

1163, 1165-67, 1181 (E.D. Wis. 1991) (invalidating a student conduct code that
prohibited hate speech at the University of Wisconsin); Doe v. Univ. of Mich.,
721 F. Supp. 852, 856-57, 867-69 (E.D. Mich. 1989) (striking down a similar
code at the University of Michigan); GoULD, supra note 39, at 123-48
(describing a number of lower court actions).

46. See, e.g., Judy Faber, CBS Fires Don Imus over Racial Slur, CBS NEWS
(Apr. 12, 2007, 7:53 AM),

47. See Erica Ritz, “To All You Black Mother****ers”; Newspaper Apologizes
After Racist Rant Gets Printed in Anonymous Comments Section, BLAZE (July
23, 2013, 2:32 PM),
comments-section/ (recounting a newspaper’s apology for printing an
anonymous letter to the editor containing offensive material); see also David
Carr, Networks Condemn Remarks by Imus, N.Y. TIMES (Apr. 7, 2007),

48. See supra notes 38-39 and accompanying text.
49. Until recently, gay rights, including the right to marry, was another.


tweets, e-mails, and chat groups circulating harshly critical
messages about their favorite targets, and much venom.̂ o

This anomaly—^hate speech proliferating in one area (the
Internet) while practically abolished everywhere else (ordinary
life)—invites attention, particularly since Internet use is advancing
rapidly, indeed threatening to overtake printed material.^i Earlier,
one of us identified a dismal “Law of Racial Thermodynamics”—
racism cannot either be created or destroyed.52 Eliminating it from
one area (say, education and pupil school assignments) merely
causes it to rise up in another (say, residential housing preferences
and white flight).^3 We may be witnessing an instance of this with
the Internet. Just as members of society have learned to speak and
treat each other respectfully in person and in writing, the Internet
has opened up new avenues for derogation and spite.^^

After reviewing this development in Part II, we offer in Part III
structural reasons why we believe it may be taking place. We also
explain why it is unlikely to give way in the face of the kind of firm
and consistent shaming that over the past twenty-five years or so
created a new norm against hate speech of the spoken or written
kind. In particular, we posit in Part IV that new scholarship and
public resistance will be necessary to counter the Internet variety, in
part because legislation does not seem a promising avenue, at least
any time soon.̂ s By the same token, organizations that run the

50. See infra Part II; see also Citron, supra note 12, at 33 (noting that
concerted “Google bombing” by sadists can elevate the location of derogatory
material accusing a victim of character deficiencies); Brian Leiter, Cleaning
Cyber-Cesspools: Google and Free Speech, in THE OFFENSIVE INTERNET: SPEECH,
PRIVACY, AND REPUTATION, supra note 5, at 155, 155 (noting that several search
engines compound this problem when they arrange sites in response to search
requests or number of links). Much Internet speech is anonymous (i.e., devoid
of self-identification). See Saul Levmore, The Internet’s Anonymity Problem, in
50, 50 (noting that defamation that is anonymous is much more damaging than
the kind that emanates from a soapbox or newspaper article).

51. Many believe that books will soon be supplanted by e-books, blogs, self-
published writings, and other online materials. See, e.g.. Do Books Have a
Future: An Interview with Robert Darnton, PIECE MONOLOGUE (Apr. 1, 2012),

52. See Richard Delgado, When a Story Is Just a Story: Does Voice Really
Matter?, 76 VA. L. REV. 95, 106 (1990).

53. Id.
54. See infra Part II. Online identities are often fiuid, even creative and

whimsical. See Michiko Kakutani, Unraveling Brothers’ Online Lives, Link by
Link, N.Y. TIMES, Apr. 24, 2013, at Al (noting “the complexities of online
identity—of the ways in which people strike poses and don masks on the Web
(which can sometimes turn into self-fulfilling prophecies), and the ways in
which the Web can magnify or accelerate users’ interests and preoccupations”).

55. Previously we thought that this avenue held promise. See DELGADO &
STEFANCIC, supra note 4, at 126-28. Now we are less sanguine because courts


Internet, including Google, Microsoft, Facebook, and Twitter, are
unlikely to take forceful and continuing action against hate speech,
at least without firm public pressure.^^


Those who wish to make life difficult for minorities, gays and
lesbians, or any other favorite target on the Internet have a number
of means of doing so. These include e-mail messages, Facebook
posts, websites devoted to the supposed deficiencies of these groups,
and blogs and tweets, usually responding to a current racially
charged event or headline and serving as a platform to vent the
author’s spleen against a favorite target.

A. E-mail

E-mail and text messages are among the most frequently used
Internet vehicles. They also find ready use as vehicles f’or hate and
contempt. Often, perpetrators of violent or terrorist acts will be
found to have used e-mail exchanges or message boards to nerve
themselves up for their actions and secure the approval of others for
what they plan to do.̂ ^ Listservs, for example, permit hate groups

have rejected regulations protecting children, see id. at 93-109, a highly
protected group, and thus seem even less likely to approve regulations
protecting minority adults from Internet-based harms. See Frederick Schauer,
Harm(s) and the First Amendment, 2011 Sup. CT. REV. 81, 83 (doubting that
Congress and the judiciary will welcome new efforts to regulate expressive
harms, including those inñicted through electronic means, such as violent video
games); Tsesis, Hate in Cyberspace, supra note 1, at 820-31, 863-72 (noting
that courts often decline to review Internet speech on the ground that
jurisdiction is lacking because cyberspace does not occupy space in a physical
sense, but nevertheless proposing legal and commercial solutions). But see
Frank Pasquale, Reputation Regulation: Disclosure and the Challenge of
Clandestinely Commensurating Computing, in THE OFFENSIVE INTERNET:
SPEECH, PRIVACY, AND REPUTATION, supra note 5, at 107, 107 (suggesting a “Fair
Reputation Reporting Act” to control slanderous material on the Internet);
Solove, supra note 17, at 21-23 (suggesting common-law tort remedies). Note
that section 230 of the Communications Decency Act, 47 U.S.C. § 230 (2012),
immunizes Internet providers and managers from lawsuits by deeming them
not publishers of material that appears on their site.

56. See Somini Sengupta, Free Speech in the Age of YouTube, N.Y. TIMES,
Sept. 23, 2012, at SR4 (noting that “Google, Facebook, and Twitter receive
hundreds of thousands of complaints about content every week” and resist most
efforts to delete material); Rosen, supra note 36 (doubting that “the big Internet
companies” will draw a line or institute “tougher rules on hate speech,” but
noting that “the quest for the perfect screening system continues”); supra notes
47, 50; see also Ruben Rodrigues, Privacy on Social Networks, in THE OFFENSIVE
INTERNET: SPEECH, PRIVACY, AND REPUTATION, supra note 5, at 237, 255
(describing the vision of Facebook as a participatory democracy that is sensitive
to consumers’ wishes, feelings, and privacy as “illusorjr”).

57. See Citron, supra note 12, at 33-34; Tsesis, Inflammatory Speech, supra
note 1, at 1167.


to “send E-mail to each other with a click of the mouse”^^ and with
relative confidence that opponents will not be listening in.
Individualized e-mail often arrives without the sender’s name on it.
Thus, the “victim of a hateful E-mail may be unable to respond,
because the sender mailed the message from a public terminal or
has turned off his

B. Websites and Pages

Websites^” may cover such mundane topics as how to groom
one’s dog or ace the SAT. They may also put forward reasons to
hate a despised group, such as blacks, immigrants, or Jews.^i
According to one estimate, about 11,000 websites, social network
pages (which resemble websites), chat forums, and microblogs fall
into the latter category.^^ The author of this estimate explains the
recent growth of these vehicles by reason of their simplicity, stating
as follows:

[T]he formula for ethnic hate is quite simple: repeat the idea
enough and people will believe it. Soon the idea becomes
socially saturated and reaches a . . . point where everyone
knows the fabrication to be “true.” Discriminatory laws may
follow, and calls for genocide can be in the offing. . . . [T]hat
none of it is true does not seem to matter. The Internet’s dark
side globally propels Dark Age beliefs to millions at the click of
a mouse.63

Other commentators, however, point out a redeeming feature of
websites, that “even if they promote hate, [websites] are incapable of
harassing or terrorizing because visitors must seek the information”
by choosing to visit

58. See DELGADO & STEFANCIC, supra note 4, at 125.
59. Id. at 126. The victim might decline to respond for fear that doing so

might spur the hate speaker to new heights of vituperation.
60. A website is a collection of electronic documents or pages served by a

single address or domain. See Web Site, MERRIAM-WEBSTER,
(last visited Apr. 1, 2014).

61. Tsesis, Inflammatory Speech, supra note 1, at 1168 (noting that “Web
pages can stay up indefinitely and affect an impressionable visitor’s
behavior . . . years afterwards”).

62. Steven K. Baum, Essay, Fiction Outsells Non-Fiction, 43 LOY. U. CHI.
L.J. 413, 423 (2012); see DELGADO & STEFANCIC, supra note 4, at 125 (listing a
similar figure reported by the Southern Poverty Law Center).

63. Baum, supra note 62, at 423-24 (emphasis omitted) (noting that
Rwandan courts understand that words can kill and observing that “[fjree
speech principles do not appear to apply to the Internet”); see also Citron, supra
note 12, at 37 (noting that Web 2.0 platforms can create a subjective feeling of
closeness among “like-minded” individuals who affirm each other’s negative
views and thus become increasingly extreme).

64. DELGADO & STEFANCIC, supra note 4, at 126.


C. Facebook

Pioneered by young Internet star Mark Zuckerberg, Facebook
arrived in 2004 and rapidly attained a preeminent role as a means
of Internet-based social communication.s» With Facebook, users
establish accounts, after which they create personal profiles and
designate a list of friends with whom they can exchange messages
and receive updates. Users may also join or create interest groups
organized by subject or a common preoccupation.^^

Although much Facebook material is unremarkable, Facebook
groups can bully unpopular targets, defame teachers,
administrators, or anyone else who arouses their ire, or engage in
concerted hate speech and criticism that can “ruinQ someone’s
life.” ‘̂’ Since material may remain in place forever, Facebook also
poses a problem of privacy, especially for the young who may not
realize how a candid disclosure or photograph could harm them
later in

D. Twitter

Twitter is a relatively recent (circa 2006) innovation that allows
subscribers to send and receive short comments, or messages, called
“tweets.”69 Many such messages are innocuous and consist of
mundane details of an individual’s day, meals, or activities. But,
tweets may also contain hurtful remarks about the writer’s enemies,
particularly ones who are in the news. Many of the same complaints
mentioned above in connection with websites, e-mails, and Facebook
posts have surfaced in connection with Twitter, ̂ o Thoughtful

65. Lindsay S. Feuer, Note, Who Is Poking Around Your Facebook Profile?:
The Need to Reform the Stored Communications Act to Reflect a Lack of Privacy
on Social Networking Websites, 40 HOFSTRAL. REV. 473, 480-82 (2011).

66. Doug Gross, Facebook to Let Users Create Personal Groups, CNN (Oct.
6, 2010),

67. Karen M. Bradshaw & Souvik Saha, Academic Administrators and the
Challenge of Social-Networking Websites, in THE OFFENSIVE INTERNET: SPEECH,
PRIVACY, AND REPUTATION, supra note 5, at 140, 144.

68. See Solove, supra note 17, at 16-22; Tsesis, supra note 30 (noting the
permanence of much Internet material). For a discussion on the way merchants
and governmental snoops comb, filter, and sort the content of Internet
messages, see, for example. Sue Halpern, Are We Puppets in a Wired World?,
N.Y. REV. BOOKS (NOV. 7, 2013),

69. See United States v. Fumo, 655 F.3d 288, 331 (3d Cir. 2011) (Nygaard,
J., concurring in part and dissenting in part) (describing Twitter, a social
networking site that allows users to send and receive short messages of up to
140 characters); Rodrigues, supra note 56, at 240.

70. See Henry Alford, Twitter Shows Its Rude Side, N.Y. TIMES, Apr. 28,
2013, at ST2 (noting that “[i]f Twitter is an excellent shopping mall full of
boutiques offering specialized news and wit and opinion, it is also a crowded
barroom that bristles with a certain kind of white male rage . . . marked by a


observers believe that the ubiquity of the medium, coupled with its
inherent space limitations, encourages a kind of flippancy, as well as
“snark, hostility, dismissiveness, and counterproductive incivility.”‘̂ !

E. YouTube

An Internet-based means of international file sharing, YouTube
is a platform for a wide variety of material—principally videos,
many of them homemade.”^ Among the videos available on this
platform are ones attacking favorite targets or even advocating
violence, such as jihad, against the video maker’s political foes.’̂ ^
Much of the material available on this medium is amateurish and
bland, though sometimes hilarious, such as a pet doing tricks. But a
camera, like a speaker, can lie, so that “[o]ne of the risks in the era
of bloggers and You Tube [sic] is that statements and actions are so
closely monitored that any particular one, taken out of context,
might seem representative of the whole, or a clue to something dark
and bad.”74

F. Blogs

Blogs are sites containing information or commentary
consisting of entries called “posts,” written by individuals or teams,
usually covering a single subject or area, such as sports or politics.’̂ ^
Blogs received a boost from the development of web tools that
enabled nontechnically trained users to operate a blog or to post
entries on one. Many blogs serve as a forum for discussion of
favorite topics, including ones of which the posters disapprove;
others are little more than diaries or records of the hosts’ activities
or thoughts on a given day. Blogs can expose political corruption”^
or stories that slip through the mainstream media, like the
inaccuracies that led to Dan Rather’s resignation.’̂ ‘̂ Most notable,

hostility toward anything poetic or naïve” as well as vicious pranks, jokes, and

71. See Kathleen Fitzpatrick, #shameonyou, CHRON. HIGHER EDUC. (Apr.
22, 2013),

72. See Viacom Int’l, Inc. v. YouTube, Inc., 676 F.3d 19, 28 (2d Cir. 2012).
73. See Tsesis, Inflammatory Speech, supra note 1, at 1169. Its availability

to the public at large, not merely those with passwords, makes it especially
suitable for disseminating destructive messages. Id. at 1170.

74. See Sunstein, supra note 20, at 104-05.
75. The term is a contraction formed from the words “Web” and “log.” See

visited Apr. 1, 2014).

76. See Jia Lynn Yang, Jilted Mistresses Exposing Chinese Officials’
Corruption, SEATTLE TIMES (July 28, 2013, 11:48 AM),

77. Howard Kurtz, Dan Rather to Step Down at CBS, WASH. POST, Nov. 24,
2004, at Al; see also Jennifer Schuessler, 4̂ Star Philosopher Falls, and a
Debate over Sexism Is Set Off, N.Y. TIMES, Aug. 3, 2013, at Al (describing how


however, is that anonymous posters can malign a favorite target
with little fear of retaliation.”^


Why does the Internet contain a seemingly limitless supply of
material singling out minorities, women, gays, and immigrants for
disrespectful treatment? This question will require an examination
of the structural conditions that promote or discourage intolerance
and racism, including two that are well known to social scientists—
social contact and confrontation. “Distancing” and “depersonalizing”
also enter in.’̂ ^ An understanding of how these mechanisms operate
requires a quick review of national values.

A. The American Creed
Social scientists who have studied race and racism in America

believe that because of America’s checkered racial history, most
citizens are ambivalent in matters of race.^” They realize, on one
level, that our official, public values are race blind, egalitarian, and
highly aspirational.81 All men are equal. We are all brothers and
sisters, equal agents in the eyes of God. Every person is a precious
moral agent. Discrimination of any kind is wrong. ̂ 2

In short, Americans hold to a formal set of higher values
according to which racism and discrimination are anathema. We

indignant bloggers called attention to a famous professor who exploited female

78. See Martha C. Nussbaum, Objectification and Internet Misogyny, in
68, 85; Solove, supra note 17, at 23-25; Sunstein, supra note 20, at 104. The
online comments sections of many blogs “are . . . full of vitriol.” Michael Erard,
Riff: “Wild Back Alleys Where People Sound Their Acid Yawps,” N.Y. TIMES
MAG., Sept. 22, 2013, at 50; see also Katrin Bennhold, Bid to Honor Austen Is
Not Universally Acknowledged, N.Y. TIMES, Aug. 5, 2013, at A5 (noting that the
proposal that Britain honor Jane Austen by placing her likeness on bank notes
evoked a fiood of misogynistic responses on Twitter). The aforementioned
vehicles are by no means the only ones for online communications. “[T]he
online world is moving so fast [that the average reader may] never catch up.”
David Pogue, A Scrapbook on the Web Catches Fire, N.Y. TIMES, Feb. 15, 2012,
at Bl. New online vehicles include Foursquare, Tumblr, Linkedin, Instagram,
Reddit, Path, and Pinterest, “a pinboard for online photos” that enables a user
to display “a tidy array of all the images that appear on [a] current web
page. . . . It’s like virtual scrapbooking.” Id. at Bl.

79. See infra note 99 and accompanying text.
80. See, e.g., GORDON W. ALLPORT, THE NATURE OF PREJUDICE 326-32 (25th

anniv. ed. 1979) (noting the attitude’s ubiquity and situation-specific nature).
81. Id. at 330.
82. 7c?.; see also Richard Delgado et al.. Fairness and Formality: Minimizing

the Risk of Prejudice in Alternative Dispute Resolution, 1985 WiS. L. REV. 1359,
1383-84 (discussing the American Creed and its role in countering


feel badly when we fall short of these ideals and resolve to do better
in the future.

B. Two Sets of Values

Despite these formal values, many of us act according to a lower
code of behavior during moments of intimacy when we believe no
one is watching.83 Thus, a typical American may insist that he or
she is color blind and act in that fashion on official occasions, such
as the Fourth of July when the bands are plajdng and local groups
are marching proudly by.̂ ^ If this person is white and finds himself
or herself standing next to a black or Latino person, he or she is apt
to smile and make a comment about the wonderful day or fine band.
He or she may even put an arm around the other’s shoulder, offering
the other a drink or an invitation to a family function.^s

This same American, however, on another occasion might
behave differently. With friends, at a bar or private gathering, he or
she might feel freer to tell a joke at the expense of blacks, gays, or
members of a minority religion. ̂ ^ ^ t work, this person might feel
free to refuse an interview or promotion to an otherwise qualified
African American if he or she believes this will not come to the
attention of others.̂ ‘̂

In short, many of us act as though we subscribe to two sets of
values, one for official occasions and another for private ones. We
select the one or the other unconsciously, depending on the situation
and the company in which we find ourselves. To cite a familiar
example, many women know by a kind of instinct that male
coworkers may behave in a supportive and antisexist fashion on the
job, particularly when representing the company.̂ ^ The same men
at a party after work may feel much freer to tell an antifemale joke,
rib a female colleague, make an aggressive pass, or otherwise
behave in a fashion that makes a woman uncomfortable.^^ Women
who remain at a gathering after “the cigars come out” often are
aware that the conversation is apt to take on a sharper edge and
that they may find it necessary to defend themselves.

Because we acquire racial attitudes at a very early age,9o often
from parents or playground companions, countering the dark

83. Delgado et al., supra note 82.
84. Id. at 1383-84, 1387-88.
85. Cf. id.
86. See id. at 1385.
87. See id.
88. The reader may recall how President Bill Clinton exhibited a proper,

even gallant, attitude toward women in public.
89. The same reader will no doubt recall how Clinton behaved quite

differently on at least a few private occasions.

ed. 1970).


impulses that we all harbor must take into account their deep-
seated nature. Social scientists have devised two principal
strategies—a social contact theory and a confrontation theory.
Often these work together.

C. The First Strategy: Social Contact
The social contact theory, as its name suggests, aims to reduce

racial prejudice by providing opportunities for members of different
races to interact, often in group settings, such as school or sports.^i
The theory holds that much of racial prejudice and friction is a
function of incorrect cognition: the individual internalizes the belief,
often from parents or playmates, that members of other races are
not to be trusted.92 To counter these beliefs, society may arrange
many opportunities for youngsters to interact with members of
different races.^^ Through frequent contact, they will realize that
those with skin colors different from their own are much like them
and their friends: some smart, others less so; some nice, some not;
some trustworthy, others conniving; and so on.

The ideal form of social contact for reducing discriminatory
beliefs is that which occurs among equals in pursuit of a common
objective.94 Sports and the military are prime examples.^^ Much
evidence suggests that the theory is highly efficacious; individuals
who grow up with many opportunities to interact with members of
different races and ethnicities are much more comfortable in mixed-
race settings than ones who do not.̂ ^ The social contact theory
formed the basis of institutional desegregation during the sixties^’
and is a mainstay of antidiscrimination scholarship and practice.
People who grow up with others of different t3^es are more
comfortable with diversity,^^ choose friends of different hues, and
feel more comfortable working with or living next to members of
racial groups other than their

91. See ALLPORT, supra note 80, at 281; Delgado et al., supra note 82, at
1385—86. Simple two-way discussion causes many individuals to moderate
negative views that they may harbor of other groups. See Eugene Burnstein &
Amiram Vinokur, Testing Two Classes of Theories About Group Induced Shifts
in Individual Choice, 9 J. EXPERIMENTAL Soc. PSYCHOL. 123, 132-33 (1973);
David G. Myers & George D. Bishop, Discussion Effects on Racial Attitudes, 169
Sci. 778, 778 (1970).

92. See Delgado et al., supra note 82, at 1380-81.
93. See id. at 1385-86.
94. Id. at 1386.

96. See Delgado et al., supra note 82, at 1385-86.
97. Id. at 1385.
98. Id. at 1385-86.
99. Id.; see also Grutter v. BoUinger, 539 U.S. 306, 343 (2003) (approving

university affirmative action programs that aim to produce a critical mass of
diverse students). Distancing, by contrast, decreases fellow feeling and makes


D. The Second Strategy: Confrontation

Because social contact is not always feasible, social planners
and architects can often turn to a second strategy. The
confrontation theory builds on the insight, mentioned earlier, that
many Americans are ambivalent in matters of race, harboring
highly aspirational convictions (“all men are equal”) but acting on a
lower set in moments of intimacy, when their guard is down and no
one is watching. 100

The confrontation theory suggests that in order to trigger the
higher set of values, it is wise to provide reminders that society
expects and hopes for this form of behavior, loi These reminders put
people on notice that society expects them to behave in a fair-
minded, nonracist fashion and that falling short of this will bring
social disapproval or sanction.

1. Examples of the Confrontation Strategy: The “Fairness and
Formality” Thesis in Operation

a. The Military

The military is a prime example of the confrontation theory. 102
With a highly diverse workforce, the military requires that
noncommissioned and commissioned officers demonstrate the ability
to work effectively with subordinates whose skin colors are different
from theirs.103 A formal set of rules and expectations for nonracist
conduct enter into a person’s evaluation for promotions. These
expectations are well known; persons who cannot bring their
conduct into conformity with them are not promoted and may be

it more likely that an individual will feel little remorse at injuring or
(2013) (noting that the operators of military drones feel less remorse over
ordering attacks that may take innocent civilian lives than those who drop the
bombs directly from the air). For information on how social media may
depersonalize and distract, see Leonard Pitts, Jr., Leonard Pitts: Social Media
Can Be Deadly, MIAMI HERALD (Oct. 12, 2013),
/2013/10/12/3684283/leonard-pitts-social-media-can.html. See also Katrin
Bennhold, Behind Flurry of Killing, Potency of Hate, N.Y. TIMES, Oct. 13, 2013,
at A6 (noting that publicity about the defects of one’s enemies can make
“murder not just permissive but obligatory. We should kill vermin or

100. See Delgado et al., supra note 82, at 1385-87.
101. Id. at 1387-88.
102. See MoSKOS & BUTLER, supra note 95.
103. Grutter, 539 U.S. at 308, 331.
104. See MoSKOS & BUTLER, supra note 95.


b. Alternative Versus Formal, In-Court Adjudication
Dispute resolution is a second example. Studies show that

formal, in-court disputes offer a fairer forum for a relatively
disempowered individual—such as a woman of color confronting (for
example, in a divorce action) a person of higher prestige, say, a
white male executive—than the informal kind of forum, such as
mediation or arbitration, lô Many features of the courtroom setting,
such as the judge sitting on high, the robes, the seal of the state or
federal authority, the prescribed time and manner of speaking, and
the explicit instructions the judge issues to the jurors on deciding
the case in accord with the law, reduce inequalities of initial
position. 106 Alternative dispute resolution, which takes place in an
informal, relaxed setting with fewer rules, includes fewer such
features, so that racial and other power differentials are apt to
infiuence the outcome even more than when a dispute occurs in a
formal courtroom. lO”

E. The Two Theories Applied to the Internet
With cyberspace, neither form of control is readily available. As

a result, the Internet offers a fertile breeding ground for racial
vituperation and contempt.

1. Social Contact
As mentioned, social contact among equals in pursuit of

common objectives moderates racial animosity.lo^ The Internet,
however, separates people from one another, los When it does enable
people to join together in some form of virtual community, it is often
a community of the like-minded.”o It does not constitute the kind of

105. See Delgado et al., supra note 82, at 1391-96; Trina Grillo, The
Mediation Alternative; Process Dangers for Women, 100 YALE L . J . 1545, 1549-
50 (1991).

106. Delgado et al., supra note 82, at 1387-88.
107. Id.
108. See supra notes 91-104 and accompanying text.
109. See Sunstein, supra note 20, at 96-104 (discussing how the Internet can

polarize and divide society into small groups of like-minded individuals who
reinforce each other’s beliefs and values); see also Frank Bruni, Op-Ed., Our
Hard Drives, Ourselves, N.Y. TIMES, NOV. 17, 2012, at SR3 (noting that
cyberspace gives some the illusion of protection, “freeing them to engage in a
kind of explicit and assertive dialogue that two people sitting across from each
other, or even talking on the phone, would in most cases be too shy to
broach. . . . After all, how could a communion so faceless prompt a brutal

110. See Citron, supra note 12, at 43-47; Sunstein, supra note 20, at 96-104;
Tom Hays, Cannibal Case Involving New York City Police Officer Blurs Lines of
Fantasy, Reality, DAILY BREEZE (Oct. 26, 2012),
blurs-lines-of-fantasy-reality (noting that Internet chats enabled members of a
website “devoted to a fetish called ‘vore'” to reinforce their common interest in


setting that, like sports, the Boy Scouts, public schools, or the
military, places people of different backgrounds together and allows
them to conquer their fears and work together in a cooperative
fashion. On the contrary, the Internet heightens one’s sense of
separation from the momentary target of one’s venom and, by
immersing the user in a community of the like-minded, increases
the feeling that the world comes divided into two groups—us and
them.Ill Individuals who begin a session in cyberspace are apt to
find little there that challenges their preconceptions about people of
different races, religions, nationalities, or sexual orientations. 112 If
they began disliking a group, say, Jews, they are likely to end the
session disliking them even more.

2. Confrontation Theory

By the same token, individuals who harbor dislikes of members
of other racial groups are apt to find little on the Internet that
challenges their pet beliefs.^^ Unlike litigation, the military, or a
seminar on diversity, little reminds the person navigating the
Internet that this area is one where society expects everyone to treat
others with respect and dignity, n* Unlike a courtroom, where

sexuahzing cannibalism, “the idea of being eaten whole and alive, eating
another alive, or watching this process,” and discussing a chat room that “spells
out strict rules for participation”).

111. See Sunstein, supra note 20, at 96-104 (describing group polarization).
112. See, e.g., id.; Jenna Wortham, Facebook Made Me Do It, N.Y. TIMES,

June 16, 2013, at SR5. Confirmation bias, a well-known mechanism in social
science, may be operating as well. See Confirmation Bias, Sci. DAILY, (last visited Apr.
1, 2014). In confirmation bias, a person pays particular attention to evidence
that confirms his or her preexisting beliefs and devotes less attention to
countervaihng evidence. See id. With Internet searching and browsing, one is
apt to come across opinions and messages that correspond to one’s search
description. For example, a searcher looking for evidence that vaccinations are
dangerous and likely to cause the very disease that they ostensibly guard
against is apt to find it. Even bizarre or deranged material is apt to find camp
followers. See Steven Schlozman, The Harvard Doctor Who Accidentally
Unleashed a Zombie Invasion, N.Y. TIMES MAG., Oct. 27, 2013, at MM46.

113. See Delgado et al., supra note 82, at 1386-89, 1391-98 (explaining how
informal settings, which lack explicit rules governing action, are apt to invite
lawless and antisocial behavior, particularly racism).

114. See Leonard Pitts, Jr., Technology Enhances Beer Muscles, MILWAUKEE
J. SENTINEL (May 29, 2012),
/155480285.html (noting that the Internet’s privacy encourages
“fiaming . . . with gleeful abandon you know they’d never dare display in the
flesh and mortar world . . . . You can now frighten and alarm someone without
leaving the comfort of your bed . . . [, creating a form of] Internet courage” and
concluding that “[tjechnology will not make us better. There is no app for
that”); see also Bruni, supra note 109 (noting that there is “a thrilling sense of
isolation and permission.. . . Cyberspace gives people more than an illusion of
protection. It gives them nerve, freeing them to engage in a kind of explicit and


physical features and rituals remind the participants that this is a
place where they are expected to exhibit behavior consistent with
the American Creed, us persons who enter cyberspace encounter few,
if any, such reminders. “^ Little conjures up public values of
fairness and equality. Users believe they can get away with
speaking their minds, and if those minds contain, at the moment,
harsh thoughts or judgments of others of different complexions from
their own, they feel free to speak them.i”

F. Summary: A Trend Toward Coarsening

The advent of the Internet, beginning around 1994, has not led
to an improvement in civility or the quality of interpersonal
relations.11^ If subjective experiences are a guide, heavy users feel
freer to take positions or espouse views that depart from our public
values.119 We posit that the Internet, for the reasons mentioned
above, may be playing a role in this change. Implicit association
tests are showing that a majority of Americans harbor feelings and
attitudes that surprise even them. 120 Might the large amount of
time many of us spend on our computers be playing a part in this

Since hate speech on the Internet is pervasive, rising, and
unlikely to subside unaided, what should society do to counter it?


Possible responses to Internet hate speech include those that
one commonly hears with this type of utterance, including talking
back to the aggressor. As we shall see, most of these avenues are
even less efficacious than they are with ordinary speech. Because
courts and legislatures are unlikely to tackle the Internet any time
soon, 121 while Internet providers and companies are even less likely
to rein themselves in, 122 new approaches are in order.

assertive dialogue that two people sitting across from each other, or even
talking on the phone” would avoid).

115. See supra notes 80-82 and accompanying text.
116. Cyberspace has few rules; those that it has (no trolling, no lurking) are

mostly voluntary.
117. See, e.g.. Citron, supra note 12; Levmore & Nussbaum, supra note 20,

at 2; Sunstein, supra note 20; Tsesis, Inflammatory Speech, supra note 1, at
1168-70, 1173.

118. See Editorial Board, supra note 17; Gedye, supra note 2 (noting a
generalized “change in the portrayal of women in the media”).

119. See supra notes 1-3, 50-54 and accompanying text (noting this trend).
120. See Jerry Kang, Trojan Horses of Race, 118 HARV. L . REV. 1491, 1491-

93 (2005).
121. See supra note 55 and accompanying text.
122. See supra note 56 and accompanying text.


A. Conventional Responses Associated with the Free Speech Position

Some of the usual means that First Amendment devotees urge
to counter ordinary—spoken or written—^hate speech are even less
promising with the Internet variety. Talking back to the aggressor
(the favorite approach of First Amendment absolutists^^s) is
impossible for speech that is anonymous or that occurs in a forum of
the like-minded.124

The same is true for the suggestion that victims of hate speech
should tolerate it because it serves as a pressure valve that enables
hate speakers to harmlessly air feelings that, if bottled up, could
explode in even more harmful forms later, î s The idea that racist
feelings will cease once a speaker expresses them is unfounded even
with ordinary speech;i26 with the Internet variety, it holds even less.
Most Internet speech, as mentioned, takes place anonymously or
among the like-minded, î v The first tirade eggs an audience on. Far
from producing a pacified speaker and audience, the speech incites
another and another. A chorus of “right ons” or “likes” encourages
the speaker to believe that his or her attitude is widely shared,
when it in fact is not.i^s

A further riposte from the free speech side is that minorities
ought to toughen up and not run to the authorities every time they

123. For more information on the talking-back approach, see, for example.
DELGADO & STEFANCIC, supra note 4, at 35, 207.

124. See id. at 207 (discussing a more-speech approach); see also Baum,
supra note 62, at 424 (“One may try to replace ‘bad’ web pages with ‘good’ web
pages, but research suggests that people will choose to seek the bad websites
out and that few are interested in obtaining a balanced view. Hateful social
beliefs will endure because as a species, we remain hopelessly more fascinated
by the salacious than by the salubrious.”).

125. See Richard Delgado & David H. Yun, Pressure Valves and Bloodied
Chickens: An Analysis of Paternalistic Objections to Hate-Speech Regulation, 82
CALIF. L . REV. 871, 878-80 (1994).

126. Id.
127. See supra notes 23, 50 and accompanjdng text (discussing this feature

of Internet speech); see also Levmore & Nussbaum, supra note 20, at 3;
Pasquale, supra note 55, at 113 (suggesting a “Fair Reputation Reporting Act”
to control anonymous slander on the Internet).

128. See supra notes 46-47 and accompanying text (describing the new norm
against hate speech); see also Wortham, supra note 112 (noting how internet
use generates its own positive reinforcement through a chorus of “likes” and
“right ons” that increases the likelihood that viewers will act on these feelings
later). At times, the positive reinforcement loop works via a multiplier effect.
An individual defaces a wall or park bench with a hate-filled graffito
disparaging Jews, say, or blacks. Someone takes a photograph of the graffito
and puts it on the Internet, thus multiplying its exposure tens of thousands of
times compared to the relatively small number of passersby who might
otherwise see the graffito while on a walk in the park. See Felicity Barringer,
As Vandals Deface U.S. Parks, Some Point to Online Show-Offs, N.Y. TIMES,
June 5, 2013, at Al (explaining that social media provide instant gratification,
which “could stimulate the impulse to deface”).


hear something that offends their feelings. 129 But it is hard to put
this approach into effect with hate speech in cyberspace, since one is
often unaware that it is taking place or that one’s identity or good
name has been impugned. And with the kind that shows up
suddenly, unbidden on one’s computer screen, one has little
opportunity to harden oneself in advance. Moreover, this approach
places the burden of countering hate speech on those who bear the
brunt of it in the first place.13°

A final rejoinder is what is known as the ‘Tjellwether”
argument, which holds that the racist who is known is better—that
is, safer—than one who is not known, î i This argument, even with
ordinary speech, is fiawed, since it ignores a third alternative—that
the racist who is deterred by firm application of rules and norms is
even safer than one who spews it constantly. i32 With cyberspeech
the bellwether argument holds with even less force than it does with
ordinary speech, since much of Internet speech proceeds in privacy,
arriving suddenly and without identifying the source. i33

B. Legal Responses to Internet Hate Speech

Hate speech has received a tepid response from the lower
courts. 134 Unless the speech is dramatic enough to support a sui
generis prohibition Gike cross burningi35) or resembles an existing
exception to First Amendment coverage, such as words of threat, i36

129. See Richard Delgado & David Yun, The Neoconservative Case Against
Hate-Speech Regulation—Lively, D’Souza, Gates, Carter, and the Toughlove
Crowd, 47 VAND. L . REV. 1807, 1816 (1994) (describing the “toughlove” position
favored by some defenders of hate speech); see also WALDRON, supra note 1, at
154-55 (noting that hate can induce minorities to avoid participation in public
life so that the system of free speech ends up less vibrant than before).

130. That is, the victim first endures an affront then receives the unwelcome
news that he or she is expected to toughen up in preparation for more of the

131. See DELGADO & STEFANCIC, supra note 4, at 210-11 (discussing this

132. Id.
133. See supra notes 23, 50, 114 and accompanying text (discussing this

feature); see also WALDRON, supra note 1, at 95 (urging that even if
counterpressure drives hate speech underground, that is a good thing).

134. The Supreme Court is an exception, supporting regulation on the few
occasions when it has considered such speech. See Virginia v. Black, 538 U.S.
343, 363 (2003) (upholding a Virginia law that “outlaw[s] cross burnings done
with the intent to intimidate”); Beauharnais v. Illinois, 343 U.S. 250, 251, 266-
67 (1952) (upholding an Illinois law that made it a crime to “exhibit in any
public place” any publication that “portrays depravity, criminahty, unchastity,
or lack of virtue of a class of citizens, of any race, color, creed or religion,” which
“exposes the citizens of any race, color, creed or religion to contempt, derision,
or obloquy”).

135. Black, 538 U.S. at 343.
136. See Delgado & Yun, supra note 125, at 883.


defamation,i3” intentional infliction of emotional distress,i^s or a
statutory remedy such as speech that creates a hostile workplace, î s
courts are apt to find hate speech constitutionally protected. Since
courts believe that Internet speech qualifies for constitutional
protection, 140 courts are likely to look unfavorably at legislative
efforts to limit speech in cyberspace.

Since vituperative speech is self-reinforcingi^i and apt to spread
among like-minded groups^^^ or loners operating in secrecy,i’̂ ^ new
approaches must be sought. Three approaches that commend
themselves are unmasking, group condemnation, and economic

1. Unmasking

One approach would focus on depriving groups and individuals
who take advantage of the cloak of secrecy that the Internet
provides. Diligent detective work will often unearth the source of a
hurtful message, blog posting, or website.i*”* Denouncing the group
or individual publicly can demonstrate to users of the Internet that
disseminating hate through this medium brings consequences and
can give pause to others who might be tempted to follow suit. This
approach would apply the confrontation theory by holding the
possibility of social sanctions over the heads of individuals tempted
to engage in anonymous name-calling and disparagement over the
Internet.145 In a few celebrated cases, courts have ordered such an
unmasking, usually in connection with civil discovery, î e

2. Group Condemnation

Norms of nondiscrimination run highest in two large groups:
those who lived through the civil rights era of the sixties and those

137. See DELGADO & STEFANCIC, supra note 4, at 61.
138. Id. at 62.
139. Id. at 63-66.
140. See id. at 127 (explaining that “most hate speech on the Internet will

not be considered threats, harassment, fighting words, or libel, since it is
generally directed broadly and not at a particular person”).

141. See supra notes 23, 109 and accompanying text.
142. See supra notes 50, 110-12 and accompanying text.
143. See supra notes 22-23, 117 and accompanying text.
144. See Eric Pfanner & Somini Sengupta, A Battle to Unmask Twitter

Users, N.Y. TIMES, Jan. 25, 2013, at B4 (describing one such effort “prompted by
a spate of anti-Semitic writing on Twitter last year, including hashtags, or
topical themes, like ‘a good Jew is a dead Jew'” and “jokes about the

145. See Levmore, supra note 50, at 56-59 (noting that the anonymous
quality of much Internet communication places the speaker beyond
accountability or sanction).

146. See Scott Talkov, 9th Circuit Unmasks Anonymous Online Speakers,
CAL. LITIG. ATT’Y BLOG (Jan. 16, 2011),


who arrived at adulthood in the early years of the current century.
Sometimes named the millennial generation, this latter cohort
considers itself fair-minded and color blind, even antiracist.^*”
Steeped in technology, this generation considers the Internet
practically an extension of their bodies and minds. Both groups are
apt to find Internet hate speech distasteful and worth fighting
against. With a small amount of encouragement, imaginative
members of these groups could join in applying pressure to racist
websites, broadcasters, and other purveyors of vitriol. Many do
already. They should find sympathetic allies among their peers.

3. Economic Sanctions

Even if a disseminator of cyberhate is beyond the reach of the
first two remedies, he or she will often operate a business, agency, or
other group that requires funds, donations, a tax deduction, an
advertising account, or another monetary benefit to keep operating.
With these individuals, economic sanctions may gain their attention.
These sanctions can take the form of a consumer boycott,
withdrawal of a business license, libel accusation, or letter to a tax
authority calling the group’s status into question. Since most
individuals and groups zealously guard their own financial fortunes,
approaches of this type may nudge them to moderate their antisocial
activity and behave in a less socially pernicious fashion.


If structural features of Internet speech—including secrecy, self-
selection, group reinforcement, and a sense of righteous potency—lie
behind an unsettling rise of derogation and hate in that medium,
society ought to consider measures to combat it or at least reduce
the damage it infiicts. The usual responses to hate speech,
including the talking-back approach, the pressure-valve argument,
and the notion that hate speech serves as a useful bellwether
alerting minorities to their foes, are even less promising in this
realm than they are in society at large. Since legislatures and the
lower courts are, for now, unpromising sources of remediation,
concerted, popular action is the most likely avenue for redress. Two
cohorts, young users from the millennial generation and those who
lived through the civil rights era of the 1960s seem most likely to

147. Young people, however, are as susceptible to the herd instinct as
anyone else and may reflexively adopt group values and loyalties. See, e.g.,
Vivian Yee, Statutory Rape, Twitter and a Generational Divide, N.Y. TIMES,
Apr. 5, 2013, at A16 (noting that teens in a Connecticut high school employed
Twitter to back football players who allegedly raped two thirteen-year-old girls
and to depict the girls as whores bent on destroying the football players’ lives).
For a discussion on the manner in which homogeneous groups tend to become
more extreme through deliberation and exchange of views, see Citron, supra
note 12, at 36-37.


mobilize against this form of conduct. The most likely actions
include unmasking, shaming, and economic sanctions against

Both the Internet and those who use it to disseminate odious
remarks are relatively young. With luck, the onset of maturity will
moderate some of the excesses of both. In the meantime, forceful,
concerted action by those who prefer a more civil society can,
perhaps, provide a nudge in that direction.

Copyright of Wake Forest Law Review is the property of Wake Forest Law Review and its
content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder’s express written permission. However, users may print, download, or email
articles for individual use.

Expert paper writers are just a few clicks away

Place an order in 3 easy steps. Takes less than 5 mins.

Calculate the price of your order

You will get a personal manager and a discount.
We'll send you the first draft for approval by at
Total price: