Bullying, Romantic Rejection, and Conflicts with Teachers: The Crucial Role of Social Dynamics in the Development of School Shootings – A Systematic Review

Abstract

A plethora of studies have appeared which argue that, prior to their attack, the perpetrators of school shootings had experienced intense conflicts and problematic relations (e.g. bullying) with peers and teachers, and were on the periphery of the schools' social life. This in turn resulted in the perpetrators' view of themselves as marginalized victims. However, methodological problems and inconsistencies mark many studies, and findings vary. In an attempt to clarify the role of perpetrators' negative social experiences with peers and teachers prior to their attack, we have undertaken a systematic search of the literature, including 35 international primary studies on school shootings. In selecting the studies we limited ourselves to ones which deal with at least two (Range: 2–39) cases in which a violent targeted attack was carried out by a current or former student who chose their school or university as the site of the attack. A total of 126 cases (128 perpetrators) from 13 countries (Sweden, and Thailand) were examined. The mean age of the perpetrators was 19 (Range: 6–62 years, SD = 8.72), and in 121 cases the perpetrators were males. Detailed information relating on the social dynamics that contributed to the attack was found in 67 case reports. Our analysis revealed that in 88.1% of cases the future perpetrator experienced social conflict within the school environment. A minority of perpetrators (29.9%) were physically bullied, while 53.7% experienced peer rejection, verbal and otherwise. Romantic rejection was only found in 29.9% of cases. Conflicts with teachers (43.3%) proved a decisive factor. In order to better understand the role of social dynamics in the developments leading up to school shootings, it is necessary to analyze the perpetrators' position within their social network and the ways in which they experienced interaction with their peers and others at the school. In addition we must obtain precise information on their views of themselves as victim over a period of time.

Full text

International Journal of Developmental Science 8 (2014) 3–24
DOI 10.3233/DEV-140129
IOS Press
Systematic Review – Target Article
Bullying, Romantic Rejection, and Conflicts
with Teachers: The Crucial Role of Social
Dynamics in the Development of School
Shootings – A Systematic Review†
Friederike Sommer∗, Vincenz Leuschner and Herbert Scheithauer
Department of Educational Science and Psychology, Freie Universit¨at Berlin, Berlin, Germany
Abstract
A plethora of studies have appeared which argue that, prior to their attack, the perpetrators of school shootings had experienced intense conflicts
and problematic relations (e.g. bullying) with peers and teachers, and were on the periphery of the schools’ social life. This in turn resulted in
the perpetrators’ view of themselves as marginalized victims. However, methodological problems and inconsistencies mark many studies, and
findings vary. In an attempt to clarify the role of perpetrators’ negative social experiences with peers and teachers prior to their attack, we have
undertaken a systematic search of the literature, including 35 international primary studies on school shootings. In selecting the studies we limited
ourselves to ones which deal with at least two (Range: 2–39) cases in which a violent targeted attack was carried out by a current or former
student who chose their school or university as the site of the attack. A total of 126 cases (128 perpetrators) from 13 countries (USA, Canada,
Germany, Finland, Brazil, Argentina, Australia, Bosnia, Greece, Hungary, Netherlands, Sweden, and Thailand) were examined. The mean age
of the perpetrators was 19 (Range: 6–62 years, SD = 8.72), and in 121 cases the perpetrators were males. Detailed information relating on the
social dynamics that contributed to the attack was found in 67 case reports. Our analysis revealed that in 88.1% of cases the future perpetrator
experienced social conflict within the school environment. A minority of perpetrators (29.9%) were physically bullied, while 53.7% experienced
peer rejection, verbal and otherwise. Romantic rejection was only found in 29.9% of cases. Conflicts with teachers (43.3%) proved a decisive
factor. In order to better understand the role of social dynamics in the developments leading up to school shootings, it is necessary to analyze the
perpetrators’ position within their social network and the ways in which they experienced interaction with their peers and others at the school. In
addition we must obtain precise information on their views of themselves as victim over a period of time.
Keywords
school shooting, bullying, social rejection, severe targeted school violence, systematic review
Introduction
The main question asked in the wake of school shoot-
ings is always why they happened, and mental health
professionals, policy makers and researchers have
∗Address for correspondence
Friederike Sommer, Freie Universit¨
at Berlin, Department of Edu-
cational Science and Psychology, Habelschwerdter Allee 45, 14195
Berlin, Germany. E-mail: friederike.sommer@fu-berlin.de
†See endnote at the end of the manuscript.
worked hard to explain why adolescents commit mass
murder at their schools. While some have approached
the question from the angle of individual pathology
(Harding, Mehta, & Newman, 2003; Langman, 2009) or
culture (Brown, Osterman, & Barnes, 2009; Kimmel,
2008), others have focused on the “social dynamics”
that eventually culminated in shootings (Newman, Fox,
Harding, Mehta, & Roth, 2004).
Empirical evidence shows that violent attacks at
schools and university campuses are rarely sudden,
ISSN 2192-001X/14/$27.50 © 2014 – IOS Press and the authors. All rights reserved 3

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
impulsive acts, but are rather the result of a long
development involving a multitude of interacting fac-
tors (Leuschner et al., 2011; Vossekuil, Fein, Reddy,
Borum, & Modzeleski, 2002). Verlinden, Hersen, and
Thomas (2000) reviewed studies which looked at youth
violence in general and the risk factors which can lead
to it. The list they proposed included amongst others
impulsiveness, hyperactivity, abuse in childhood, aca-
demic failure, access to weapons, and the experience
of being bullied. In examining the role of these fac-
tors, they selected nine cases involving violent mass
assaults carried out in schools. In addition to individual
risk factors such as the feeling of being let down by
their families, adverse social and environmental condi-
tions, and acute strain arising from various sources, they
found problem-laden social relationships in the school
environment in the perpetrators’ histories, for which we
use the term social dynamics.
Detailed investigation of perpetrators’ social rela-
tionships in the school they attended makes sense, as
the perpetrators deliberately chose it as the scene of
their violent act. When it comes to social dynamics, the
primary factor discussed is the perpetrators’ perception
of themselves as victims of physical or verbal bullying
prior to their attack. Bullying is defined as long-term
repeated victimization, with an imbalance of power
between the bully and victim (Olweus, 1994). It can
be physical, verbal or psychological in nature. A multi-
tude of studies have revealed that bullying can result in
feelings of helplessness, loneliness, anxiety and depres-
sion, and can lead to psychosomatic disorders, eating
disorders or even suicide (also referred to as bully-
cide). Truancy and a decline in academic performance
are also common as are problems in relationships;
aggressive, delinquent and anti-social behavior; dating
violence and excessive risk-taking (for a summary see
e.g. Hess & Scheithauer, in press; Scheithauer, Hayer,
& Petermann, 2003; Ttofi, Farrington, & L¨
osel, 2012).
Research has confirmed that being bullied has a strongly
negative effect on children’s psychosocial development,
thus it is reasonable to assume that it plays a role in
school shootings. On the website StopBullying.gov it
is stated that “a very small number of bullied children
might retaliate through extremely violent measures. In
12 of 15 school shooting cases in the 1990s, the shooters
had a history of being bullied” (“Effects of Bullying”,
2014). Indeed, many case studies, especially from the
U.S., have indicated that peer rejection is most likely
the key risk factor in the development of future per-
petrators (Fast, 2009; Kidd & Meyer, 2002; McGee &
DeBernado, 1999). According to Larkin (2009), more
than half of the perpetrators he studied were seeking
revenge for mistreatment they had suffered at the hands
of their peers and that they specifically targeted peo-
ple who had tormented them. The perceived social
rejection might have been linked to a lack of social
skills, which caused the future perpetrator to founder
in social situations, setting him or her off on a down-
ward spiral of increasing isolation and rejection (Kidd
& Meyer, 2002; Nansel et al., 2001). This assumption
is confirmed by results obtained in a case study pre-
sented by Leary, Kowalski, Smith, and Phillips (2003),
who described the strong link between interpersonal
rejection and aggressive behavior in general, thereby
supporting the thesis that school shootings are com-
monly provoked by real or imagined rejection in the
form of teasing, ostracism, or unrequited love. In only
two of the fifteen incidents they analyzed there was no
perceivable pattern of ongoing ostracism, bullying, or
malicious teasing. Thus, the authors argued that rejec-
tion alone does not suffice as a cause of violence, but
when it appears along with other factors such as psy-
chological problems, fascination with weapons and/or
themes of death, the risk of violence is higher.
These findings are consistent with those obtained by
the U.S. Secret Service, who found evidence of bully-
ing, ostracism, and social rejection in over two-thirds
of the 37 cases they reported in their Safe School Initia-
tive (Vossekuil et al., 2002). The authors argued that the
experience of being bullied most likely had a significant
impact on the perpetrators’ motivation in committing
their attack. Leary et al. (2003) agreed on this, stating
that the key findings in the U.S. Secret Service report
suggest that prior to the shooting, the future perpetrators
had experienced repeated bullying, reportedly border-
ing on torment. In a study carried out by Newman et
al. (2004), social marginality – the shooter’s percep-
tion of himself as extremely marginalized in the social
world that matters to him – was discussed as one of five
key factors on the path to school shootings. There was
evidence of social marginality in all but one of the 25
cases studied by Newman et al. (2004): 67% had felt
marginalized and 63% had been bullied or teased. One
particular type of bullying is discussed by Kimmel and
Mahler (2003), who support the thesis that future perpe-
trators failed to measure up to prevailing norms of mas-
culinity. They found that nearly every one of the 31 per-
petrators they analyzed in their study had been accused
of being gay because they were not tough enough. This
finding lends credence to the claim that most of the
shootings represented a violent response to what the
perpetrator perceived as attacks on his masculinity.
4 International Journal of Developmental Science 8/2014, 3–24

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
Since most school shootings have been carried out
by males, it is important to investigate the role of gen-
der in school shootings. In the work by Leary et al.
(2003) mentioned above, it is argued that one particular
form of rejection, namely unrequited love, can serve
as a risk factor culminating in a violent act. Klein’s
(2012) finding revealed that in about half of the cases
she discussed, the future perpetrator had suffered from a
romantic breakup or unrequited love. Klein found that
the shooters’ motives were at least partly linked with
rejection, jealousy, frustration, perceived failure with
girls or a desire to protect them. Thus, in numerous
incidents, the perpetrators targeted and killed girls they
felt had rejected them.
In general, school shooters have been commonly
described as loners (Meloy et al., 2001; Newman et al.,
2004) with few friends or none at all. As Levin and
Madfis (2009) stated in their description of mounting
strain in the personal development of school shooters,
long-term frustration experienced early in life can lead
to social isolation. With social support systems lacking,
strain can become unbearable. Newman et al. (2004)
argued that the perpetrators were not loners so much
as “failed joiners, who always tried to fit in” (New-
man & Fox, 2009, p. 19). By contrast, in shootings
that occurred on college campuses the perpetrators were
apparently more disconnected and had stopped trying
to integrate into groups (Newman & Fox, 2009).
Recently, Dumitriu (2013) divided school shoot-
ers into three distinct groups. Interestingly 71 of the
163 school shooters she analyzed had no friends,
experienced difficult relationships with girls, and suf-
fered repeated bullying at school. In contrast, 22 were
described as “perfect students” who had friends and
generally normal relationships, and who did not suffer
bullying. The third type (n= 70), typically 31 years old
or over, a former student who had been an outsider at
the school, is not discussed in detail by Dumitriu’s, who
drew the following conclusions from her study:
“School shootings are more complex than usually
portrayed in many studies and especially in media
accounts which proceed from the assumption that a
school shooter is a shy student who had been bullied
by some of his peers and whom he or she shot in an
episode of rage” (Dumitriu, 2013, p. 306).
The precise role of social rejection in school shoot-
ings remains a matter of controversy (Rocque, 2012).
While the studies cited above give evidence of the cru-
cial role social rejection plays in the development of
school shooters, others do not consider social dynamics
in their list of key risk factors. For example, Langman
(2009) argued that only 6% of the 35 attackers in his
case study targeted specific students who had picked on
them. Thus, even if the shooters were victims of bully-
ing, this did not necessarily mean that the mistreatment
caused the attack. Dutton, White, and Fogarty (2013)
found evidence of paranoid thinking in self-reports of
mass shooters who explained that they had suffered
psychologically from rejection. While there was little
evidence of actual bullying incidents, offenders seemed
obsessed with the idea that they had been rejected by
their peers (Dutton et al., 2013). The three types of rejec-
tion listed by Leary et al. (2003) – ostracism, bullying,
and romantic rejection – were less evident in inves-
tigations of more recent school shootings (Weatherby,
Strachila, & McMahon, 2010). In fact, some researchers
suggested that in a number of cases school shooters had
been bullies themselves, and that others seemed quite
popular among their peers (Dumitriu, 2013; Fast, 2009;
Langman, 2009; Newman et al., 2004). At least 41% of
the shooters in the study done by Vossekuil et al. (2002)
socialized with mainstream students or were considered
mainstream students themselves.
Another form of marginalization still not properly
understood involves teachers and school administra-
tors (see also Bond¨
u & Scheithauer, 2014). Given the
fact that at least 150 parents, teachers, administrators,
coaches and other adults were killed or wounded in
166 school shootings that occurred between 1979 and
2009 (Klein, 2012), we need to explore more closely
the relationship between the perpetrators and their adult
victims. The conflicts between perpetrators and their
teachers and other school staff varied in manner and
intensity. In some cases teachers and administrators had
merely ignored or dismissed the bullying suffered by the
future schoolyard assailant and had failed to intervene,
while in others the teachers played a more active role, at
least in the eyes of the perpetrators. Klein (2012) stated
that in at least 24 of the 166 incidents she had studied,
the perpetrators said they were responding to what they
saw as academic or disciplinary injustices inflicted on
them. In particular, European shooters and perpetrators
on college campuses had frequently received negative
school reports or punishment of various kinds (Bond¨
u,
2012; Bond ¨
u & Scheithauer, 2014; Fox & Savage, 2009;
Hoffman, Roshdi, & Robertz, 2009). Thus, being sus-
pended from school may have been the reason for some
shootings in which the attackers targeted teachers and
administrators (Newman et al., 2004).
As most students who experience rejection, even
those who are bullied and ostracized by peers or teach-
International Journal of Developmental Science 8/2014, 3–24 5

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
ers, do not resort to lethal violence, it seems likely that
additional risk factors play a role. Nevertheless most
of the primary studies conducted thus far confirm that
many perpetrators had experienced some kind of social
rejection.
The purpose of this paper is to systematically review
and summarize research findings on school shootings,
focusing on the role of social dynamics. We have con-
sidered all primary studies which deal with at least two
cases of school shootings, defined as offenses commit-
ted by a current or former student who deliberately
chooses his or her school or university as the site of an
attempt to kill one or more people (cf. Bond¨
u, 2012).
Since findings relating to social dynamics in the research
on school shootings remain inconsistent, we have not
formulated any hypotheses in advance, but remain open
to any links that might become apparent in the data.
Method
Our review of the literature is based on criteria for-
mulated in the PRISMA statement (Moher, Liberati,
Tetzlaff, Altman, & the PRISMA Group, 2009), our aim
being to identify and examine all the relevant research
studies in a systematic manner. Hopefully this will help
resolve disagreements arising from contradictory find-
ings in the various studies (Klassen, Jadad, & Moher,
1998) and will contribute to a more objective appraisal
of research results (Egger, Smith, & O’Rourke, 2001).
Literature Search
We only included those primary studies in our review
which (1) focused on school shootings, (2) contained
information on the personal development of the perpe-
trators, providing both qualitative as well as aggregated
data, (3) appeared between January 1990 and Decem-
ber 2013. We began our search with the help of the
electronic databases PsychNET, PsychINFO, Pubmed,
Scopus, Google Scholar, and ScienceDirect, using the
search terms school shooting, homicidal violence, ram-
page, severe targeted school violence, and amok.Next
we went through the reference lists we found in those
studies which met our criteria, and contacted experts in
the field. In the end we collected a total of 454 journal
articles.
Inclusion Criteria and Study Selection
In a second step two raters read the abstracts of the
454 publications and gathered further information from
the full publication to decide whether a certain publi-
cation was eligible. Studies were included if they met
the following criteria: (1) they dealt with at least two
school shootings committed by a current or former stu-
dent who deliberately chose their school or university as
the site of their planned attack, (2) they were published
in English or in German, (3) they met certain research
standards (we excluded articles and papers which sim-
ply listed incidents or had appeared in newspapers or
on the Internet). Studies were excluded if (1) the cases
discussed did not involve at least two incidents that
occurred in an academic institution, but instead focused
in incidents like family killings, (2) secondary analyses
were not based on a sampling of cases, but simply sum-
marized previous research, (3) they were dissertations,
(4) they focused exclusively on global questions con-
nected with school shootings, like prevention efforts or
prevalence.
All the studies that fulfilled our criteria were marked
for review. The final decision on whether to include a
particular article was made after reading the full text. In
the end a total of 35 studies were selected. Table 1 gives
reference information and lists the key characteristics
of the studies.
Characteristics of Primary Studies Reviewed
Twenty nine of the 35 primary studies were in English
and dealt primarily with cases in the U.S., only occa-
sionally discussing cases elsewhere [primary studies1:
1–18, 20–22, 24, 28, and 30–35]; five were in German,
four of which focused exclusively on cases in Germany
[19, 23, 25, and 29], one on cases in other countries [26].
One of the studies in English focused on Finnish cases
[27]. Twenty five primary studies [1, 4, 7, 8, 10–14,
16, 17, 20–27, 29–33, and 35] contained detailed qual-
itative data for each incident analyzed, while ten other
studies named the cases they analyzed, but did not pro-
vide detailed information on the incidents [2, 3, 5, 6,
9, 15, 18, 19, 28, and 34]. In examining these studies
we only considered aggregated information relating to
social conflicts experienced by the perpetrators.
The data sources referred to in the primary stud-
ies varied considerably. In four studies no information
could be obtained on data sources [2, 8, 15, and 23]. The
analyses in six primary studies relied solely on media
accounts [6, 10, 12, 22, 24, and 33], while in six studies
the data was collected from the media, scientific articles
and previous research [11, 16, 21, 26, 30, and 32]. In a
total of 19 studies [1, 3–5, 7, 9, 13, 14, 17–20, 25, 27–29,
1Number of primary study according to Table 1
6 International Journal of Developmental Science 8/2014, 3–24

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
Table 1
Included Primary Studies
Reference Data source for analysis Definition for incidents Number of cases analyzed General position on social dynamics in
information school shootings
1. Mc Gee &
DeBernado
(1999)
“Limited amount of fragmentary data
and unconfirmed anecdotal accounts
derived from official police reports and
the popular media” (p. 7)
Non-traditional shooting incidents which
occurred in the U.S. middle and high
schools
18 cases (1993–2001) Precipitant discipline/rejection, bullying:
100%; social outcast: 100%; teased/felt
victimized: 100%
2. Band &
Harpold (1999)
No information available No information available 8 cases “They felt rejected by others and sought
revenge or retaliation for real or perceived
wrongs done to them. They appeared to be
loners, average students, and sloppy or
unkempt in dress. They had a propensity to
dislike popular students or students who
bully others.” (p. 14)
3. O’Toole (1999) “A summary of the incident, tapes or
transcripts of interviews with the
offender(s), witness statements,
interviews with persons who knew the
students and families and provided
information about offenders’
background, counseling and
psychiatric reports and evaluations,
school records and class work,
interviews with law enforcement and
school personnel.” (p. 34)
“18 school shooting cases around the
country (in 4 the student or students
involved planned a shooting and made
significant preparations, but were
detected and preempted by law
enforcement). The cases involved
single and multiple offenders.” (p. 34)
18 cases (14 actual, 4
preempted)
The threat assessment model lists certain
types of behavior, personality traits, and
circumstances in the family, school, and
community environment that should be
regarded as warning signs. Social
dynamics connected to school shootings
might be: Failed love relationship,
injustice collector, closed social group,
pecking order among students, peer
groups.
4. Verlinden,
Hersen, &
Thomas (2000)
Convictions in 4 cases, information
obtained directly from the courts
(interviews, videotapes, court
documents) in 2 cases, search of local
and national media
“Multiple victim violent assaults in
schools that have occurred during the
past 3 school years and have involved
use of firearms.” (p. 27)
10 cases (1996–1999) In nine of ten cases the perpetrator felt
rejected by peers and persecuted; in eight
of ten cases the perpetrator was socially
isolated.
5. Meloy, Hempel,
Mohandie,
Shiva, & Gray
(2001)
Courtroom testimony, scientific articles,
academic books, video and audiotapes,
interviews, media
Adolescent mass murder: “Intentional
killing of at least three victims (other
than the perpetrator) in a single
incident by an individual 19 years old
or younger. Individuals who used a
firearm, cutting instrument, or blunt
object with or without other weapons.”
(p. 720)
34 (8 school shootings in the
sample) adolescent mass
murder (1958–1999)
70% loners (“self-labeling, if he showed a
marked tendency to spend time alone
rather than with others”), 17% history of
bullying others, 43% were bullied by
others (“long-term victimization of a
student by his peers, included both
physical and psychological attacks”).
(p. 721)
6. Danner &
Carmody (2001)
Major newspapers, articles published
within two weeks of the event
Sample of infamous school violence
cases included here in media accounts
of the Jonesboro shootings.
9 cases (1997–1999) “The most frequent explanation for the
shootings was “response to bullying”. This
frame captures explanations that described
the offenders’ violence as a response to
being picked on or bullied by fellow
students.” (p. 103)
International Journal of Developmental Science 8/2014, 3–24 7

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
Table 1
(Continued)
Reference Data source for analysis Definition for incidents Number of cases analyzed General position on social dynamics in
information school shootings
7. Kidd & Meyer
(2002)
“Interviews with offenders, family
members, victims, witnesses, or others
closely involved in the case, reports
from the more reputable national,
regional, or local news sources.” (p. 5)
Offenses in a rural or small town
community: “Cases were limited to
those which involved multiple student
fatalities on school campuses during
normal operating hours.” (p. 3)
8 cases (1996–1999) “A feeling of peer rejection was a
characteristic shared among six of the
eight offenders. A few offenders targeted
particular victims who had teased them, or
they reported seeking revenge for the lack
of respect and maltreatment they received
from peers. Feelings of rejection also
occurred after several offenders were
apparently spurned by a love interest.”
(p. 7)
8. Waterman
(2002)
No information available “The recent school shootings were
premeditated high-profile affairs in
which victims were more or less
chosen at random.” (p. 4)
8 cases “In each case, bullying and tormenting
occurred on a daily basis, yet teachers and
administrations did nothing, or even
worse, punished the victims for
retaliating.” (p. 19)
9. Vossekuil, Fein,
Reddy, Borum,
& Modzeleski
(2002)
Investigative, school, court, and mental
health records, supplemental
interviews with 10 of the perpetrators.
“Incidents of targeted violence in school
settings: School shootings and other
school based attacks where the school
was deliberately selected as the
location for the attack and was not
simply a random site of opportunity.
Current student or recent former
student attacked someone at his or her
school with lethal means (a gun or a
knife) and the student attacker
purposefully chose his or her school as
the location of the attack.” (p. 13)
37 cases, 41 perpetrators
(1974–2000)
“41% socialize with mainstream students or
were considered as mainstream student
themselves, 27% were considered to be a
part of a fringe group, 12% had no close
friends, 24% loners, 71% felt persecuted,
bullied, threatened, attacked or injured by
others prior to the incident.” (p. 24)
10. Harter, Low, &
Whitesell
(2003)
Media accounts Major School shootings since 1996 10 cases, 12 perpetrators “A common feature in the histories of the
school shooters has been that they each
had a history of being humiliated by peers,
a romantic other, or a teacher. In certain
cases, the actual precipitating event was
such victimization, leading to revenge.”
(p. 5)
11. Kimmel &
Mahler (2003)
Analysis of the extant commentary and
literature on school violence and
secondary media reports
“Random school shootings: A young
student opens fire on school grounds,
apparently randomly, and shoots
teachers and students.” (p. 1456)
31 cases (1982–2001) “Nearly all of the school shooters had stories
of being constantly bullied, beat up, and,
most significantly for the analysis
“gay-baited”. Five of the school shooters
had what they felt was serious girl trouble,
especially rejection.” (p. 1445, p.1454)
8 International Journal of Developmental Science 8/2014, 3–24

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
12. Leary,
Kowalski,
Smith, &
Phillips (2003)
National news media, newspapers from
the local area, world wide web sites
“All well-documented cases of school
violence in the U.S. Shooting incident
must have occurred at a school during
the school day, must have been
perpetrated by students and resulted in
injury or death to at least one student.
Incidents in which the only victims
were nonstudents were not considered
because we were explicitly concerned
only with student’s aggression toward
their peers.” (p. 204)
15 cases (1995–2001) “Most of the shooters (12 of 15) had
experienced an unusually high amount of
bullying and ostracism, that was
particularly relentless, humiliating, and
cruel. In addition many cases involved
ongoing ostracism that left the perpetrator
on the periphery of the school’s social life.
In about half of the episodes the
perpetrator had also experienced a recent
rejecting event. In only two cases was
there no evidence whatsoever that the
perpetrator had been rejected or mistreated
by other people.” (p. 210)
13. Moore, Petrie,
Braga, &
McLaughlin
(2003)
Records, court reports “Young people arming themselves and
opening fire on their schoolmates and
teachers, killing or seriously injuring
them. The incidents occurred in the
hallways and common areas of schools
and at a school-sponsored event.”
(p. 249)
8 cases (1991–1999) “In 4 of 8 cases the perpetrator was a victim
of bullying, in the same amount of cases
he was a bully himself, and in three cases
he suffered from a recent peer rejection.
Yet the shooters perceptions seemed to
have little basis in reality, or, if they were
real, they were not widely understood and
shared by others. They were not being
threatened with physical violence at the
time they shot. In three cases there have
been disciplinary problems with teachers.”
(p. 251)
14. Newman, Fox,
Harding, Mehta,
& Roth (2004)
National Database of school-associated
violence deaths by the Centers for
Disease Control and Prevention, report
by the U.S. secret service, own dataset
by media accounts and other case
studies
“Deadly assaults on an institution-the
school. An institutional attack takes
place on a public stage before an
audience, is committed by a member
or a former member of the institution,
and involves multiple victims, some
chosen for their symbolic significance
or at random.” (p. 232)
29 cases (1974–2002) A school shooters are not all loners and they
are not all bullied, but nearly all
experienced ostracism and social
marginality: 5% were popular, preppies,
jocks, or athletes; Loner 11–34%; no close
friends: 12%; Fringe or outcast: 52%;
Marginal 78–96%; Victimized: 53%;
Bullied or teased: 63%; Felt
victimized/marginalized: 67–71%;
Masculinity challenged: 63%. There is
another form of marginalization: Being
pushed out of the institution altogether,
which explain some shootings in which the
attackers more explicitly targeted teachers
and administrators.” (pp. 239–242)
15. Fox, Levin, &
Quinet (2005)
No information available Table with selected school massacres:
Adapted and reprinted with permission
from Time Magazine, 31.5.1999.
8 cases There has been teasing in six of eight cases,
romantic rejection in half of the cases and
in one case the perpetrator was expelled
from school before he went on the
massacre.
International Journal of Developmental Science 8/2014, 3–24 9

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
Table 1
(Continued)
Reference Data source for analysis Definition for incidents Number of cases analyzed General position on social dynamics in
information school shootings
16. Klein (2006) “All news coverage and research was
systematically gathered on school
shooting cases.” (p. 45)
“Cases were at least two people have
been killed and/or there were four or
more victims.” (p. 45)
10 cases (1996–2001) “Gay harassment consistently appears as a
motivating factor. In five of the ten cases,
students were routinely called gay, faggot,
and the like. In nearly all of the shootings,
boy’s masculinity was systematically
challenged via harassment and/or
name-calling.” (p. 46)
17. Fast (2008) Documentary evidence, articles from
community newspapers, legal
documents, and personal accounts
“The shooter and the victims had to be
on school grounds during the crime,
the assailants had to be adolescents,
and the victims had to number two
besides the shooter, had he committed
suicide, too.” (p. 14)
15 cases “Nearly every case involves bullying, and the
obvious chain of causation, from bullying
to humiliation to rage to revenge, made
this one of the first characteristics of
school rampage shooters to be identified.”
(p. 13)
18. Fox & Savage
(2009)
“Data gleaned from the FBI’s Uniform
Crime Reporting program and the U.S.
Department of Education’s records
mandated by the Clery Act, as well as
detailed media reports gathered from
searching electronic newspaper
databases.” (p. 1467)
“Shooting involving multiple fatalities
on college campuses in the United
States.” (p. 1476)
76 cases (1990–2008), 13 are
named
“Shootings at high schools often precipitated
when students feel bullied or persecuted by
their classmates and/or teachers. However,
the perpetrators of mass shootings at
colleges and universities are often graduate
students-older individuals who turn to
violence in response to what they perceive
to be unbearable pressure to succeed or the
unacceptable reality of failure.” (p. 1475)
19. Hoffmann,
Roshdi, &
Robertz (2009)
Crime and court records Severe targeted school violence is
defined as an intentionally deadly
offense against specific persons or
group of people. The school as the site
to carry out the attack is chosen by
purpose.
7 cases (1999–2006) 71.4% of the German offenders have been
teased and humiliated (and have been
sensitive to this), 57.1% have been
perceived as loners (but had few friends
from time to time), 57.1% withdrawal
from their social contacts prior to the
attack, 85.7% had disciplinary conflicts
with teachers. Imminent to the homicide
the perpetrator experienced a rejection (by
a girl, classmates or suspension from
school).
20. Newman &
Fox (2009)
“Lists of shootings from the Virginia
Tech Review Panel’s compendium of
fatal school shootings in the United
States and a variety of media
compilations of such events.” (p. 3)
“The location of the incident is a public
stage, either on the school property or
at a school related function, the
shooters must be current or former
students of the school, there must be
multiple victims (although the injuries
do not have to be fatal) or at the very
least, multiple targets.” (p. 2)
9 cases (2002–2008) Social marginalization is one of the five key
factors: “The protagonists were loners or
fringe figures (certainly not popular),
teased/bullied, subjected to masculinity
tests that they failed, or felt marginalized
even if there is little evidence to this
effect.” (p. 9)
10 International Journal of Developmental Science 8/2014, 3–24

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
21. Larkin (2009) Academic research, various media
outlets and Internet sites, archives of
local and regional newspapers
“A student or a former student brings a
gun to school with the intention of
shooting somebody, the gun is
discharged and at least one person is
injured, and the shooter attempts to
shoot more than one person, at least
one of whom was not specially
targeted.” (p. 1310)
36 cases, 38 perpetrators, 2
are named (1974–2012)
52.6% were motivated by revenge against
bullying, harassment, and intimidation by
their peers. The shooters peer status was
determined in 25 of 38 cases. Of those
84% were either outcasts or loners who
were rejected by their peers either as
individuals or as members of identifiable
low-status collectivities.
22. Kalish &
Kimmel (2010)
Sampling of media reports “A white boy (or boys) brings
semi-automatic rifles or assault
weapons to school and opens fire on
classmates and teachers, seemingly at
random. School shootings that
culminate in the suicide of the
assailant(s).” (p. 452)
3 cases, 4 perpetrators
(1999–2008)
“Nearly all had stories of being constantly
bullied, beaten up, and “gay baited”.”
(p. 455) “In all three cases, the shooters
felt both victimized by others and superior
to them.” (p. 459) “It was not just that they
were bullied and harassed and intimidated
every day, it was the administrations,
teachers, and community colluded with it.”
(p. 462)
23. Wieczorek
(2010)
No information available Deadly violence that occurred in schools 9 cases (1964–2009) In seven of nine cases social risk factors
(mostly disciplinary problems) could be
conducted.
24. Weatherby,
Strachila, &
McMahon
(2010)
Newspapers from the day of the shooting
to seven days later
Each case examined was committed by a
middle school or a high school student
in a learning environment within the
United States.
12 cases (2001–2008) “There is less evidence of extreme teasing or
ostracism than Leary, et al. (2003) found
in their study. While these findings differ,
both studies do illustrate that rejection,
teasing, and ostracism can be prominent
factors in school shootings. There may be
other factors, however.” (p. 11)
25. Roshdi &
Hoffmann
(2011)
Court records, interviews See Vossekuil et al. (2002) 10 cases (1999–2009) Six out of ten perpetrators faced problems
within the school context (disciplinary
problems, conflicts with teachers), two
perpetrators experienced bullying, and
three perpetrators were perceived as
loners.
26. Brumme
(2011)
Scientific research, media accounts Current or recent former student, who
purposefully selected and targeted his
or her school. Multiple victims
(targeted or killed by random), that
were chosen by their symbolic
connection to the institution.
14 cases (1995–2009) Most of the perpetrators had been teased or
humiliated by peers. Not all were loners,
mostly they had a few but lukewarm
friendships. Offenders felt isolated and
marginal.
27. Kiilakoski &
Oksanen (2011)
“Pre-investigation reports by the Finnish
police and the reports by the
government commissions created to
investigate the shootings provided
background material” (p. 32)
No information available 2 cases (2007–2008) “All Finnish school shootings were
associated with negative and violent
experiences in school. The shooters tended
to feel marginalized and to lack peer group
approval in their school careers.” (p. 33)
International Journal of Developmental Science 8/2014, 3–24 11

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
Table 1
(Continued)
Reference Data source for analysis Definition for incidents Number of cases analyzed General position on social dynamics in
information school shootings
28. Lankford
(2012)
Previous scholarship, government
reports, media reports
“The study was designed to analyze
terrorism, rampage, workplace, and
school attacks that involved suicide
attempts and occurred in the U.S.”
(p. 257) Definition of school
shootings: “Individuals who attacked
at a school, college, or university that
they had ever attended.” (p. 258)
16 cases (1991–2008) 75% Social Marginalization (loners, bullied,
teased, did not have close friends, felt
socially marginalized or socially isolated);
88% Work/school Problems (struggling to
succeed in work or school, were angry or
upset about sth. at work or school, or had
been suspended, fired or otherwise
disciplined at work or school).
29. Bannenberg
(2012)
Police and court records Rampage shootings of adolescents with
unknown motive
18 cases of adolescent mass
murderer, 11 cases
occurred in an institution
(1992–2009)
Perpetrator felt humiliated and bullied, but
also withdrawn from social contacts and
did not want to fit in. In some cases
disciplinary problems with teachers
occurred. In four of eleven cases, shooters
were perceived as loners, two experienced
a romantic rejection prior to the attack.
30. Schiller (2013) Articles, biographies, books, media,
interviews, documentaries, reports,
and Secret Service findings
“Any incident where a current student or
recent former student attacked
someone at his school with lethal
means, and where the student attacker
purposefully chose his school as the
location of the attack.” (p. 102)
12 cases (1978–2000) “Shooters are typically young men who feel
marginalized. Many of the shooters state
that they were bullied or humiliated at
school by their peers or belittled by adults
who were close to them. In 11 of 17 cases
bullying was found, in one case the
perpetrator was seeking revenge for the
punishment of a teacher.” (p. 105)
31. Langman
(2013)
Student journals, police records, court
documents, official reports, books, and
articles in both scholarly journals and
news outlets
“Attacks at schools in which there were
multiple victims. The victims included
people who were shot randomly, as
well as some who were specifically
targeted. The perpetrators in almost
every case were either current or
former students at the schools they
attacked. In two cases the perpetrators
were not students of the schools they
attacked.” (p. 131)
35 cases (1975–2011) “Only 6% targeted a specific student who
had picked on them, more frequently they
targeted family members, females, and
school personnel. Even if they were teased
and/or bullied, this does not mean that the
mistreatment caused the rampage.”
(p. 145)
12 International Journal of Developmental Science 8/2014, 3–24

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
32. Dutton, White,
& Fogarty
(2013)
Perpetrators’ diaries and web-sites No information available 4 cases, 3 of them school/
university shooter
(1999–2011)
“There are several elements in the thinking
of school shooters that suggest a paranoid
personality disorder blended with
malignant narcissism. Paranoid
personalities have a pervasive mistrust of
others, are suspicious and hyper-vigilant
for “disrespect” and actively seek
“evidence”. They persistently feel
wronged - a victim of social injustice and
are seething with anger when they recount
these injustices. There is little reference to
specific experiences of being a victim of
bullying.” (p. 551)
33. Madfis &
Levin (2013)
Media accounts “Analyses includes only those
perpetrators who themselves, at the
time of the attack, were enrolled in or
were recently withdrawn from the
middle school, high school or college
that they targeted. Multiple human
targets were killed or injured on school
property by a student or recent former
student of the targeted school.” (p. 80)
12 cases (1999–2011) “The catalyst for school shooters is most
often a humiliating loss of face, a rejection
by a girlfriend, a loss of academic
standing, an eviction from a community of
peers, or even a major illness. At least 7 of
the 12 incidents involved long-term
bullying as an expression of the killer’s
chronic strain and longstanding desire to
get even. Bullying was not, however, the
only form of long-term frustration
experienced by school rampage shooters
in the international sample.” (pp. 82 ff.)
34. Dumitriu
(2013)
Governmental reports of inquiries into
these events, police reports, court
records, school records, newspaper
articles and archival documents, field
data (interviews and/or focus-groups)
“Multiple-victim act of extreme violence
perpetrated on school premises,
generally by a school-related
perpetrator who carefully plans the act
in advance.” (p. 301)
160 cases, 10 selected for in
depth-qualitative case
studies (1900–2013), 31
perpetrators are referred
“The results revealed that school shootings
are much more complex than are portrayed
in most research studies and official
reports in the field, and according to which
the central character is a shy student who
had been bullied by some of his peers and
whom he shot in an episode of rage.”
(p. 306)
35. Malkki (2013) “Publicly available communication by
the perpetrators, official reports and
police investigation documents in the
incidents, media reports as well as
previous research on the incidents.”
(p. 188)
“At least partly indiscriminate shootings
perpetrated by former or current
students of that school. The student
has brought the gun to the school with
the intention of shooting somebody, at
least one person (other than the
shooter) has to be injured by a bullet,
and that the shooter attempts to shoot
more than one person, at least one of
whom was not specifically targeted.”
(p. 188)
28 cases (1999–2011) Three types of school shooters: “1. Political
school rampage shootings: The
perpetrators express angrily their
resentment of the way they have been
treated, arguing that they were pushed to
their act of vengeance. From the
description of the personal situation
emerges an image of a failed joiner, who
tried to fit in but was bullied, teased, and
rejected. At some point, they understood
that they did not even want to fit into the
majority and live in a world like that.”
(p. 200) 2. Columbine-Influenced
Shootings. 3. Isolated shootings.
International Journal of Developmental Science 8/2014, 3–24 13

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
31, 34, 35] analyses of cases were based on official
primary sources including police-, school-, court- and
mental health records. In addition, personal accounts
were consulted, along with interviews with family
members, school staff and perpetrators themselves.
In reviewing the 35 selected studies, we identified 126
different cases of school shootings, involving 128 per-
petrators. Of these 126 attacks, 112 attacks occurred in
schools, (88.9%), 14 (11.1%) on university campuses.
The earliest case included in the primary studies took
place in Volkhoven (Germany) in 1964, the latest in
Newtown (USA) in 2012. The mean age of perpetrators
was 19 (SD = 8.72) and the median age of perpetrators
was 16. The youngest perpetrator was six years old, the
oldest 62. A total of 121 shooter were males (94.5%)
while seven offenders were females (5.5%). Most of the
attacksoccurred(n= 97)inthe U.S. (75.3%).There were
14 in Germany (10.9%), four in Canada (3.1%), three in
Finland (2.3%), two in Brazil, and one shooting each in
eight additional countries (Argentina, Australia, Bosnia,
Greece, Hungary, Netherlands, Sweden, and Thailand).
One important task was to determine how often indi-
vidual perpetrators were included in the various primary
studies selected. Our results showed an unequal dis-
tribution of cases: 63 of the 128 perpetrators (49.2%)
were included in one study only, 22 (17.2%) in two,
33 in three to ten studies (25.8%), and, finally, ten
perpetrators were included in more than ten studies
(7.8%). The ten most frequently reported cases were
Eric Harris: Columbine 1999 (included in 22 stud-
ies); Dylan Klebold: Columbine 1999 (20 studies);
Luke Woodham: Pearl 1997; Michael Carneal: West
Paducah 1997; Andrew Golden and Mitchell Johnson:
Jonesboro 1998; Kipland Kinkel: Springfield 1998 (17
studies); Barry Loukaitis: Moses Lake 1996 (14 stud-
ies); Thomas Solomon: Conyers 1999 (13 studies) and
Evan Ramsey: Bethel 1997 (11 studies).
Categories and Content Analysis
In order to analyze the social dynamics which might
have played a role in the various cases, we chose those
which contained detailed information on social posi-
tion, as well as on personal relationships and conflicts–
questions like whether the perpetrator had been a loner,
had been bullied or had experienced conflicts with
teachers. As ten studies presented only aggregated data
[2, 3, 5, 6, 9, 15, 18, 19, 28, and 34] particular categories
were not applicable in all of the 126 cases. We excluded
those studies which contained either no detailed quali-
tative information on the cases or only aggregated data.
In addition, four cases were excluded from analysis,
either because they did not fit our definition of a school
shooting. In addition, four cases were excluded from
analysis, either because they did not fit our definition of
a school shooting or the attack had been thwarted. As
stated in the introduction, we define school shootings
as offenses committed by a current or former student
who purposely chooses his or her school or university
campus as the site of an attempt to kill one or more
persons.
The list was narrowed down to 65 school shootings,
committed by 67 perpetrators discussed in 25 studies,
including 59 attacks in schools (90.8%) and six on uni-
versity campuses (9.2%). The mean age of perpetrators
included in this reduced sample was 18 (SD = 7.38)
and the median age of perpetrators was 16 years. The
youngest perpetrator was 11 years of age, the oldest
62. Sixty three perpetrators were males (94.0%) and
four were females (6.0%). The percentage of cases
from the USA was slightly lower compared to the ini-
tial sample (46 cases = 70.8%), and the perpetrators in
the reduced sample came from a smaller number of
countries (Germany: nine cases = 13.8%; Finland: three
cases = 4.6%; Canada and Brazil: two cases = 3.0%;
Australia, Greece, and Thailand: one case each). The
average case involved 2.83 fatalities (excluding perpe-
trators’ suicide) and 5.81 injuries. Twenty perpetrators
(29.9%) committed suicide.
In the most recent list to be published, B¨
ockler,
Seeger, Sitzer, and Heitmeyer (2013) attributed 63% of
school-shootings worldwide to the U.S. In our sample
the proportion is slightly higher, at 70.8%. In contrast
to many primary studies (e.g. McGee & DeBernardo,
1999; Newman et al., 2004; Vossekuil et al., 2002)
we included female perpetrators in our sample (n= 4).
The majority of perpetrators (83.6%) were adolescents
(between 12 and 21 years of age) – a common find-
ing in the literature on school shootings. The average
of 2.83 dead and 5.81 injured victims per attack corre-
sponds roughly to the average rate of victims per offense
between 2000 and 2010 as reported by B¨
ockler et al.
(2013). Likewise, the proportion of perpetrators in our
sample who committed suicide (29.9%) was approx-
imately the same as that reported by B¨
ockler et al.
(2013), with 33 out of 123 perpetrators (27%) taking
their own lives. Thus we can conclude that our sample
is in line with findings which are commonly found in
the literature on school shootings.
In analyzing the contents of case reports, we devel-
oped a coding scheme based on descriptions of risk
factors which frequently appear in the literature. All
14 International Journal of Developmental Science 8/2014, 3–24

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
categories and concepts bearing on the perpetrator’s
social position, personal relationships and conflicts
were collected from the literature and consolidated in
a coding scheme.
The first step in our analysis was to select from
the primary studies all qualitative information relating
to social dynamics and to list it according to case. If
two perpetrators carried out an attack, we analyzed the
social dynamics observable in each one. Thus, in the fol-
lowing remarks the term “cases” refers to the number of
perpetrators, rather than the number of incidents. Sec-
ondly, two experts on school-shooting research worked
independently, using our coding scheme in rating infor-
mation. Subsequently they discussed the ratings they
obtained, case by case, until they reached consen-
sus. Thus, analyzing the reliability of ratings was not
appropriate. In reconstructing the perpetrator’s profile,
specific categories like “loner” or “bullying victim”
were described as either “present” or “not present”. We
applied the classification “inconsistent data” in cases
where the information contained in the primary studies
was contradictory.
In examining the social dynamics in each case, we
looked at the following categories:
•Loner (self-report): The perpetrator sees him- or
herself as someone who prefers to be alone and
consequently either does not actively seek human
interaction or deliberately avoids it. The sources we
consulted included diaries and interviews.
•Loner (external report): Here we consulted external
sources, including the accounts of peers, teachers,
and other individuals in the case descriptions.
•No friends: In describing the social relationships of
school shooters, many researchers have stated that
they had no friends at all. As cultural differences can
be reflected in definitions of what a friend is, we only
applied this category when there was evidence in
case descriptions that the perpetrator had no friends.
•Physical bullying: Any kind of bullying resulting in
physical pain or damage to property. There was no
way of knowing exactly how the term was defined in
the various primary studies we examined, (e.g. as it
has been defined by Olweus, 1994), so we accepted
the researchers’ use of the word in describing a per-
petrator or in referring to beatings suffered by the
perpetrator at the hands of peers.
•Verbal and/or relational peer rejection: This
includes teasing, name calling, gay bating, and
ostracism. Whether or not these forms of peer rejec-
tion are covered by the definition of bullying used
in the individual case studies remains unclear. Most
give no detailed information on the frequency or
duration of such incidents or on the relationship of
the perpetrator to those who rejected him or her. We
applied the term if the perpetrator had been teased,
picked on, “gay-baited”, or otherwise rejected by his
or her peers according to the information contained
in the primary studies.
•Romantic rejection: This describes a romantic
breakup or a case of unrequited love prior to the
school shooting, which apparently had a bearing on
it.
•Conflict with teachers or school system: Open con-
flicts with an individual teacher or other school
authorities, including the school principal (disci-
plinary or academic injustices, personal conflicts,
suspension from school).
•Bully: Evidence that the shooter had bullied others.
•Model student: We borrowed this category from
Dumitriu (2013), who described perfect (honored)
students as well-socialized individuals (male), who
have friends as well as relationships with girls,
and who are generally regarded as polite, mild-
mannered, respected, model students.
•Social marginality (self-report): Social marginality,
the most important feature postulated in the work
of Newman et al. (2004), refers to the perpetrator’s
own self-perception as extremely marginal in the
social world that matters to him.
•Urge for revenge: Revenge as a motive can provide
information on perpetrators’ suffering as a result
of social exclusion and/or personal conflicts. This
feature was included if studies explicitly mentioned
that revenge was a motive in the shooting.
In addition to these categories, we also reviewed
evidence for concepts often used in school shooting lit-
erature. Concepts combined several categories within a
theoretical framework:
•Any marginalization: Under this category, Newman
et al. (2004) list evidence that the “protagonist was
(a) loner or fringe figure (certainly not popular), or
(b) teased/bullied, or (c) subjected to ‘masculinity
tests’ that he failed, or (d) felt marginalized even
if there is little evidence for this effect” (Newman
& Fox, 2009, p. 9). In order for us to give a rat-
ing on this, information relating to the categories
“loner” (either self or external description) and/or
“physical bullying” and/or “verbal and/or relational
peer rejection”, and/or “social marginality” had to
be evident from the case descriptions.
International Journal of Developmental Science 8/2014, 3–24 15

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
•Peer rejection: Here Leary et al. (2003) include
physical bullying, teasing and other forms of peer
rejection (verbal, romantic, and other). We con-
cluded that this applies when there is evidence of
“physical bullying”, and/or “verbal and/or relational
peer rejection”, and/or “romantic rejection” in the
case description.
•Any interpersonal conflicts: This includes all inci-
dents that we interpreted as personal conflicts
(“physical bullying”, and/or “verbal and/or rela-
tional peer rejection”, and/or “romantic rejection”,
and/or “conflict with teachers or school system”).
Findings
The rating of information – and subsequently the inter-
pretation of our findings – was hampered by the fact that
the primary studies we selected did not always include
information for every category which was of interest to
us. In such cases there was no way of knowing whether
the researchers were unable to access the information
we wanted or whether they had found no evidence of
certain features relating to social dynamics. For this
reason we decided to pursue a two-pronged strategy
in analyzing individual studies. In a so-called “con-
servative analysis” we interpreted only the information
which the authors of the primary studies provided. In a
“progressive analysis” we assumed that if the authors
of the primary studies did not report any informa-
tion which fell under a specific category that meant
that they did not find any evidence of its presence.
In this case we rated “no information”, meaning “no
evidence”. Assuming that the authors of such primary
studies would normally report all of the information
that emerged and was regarded as relevant, we use the
“progressive analysis” in the following description and
interpretation.
Regarding the perpetrators’ social position within
their respective social networks, discrepancies were
discovered between their self-appraisals and those
given by others. In 47.8% of cases, perpetrators were
described as loners or social outcasts. Some degree of
social marginalization, as reported in external descrip-
tions, was found in 85.1% of cases. In contrast to these
external reports, only in 23.9% of cases did perpetra-
tors describe themselves as loners, and in only 55.2% of
cases was there evidence that perpetrators themselves
had felt socially marginalized. In addition, in eight cases
(11.9%) feelings of marginalization, in five cases feel-
ings of being a loner were explicitly excluded in the
primary studies. Thus, an all-encompassing social iso-
lation/marginalization of the perpetrator is apparently
relatively rare. Additionally, there was almost no case
in which the perpetrator had no friends at all (4.5%,
n= 3). However, this finding does not mean that perpe-
trators were popular among their peers. In most studies
researchers found that the perpetrators had few friends,
with only two out of 67 perpetrators (3.0%) described
as being a model student (cf. Dumitriu, 2013).
We found little evidence that there had been physical
bullying (29.9%), and in 31.3% of the cases, physical
bullying was explicitly excluded. Other forms of peer
rejection such as teasing, provoking, name calling and
“gay bating” were reported more often (53.7%). In only
10 cases (14.9%) they were explicitly excluded. These
results indicate that there are school shooting cases in
which bullying definitely played no role, while in nine
cases (13.4%) perpetrators were described as bullies,
without being bullied themselves.
In connection with peer rejection, romantic rejection
must also be considered. Leary et al. (2003) mentioned
romantic rejection as a third category of peer rejection
– alongside teasing and ostracism – and found this form
of rejection in about half of the cases analyzed. In our
review we found evidence for romantic rejection in only
29.9% of the cases. It is worth noting that in some cases
(n= 6), romantic rejection was apparently the only form
of interpersonal conflict within the network of relation-
ships at school prior to attack. Altogether, in 67.2%
of the cases, we concluded that peer rejection cursive
instead in some form – whether physical bullying, ver-
bal abuse, or romantic rejection – was present.
Researchers have commonly neglected the signif-
icance of student-teacher problems and interaction
(Bond¨
u & Scheithauer, 2014). In our review we found
that conflicts with teachers (in 43.3% of the cases)
prior to the attack might have been a key factor influ-
encing the negative development of a future assailant.
In 88.1% of the cases there were interpersonal prob-
lems/conflicts in some form (bullying, teasing, romantic
rejection, and/or conflicts with teachers) within the
school environment. In four cases (6.0%), findings were
inconsistent, and in only two cases was there no evi-
dence of such problems or conflicts.
Analyzing possible motivating factors can also tell us
something about the self-perception of the perpetrators
with regard to social dynamics prior to their attack. If
perpetrators saw their shooting as revenge for mistreat-
ment suffered at the hands of significant people in their
social environment, we can assume that social relation-
ships were a source of suffering. In our review, scholars
described urge for revenge as a motive for 26 perpe-
16 International Journal of Developmental Science 8/2014, 3–24

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
trators (38.8%). In one case revenge as a motive was
explicitly excluded. In the majority of cases (59.7%)
no information was given on the perpetrator’s motives.
Discussion
Our study indicates that, in analyzing the impact of
social dynamics prior to a school shooting, it is impor-
tant to include different kinds of relationships and to
examine how they are interconnected. As stated above,
bullying and peer rejection are very important, but they
are not the only forms of conflict. Conflicts with teach-
ers – and with the school system as a whole – are
also of key significance. In addition, the perpetrators’
social position and subjective perception of victimiza-
tion and marginality can be crucial factors on the path
toward violent school attacks. For our content analysis
of school shooting cases reported in primary studies, we
used common categories and concepts of risk factors
representing significant experiences within the domain
of “social dynamics”. From our perspective, these cate-
gories rated with our coding scheme can be allocated to
– at least – three different superior classes of concepts:
Under one superior class we place conflicts and other
forms of negative interaction experienced by the per-
petrators (conflicts with peers and teachers, bullying,
romantic rejection); a second class includes their social
position within the social network at school (whether
loner, outcast, or model student) and finally perceptions
of the perpetrators themselves serve as a third superior
class of concepts (feelings of marginalization, of being
a loner, of wanting revenge).
From our perspective it seemed useful to discuss our
findings along these classes in order to allow a greater
understanding of the social dynamics prior to an inci-
dent of school shooting and to find categories which
could be functional equivalent.
Negative Interaction and Conflict
There is general consensus in the literature on school
shootings on the key role of negative events and dis-
turbed personal interaction prior to school shootings.
Our findings suggest that a broad constellation of events
and developments must be included in our analyses.
Only a minority of perpetrators (29.9%) experienced
actual physical bullying and/or other extreme forms
of mistreatment. More frequently perpetrators suffered
from a general feeling of being rejected by peers.
However, the overall rate of verbal and relational peer
rejection we discovered was slightly lower (53.7%) than
reported in some of the individual studies included in
our review. Thus our findings are more in line with those
of Weatherby et al. (2010), who found ongoing teasing
or ostracism in only half of the cases in their sample.
Considering the prevalence of rejection in general,
expressed in various ways, (e.g. Hess & Scheithauer, in
press; Scheithauer, Haag, Mahlke, & Ittel, 2008; Scheit-
hauer, Hayer, Petermann, & Jugert, 2006), the question
arises as to whether this particular kind of negative inter-
action with peers might be a significant and necessary
risk factor in the development of the school shooter, as
is commonly maintained in the literature on the sub-
ject and in the media. It is important to note here that
some of the perpetrators were described as being bul-
lies themselves (13.4%). This finding casts doubt on
models which place sole emphasis on the victim sta-
tus of perpetrators (Levin & Madfis, 2009). Research
on bullying confirms that bullies are or were also vic-
tims (so called bully-victims; e.g. Hess & Scheithauer,
in press; Scheithauer et al., 2003; Scheithauer et al.,
2006). In our sample we found only one perpetrator out
of nine who was described as a bully and whose actions
included physical abuse, while five had experienced
peer rejection in some form.
Another form of negative social interaction which
might play a role is romantic rejection, often catego-
rized under the term “peer rejection”. In the cases we
investigated, romantic rejection (29.9%) did not appear
to be as widespread as Leary et al. (2003) stated in
their study (in about half of the cases). Nevertheless,
in some of the cases, romantic rejection and unrequited
love were the only forms of interpersonal conflict in
the perpetrator’s history (n=6). Nor did our findings
on peer rejection (in 67.2% of the cases) correspond to
those of Leary et al. (2003), who found peer rejection in
13 of 15 cases (86.6%). Instead they were in line with
the findings of Vossekuil et al. (2002), who discovered
evidence of bullying, ostracism, and social rejection in
over two-thirds of the cases analyzed.
Apart from peer rejection or disturbed relations
with friends, our findings suggest that another type
of negative social interactions within the school envi-
ronment must be taken into account (cf. Bond¨
u&
Scheithauer, 2014): In 43.3% of the cases, perpetra-
tors experienced ongoing conflicts with teachers and
school officials. This finding was remarkable, con-
sidering that researchers have commonly overlooked
student-teacher problems and interaction as a signifi-
cant risk factor in the development of later perpetrators
leading up to a school shooting. Furthermore, despite
International Journal of Developmental Science 8/2014, 3–24 17

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
Table 2
Findings1
Categories No Any Inconsistent Sum No Sum
evidence evidence finding2conservative information progressive
Loner (self report) N 5 16 21 46 673
Conservative 23.8% 76.2% 100.0%
Progressive 7.5% 23.9% 68.7% 100.0%
Loner (external report) N 8 32 4 44 23 67
Conservative 18.2% 72.7% 9.1% 100.0%
Progressive 11.9% 47.8% 6.0% 34.3% 100.0%
No friends N 29 3 4 36 31 67
Conservative 80.6% 8.3% 11.1% 100.0%
Progressive 43.3% 4.5% 6.0% 46.3% 100.0%
Physical bullying N 21 20 7 48 19 67
Conservative 43.8% 41.7% 14.6% 100.0%
Progressive 31.3% 29.9% 10.4% 28.4 100.0%
Verbal or relational peer rejection N 10 36 8 54 13 67
Conservative 18.5% 66.7% 14.8% 100.0%
Progressive 14.9% 53.7% 11.9% 19.4% 100.0%
Romantic rejection N 5 20 1 26 41 67
Conservative 19.2% 76.9% 3.8% 100.0%
Progressive 7.5% 29.9% 1.5% 61.2% 100.0%
Conflicts with teacher/school N 3 29 32 35 67
Conservative 9.4% 90.6% 100.0%
Progressive 4.5% 43.3% 52.2% 100.0%
Bully N 8 9 1 18 49 67
Conservative 44.4% 50.0% 5.6% 100.0%
Progressive 11.9% 13.4% 1.5% 73.1% 100.0%
Model student N 10 2 2 14 53 67
Conservative 71.4% 14.3% 14.3% 100.0%
Progressive 14.9% 3.0% 3.0% 79.1% 100.0%
Urge for Revenge N 1 26 27 40 67
Conservative 3.7% 96.3% 100.0%
Progressive 1.5% 38.8% 59.7% 100.0%
Social marginality (self report) N 8 37 4 49 18 67
Conservative 16.3% 75.5% 8.2% 100.0%
Progressive 11.9% 55.2% 6.0% 26.9% 100.0%
Any marginalization N 4 57 61 6 67
Conservative 6.6% 92.4% 100.0%
Progressive 6.0% 85.1% 9.0% 100.0%
Peer rejection N 9 45 6 60 7 67
Conservative 15.0% 75.0% 10.0% 100.0%
Progressive 13.4% 67.2% 9.0% 10.4% 100.0%
Any interpersonal conflict N 2 59 4 65 2 67
Conservative 3.1% 90.8% 6.1% 100.0%
Progressive 3.0% 88.1% 6.0% 3.0% 100.0%
1Findings are presented in the table as follows: The first line of every category shows the frequencies. The second line (conservative) indicates the
percentage for a conservative analysis of frequencies, interpreting “no information” to mean that information was lacking. The second line shows
the percentage for a progressive analysis of frequencies, interpreting “no information” as no evidence for the particular category. 2“Inconsistent
finding” = scholars made divergent statements about the evidence of an individual factor. 3N = 67 cases (perpetrators) with detailed information
regarding social dynamics.
all the literature on school shooters and their relation-
ship to the school system in general (Adams, 2000;
Staples, 2000; Thompkins, 2000), little attention has
been given to teacher-perpetrator relations in case stud-
ies on school shootings in the U.S. In contrast, in various
German studies (Bond¨
u 2012; Bond¨
u & Scheithauer,
2014; Hoffmann et al., 2009; Wieczorek, 2010) con-
flicts between teachers and future perpetrators, along
with feelings of injustice suffered at the hands of school
staff, are well documented. Thus, the German cases sug-
gest that loss of status within the school environment
as reference system is an important key factor. In our
review, we found evidence of this in U.S. cases as well
(41.7% of U.S. cases).
In conclusion, in most of the cases analyzed (88.1%),
perpetrators experienced conflicts and other forms of
18 International Journal of Developmental Science 8/2014, 3–24

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
negative social interaction within the school environ-
ment as reference system. Such negative interaction
included (1) physical mistreatment and/or social exclu-
sion by peers, expressed verbally or otherwise, (2)
romantic rejection, (3) feelings of being treated unjustly
by teachers and school officials. While in a few cases
(n= 7) all three types of negative interaction were
present, in 41 of the 67 cases examined only one
was found (bullying/verbal and relational peer rejection
n= 19; romantic rejection n=6; conflicts with teachers
n= 16). As these results indicate, it remains to be seen
whether these three types of negative interaction are
functionally equivalent on the development of perpetra-
tors. This is a hypothesis which might be investigated
in further research, in which case analysis should focus
on the personal relationship between the perpetrator and
selected victims before and during the act of violence.
Social Position
Typically school shooters are described as marginalized
loners, or social outcasts. Our results only partially con-
firmed these descriptions. While 47.8% of perpetrators
were characterized as loners or outcasts, 11.9% were
not. Thus, perpetrators were not completely isolated,
but in fact had social contacts and were not as lonely as
one would normally expect a “loner” to be. Moreover,
there is evidence that, in some cases, the perpetrators’
friends could be labeled “deviant”. Sometimes these
friends played a significant role in the social dynamics
that lead up to the attack, in that they either directly
or indirectly expressed support for the future perpe-
trator, either by procuring weapons, placing bets on
whether the attack would be carried out and/or increas-
ing the pressure to act by making light of threats. Thus,
in line with Vossekuil et al. (2002), we can conclude
that peer interaction and friendship patterns varied all
the way from being socially isolated to being popular.
On the other hand we disagree with Dumitriu (2013),
who claimed that a relevant number of school shoot-
ers were “perfect (honored) students”: in our sample
only two perpetrators fitted the description model stu-
dent. In contrast, Dumitriu identified 22 (13.5%) out
of a total of 163 school shooters as model students. It
should be noted, however, that methodological differ-
ences between her study and our review2make direct
comparison difficult.
2Her study did not allow for a direct assignment from character-
istics to cases, she included cases from 1900 to 2013.
Our findings confirmed the concept of “marginaliza-
tion” developed by Newman et al. (2004), with 85.1%
of the cases in our review fitting that concept. At the
same time, Newman combines self-reports of marginal-
ization and bullying on the one hand, and external
descriptions of the perpetrator as being a loner, on
the other, in one single concept. In our opinion, this
is problematic, as it prevents us from seeing how the
two are interrelated. It becomes impossible to deter-
mine whether the marginal position of the perpetrator
was the result of peer rejection or whether it was “self-
chosen withdrawal”. Thus, marginalization becomes a
“catch-all” category, which is very difficult to interpret.
Overall, there is some evidence that a majority of attack-
ers were marginalized among their peers, but we must
be careful in making generalizations. In our review we
even found descriptions which suggested that the per-
petrators were by no means loners or outcasts and were
not at all marginalized.
Subjective Perception of Perpetrator
Some authors argued that there was no objective
evidence of peer rejection in some cases, while the per-
petrators themselves felt rejection, a factor which might
have played a role in the decision to attack (e.g. Dutton
et al., 2013; Langman, 2013). This divergence might be
explained by personality traits such as narcissistic ten-
dencies or excessive sensitivity to perceived injustice
(Hoffmann et al., 2009). Following this presumption
one would expect reports of self-reported victimization
to outnumber external ones. However, our findings sug-
gest the opposite: Perpetrators described themselves as
loners less often than others perceived them as such
(23.9% self-report vs. 47.8% external report), and less
often as marginalized (55.2% self-report of marginal-
ization vs. 85.1% report of any marginalization; cf.
findings by Newman et al., 2014, who found the self-
description “loner” in 34% of the cases, “marginalized”
in 67%). Thus, from the perspective of the perpe-
trator, revenge resulting from interpersonal conflicts
might have been a driving motive. In fact, in 38.8%
of cases perpetrators cited revenge as a motive – a
somewhat lower rate than that reported by Vossekuil
et al. (2002), who concluded that more than half of
the attackers were driven by revenge. In our analysis
we were unable to determine, in retrospect, whether
perpetrators themselves cited “revenge” as a motive or
whether researchers did so. Here we must remember
that in many cases there were no self-reports, as the
perpetrators had died at the scene of the violent act.
International Journal of Developmental Science 8/2014, 3–24 19

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
Additional Findings
Additionally to the content-related findings, we have
found some other noteworthy aspects in the primary
studies:
First, as conceded by the researchers themselves (e.g.
Harding, Fox, & Mehta, 2002), studies on school shoot-
ings have been largely limited to small samples, with
all the consequent methodological difficulties. Many
of the primary studies we reviewed investigated more
or less the same ten cases (Moses Lake, Bethel, Pearl,
West Paducah, Erfurt, Jonesboro, Littleton). All of these
cases dated from the late nineties, occurred in the U.S.
and were analyzed in the first wave of school shooting
research published between 1999 and 2004. Thus most
of our knowledge of school shootings stems from just
a few case studies from the same time and place. Sixty
three out of a total of 128 perpetrators (49.2%) were
included in only one single study, which means that
little consideration was given to half of all cases that
qualified. Integrating these cases into one comprehen-
sive review might lead to different results, especially
as it is not clear why some cases were included in the
primary studies and others were not. It is possible that
those cases were selected that were prototypical, while
those which deviated from the perceived norm were
disregarded.
Second, the mode of data collection is potentially
biased, as information was mainly obtained from press
reports, which may reflect reporters’ personal theories
regarding the causes and dynamics of the shootings,
instead of delivering objective information. Journalists
tend to look for the typical, for prototypes that rein-
force existing images and stereotypes. In doing so, they
sometimes ignore or downplay other explanations of
violence (Danner & Carmody, 2001). These limitations
affect a total of 12 primary studies in our review which
mostly relied on media accounts. Furthermore, due to
the need for retrospective analysis focusing on this phe-
nomenon, the data that was obtained by court-, police-,
or school reports may also be biased.
Third, definitions proved a problem in carrying out
our review. In most of the primary studies, the term
“school shootings” is defined in advance, and cases
are selected accordingly. The definitions vary widely
– resulting in very different lists of cases regarded
as “relevant”. There was disagreement with regard to
the number of victims (multiple vs. single casualties),
choice of weapon (solely firearms vs. lethal means in
general), the question whether the attack was premed-
itated or spontaneous (school deliberately chosen vs.
random selection), the status of perpetrator (student
or former student vs. non-student), the outcome of the
attack (thwarted vs. carried out), and the country where
it took place. The only feature common to all the defi-
nitions is the site of the violent act, which is a school or
university. In short, we realized that the cases we under-
took to compare probably differed in many respects.
Thus we were careful to include only those cases which
fit our basic definition of a school shooting. We left open
the questions as to the country where the attack took
place, the choice of weapon, the number of victims and
the outcome.
Conclusion: From Risk Factors Towards a
Developmental Perspective
The comprehensive review we have undertaken reveals
that the social dynamics which can play an important
role in school shootings are much more varied than
indicated in any one primary study. Although there are
many similarities between the individual cases (perpe-
trator felt marginalized, experienced peer rejection, and
saw his or her attack as an act of revenge) there were
also cases which do not fit this description (perpetra-
tor was a bully, had not experienced peer rejection, had
been characterized as model student). These atypical
cases merit particular attention and analysis. Thus, in
examining the social dynamics as they developed prior
to a school shooting, we found numerous differences
among perpetrators and discovered that differing paths
had led them to the shooting incident (equifinality; cf.
Cicchetti & Rogosch, 1996). Having taken a close look
at the literature, we concluded that, among the factors
investigated, there was no single one which was present
in all cases, which means that these are no necessary
conditions when it comes to school shootings. What
we see instead are multiple risk factors from differ-
ent domains, interacting with each other in a complex
manner, including such things as individual person-
ality traits and specific situations. Grouping together
individual factors into more “general” concepts such
as “social marginality of the perpetrator” (85.1% of
cases) or “negative social interaction within the school
environment” (88.1% of cases) we can conclude that
most of the cases fulfilled the criteria. Further studies
should investigate if all social dynamics perform the
same function in the developments which culminate
in a school shooting, including factors like bullying,
romantic rejection, and conflicts with teachers.
With the exception of several in-depth case studies
(e.g. Fast, 2009; Larkin, 2009; Newman et al., 2004)
20 International Journal of Developmental Science 8/2014, 3–24

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
the primary studies we included approached risk fac-
tors as “existent” or “non-existent”, without asking for
example who had assessed them (perpetrator, special-
ists, scientist, etc.) or what time frame was involved
(early childhood, the period leading up to the shoot-
ing). All of the risk factors are looked at individually
and analyzed separately in determining whether they
increase or reduce risk. Thus, complex social dynamics
are divided up into isolated parts in an “a-theoretical”
manner (Fox & Burstein, 2010). Real understanding
of the factors leading to violent acts such as school
shooting we will require a careful analysis of the links
connecting social position, social interaction, and the
subjective self-perception of the future perpetrator over
a period of time. Whether or not romantic rejection, for
example, can be considered a contributing factor, will
depend on the perception of the perpetrator. Thus, an
approach is needed which takes into account the vari-
ous contributing factors and their respective impacts on
the psychosocial development of the individual.
Limitations and Prospects
Research on school shootings has been commonly
based on retrospective, incomplete data, and these lim-
itations hamper any content analysis of the primary
studies. Some researchers supply a great deal of infor-
mation on certain cases, while others provide little or
none. Thus, giving ratings on the basis of a coding
scheme was quite simple in some cases, but challeng-
ing in others. Without knowing whether, for example,
bullying involved repeated physical, verbal or psycho-
logical attacks or some other form of intimidation which
went on for a long period and reflected an imbalance
of power, we cannot speak of bullying per definition.
The extent to which the victimization experience neg-
atively impacts on the psychosocial development of
an individual depends in part on his or her coping
strategies, emotional regulation, and further contextual
factors (Mahady Wilton, Craig, & Pepler, 2000). Fur-
ther research on the impact of bullying experiences or
other social difficulties on the perpetrator’s develop-
ment to violence should therefore focus more on the
impact of victimization perceived by the perpetrator
than on the bullying experience itself.
The present comprehensive review was conducted
with the aim of shedding light on the crucial role of
social dynamics in school shootings. Other individ-
ual and contextual factors were left out of the picture,
although they no doubt also contribute to social con-
flicts and emotional strain. Investigating these factors
and the ways in which they interact over time will be
necessary if we are to gain a comprehensive understand-
ing of school shooting cases. Finally, it is likely that we
have not given due consideration to cultural factors.
What is seen as social marginalization and conflict in
one country or culture might not be seen as such in
another, a fact which should be born in mind when the
media and court reports in one country register fewer
signs of social marginalization and conflict than else-
where, in attempting to explain school shootings (cf.
Lankford, 2012). Further analyses should address the
differences between cases in the U.S. and those that
occurred in other countries, in attempting to clarify the
role of cultural influences in the developments that lead
to school shootings. Recent research has delivered anal-
yses of school shootings, comparing them with other
acts of homicide (Lankford, 2012). These are promis-
ing efforts which should contribute to our understanding
of the phenomenon, by specifying the particular con-
stellation of causes and contributing factors which turn
young people into school assassins.
Acknowledgments
The present study is part of the interdisciplinary three-
year research project TARGET (2013–2016), funded
by the Federal Ministry of Education and Research
(BMBF) of Germany (funding code 13N12646).
Conflicts of Interest
There are no conflicts of interest.
References3
Adams, A. T. (2000). The status of school discipline and violence.
The ANNALS of the American Academy of Political and Social
Science, 567, 140-156.
*Band, S. R., & Harpold, J. A. (1999). School violence. Lessons
learned. FBI Law Enforcement Bulletin, 68, 9-16.
*Bannenberg, B. (2012). Sogenannte Amokl¨
aufe junger T¨
ater –
Mehrfacht¨
otungen aus unklarem Motiv. [So-called rampage
shootings of young perpetrators – Multiple homicides with
unknown motive]. In H. Remschmidt (Ed.), T¨otungs- und Gewalt-
delikte junger Menschen. Ursachen, Begutachtung, Prognose(pp.
77-104). Berlin, Heidelberg: Springer.
B¨
ockler, N., Seeger, T., Sitzer, P., & Heitmeyer W. (Eds.). (2013).
School shootings. International research, case studies, and con-
cepts for prevention. New York: Springer.
Bond¨
u, R. (2012). School Shootings in Deutschland. Inter-
nationaler Vergleich, Warnsignale, Risikofaktoren, Entwick-
lungsverl¨aufe [School shootings in Germany. International
3References with a * were included in the systematic review
International Journal of Developmental Science 8/2014, 3–24 21

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
comparison, warning signs, risk factors, developmental path-
ways]. Retrieved from Freie Universit¨
at Berlin’s electronic The-
ses and Dissertations Archive: http://www.diss.fu-berlin.de/diss/
receive/FUDISS thesis 000000037683
Bond¨
u, R., & Scheithauer, H. (2014). Peer and teacher relationships in
German school shooters. International Journal of Developmental
Science, 8, 57-63.
Brown, R., Osterman, L., & Barnes, C. (2009). School violence
and the culture of honor. Psychological Science, 20, 1400-
1405.
*Brumme, R. (2011). School Shootings. Soziologische Analysen
[School Shootings. Sociological analysis]. Heidelberg: Springer.
Cicchettia, D., & Rogoscha, F. A. (1996). Equifinality and mul-
tifinality in developmental psychopathology. Development and
Psychopathology, 8, 597-600.
*Danner, M. J. E., & Carmody, D. C. (2001). Missing gender in cases
of infamous school violence: Investigating research and media
explanations. Justice Quarterly, 18, 87-114.
*Dumitriu, C. (2013). School violence around the world: A social
phenomenon. Procedia –Social and Behavioral Sciences, 92,
299-308.
*Dutton, D. G., White, K. R., & Fogarty, D. (2013). Paranoid think-
ing in mass shooters. Aggression and Violent Behavior, 18, 548-
553.
Effects of Bullying (2014). Retrieved April 16, 2014, from
http://www.stopbullying.gov/at-risk/effects/
Egger, M., Smith, G. D., & O’Rourke, K. (2001). Rationale, potentials
and promise of systematic review. In Egger, M., Smith, G. D., &
Altman, D. G. (Eds.) Systematic reviews in health care: Meta-
analysis in context (pp. 3–19). London: BMJ Books.
*Fast, J. (2008). Ceremonial violence. A psychological explanation
of school shootings. Woodstock: The Overlook Press.
Fox, J. A., & Burstein, H. (2010). Violence and security on campus:
From preschool to college. Westport: Praeger.
*Fox, J. A., Levin, J., & Quinet, K. (2005). The will to kill: Making
sense of the senseless murder. Boston: Allyn and Bacon.
*Fox, J. A. & Savage, J. (2009). Mass murder goes to college: An
examination of changes on college campuses following Virginia
Tech. American Behavioral Scientist, 52, 1465-1485.
Harding, D. J., Fox, C., & Mehta, J. D. (2002). Studying rare events
through qualitative case studies: Lessons from a study of rampage
school shootings. Sociological Methods and Research, 31, 174-
217.
Harding, D., Mehta, J., & Newman, K. (2003). No exit: Mental illness,
marginality, and school violence in West Paducah, Kentucky. In
M. H. Moore, C. V. Petrie, A. A. Braga, & B. L. McLaughlin
(Eds.), Deadly lessons: Understanding lethal school violence (pp.
132-162). Washington, DC: The National Academies Press.
*Harter, S., Low, S. M., & Whitesell, N. R. (2003). What we have
learned from Columbine. The impact of the self-system on sui-
cidal and violent ideation among adolescents. Journal of School
Violence, 2, 3-26.
Hess, M., & Scheithauer, H. (in press). Treatment and prevention
of bullying. In T. Gullotta, M. Evans, & R. Plant (Eds.), The
handbook of adolescent behavioral problems, 2nd edition.New
York: Springer Academic Publishing.
*Hoffmann, J., Roshdi, K., & Robertz, F. J. (2009). Zielgerichtete
schwere Gewalt und Amok an Schulen. Eine empirische Studie
zur Pr¨
avention schwerer Gewalttaten. [Severe targeted violence
and rampage at schools. An empirical study for the prevention of
severe violence.]. Kriminalistik, 63, 196-204.
*Kalish, R., & Kimmel, M. (2010). Suicide by mass murder: Mas-
culinity, aggrieved entitlement, and rampage school shootings.
Health Sociology Review, 19, 451-464.
*Kiilakoski, T., & Oksanen, A. (2011). Cultural and peer influences
on homicidal violence: A Finnish perspective. New Directionsfor
Youth Development, 129, 31-42.
Kimmel, M. S. (2008). Profiling school shooters and shooters’
schools: The cultural contexts of aggrieved entitlement and
restorative masculinity. In B. Agger, & D. Luke (Eds.), There
is a gunman on campus: Tragedy and terror at Virginia Tech (pp.
65–78). Lanham: Rowman and Littlefield.
*Kimmel, M. S., & Mahler, M. (2003). Adolescent masculinity,
homophobia, and violence. American Behavioral Scientist, 46,
1439-1458.
*Kidd, S. T., & Meyer, C. L. (2002). Similarities of school shootings
in rural and small town communities. Journal of RuralCommunity
Psychology, E5 (1).
Klassen, T. P., Jadad, A. R., & Moher, D. (1998). Guides for reading
and interpreting systematic reviews: I. Getting started. Archives
of Pediatrics &Adolescent Medicine, 152, 700-704.
*Klein, J. (2006). Cultural capital and high school bullies: How social
inequality impacts school violence. Men and Masculinities, 9,
53-75.
Klein, J. (2012). The bully society. School shootings and the crisis of
bullying in America’s schools. New York: New York University
Press.
Langman, P. (2009). Rampage school shooters: A typology. Aggres-
sion and Violent Behavior, 14, 79-86.
*Langman, P. (2013). Thirty-five rampage school shooters: Trends,
patterns, and typology. In N. B¨
ockler, T. Seeger, P. Sitzer, & W.
Heitmeyer (Eds.), School shootings: International research, case
studies, and concepts for prevention (pp. 131-156). New York:
Springer.
*Lankford, A. (2012). A comparative analysis of suicide terrorists
and rampage, workplace, and school shooters in the United States
from 1990 to 2010. Homicide Studies, 17, 255-274.
*Larkin, R. W. (2009). The Columbine legacy: Rampage shootings
as political acts. American Behavioral Scientist, 52, 1309-1326.
*Leary, L. M., Kowalski, R. M., Smith, L., & Philips, S. (2003). Teas-
ing, rejection, and violence: Case studies of the school shootings.
Aggressive Behavior, 29, 202-214.
Leuschner V., Bond ¨
u, R., Schroer-Hippel, M., Panno, J., Neumetzler,
K., Fisch, S., Scholl, & J., Scheithauer, H. (2011). Prevention
of homicidal violence in schools in Germany: The Berlin Leak-
ing Project and the Networks Against School Shootings Project
(NETWASS). New Directionsfor Youth Development, 129, 61-78.
Levin, J., & Madfis, E. (2009). Mass murder at school and cumulative
strain: A sequential model. American Behavioral Scientist, 52,
1227-1245.
*Madfis, E., & Levin, J. (2013). School rampage in international per-
spective: The salience of CumulativeStrain Theory. In N. B¨
ockler,
T. Seeger, P. Sitzer, & W. Heitmeyer (Eds.), School shootings:
International research, case studies, and concepts for prevention
(pp. 79-104). New York: Springer.
Mahady Wilton, M. M., Craig, W. M., & Pepler, D. J. (2000). Emo-
tional regulation and display in classroom victims of bullying:
Characteristic expressions of affect, coping styles and relevant
contextual factors. Social Development, 9, 226-245.
*Malkki, L. (2013). Political elements in post-Columbine school
shootings in Europe and North America. Terrorism and Political
Violence, 26, 185-210.
22 International Journal of Developmental Science 8/2014, 3–24

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
*McGee, J. P., & DeBernardo, C. R. (1999). The classroom avenger.
The Forensic Examiner, 8, 1-16.
*Meloy, J. R., Hempel, A. G., Mohandie, K., Shive, A. A., & Gray, B.
T. (2001). Offender and offense characteristics of a non-random
sample of adolescent mass murderers. Journal of the Ameri-
can Association of Child and Adolescent Psychiatry, 40, 719-
728.
Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., & the PRISMA
Group (2009). Preferred reporting items for systematic reviews
and meta-analyses: The PRISMA Statement. Annals of Internal
Medicine, 151, 264-269.
*Moore, M. H., Petrie, C. V., Braga, A. A., & McLaughlin, B. L.
(2003). Deadly lessons: Understanding lethal school violence.
Washington, DC: National Academies Press.
Nansel, T. R., Overpeck, M., Pilla, R. S., Ruan, W., Simons-Morton,
B., & Scheidt, P. (2001). Bullying behaviors among US youth:
Prevalence and association with psychosocial adjustment. Jour-
nal of the American Medical Association, 285, 2094-2100.
*Newman, K. S., & Fox, C. (2009). Repeat tragedy: Rampage shoot-
ings in American high school and college settings, 2002-2008.
American Behavioral Scientist, 52, 1286-1308.
*Newman, K. S., Fox, C., Harding, D. J., Mehta, J., & Roth, W.
(2004). Rampage. The social rootsof school shootings. New York:
Perseus Books.
Olweus, D. (1994). Annotation: Bullying at school: Basic facts and
effects of a school based intervention program. Journal of Child
Psychology and Psychiatry, 35, 1171-1190.
*O’Toole,M. E. P. D. (1999). The school shooter: A threatassessment
perspective. Quantico, VA: National Center for the Analysis of
Violent Crime, Federal Bureau of Investigation.
Rocque, M. (2012). Exploring school rampage shootings: Research,
theory, and policy. The Social Science Journal, 49, 304-313.
*Roshdi, K., & Hoffmann, J. (2011). Ein Vergleich von Taten ziel-
gerichteter Gewalt an Schulen mit und ohne Amok-Dynamik. [A
comparison of cases of severe targeted school violence with and
without amok dynamics]. In C. Lorei (Ed.), Polizei und Psycholo-
gie 2009: Kongressband der Tagung Polizei und Psychologie am
27. und 28. Oktober 2009 in Frankfurt am Main (pp. 83-104).
Polizeiwissenschaft: Frankfurt.
Scheithauer, H., Haag, N., Mahlke, J., & Ittel, A. (2008). Gender-
differences in the development of relational aggression –
Evidence for an age effect? Preliminary results of a meta-analysis.
European Journal of Developmental Science, 2, 176-189.
Scheithauer, H., Hayer, T., & Petermann, F. (2003). Bullying unter
Sch¨ulern –Erscheinungsformen, Risikobedingungen und Inter-
ventionskonzepte [Bullying among students – forms, risk factors,
and interventions]. G¨
ottingen: Hogrefe.
Scheithauer, H., Hayer, T., Petermann, F., & Jugert, G. (2006). Phys-
ical, verbal and relational forms of bullying among students from
Germany: Gender-, age-differences and correlates. Aggressive
Behavior, 32, 261-275.
*Schiller, J. (2013). School shootings and critical pedagogy. The
Education Forum, 77, 100-110.
Staples, J. S. (2000). Violence in schools: Rage against a broken
world. The ANNALS of the American Academy of Political and
Social Science, 567, 30-41.
Thompkins, D. E. (2000). School violence: Gangs and a culture of
fear. The ANNALS of the American Academy of Political and
Social Science, 567, 54-71.
Ttofi, M., Farrington, D., & L¨
osel, F. (2012). School bullying as
a predictor of violence later in life: A systematic review and
meta-analysis of prospective longitudinal studies. Aggressionand
Violent Behavior, 17, 405-418.
*Verlinden, S., Hersen, M., & Thomas, J. (2000). Risk factors in
school shootings. Clinical Psychology Review, 20, 3-56.
*Vossekuil, B., Fein, R. A., Reddy, M., Borum, R., & Modzeleski, W.
(2002). The final report and findings of the Safe School Initiative:
Implications for the prevention of school attacks in the United
States. Washington DC.
*Waterman, B. T. (2002). Terror in our schools: An exam-
ination of the recent school shootings. Retrieved from
http://www.bedrugfree.net/Terror.pdf
*Weatherby, G. A., Strachila, S., & McMahon, B. (2010). School
shootings: The deadly result of teasing and ostracism. Journal of
Criminal Justice Education, 2, 1-15.
*Wieczorek, A. (2010). Sch¨
ulerattentate an deutschen Schulen.
Mythen, Fakten und Schlussfolgerungen f¨
ur die polizeiliche
Praxis. [Student attacks on German schools. Myths, facts and
implications to police practice.]. Kriminalistik, 3, 153-160.
Bio Sketches
Friederike Sommer is a psychologist working as a research asso-
ciate in the TARGET project at Freie Universit¨
at Berlin, Division
Developmental Science and Applied Developmental Psychology.
Dr. Vincenz Leuschner is a social scientist working as coordi-
nator of the TARGET project at Freie Universit¨
at Berlin, Division
Developmental Science and Applied Developmental Psychology.
Dr. Herbert Scheithauer is Professor for Developmental and Clin-
ical Psychology at Freie Universit¨
at Berlin and Head of the Division
Developmental Science and Applied Developmental Psychology.
Endnote: Errata
In an earlier version of this target article we described
the category “conflicts with teacher and the school” as
follows:
“Open conflicts with an individual teacher or other
school authorities, including the school principal,
which were not related to academic performance.”
Consequently, Jessie Klein criticized in her commen-
tary on the target article:
“In short the paper misses (…) (3) The impact of
perceived academic failure (…)”
“Sommer et al. include school shooting cases related
to rage at being suspended or otherwise punished; but
exclude cases related to academic performance.”
“Nonetheless, in their study the authors only
included: “Open conflicts with an individual teacher
or other school authorities, including the school princi-
pal, which…” were explicitly “not related to academic
performance”
International Journal of Developmental Science 8/2014, 3–24 23

F. Sommer et al. / Bullying, Romantic Rejection, and Conflicts with Teachers
After reading Klein’s critique, we realized that we used
a misleading description of the category. Thus, we mod-
ified the description in the final version of the target
article to make sure that the category is not describ-
ing the decline in school performance only (which is
often described as a risk factor present in later perpe-
trators), but rather also includes conflicts with teachers
that result from poor academic performance.
Thus, we changed the description of the category
in the present version of the target article as follows:
“Open conflicts with an individual teacher or other
school authorities, including the school principal (dis-
ciplinary or academic injustices, personal conflicts,
suspension from school)”.
24 International Journal of Developmental Science 8/2014, 3–24