|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
PORTRAYAL OF SCHOOL SHOOTINGS BY AMERICAN NEWSPAPERS
JOHN ROBERT BENNETT
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF ARTS
UNIVERSITY OF FLORIDA
This document is dedicated to my wife, Tosha L. Bennett. Without her support, I
would not have been able to complete this study. Special thanks go to Charles F.
Gattone, Ph.D. and Joe R. Feagin, Ph.D.; without their guidance, support, and providing a
committee for this thesis project none of this would be possible. I thank Terry L. Mills,
Ph.D., and Karen F. Parker, Ph.D., for their initial advice on how to go about developing
my ideas into a working project which ultimately became my thesis.
TABLE OF CONTENTS
A C K N O W L E D G M E N T S ......................................................................... .....................ii
LIST OF TABLES .................................. .. ... .... ................. .v
A B STR A C T................................................... vi
INTRODUCTION .......................... ........ .. ... .... ........ ...............
1.1 Statem ent of the R research Problem ..................................................................
1.2 B background ...................................................... 2
1.3 K ey Term inology and Concepts ........................................ ......... ............... 3
1.4 Theoretical Perspectives ............................................................................. 5
1.4.1 Conflict-Labeling Crim inology.................................................................5
1.4.2 Pow er-Threat H ypothesis ........................................ ......... ............... 7
1.4.3 M edia C rim inology............................................................ ............... 9
1.4.4 Critical-M edia Perspective ........................................ ...... ............... 12
1.4.5 Gaps in the Literature ................................................... 16
METHODOLOGY .................. ................................... ........... ............... 19
2 .1 Su m m ary of F in ding s................................................................. .....................2 0
2.2 Discussion of Data Characteristics .............. ........... .................................. 22
FIN D IN G S AN D AN ALY SIS.....................................................................................24
3.1 R research Findings .................. .................. .......................................24
3.1.1 Q ualitative .................................................. ...................... .... 24
3.1.2 Q uantitative ................................ ............. ........ .............. 27
220.127.116.11 N on-aggregate D ataset................................................................... 28
18.104.22.168 A ggregate D ataset...................................................... ............ 34
3.2 A analysis of Findings ........................... .... ................ ... .... .. ........... 4 1
3.2.1 Q ualitative A nalysis......................................................... ............... 4 1
3.2.2 Q uantitative A nalysis........................................... ........................... 48
DETAILED DISCUSSION AND ANALYSIS OF SPECIFIC INCIDENTS .................57
4.1 C olum bine H igh School Shooting .......................................................................57
4.2 Thurston H igh School Shooting................................................. ....... ........ 65
4.3 John Batram High School Shooting................................... ....................... 67
4.4 Lew W allace High School Shooting................................... ...................... 67
CONCLUSIONS.................. ..... .. ....... .... ....................69
5.1 Overview of Theories Utilized in this Analysis........................................70
5.2 Overview of Important Findings........................ .......................70
5.3 Critique and Suggestions for Future Research of Differential Newspaper
C ov erag e ...................................... ................................................ 7 5
L IST O F R E FE R E N C E S ......................................................................... ....................78
B IO G R A PH ICA L SK ETCH ............................................................................ ........... 80
LIST OF TABLES
2-1. Number of articles written per case. ........................................ ....... ............... 21
3-1. Basic figures for non-aggregate variables. .................................... .................28
3-2. Basic figures for aggregate variables................................... ......................... 35
3-3. M atrix of cases with exceptional qualities..................................... ............... 49
3-4. List of cases from lowest to highest affirmation of NTSV status.............................50
3-5. Correlations of aggregate variables. .............................................. ............... 53
3-6. Table of significant correlations for aggregate variables...............................54
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Arts
PORTRAYAL OF SCHOOL SHOOTINGS BY AMERICAN NEWSPAPERS
John Robert Bennett
Chair: Joe R. Feagin
Cochair: Charles F. Gattone
Major Department: Sociology
The purpose of this analysis is to investigate the way American newspapers portray
domestic school shootings. To this date, there have been no investigations of media
portrayal of school shootings in America. School shootings make up a small percentage
of school violence in America; however, the damage inflicted on the victims of these
crimes is devastating. This study is concerned with the coverage and portrayal of these
shootings by the news publishers in America.
There are no current data sources listing all of the variables of interest to this study.
I relied on the Lexis Nexis database of newspapers to find a sample of articles that
address this issue. The time frame for the search has been limited to include only articles
which were published from 1st May 1997 through 30th April 2002.
These cases were analyzed using the structural functional and conflict paradigms
with a specific interest in the conflict-labeling criminology, power-threat hypothesis,
media criminology, and critical-media perspectives.
In conclusion, two key factors seem to explain the relationship between the
variables used in this analysis to determine if there is differential coverage of NTSV and
TSV. The two key factors are the level of sensational behavior exhibited in a given case
and whether or not the behavior exhibited was considered to be expected based upon
socially acceptable norms of American society.
1.1 Statement of the Research Problem
The purpose of this analysis is to investigate the way American newspapers portray
domestic school shootings. To this date, there have been no formal sociological
investigations of media portrayal of school shootings in America. School shootings make
up a small percentage of school violence in the United State; however, the damage
inflicted on the victims of these crimes is devastating. This analysis is not necessarily
concerned with the effects of each school shooting. It is more concerned with the
coverage and portrayal of these shootings by the news publishers in the U.S.
There are no current data sources listing all of the variables of interest to this study.
Therefore, I relied on the Lexis Nexis database of newspapers to find a sample of articles
that address this issue. The time frame for the search has been limited to include only
articles which were published from 1st May 1997 through 30th April 2002. I gathered and
analyzed articles that span a five year period of time. During this period, 578 articles
were published in U.S. newspapers on school shooting. Of these 578 articles, I selected
one hundred at random for inclusion in the sample and coded each for the following
variables: title, case name, number of words, date of occurrence, date of publication,
section of the newspaper in which the article was published, time difference between date
of publication and date of occurrence, race of offender and victim, socioeconomic status
of offender and victim, number of offenders and victims, number of offenders and
victims killed, type of weapons used, number of weapons used, relation of victim to
offender, and level of unique interest. I also developed qualitative descriptions of each
Of the one hundred articles included in the sample, twenty seven cases emerged. I
organized these articles on the basis of case name so that key variables from the
individual articles could be represented in a combined form. The variables included in
the aggregate dataset are as follows: new trend vs. tradition, case name, number of
articles per case, number of articles written within first year after the occurrence,
socioeconomic status of offenders and victims, quantity of weapons used, type of
weapons used, race of victims and offenders, number of offenders and victims, number of
offenders and victims killed, level of unique interest controlling for time, racial make-up
of the city where shooting occurred, median family income of the city where shooting
occurred, and total population for the city where the shooting occurred. These cases
were analyzed using the structural functional and conflict paradigms with a specific
interest in the conflict-labeling criminology, power-threat hypothesis, media criminology,
and critical-media perspectives.
Government agencies, criminal justice departments, and scholars in the U.S. have
not studied school shootings in the way that is proposed here. Among these institutions
the tendency has been to focus on data regarding the number of shootings and number of
victims, the types of weapons used in offenses, and components related to offenders'
psychological makeup. The Federal Bureau of Investigation has been involved in
compiling profiles for offenders in school shooting cases. Other agencies have developed
' The 2000 U.S. Census American FactFinder was used to find the racial makeup of the city
where the shooting took place and the median family income of the city where the occurred.
programs aimed at trying to prevent school violence and teaching young people how to
deal with stress and violent feelings. Most of the academic resources have been invested
in the psychological characteristics of the offenders and, at the same time, teaching others
methods of anger management. Psychologists and social workers have been involved in
providing services to those who have been victims of school shootings and those who
have had to live in the aftermath of school shootings. No study to this date has been
conducted by social science researchers with regard to news coverage of school
shootings. There is a void that exists in the academic study of news coverage of school
1.3 Key Terminology and Concepts
The specific goals of this project are to determine the nature of media portrayal of
school violence by domestically owned and published newspapers. School violence is
classified as being of two types: traditional and new trend. Traditional school violence
involves actors who are of a low socioeconomic status, unemployed, or working class.
The target of the violence is generally other students and violence may be related to gang
or drug activities. Types of weapons used vary from assault with hands to assault with
guns that are not mass-casualty producing. The offender usually targets one or two
persons, and violence seems to happen in more of a spontaneous nature. There also may
be planned acts of violence in this category. New trend school violence is violence in
which the offender is usually a white male of a middle class or higher socioeconomic
status, and the target of the violence tends to be not only peers but also teachers, school
officials, and parents. Violence is likely related to feelings of revenge or retribution and
the types of weapons used are mass-casualty producing and tend to be excessively
violent. The offender usually plans to cause many casualties but may actually kill fewer,
and violence seems to be planned, premeditated, and organized. Some key distinctions
between these two types of school violence offenders are their level of socioeconomic
status and the relation and status of their targeted victims. This will not be measured at
the individual level in this study due to lack of data. The SES data will be organized on
the basis of the regional location in which the incident occurred.
Another key concept in this project is the expected normal behavior of the offender.
Expected normal behavior is applied in the following way: in general, white American
news journalists and editors share common ideas about normal behavior that they expect
to find in any given community.2 In cases of school shootings that occur where the
violent behavior is expected, the offender will be labeled as exhibiting normal behavior,
and white American journalists will not likely give any special attention to these cases
unless there is something unusual about them. For example, in a case where several
victims were involved, if unusual weapons were used to commit the violence or some
other anomaly along these lines, white journalists may exhibit special interest in a case.
In cases of school shootings that occur where the behavior is not expected, the offender is
labeled as exhibiting abnormal behavior and most white American journalists will give
additional attention to these cases, covering them more thoroughly and are expected to
show signs of unique interest. Some characteristics of unique interest that the white
American news journalists are expected to exhibit in these abnormal cases are the
following: an invocation of a psychological-medical model to explain the occurrence is
possible. For example, in instances in which the offender is likely to be labeled as
2 This study uses the terms "white American journalists and editors" to convey the assumption
that the majority of articles written in this study were written by the "average" white American
journalist who is expected to present each case in a certain way. There are editorials in this
psychologically ill or portrayed as an outsider from the "normal" student population at
the school, words such as "understand," "sympathy," "empathy," "cope" are expected to
be used in this type of article. In instances in which readers assume that the violence was
not supposed to happen in the community of the shooting, it is expected that the article
will present a sense of urgency about finding a solution. The article may also describe
steps that are being taken by authority figures in the community where the violence took
place to rectify the situation, and there will likely be a description involving the aide that
will likely be given to members of the community to help them cope with the act of
1.4 Theoretical Perspectives
Analysis of the literature shows that theory has not been applied to the issue of
school shooting. Sociologists have not formally studied the ways newspapers present
information and cover stories of school shootings. The following literature review is a
compilation of texts written from the conflict perspective, labeling theory, power-threat
hypothesis, with contributions from media criminology and critical-media perspectives.
1.4.1 Conflict-Labeling Criminology
This section illustrates findings from two conflict-labeling theorists that describe
the relationship between a criminal offender's status in society and the way that he or she
is likely to be labeled by authority figures that preside within the criminal justice system.
Both Turk (1964) and Quinney (1970) share similar ideas about the ways people are
labeled as criminal. Both new trend and traditional school violence cases are labeled as
criminal by newspapers. The difference is that traditional school violence (TSV) tends to
sample which are not likely to have been written by white American journalists, but it is assumed
that the majority of the editorials are written by white Americans.
be labeled as more criminal relative to new trend school violence (NTSV). For both Turk
(1964) and Quinney (1970), society is made up of a set of laws that are produced by
people in power to maintain order in society. If a person violates those laws, the less
power that individual has, the more likely he or she is to be punished and labeled in a
negative way. In this argument, the critical component of power, as measured by SES or
some other economic indicator, comes into play. School shooting offenders who have
lower SES are typically labeled as criminal more readily than those of higher relative
How is this translated into the coverage of school violence by newspapers? One of
the first points to realize is that the criminal justice system in the U.S. deals with the
offenders of school shootings. In this system, police officers, attorneys, and judges are
the authority figures who officially label the offenders in the legal sense. At the same
time, the authors and editors of newspapers have an opportunity to further label the
offenders by providing coverage of the case in the newspaper. Examining the work of
Turk (1964) and Quinney (1970) might lead one to expect that the media would cover all
school shooting cases. However, one would expect to find that the coverage of cases that
are committed by offenders of low or working class SES will be more harsh toward the
offender than those articles that cover the cases in which the offender is of middle or high
SES. This differential coverage is likely because Turk (1964) and Quinney (1970) found
that authority figures will treat criminals more harshly if they have less power. In this
case, the lower class and working class school violence offenders are more likely to
receive unforgiving coverage in relation to the middle class and upper middle class
1.4.2 Power-Threat Hypothesis
This section introduces two concepts that Hubert Blalock (1967) developed within
the structure provided by the power-threat hypothesis (Tolnay, Beck, & Massey, 1989).
The two concepts offer further guidance in determining what type of coverage one would
expect U.S. media sources to provide for the two types of school violence. Blalock
(1967) identified two types of discrimination that can motivate the white majority to
discriminate against the black minority: symbolic segregation and threat-oriented
ideologies. He examined U.S. society during the 1960s, where he divided America into
North and South and observed trends of discrimination in both regions of the country.
Blalock (1967) developed the idea symbolic segregation in which white Americans create
belief systems that promoted segregation between white Americans and minority groups,
in particular black Americans. Blalock's (1967) idea of symbolic segregation would lead
one to expect that the media may not present coverage of all school shooting cases, and
that the news articles are likely to differ in their coverage of the two types of school
violence. By presenting TSV cases as though they were normal occurrences relative to
NTSV cases, or by not providing coverage of the TSV cases at all, the authors and editors
of national newspapers seem to be symbolically segregating TSV offenders and their
families from the rest of the society. By providing this type of coverage for TSV cases,
the authors and editors keep TSV cases from entering the consciousness of most
At the same time, newspaper coverage provides a sense that NTSV is abnormal
which encourages the average American to think about these offenders and their families,
the victims of such offenses, and the circumstances around which the offense occurred. It
is deducted from Blalock's (1967) theory that the average American may develop
sympathy in an attempt to understand what happened, and may try to determine what
society can do to stop this new trend of school violence. This type of sympathy,
compassion, and understanding is not extended to the TSV offenders, families, and
victims since they are not within the realm of the average American's consciousness.
The symbolic segregation created by the editors and authors of newspapers in the
coverage of TSV offenders leads to neglect and reminds the offenders and their families
where they belong. The implied message is that they do not belong in the newspapers
portrayed in a similar way as the NTSV offenders and that they do not deserve the same
sympathy and respect that is given to the NTSV offenders, family members, and victims.
These findings tie into the second concept that Blalock (1967) developed termed
"threat-oriented ideology." He argued that southern white Americans tended to
exaggerate the threat that black Americans posed to the white community of recent
history. Similarly, the average white American today tends to expect violence that occurs
in the lower class and working class communities where TSV takes place. The average
white American expects to see violent behavior to an exaggerated degree in these
communities. This sentiment has been recorded by media interviews with many people
after recent incidents of NTSV have occurred. The people who are interviewed state
things like: "How can this happen here?" "I can't believe this happened here." "This is
not supposed to happen here." These statements made many times by many, mainly
white, people in America have a logical complementary statement: school violence is
supposed to happen somewhere else. That somewhere is in communities unlike those
middle and upper class communities where NTSV is occurring. It is supposed to happen
in communities where TSV occurs. This line of reasoning leads to the culmination of the
thought that there is a subconscious idea in the minds of the average white American that
there are communities "out there" where violence is supposed to occur; and the
newspapers do not cover this violence with the same care and understanding that is used
in the coverage of NTSV because it is expected from "those" communities where "those"
people live. This finding leads one to conclude that authors and editors of newspapers
will cover NTSV cases with more emphasis placed on the abnormality of the occurrences
relative to TSV cases. Additionally, only a portion of the TSV cases will be covered in
A different way in which media organizations do portray TSV cases is introduced
in the next section and is centered on the idea that really horrific cases of TSV that
perpetuate the idea of threat-oriented ideologies will be published because it does
reinforce thoughts that are currently socially acceptable with regard to TSV.
1.4.3 Media Criminology
Media-criminologists conduct research on the proposed effects that news coverage
has on the reader. Howitt (1998), Barak (1994), and McQuail (1992) all share similar
ideas about the media's role in maintaining the social order that is made up of inequality
and power struggles. "The media present a world view which continually and
pervasively regenerates the ideological structures required for the maintenance of existing
power hierarchies" (Howitt 1998, p. 13). With this in mind, this paper now tries to
explain why one should expect to find little national news coverage of TSV. Howitt
(1998) found that the national news coverage of crime does not necessarily coincide with
the number of occurrences that the crime took place. For example, "lynchings were more
likely to be reported if: (a) they were in areas with easy communications; (b) they
involved rape or sexual assault; and (c) there were several victims" (p. 25). This shows
that the media covered only the sensational cases of lynchings and gives additional
support to the idea that the national newspapers will cover TSV cases on a limited basis.
Cases that are sensational will be presented in news articles. From this perspective, one
would explain the heightened media coverage of NTSV cases because they are
sensational. On average, the level of intended violence is sensational. It seems to be a
relatively new phenomenon in which children of middle and higher SES are committing
horrible acts of violence against fellow students and authority figures. This raises the
question: "Is the divide in the coverage of the two types of school violence simply a
matter of the media going after the sensational 'big scoop'?" One of the goals of this
project is to address this question.
Another goal of this research is to examine the claim that the media present
sensational articles, and those that maintain and reinforce current ideological structure.
One could argue that the coverage ofNTSV and the coverage of sensational TSV cases
do hold true to this finding as well. The NTSV cases are covered in such a way as to
separate the NTSV offenders from the "normal" group of students. It is expected that
newspaper publishers will portray NTSV offenders as psychologically ill, or as part of
some deviant sub-group of juveniles that committed these heinous acts of violence. This
is a way in which the authors of these articles do maintain and reinforce current
ideologies about the "normal" population. In a similar fashion, the TSV offenders who
commit sensational acts of violence that seem to perpetuate current ideological beliefs
will be covered in news articles. However, newspaper publishers will not attempt to
separate the TSV offender from his or her peers because that would not reinforce and
maintain current ideological beliefs about lower and working class SES communities. In
both cases, though, the news articles tend to portray the offender in such a way to
maintain status quo. However, the question still begs to be answered, is this simply a
product of news journalists' attempts to get the "big scoop"? One may argue that the
journalists have gone after the "big scoop." Indeed, they must approach their work in this
manner in order to compete and succeed as journalists, but from the theory, one would
counter that they have done so selectively in a way that perpetuated current socially
accepted beliefs about both types of offenders. However, an answer from the data has
not yet been established.
Based upon the previously mentioned conclusion, let us now discuss Howitt's
(1998) and Barak's (1994) finding. They found that media sources tend to not only
interpret stories but oversimplify stories into false dichotomies, either black and white.
There is no middle ground. They also found that the media tend to focus on the
individual for his or her situation rather than placing partial or total blame on social
policy or some social institution (Howitt, 1998; Barak, 1994). How is it that the
newspaper tends to blame individuals? When the article presents the story in terms of
either black or white, wrong or right, the reader may be left with the impression that the
offender is either wrong or right for doing what he or she has done. There is not a middle
ground. There is no room to introduce the idea that the offender did act a certain way, or
that he or she may have acted differently if the contextual or situational characteristics of
the occurrence where different. One reason to place total responsibility on the suspects
involved is that it is easier for society to see these actions as an example of abnormal
behavior rather than as a social problem.
Barlow, Barlow, and Chiricos (1995) are a group of media-criminologists, as were
Barak (1994), Howitt (1998), and McQuail (1992), who argue that the media rarely give
coverage to cases involving working class and unemployed communities. This is
primarily because these cases are not seen as noteworthy news. Again, this leads back to
the previous finding that the average white American expects to see crime and violence in
neighborhoods and communities where the working class and unemployed reside. An
article on TSV is most likely to be published by the newspapers when the story is
sensational and maintains the status quo ideologies of American society.
1.4.4 Critical-Media Perspective
David Paletz and Robert Entman (1981) provide a general analysis of the media
system in the U.S. In their analysis, they include discussion of both the sources of media
and the factors that influence what is offered by media organizations and how it is
presented. A key point described in their analysis is that the owners of media agencies
are wealthy Americans with "traditional American values [such as] individualism,
free enterprise, competitiveness, and materialism" (Paletz & Entman 1981, p. 10). These
owners operate media agencies the same as any other business concerned about profits
and prestige. Paletz and Entman (1981) note that the search for profit frames or places
limits on the content that media agencies will present. Owners and producers of news
coverage do not want to offend powerful and influential readers and advertisers because
it is not good for business (Paletz & Entman, 1981). Linking this in with findings from
this literature review, one may state that this gives further evidence as to why news
agencies are not likely to offer coverage that is counter to main-stream American
ideologies. Not only would news producers potentially lose readers, but they could also
lose advertisers, the support of powerful individuals in society, and potential investors.
News agencies are businesses and must take into consideration the various consequences
of the news that they provide before they offer coverage of certain topics.
Paletz and Entman (1981) provide additional insight into the discussion of what
makes something newsworthy. They state that journalists usually have a set of guidelines
provided by editors and news executives that provide them with direction on how to
discern what is newsworthy. According to Paletz and Entman (1981), some of the key
components of newsworthy events are the following questions: does the event show some
degree of drama and does it show some threat to American citizens either in a physical or
a social sense? If the newsworthy event does threaten American citizens, it must have
some resolution or solution so that the American public will be reassured of their security
(Paletz & Entman, 1981). The idea of threat is linked with cases of NTSV because these
occurrences provide a potential physical threat for American children that are sent to
school everyday and may come into contact with another student that is the next NTSV
offender. The NTSV phenomenon is also a social threat.3 News articles ask rhetorical
questions about the future of American society pondering "what is the matter with
America's youth"? The model, provided by Paletz and Entman (1981), is not complete
without the reassurance to American citizens that everything will be all right. Most news
3 Social threat refers more to the idea that the threat is societal in scale and may involve non-
physical threats to the average white American in general. For example, in the case of NTSV, the
phenomenon obviously provides a physical threat that is presented as though it is exists on a
national scale, but it also provides a sense that something is wrong with average white American
youths and the average white American way of life in general. This is what generates a sense of
newsworthiness in the case of NTSV. TSV, on the other hand, even if it is recognized as a
nationwide threat of physical violence, the average white American does not necessarily feel a
direct physical danger because the school system is segregated for the most part. The schools
where the TSV is happening are not likely to have a direct affect on the children of average white
Americans. Likewise, TSV does not provide any sort of social threat to the average white
American. It is usually seen and presented as instances of "black on black" crime and has no
bearing on the overall health and wellbeing of average white American youths.
agencies that offer coverage of NTSV provide this reassurance with the help of
government agencies in the form of expert advice about warning signs of disturbed
youths, counselors for families of victims and survivors of NTSV attacks, political and
police officials' promise that everything is being done to resolve the problem and that the
offenders have been apprehended. This model of "human interest" stories, as Paletz and
Entman (1981) refer to them, can also be extended to include the TSV phenomenon (p.
18). Those cases which offer an opportunity for drama, yet also give an opportunity for
resolution in a socially-acceptable way will be considered newsworthy and are more
likely to receive coverage. The critical component in this model in the way that it
explains the coverage of TSV cases is the idea of the "reassurance" portion of news
coverage. In order to reassure readers that a particular case of TSV is no longer a threat
to their physical or social wellbeing, journalists must provide a socially acceptable
conclusion in their coverage of TSV cases. From this reading, I hypothesize that a
socially acceptable resolution from the point of view of the average American will
include a situation where the offender has been detained and locked away. This is the
same conclusion that would be drawn in the coverage of NTSV cases. There too, one
would expect to see that the offender has been apprehended and is in jail. The difference
in the type of coverage that one would expect to see is that the TSV cases are not likely to
be presented as a social threat to the average white American. In the case of NTSV, it is
exactly the opposite. The average white American reader is likely to have a child who
attends public school and who is potentially endangered by this recent form of school
violence. This is a major explanatory factor in why NTSV is expected to be portrayed in
a different way than TSV. News agencies are compelled to cover NTSV differently
because of the affect that it has on their main audience, the average white American.
In summary, the literature appears to suggest that news agencies should portray
NTSV and TSV differently. NTSV provides an additional component of newsworthiness
that TSV lacks, and that extra component is the threat to personal wellbeing of the
average white American reader in both a physical and social way. When one considers
news agencies from a business perspective, and considers that the average white
American reader is the target audience of news producers, it becomes clear why news
agencies would likely provide different coverage for TSV and NTSV. By providing
detailed and resolute coverage of NTSV, news agencies are protecting their bottom line.4
News producers are attempting to create solidarity among their readers and assure them
that this new problem (NTSV) is under control; it is limited to a few "sick" and mentally
ill young people who exist at the fringe of the population of young people in America.
One could speculate that news agencies appear to provide the reassurance to maintain
4 By "resolute coverage," I mean that the news article will likely portray the case of NTSV as
being an abnormal occurrence as opposed to some new main-stream social movement by the
youths of America. The offender is abnormal, he or she is psychologically ill, unable to cope as
normal children are with the stresses of growing up. Blame may be placed on the parents of the
children. Expert advice will be given along with reassurance from various officials that
everything is being done that can be done to ensure the safety of America's schools. The U.S.
Congress and state legislative bodies debated various proposals with regard to the issue of gun
control in the aftermath of NTSV in order to reassure the average white American that everything
was being done to protect his or her wellbeing to include the safety of his or her family. Note
also that this kind of dialog does not take place with regard to TSV. There are not officials on
national media programs stating that they will do everything it takes to protect America against
TSV offenders. Congress has not debated laws that concern the need for resolution to systemic
problems within lower and working class communities that gives rise to TSV in schools. There
appears to be a completely different approach in the way that new agencies handle TSV.
current social beliefs that are necessary for the continuation of capitalist American
1.4.5 Gaps in the Literature
Specific gaps in the existing literature are as follows: No one has specifically
studied the relationship between media portrayal of school violence and the factors about
the school violence cases that affect the type of coverage given to the case. The conflict-
labeling literature provides a foundation for the idea that laws are created by those who
have power and those who do not have power are often labeled negatively with regard to
those laws based upon their interactions with those who have power. Power-threat
literature raises the idea that newspapers cover the different types of school violence
differently to support the current social ideologies and to reinforce those ideologies. The
differential coverage further segregates society symbolically, which perpetuates the
existing system of stereotypical beliefs. This creates a self-perpetuating cycle of
discrimination. The media criminology literature shows that media sources provide
coverage of TSV cases that are sensational and, again, perpetuate the current social
thoughts on TSV offenders. Finally, critical-media literature provides insight into what
makes a particular event newsworthy. Ultimately, the criteria are linked to the financial
prosperity of news agency. News producers are less likely to provide coverage that will
inhibit their agency's ability to make a profit and gain prestige.
5 The reassurance was necessary to provide a sense that everything is going to be all right. One
can easily see the ramifications of what a lack of reassurance and faith in the mainstream
American way of life can do for the American capitalist economy in light of current lack of
assurance and faith that the American public has in corporate America. In the face of current
uncertainties and the apparent inability of news producers to reduce the anxiety in America, the
economy is not doing extremely well. This is bad for those in power, those that have influence
over the media; those that have much to lose if American and World economies do not start to
move in a positive direction. Reassurance is a critical function of news agencies.
The real gap is that none of the researchers have tested these ideas on cases of
school violence directly. Turk (1964) and Quinney (1970) analyzed incidents in which
youths were coming into contact with authority figures in society usually through some
criminal justice system process, but not necessarily related to school violence. Blalock
(1967) developed his ideas around civil rights issues and interactions that African-
Americans had with white Americans. Howitt (1998), Barak (1994), McQuail (1992),
and Barlow, Barlow, and Chiricos (1995) studied the affects of media on the production
of crime. Paletz and Entman (1981) were interested in the process of news production in
general with no particular interest in media coverage of school violence. However, this
study seeks to focus on the relationship between the way newspapers portray school
violence and key factors about the offender's behavior that affect the way he or she is
portrayed by domestic newspapers.
The specific goals of this project are to determine (1) if domestic newspapers do
portray the two types of school violence (traditional and new trend) in different ways, (2)
if newspapers do differ in coverage of the two types of school violence, how do they
differ in coverage, and (3) finally, why do the newspapers differ in coverage of NTSV
and TSV? The differential coverage of school violence has something to do with
expected normal behavior which is linked with the discussion of newsworthiness offered
above.6 It is hypothesized that the cases of NTSV will be portrayed as abnormal or
deviant by the newspapers, where as TSV will be seen as less unusual.
6 Expected normal behavior is linked with newsworthiness in the following way: an event is
considered ne\\ s\\ orth\ if it involves drama and is entertaining. According to Paletz and Entman
(1981), drama as it relates to news coverage involves the idea that the event is threatening to the
public, both physically and socially, and that there is some resolution to the problem. Expected
normal behavior is related to the case of NTSV because these cases are sensation or involve
drama. The behavior that is exhibited by NTSV offenders is not expected and is not seen as
normal behavior for someone of their social position and race. In the case of TSV, the behavior
exhibited by the offenders is considered to be normal for someone of their social position and
race and is therefore considered to be expected normal behavior. Most likely, the only instance in
which TSV will receive special interest is where the offender's behavior is unexpected, or put in
the terms of Paletz and Entman (1981), where the behavior provides a social threat to the average
white American in general.
The purpose of this paper is to study the portrayal of domestic school shootings by
the newspapers in the U.S. using a compilation of theory and concepts developed from
the structural functionalist and conflict perspectives, labeling-theory, power-threat
hypothesis, media criminology, and critical-media perspective. There is no pre-existing
data-source with the variables of interest for this study. It is, therefore, necessary to
develop a dataset by using the Lexis Nexis search utility to find newspaper articles that
covered school shootings in the U.S. The time frame for the search has been limited to
include only articles which were published from 1st May 1997 through 30th April 2002.
Lexis Nexis allows for a search to be conducted using keywords. The keywords used in
this study were "school shooting." These keywords were used as opposed to "school
violence" because this study is specifically interest in school shootings rather than school
violence in general.1
Of the articles that were gathered for this study, the following list of twenty three
newspapers contributed either "full length" articles covering the cases of school shooting,
editorials, captions of a photograph at the scene of some school shooting or memorial
1 It is interesting to note, however, that the keyword search for "school violence" brought up a
roughly similar number of articles on the Lexis Nexis search. This search utility searches for
keywords in around fifty major newspapers worldwide. The main difference in using keywords
"school shooting" as opposed to "school violence" is that the articles related to the keywords
"school shooting" are primarily concentrated in the United States. While those articles found
using keywords "school violence" had a larger percentage of articles that were written about
incidents that occurred in other countries as well as the United States.
service, or some form of a chronological listing of school shootings with brief details on
each one: Rocky Mountain News, Sacramento Bee, The Washington Post, Los Angeles
Times, The New York Times, The Times-Picayune, The Baltimore Sun, The San Diego
Union-Tribune, Newsday, St. Louis Post-Dispatch, The Boston Globe, Milwaukee
Journal Sentinel, The Atlanta Journal Constitution, The Seattle Times, USA Today, The
Hartford Courant, The Plain Dealer, St. Petersburg Times, Chicago Sun-Times,
Information Bank Abstracts (Wall Street Journal), Star Tribune, The Buffalo News, and
The San Francisco Chronicle. Most of the newspapers offered either local or national
coverage of school shootings. Some newspapers offered both local coverage of events
that occurred within the community where the paper is published and coverage of school
shootings in other communities around the country.
2.1 Summary of Findings
News articles found using keywords "school shooting" for the five year period
were analyzed. Initially, a total of 757 articles resulted from the search. Articles
published by foreign news agencies or written about school shootings in other countries
were deleted. Articles were removed from the sampling frame if they were duplicates.2
In general, if there were duplicate articles, the more elaborate of the two was included in
the sampling frame. This method of selection for inclusion was chosen simply to have
some systematic way for choosing between duplicate articles. After this filtering process
was complete, a total of 578 articles remained. Of these 578 articles, I randomly selected
100 articles for inclusion in the sample. Of these 100 articles, 27 cases of school
A few general trends emerged within these datasets:
Table 2-1. Number of articles written per case.
Case Name Number of Articles
Albany HS (Albany, California) 1
Alva W. Dimmitt MS (Seattle, Washington) 1
Appalachian School of Law (Grundy, Virginia) 1
Buell ES Mount (Morris, Michigan) 2
Burlington HS (Burlington, Wisconsin) 1
Carter G. Woodson MS (New Orleans, Louisiana) 2
Columbine HS (Littleton, Colorado) 22
Destrehan HS (Destrehen, Louisiana) 1
Fort Gibson MS (Fort Gibson, Oklahoma) 1
Heath HS (Paducah, Kentucky) 2
Heritage HS (Conyers, Georgia) 10
John Batram HS (Philadelphia, Pennsylvania) 1
Lake Clifton Eastern HS (Baltimore, Maryland) 1
Lake Worth MS (Lake Worth, Florida) 1
Lew Wallace HS (Gary, Indiana) 1
Martin Luther King HS (New York, New York) 2
Parker MS (Edinboro, Pennsylvania) 1
Pearl HS (Pearl, Mississippi) 1
Public School 18 ES (Buffalo, New York) 1
Ridgewood HS (New Port Richey, Florida) 1
Santana HS (Santee, California) 12
Shawnee Mission South HS (Kansas City, Missouri) 1
Sumner HS (Sumner, Washington) 1
Thurston HS (Springfield, Oregon) 2
Vigor HS (Prichard, Alabama) 1
Westside MS (Jonesboro, Arkansas) 5
? ES, (East Boston, Massachusetts*) 1
Total number of articles 77
* The actual name of the school was not given in the article and other means of uncovering the
name of the school were unsuccessful.
Note: The names of the schools are given and the city and state are given. The abbreviations ES,
MS, and HS indicate elementary school, middle school, and high school respectively.
One can see from Table 2-1 that not all of the one hundred articles were written
about one specific case. Only seventy seven articles are actually written about any one of
the twenty seven cases of school shooting included in the sample. Of the thirty three
2 Duplicate articles were written about the same incident and published at two different instances
either by the same newspaper or by different newspapers. The articles were listed under the same
articles that are not written about a specific school shooting incident, one does not cover
school shooting at all, there are three different articles that list several cases of school
shootings that have occurred over a given period of time, and a majority of the twenty
nine remaining articles are editorials that offer reactions to school shooting in general but
that do not necessarily discuss a specific case of school shooting. Of the articles that do
discuss specific incidents of school shooting, most of them are concentrated in the
coverage of a few cases of school shooting. The vast majority of cases are only discussed
in one brief article.
2.2 Discussion of Data Characteristics
There are several limitations of these data that require discussion. However, I will
first provide a listing of the variables that were common among almost all articles
included in the sample. The following information was included in almost all articles:
title of article, placement of article in newspaper, number of words in article, publication
date of article, and the name of the school where the shooting occurred.
The data used in this study includes a wide variety of articles. Some are "full
length" descriptions of the events surrounding the shooting along with the reactions of
school officials and police. Others are editorial descriptions and analyses of the
shootings. Still others are very limited captions related to a photograph that was taken at
the site of the school shooting or a memorial service. There are approximately three
articles that list several incidents of school shootings that had taken place within the past
few years, and these include a short description of each case. Additionally, there are a
few articles that provide discussion of legislative and political issues that were debated
regarding school shootings and school violence in general.
title but the word count was slightly different.
The task of coding for pertinent variables was complicated by a lack of description
of these variables in a majority of the articles. Of the twenty seven cases included in the
sample, three describe race and only eight describe socioeconomic status. These two
variables are critical for this analysis. To supplement the lack of information about these
variables offered in the articles, the U.S. Census American FactFinder was utilized to find
both race and median family income figured for the cities where the school shootings
took place.3 These data were compiled at the case level of analysis along with other
composite findings from the one hundred articles in the sample. Therefore, there are two
data sets: (1) one with variables for each individual article, and (2) another for the
composite variables that provide information on each case of the twenty seven cases.
3 The American FactFinder can be found at www.census.gov.
FINDINGS AND ANALYSIS
3.1 Research Findings
As described in the previous chapter, the vast majority of cases discussed in this study
were only covered by one article. At the same time, there are a few cases that have a
relatively large number of articles describing the events surrounding a given shooting.
The articles in the sample vary greatly in their coverage of school shootings. Some of
them give in-depth descriptions of the shooting and the events surrounding the offense.
These articles tend to describe the reactions and plans of the school administrators as they
deal with school shootings. Other articles briefly describe the shooting and the
circumstances in which the shooting took place. A few "articles" were captions explaining
the content of photographs, which were not provided by Lexis Nexis. Editorials included in
the sample provide less descriptive coverage of the shootings, but they do provide insight
about the way the authors of these editorials think administrators should deal with school
shootings. Selective listings of school shootings are included in the sample. It is assumed
that these listings are those cases that the news publishers considered to be newsworthy
The general tendency of the in-depth articles written on school violence was that they
covered the more popular cases of school shootings, or put another way, they covered the
more newsworthy cases of shootings. These detailed articles provided at least the basic
information about each case, including the descriptions of the situation where the shooting
took place, the age, name and gender of the offender, and the number of offenders and
victims involved. The article would usually include a description of the surrounding
community, the neighborhood where the offender lived. They were typically based on
providing interviews with friends, teachers of the offenders, or survivors of the shooting and
sometime all three. There is likely to be some description of the school administrator's plans
for dealing with the shooting, some description of aide to the victims and survivors, and
possibly some discussion of legislation and gun policies. There are only a handful of these
articles in the sample and they cover only the few cases which seem to have received the
highest percent of coverage from the sample of articles used in this study.
Another slightly different type of article described the basic information of a case
briefly, and then, moved into a debate regarding legislation on gun control and school
violence.1 Some articles of this type discussed the political stances of the presidential
candidates with regard to gun control policy as they were campaigning for the 2000
elections. Others discussed the implications of the age of the school shooting offenders and
whether or not they would be tried by the judicial system as adults or juveniles. The
underlying theme of this type of article is that it usually revolves around the ability of school
officials to deal with the legal and political aspects of school shooting.
A third type of article reviewed in this study briefly describes the circumstances of the
shooting and provides some of the basic information for the case. Of the incidents in the
sample that received coverage by one article, a majority of those articles were of this type.
They were very simple descriptions of the case with limited description or discussion of any
Captions that describe the content of a photograph were included in this sample. There
are only a few of these captions in the sample itself. The captions provide a brief outline of
the case to include most of the basic information and then a brief description of the
photograph is given. Some of the captions only describe the photograph and do not describe
the basic information about the case. Therefore, one can see that the information given in
these captions varies greatly.
Editorials included in the sample provide a plethora of information about school
shootings. Most do not provide the detailed information about the case itself. They do not
given descriptions of the offenders, victims, and crime scene. They do, however, provide
information about the way that a variety of U.S. citizens see and interpret school shootings.
In essence, they provide a crude indication of how newspapers and other media sources
portray school shootings under the assumption that people formulate at least some of their
opinion about school shootings based upon what they have learned from news coverage. The
range of editorial opinions on this subject is so great that it is not feasible to describe each in
detail. However, a few general tendencies do arise. Some editorials focus on reactions to
school violence and tend to highlight a trend that the author feels is occurring in the U.S.
These editorials discuss policy initiatives, gun control laws, and other alternatives for putting
an end to school violence in the U.S. Another type of editorial, although rare, gives insight
into the portrayal of school shooting and describes that school shooting is not necessarily a
new phenomenon nor does the coverage of school shooting by the media give an accurate
representation of what is actually happening in U.S. schools. Most editorials fall along these
two types: (1) either providing a discussion on policy and legislation related to school
SBasic information includes variables like, number of offenders and victims, age of offenders and
victims, relation of offenders and victims, and name of school. Other than this the information
shootings and gun control, or (2) providing information concerning the actual issue of school
shootings in the U.S. and whether or not they are accurately described by media sources.
Finally, there are three articles that provide listings of school shootings in the U.S. over
certain periods of time. These listings are provided by journalists that work for a newspaper
company. They are therefore the listings of shootings that are considered newsworthy by the
news industry. The listings do not include all of the school shootings that occurred during
the period covered by the article. The primary utility of these listings is that it allows for a
comparison of findings from this study with regard to level of unique interest.
In summary, the articles in this sample tend to vary greatly in their coverage of school
shootings. Some of them offer in-depth descriptions of the shooting and the surrounding
events. Additionally, they may describe the reactions and plans of the school administrators
as they deal with school shootings. Other articles provide only a brief description of the
shooting and the circumstances. Editorials included in the sample provide less description of
the shootings, but more insight into the way that the authors of the editorials think about
school shootings. Listings of school shootings are included in the sample. The listings,
taken as a whole, provide a list of cases which the news industry believes to be newsworthy.
The following list of variables has been compiled in the dataset for this study. Each
article included in the sample has been coded for these variables. In the dataset that is non-
aggregated, these articles are coded. Several of these variables are used to build an aggregate
dataset that is coded for each case. Additional information has been collected for this
aggregate level data. The sources for these data are both the articles provided by Lexis
provided varied and even this basic information is not always given in each article.
Nexis, the year 2000 data from U.S. Census Bureau American FactFinder utility, and the
World Wide Web.
22.214.171.124 Non-aggregate Dataset
Table 3-1. Basic figures for non-aggregate variables.
Variable N Minimum Maximum Mean S.
section of the newspaper 99 1 4 2.48 0.98
number of words 100 20 2700 545 471
number of offenders 50 1 7 1.58 1.11
number of offenders killed 50 0 2 0.14 0.50
number of victims 47 0 45 8.70 9.63
number of victims killed 45 0 23 2.82 4.48
race of victim 5 1 3 1.80 1.10
race of offender 3 1 3 1.67 1.15
type of weapon 21 0 3 2.10 0.62
quantity of weapons 26 0 30 2.96 5.99
SES victim 10 1 2 1.80 0.42
SES offender 11 1 2 1.73 0.47
relation of victim to offender 47 1 2 1.28 0.45
unique interest 89 0 7 2.47 1.91
type of article 99 1 4 2.04 1.28
Section of the newspaper. Each article is published in a certain section of the
newspaper and this variable refers to that section. The variable was broken down into several
possible replies: (1) front page of news paper; (2) front page section but not on front page;
(3) front page of a section other than the front page section; and (4) in a section other than the
front page section not on front page of the section. On average, news articles were coded
2.48. The largest percent of articles were in the second category at 56.6%. Twenty five
point three percent of the articles fell within the third category. The vast majority of articles
do not make front page news. Of those articles that do appear on the front page, six out of
ten, cases are related to NTSV shootings.
Number of words. The number of words in the article is recorded as given by Lexis
Nexis. The mean number of words for the set of articles was 544.7 with a standard deviation
of 471.5. Of those articles which are in the top quartile of this variable, most of them are
NTSV cases. Most of the articles in the top quartile tend to score above the mean on the
variable that measures level of unique interest.
Case of school violence. This variable is simply the name of the school where the
school shooting took place and city and state where the school is located. Thirty three
articles in the sample did not have any entry for this variable. This was due to the fact that
they did not address a specific case, they did not focus on school shootings, or they covered
several school shootings.
Number of offenderss).2 The number of offenders involved in the case is taken
directly from the article. Fifty out of the one hundred articles include information on this
variable. Of the fifty articles that do include information on this variable, the average
number of offenders is 1.58 as given per article.
Number of offenders) killed.* This is the number of offenders reported in a given
article as being killed. Only the Santana High School and Columbine High School shootings
were reported as having offenders that were killed during the act of committing the school
shooting. Both of these can be classified as NTSV cases.
Number of victimss)* The number of victims involved in each case is taken directly
from the article. Forty seven out of the one hundred articles included in the sample give
information about the number of victims for a given case. The average number of victims
involved in a school shooting is 8.70 as given per article. However, 63.8% of the forty seven
articles including information about the number of victims include ten victims or less. The
2 The variables that are marked with asterisks were originally gathered at the article level of analysis
but were later compiled at the case level of analysis. A description of the data for these variables will
be given at both levels of analysis.
Columbine High School incident makes up 23.4% of the forty seven articles with a total of
forty five victims reportedly involved.
Number of victims) killed.* The average number of victims that are reported a
killed in the incident is 2.82. Of the forty five articles that include data on the number of
victims killed in a school shooting, 91.1% of them report five victims or less. The two cases
in which the number of victims killed was the highest are Columbine High School and
Westside Middle School. Both of these cases are considered NTSV cases given their level of
unique interest derived from the articles in this sample.
Race of the victim and race of the offender.* The race variable is coded as follows:
(1) white, (2) black, and (3) other. The race of the victim and the race of the offender are
both inadequately discussed in the news articles. Of the one hundred articles in the sample,
five gave information about the race of the victim. Three articles stated that the victims)
were white and two articles gave very strong indications that the victimss' race was not
white. Three articles specifically state the race of the offenderss. Of these three articles,
two reported the offenders) as being white while one article gave an indication that the
offender was not white.
Type of weapon used in the shooting.* This variable is used to record the type of
weapon that an offender is reported to have used in a school shooting incident. The variable
is coded in the following manner: (0) no weapon is used, (1) a weapon other than mass-
casualty producing weapon is used, (2) a gun other than mass-casualty producing gun is used,
and (3) a mass-casualty producing weapon is used.3
3 Weapons other than mass-casualty producing weapons are defined as those weapons which can be
used to assault and murder someone but this category does not include guns of any kind, examples
include knives, brass knuckles. Guns other than mass-casualty producing weapons includes guns that
can be used to kill people but do not use a high caliber ammunition and do not fire at a high rate of
Of the one hundred articles in the sample, twenty one reported specific information
about the type of weapon used in the school shooting. Of these twenty one articles, 76.2%
reported that offenders) used guns that were not mass-casualty producing to commit the
shooting. A total of four out of the twenty one articles reported that the offenders) used
mass-casualty producing weapons. Three cases were associated with the reported use of
mass-casualty producing weapons: Columbine High School, Thurston High School, and
Heritage High School. Of these cases, only Columbine High School appears to be a NTSV
case according to its level of unique interest.
Quantity of weapons) used by the offenderss)* This variable simply indicates the
number of weapons reported by an article as being used in the shooting. Thirty articles
report information regarding the number of weapons used in a shooting incident. Seventy
three point one percent of the articles report that one weapon was used in the offense.
Columbine was reported to have the highest count of weapons at thirty weapons to include
guns and bombs. The next highest count of weapons was nine for both Buell Elementary
School and Thurston High School. Westside Middle School shooting came in with the third
highest count reportedly involving the use of six weapons.
Socioeconomic status of the victims) and offenderss)* This variable was coded
as follows: (1) low, (2) medium, and (3) high. It was originally thought that the SES of the
offenders would be the same as the SES of their parents. The SES of the victims) and
offenders) was to be recorded if it was reported in the news articles; however, only a small
percent of the articles in the sample reported SES.
speed, examples include .22 caliber handgun, and .22 caliber rifle. Mass-casualty producing weapons
include weapons that can shoot large caliber ammunition and/or can fire at a high rate of speed. This
category also includes explosive devices that can produce a large number of casualties, examples
include shotgun, machine gun, high-powered rifle, and bombs.
Ten of the one hundred articles in the sample reported any information on the victims
SES. Meanwhile, eleven articles in the sample reported information on the SES of the
offenders. Of those articles that do report on the level of SES of victims and offenders, 80%
and 72.7% of the articles list the SES of medium, respectively.
Relation of victim to offender. This variable indicates the relation of the victim to
the offender. The main relations that are theoretically significant are the following: (1) is the
victim a peer to the offender or (2) is the victim an authority figure in relation to the
Forty seven out of the one hundred articles in the sample reported information on the
relation between the victim and offender. Of these forty seven articles, 72.3% reported that
the relationship between the victim and offender was a peer relation. The number reporting
that that victim was an authority figure in relation to the offender was 27.7%. Of this 27.7%
of articles reporting that the victim was an authority figure, there is no clear indication of
whether the majority of these articles are NTSV or TSV in nature.
Level of unique interest. The level of unique interest is a composite variable which
is an overall indication of whether or not the author of the article showed a level of interest in
the case that would be considered high with regard to this type of news coverage. There are
seven variables that were used to determine the level of unique interest (UI). Each of these
variables is coded (0) for a no response and (1) for a yes response indicating that "yes" the
article does portray this variable. The seven variables are described below: (1) is the
psychological-medical model used implying that the offender is somehow mentally unstable?
(2) Is the offender labeled as an outsider or portrayed as an outsider from the majority of
students? (3) Does the article used words such as empathy, sympathy, and understanding to
describe the school shooting and the events surrounding the shooting? This variable has two
potential positive answers: in one way, the article may be written in a way so that it makes an
attempt to understand what has happened, in another way the article may express sympathy
toward the victims and families of the victims. Either of these two attempts by the author
would result in an affirmative response for this variable. (4) Does the article portray a sense
that one should not expect a school shooting to happen in the community where it did? (5)
Does the article convey a sense of urgency that the problem of school shootings in America
is a critical problem that must be resolved right now? Also, is there a specific indication that
the authority figures involved in the case under analysis reacted promptly to alleviate the
shooting incident and any problems that may have arisen regarding the incident? Either of
these two indications in the article would have given an affirmative response for this
variable. (6) Is there a description of the actions that the school administrators and other
authority figures involved in the case have taken to rectify the problem? (7) Does the article
describe any aide that was provided for the victims and other students involved in the
incident? Each of these seven variables was coded and the scores were tallied, and the
resulting figure was the level of unique interest (UI) that was recorded for a given article.
The range of score for the level of UI of any given article was zero to seven. If none of these
variables were exhibited in the article, then a total score of zero UI would be recorded for the
article. If all of the variables exhibited in the articles, then a score of seven would be
recorded for the article.
A total of eighty nine out of the one hundred articles in the sample were given a score
on this variable. The actual range provided by the data for this variable was from zero to
seven. The average score for any given article was 2.47. Of those articles that scored higher
than the mean of 2.47, Columbine High School has eight hits, Santana High School has six
hits, Carter G. Woodson Middle School has two hits, Westside Middle School has two hits,
Heritage High School has two hits, Albany High School has one hit, Alva W. Dimmitt
Middle School one hit, Burlington Middle School has one hit, Ridgewood High School one
hit, Thurston High School one hit, and there are ten additional articles that scored above the
mean on the UI scale but are not associated with any particular case of school violence.4
Type of article in newspaper. This variable refers to the type of article as one of the
following choices: (1) article related to one case of school shooting, (2) editorial, (3) citation
of a photograph, (4) other which includes articles that do not fall into any of the other three
Ninety nine of the one hundred articles in the sample were coded for this variable. The
one article that was not coded is the article that is not related to school shootings at all. Of
these ninety nine articles, fifty two are "full length" articles that discuss one case of school
shooting. There are a total of seventeen editorials in the data sample. Four photographic
citations are included in the sample. A total of twenty six articles were listed as other. Of
the articles listed in the "other" category, there appears to be no trend in the characteristics of
these articles other than the fact that most of them are missing information on basic variables
in this dataset and many of these articles have a small word count.
126.96.36.199 Aggregate Dataset
Number of offenderss)* This variable refers to the number of offenders that are
reported for each case. Twenty five of the twenty seven cases have reported figures on the
number of offenders. Sixty four percent of the cases are reported as having one offender.
4 Hit indicates that there is one article written about the case that has a UI score above the mean of
Twenty four percent of the twenty five cases reported on were listed as having two offenders.
Destrehan High School, Burlington High School, and Pearl High School were reported to
have four, five, and seven offenders, respectively.
Table 3-2. Basic figures for aggregate variables.
Variable N Minimum Maximum Mean Std. Deviation
number of offenders 25 1 7 1.76 1.48
number of offenders killed 25 0 2 0.12 0.44
number of victims 25 0 45 6.00 10.25
number of victims killed 24 0 13 1.58 2.81
race of victim 3 1 3 1.67 1.15
race of offender 3 1 3 1.67 1.15
revised race of victim 8 1 4 2.00 1.41
revised race of offender 13 1 3 1.38 0.65
type of weapon 15 0 3 2.07 0.70
quantity of weapons 18 0 30 3.78 7.10
SES victim 8 1 2 1.75 0.46
SES offender 9 1 2 1.67 0.50
number of articles per case 27 1 22 2.85 4.69
articles written within 1 year 27 0 20 2.41 4.38
unique interest within 1 year 23 0 5.0 2.24 1.49
median family income 27 $23,519 $66,870 $53,226 $11,599
percent white 27 11.90 95.90 66.87 24.83
percent black 27 0.4 84.50 23.18 25.55
percent other 27 1.30 29.30 9.74 8.67
total population of city 27 1,105 8,008,278 481,685 1,541,693
Number of offenders) killed.* This variable records how many of the offenders
were killed as a result of their involvement in the shooting. Columbine High School and
Santana High School are the only two cases in this dataset that involve the murder of the
offenders. No other offenders in this dataset were killed as a result of their involvement in
Number of victimss)* Forty eight percent of the twenty five cases reported that the
number of victims were one or less. A total of 20% of the twenty five cases reported no
victim at all. Eighty percent of the twenty five cases did not have more than eight victims.
Santana High School, Westside Middle School, Thurston High School and Columbine High
School had a reported total of 15, 15, 26, and 45 victims, respectively.
Number of victims) killed* This variable refers to the number of victims that were
reportedly killed as a result of the school shooting. Forty five point eight percent of twenty
four cases report that no victims were killed. Twenty five percent of twenty four cases show
that one victim was killed. Columbine surpasses all of the other cases in this dataset with a
total of thirteen victims dead.
Race of the victim and race of the offender.* The races reported in this dataset are
recorded directly from each news article. In instances when the article did not explicitly state
the race of the victim or offender, no race was recorded. A total of three cases in this dataset
had information that indicated the race of the victim: two cases had white victims and one
case had non-white victims. Twenty four cases in this set do not include information about
the race of the victim because it is not explicitly given in any of the articles in the sample.
Likewise, there are a total of three cases in this dataset that report the race of the offender:
two cases reported white offenders, while one case reported non-white offenders.
The race of offender and victim variables have been revised based upon additional
internet searches performed for the explicit reason of identifying the races of offenders and
victims. The new results of these searches are listed in Table 3-2 above. An additional race
category had to be created for the revised variable. The new category is (4) multiple races
which means that there were either victims or offenders of different races involved in a given
case. For example, in the case of the Columbine shooting, two white offenders shot and
killed several white victims and one black victim. The Santana shooting was the only one
other case involved the used of the new code. The offender in the Santana shooting was
white and he shot and killed one white victim and one Asian-American victim. Both
Columbine and Santana are cases of NTSV.
Type of weapon used in the shooting.* This variable is used to record the type of
weapon that an offender is reported to have used in a school shooting incident. The variable
is coded in the following manner: (0) no weapon is used, (1) a weapon other than mass-
casualty producing weapon is used, (2) a gun other than mass-casualty producing gun is used,
and (3) a mass-casualty producing weapon is used.
Fifteen cases reported information for this variable. Of these fifteen, one case reported
that no weapon was used. Three cases reported that mass-casualty producing weapons were
used, and the remaining cases were reported to have guns other than mass-casualty producing
guns involved in the shooting. Heritage High School, Thurston High School, and Columbine
High School reportedly involved the use of mass-casualty producing.
Quantity of weapons) used by the offenderss)* A total of eighteen out of the
twenty seven cases reported data describing the number of weapons used in the school
shooting. Of these eighteen cases, one reported that no weapons were used. This case was
actually one in which the offenders were planning an attack on the school but the police
found out about the plan before the offenders were able to carry out the attack. The offenders
had not been able to gather any weapons. Sixty six point seven percent of the eighteen cases
reported that one weapon was used in the offense. Of the remaining cases, an unknown
elementary school in Boston, Westside Middle School, Thurston High School, Buell
Elementary School, and Columbine High School reportedly had two, six, nine, nine, and
thirty weapons involved in the shootings, respectively.
Socioeconomic status of the victims) and offenderss)* Eight of the twenty seven
cases reported on the SES of the victims. Of these eight cases, the SES of the victims of two
cases was reported to be low. The remaining six cases indicated that the SES of the victims
was medium. Nine of the twenty seven cases in this dataset reportedly have data on the SES
of the offenders. Of the nine cases with SES data on the offenders, three cases have
offenders of low SES. Six of the nine cases, report that the offenders' SES is medium.
Number of articles per case. The findings of this variable were shown in chapter
two. They will be described here briefly. Sixty six point seven percent of the cases in this
dataset have one article included in the sample that corresponds directly with this case.
Eighteen point five percent of the cases have two articles in the sample the directly
correspond to the case in the dataset. Westside Middle School, Heritage High School,
Santana High School, and Columbine High School have five, twelve, twelve, and twenty two
articles, respectively, that are in the sample and that correspond directly with these four
Number of articles written within first year. This variable measured how many
articles are written about a given case up to one year after the school shooting occurred.
Three of the twenty seven cases in this dataset have no articles written corresponding to the
case within the first year. Sixty six point seven percent of the cases in this set have one
corresponding article in the sample that is written within the first year. Carter G. Woodson
Middle School, Martin Luther King High School, Buell Elementary School, Heritage High
School, Santana High School, and Columbine High School had two, two, two, ten, twelve,
and twenty two articles, respectively, written about each case within the first year after the
Unique interest for articles written within first year. This variable refers to the
average unique interest (UI) per case based upon the UI scores of all of the articles written on
a given case within the first year. This method of recording UI per case for the first year may
be misleading because those cases that have more articles written about them within the first
year following the shooting tend to have disproportionately smaller UI levels even though
some of the articles that were written about these cases do have high levels of UI.
The range of UI for twenty three of the twenty seven cases in the set is from zero to
five. Sixty five point two percent of the twenty three cases are at or below 2.2 on the UI
scale. The mean UI for these twenty three cases is 2.41. Albany High School, Westside
Middle School, Burlington High School, Thurston High School, Carter Middle School,
Ridgewood High School, and Alva W. Dimmitt High School have 3.0, 3.0, 4.0, 4.0, 4.0, 5.0,
and 5.0 levels of UI, respectively. Each of these cases has one or two corresponding articles
in the sample.
Median family income. The data for this variable were gathered from the U.S.
Census Bureau's American FactFinder search utility on the www.census.govwebsite.5
These data were organized on the basis of city name. All of the data are from the 2000
census updates. It is understood that all of the cases did not occur in the year 2000, but an
assumption has been made that the data used from the census have not changed significantly
over the period of time included in this study.
The minimum family income recorded for any of the cities where the shooting
occurred was $23,519 per year. The maximum median family income recorded for any of
5 The American FactFinder is a utility created by the U.S. Census Bureau that assists in finding
demographic, economic, and various other forms of information that are gathered by the census
bureau. These data are usually orders in geographically delimited regions. The user can search based
the cities was $66.870. The median dollar amount for the median family income is $43,266
with a standard deviation of $11,599.
Percent white. These data were also gathered from utilizing the U.S. Census
Bureau's American FactFinder. The same qualification that applies above applies here as
well. The minimum percent white for a given city where a shooting took place is 11.9%
while the maximum percent is 95.9%. The mean percent white is 66.9% with a standard
deviation of 24.8%.
Percent black. These data were collected from the American FactFinder 2000. The
smallest percent of the black population in a city where school shooting took place is 0.4%
while the largest percent is 84.5%. The mean percent black for cities within the sample
included in this dataset is 23.2% with a standard deviation of 25.6%.
Percent other. Percent of those cities with a race other than white and black was also
gathered using the American FactFinder utility. The minimum percent of other race recorded
for this sample is 1.3% while the maximum percent other is 29.3%. The mean percent other
race in this set of data is 9.7% with a standard deviation of 8.7%.
Total population of city where shooting took place. The data for this variable
originated in the American FactFinder utility. Of the cities that were included in the sample,
the smallest city had a total of 1,105 people and the largest city had a total of 8.0 million
people. The mean city population size was 481,685 people with a standard deviation of 1.5
million. If we exclude the outlier with the population of 8.0 million people the mean
population is 139,187 with a standard deviation of 218,138 people.
upon the limits for the region that he or she desires. In this study, I limited my search by city-level
3.2 Analysis of Findings
3.2.1 Qualitative Analysis
These findings will now be summarized and analyzed in the following qualitative
analysis of the descriptive data given above. A majority of articles fall within front page
section of the news paper but do not appear on the front page. Most of the articles are several
hundred words in length although a few articles do exist in either extreme: very short or very
lengthy. Carter G. Woodson Middle School, Columbine High School, Thurston High
School, Santana High School, Westside Middle School, and Ridgewood High School appear
in the top 25% of those articles with the highest word count.
The majority of cases in this dataset have only one or two articles written on them that
are included in this dataset. There are a few exceptions to this trend: Westside Middle
School, Santana High School, Heritage High School, and Columbine High School all have
five or more articles written about them that are included in this dataset. The vast majority of
cases included in this study also had most of the articles written about them within the first
year after the incident occurred. Again, there are exceptions to the rule: Heritage, Santana,
and Columbine High Schools all have nine or more articles written about them within the
first year after the incident occurred that are present in this sample.
Of the cases included in this dataset, twenty three had articles written within the first
year of the incident. A considerable percentage of these twenty three cases scored a level of
approximately 2.0 on the UI scale. However, a considerable percentage scored 0.0 and 4.0
on the UI scale as well. The following list indicates the top 25% of cases that ranked the
highest on the UI scale during the first year: Albany High School, Westside Middle School,
Burlington, Thurston High Schools, Carter G. Woodson Middle School, Ridgewood High
School, and Alva W. Dimmitt Middle School. A similar trend does appear in these results as
it does in the previous results, but there are additional cases in these results that do not appear
in the previous results.
Most of the articles report no deaths of offenders and only a few deaths among the
victims. However, there is a trend that starts to appear in this variable and will continue
throughout the paper. Columbine High School had a total of two offenders and both killed
themselves during the assault on the school. At Santana High School, the one offender
involved in the case was reported as being killed during the shooting. Destrehan High
School, Burlington High School, and Pearl High School were unique in that they had four,
five, and seven offenders involved in the planning and or actually committing the shooting,
The number of victims reportedly involved in a case range greatly from one through
forty five. Here again, the trend develops: Pearl High School, Santana High School,
Westside Middle School, Thurston High School, and Columbine High School are reported to
have greater than ten victims involved in the shooting. While Pearl High School, Westside
Middle School, Thurston High School, and Columbine High School all have the highest
number of victims that are killed during the shooting as well.
The only comment that can be made about the race of offenders and victims based
upon the data given in the articles is that race is not discussed explicitly. One could make
assumptions about the race of the offenders and victims based upon general knowledge of the
racial makeup of the U.S. in addition to the language that is used to describe the school
shooting, but an attempt has not been made to do so here. The methods used in this study do
not allow for me to clearly state the race of the offender or victim, unless it is explicitly
stated in the article. In summary, 97% and 95% of the articles do not give information about
the race of the offenders and victims, respectively.
When describing the type of weapon used in the school shooting, there is also a lack of
information available from news articles in this sample. Seventy six percent of the sample
does not include information about the type of weapon involved in the shooting. Of those
articles that do present information about the type of weapon used in the shooting, the
majority of the weapons that are used are guns other than mass-casualty producing guns e.g.,
a hand-gun. Heritage, Thurston, and Columbine High School shootings reportedly involved
the use of weapons of mass-casualty producing capabilities. This gives additional evidence
to the trend that appears to be manifesting itself in the data.
It appears that of the twenty seven cases found in this dataset, eighteen have
information about the quantity of weapons used in the assault on the school. Most of the
offenders used one weapon. From the previous description on type of weapon, it is likely
that this weapon is a handgun. However, there are several cases were the offenders used
many weapons of varying types. The following list indicates cases in which the number of
weapons used was in the top 25% reportedly used to carry out the offense: Unknown
Elementary School in Boston, Westside Middle School, Buell Elementary School, Thurston
High School, and Columbine High School. Again, a pattern is emerging from the data.
A majority of the articles included in this sample are defined as regular length articles
that describe a school shooting incident in sufficient detail. There are also, however, several
articles that are editorials presenting a variety of views and discussions on school shooting.
Columbine High School has the largest number of these "normal" length articles written with
a total often. Santana High School has the second largest number of "normal" length articles
written about the shooting that happened there with a total of eight articles. Heritage High
School has the third highest number of these "normal" length articles written with a total of
five. This supports the trend that has appeared in the study thus far.
The SES of the victims and offenders is an integral part of the theoretical argument of
this paper. There is a huge gap in this study left by the inadequate information that is
provided about this variable on the individual level by the news articles. Of the twenty seven
cases in this dataset, only eight and nine give explicit information about the victims' and
offenders' SES respectively. The majority of offenders and victims appear to be middle class
according to the vague information provided. Instead of using this information, the study
utilized the American FactFinder utility of the U.S. Census Bureau to find data on a proxy
measure of SES for the city where the school shooting took place. The proxy measure used
is median family income.
The range of the median family income of those cities were school shootings have
taken place varies greatly. The names of the schools located in a town that is in the top 25%
of median family income are listed below: Albany High School, Alva W. Dimmitt Middle
School, Appalachian Law School, Burlington, Columbine, Destrehan High Schools, Parker
Middle School, and Santana High School.
The names of the schools in the cities that are in the top 25% for each race category are
listed below. These schools that are in cities with the top 25% of whites in the population
are: Santana, Thurston, Sumner, Columbine High Schools, Parker Middle School, Buell
Elementary School, Ridgewood, and Burlington High Schools are in the top 25% of the
white percent population variable. Shawnee Mission South High School, Heritage High
School, Public School 18 Elementary School, John Batram, Lake Clifton Eastern High
Schools, Carter G. WoodsonMiddle School, Lew Wallace High School, and Vigor High
School are all in the top 25% of the black percent population among the schools included in
this dataset. Santana, John Batram High Schools, Lake Worth Middle School, unknown
elementary school in Boston, Alva W. Dimmitt Middle School, Martin Luther King High
School, Albany High School, and Fort Gibson Middle School are the top 25% of cases in the
other race category which is made up of all other racial groups beside white and black. Many
of the schools listed in the percent black and percent other categories have not been present
in the other categories and variables that have been described thus far. This indicated that
many of these other schools are not receiving the media attention that has been given to the
schools that have been present in the previous listings.
The schools that fall within the top 25% of cities with the largest populations are listed
below: Public School 18 Elementary School, Shawnee Mission South High School, Carter G.
Woodson, Alva W. Dimmitt Middle Schools, unknown elementary school in Boston, Lake
Clifton Eastern, John Batram, and Martin Luther King High Schools. These schools appear
to be the some of the same cases that have been showing up in the top 25% of black and
other racial categories.
In summary, the following general trends appear to be emerging out of the data that
exist in this sample of articles and cases: (1) the majority of articles in the sample appear in
the front page section of the news paper. (2) Most of the cases described in this dataset are
only covered by one article that is included in the sample. (3) Twenty three of the cases
included in this dataset had articles that were published within the first year after the
occurrence happened. Of these, the average score on the UI scale was approximately 2.0.
(4) The vast majority of articles reported no deaths among the offenders and an average of
two deaths among the victims in any given case. (5) The actual number of victims involved
in any given case varies significantly from case to case. (6) There is a complete lack of
explicit information about the race of the victims and race of the offenders. (7) Many of the
articles do not provide information about the type of weapons used in the offense. Of the
approximately 25% of articles that do provide information about the type of weapon used,
the primary type of weapon used was a handgun. (8) Slightly over half of the twenty seven
cases described by these data offer information about the number of weapons used in the
offense. In a majority of the cases that do provide information, the offender used one
weapon to commit the offense. (9) A majority of the one hundred articles included in the
sample provide "full length" coverage of any given case of school shooting. This is not to
say that the "full length" articles are distributed evenly among the various cases. (10)
Information concerning the SES of the victims and offenders is limited. Only several articles
provide this information in the sample. Of these articles, the majority of them indicate that
the victims and offenders are of middle class SES levels. (11) The range of the median
family income varies greatly for this dataset. (12) The range of the actual population size of
the cities included in this sample varies greatly. (13) The racial makeup of most of the cities
included in the sample is primarily homogeneous. There are a few exceptions, but the bulk
of cities in this sample consist of one racial group that consists of a large majority of the city
Table 3-3 provides a visually representation of which cases appear to be exceptional.
The cases that have a high incidence of checked boxes for variables one through twelve and
fifteen should also be those cases that can be labeled as NTSV cases. Each time a case has a
parameter that appears in the top 25% of a variable, there will be a checked box for that
variable. The case with the highest number of checked boxes means that this variable has
parameters which appeared in the top 25% of a lot of variables. The variables one through
twelve and fifteen are those that, by definition, would indicate NTSV cases. Alternatively,
those cases that have boxes marked for columns thirteen and fourteen are less likely to be
labeled as NTSV cases.
Table 3-4 provides a listing of those cases that have the largest number of markings in
the boxes that would affirm NTSV status from lowest to highest. Columbine, Santana,
Thurston, Westside, Burlington, Heritage, Albany, Alva W. Dimmitt, Pearl, and Ridgewood
are within the top 25% of the list of cases on Table 3-4. Two high schools in the top 25% of
cases are Columbine High School and Thurston High School. Two high schools in the lower
25% of cases are John Batram High School and Lew Wallace High School. To keep a more
homogeneous group for comparison, the two bottom ranking and two highest ranking high
schools will be analyzed in more detail in chapter four.
Three listings that were found by Lexis Nexis provide basic information on several
cases of school violence ("List of school shootings," 1998; "School shootings," 1999; "Other
school shootings," 2000). Based upon an analysis of the listings, Pearl High School,
Westside Middle School, Thurston High School, Heath High School, Lincoln County High
School, and Parker Middle School are likely to be portrayed as NTSV cases. These findings
are based on the assumption that news journalists and editors chose the cases for inclusion in
the listings. It is more likely that NTSV cases will appear in the listings than TSV cases.
However, it may be possible that TSV cases do appear in the listing, but it is likely that any
TSV cases that appear in the listing will have some sensational component that makes them
more newsworthy. According to the listings, the following schools are less likely to be
NTSV cases, Bishop Neumann High School, Santana High School, Fort Gibson Middle
School, a school in Deming, New Mexico, Heritage High School, a high school in Notus,
Idaho, a school in St. Charles, Missouri, a school in Onalaska, Washington, Columbine High
School, and Bethel Alaska High School. Only four of these schools have coverage in this
sample. The others are not included in the sample. Therefore, conclusions cannot be drawn
about whether or not these findings do convey that these cases should be less likely to be
NTSV cases. However, in general, for the cases that do coincide in both the sample and the
listings, there does not appear to be a strong correlation between what the listings predicted
would be NTSV and what is actually provided from the data gathered in the sample.
3.2.2 Quantitative Analysis
There are several key correlations between variables within this dataset. This
discussion starts with the analysis of a composite variable that was first represented in Table
3-4. This variable is labeled NTSV vs. TSV (NVT) it is simply the total listed in column
four of Table 3-4. The higher the score the more likely it is that the case is one of NTSV.
For example, Columbine scored ten points on the NVT scale. The relationship between NVT
and the number of articles written per case is strongly positive (0.724*). Similarly, a
moderately strong positive relationship exists between NVT and articles written within first
year, unique interest for first year, median family income, percent white, number of victims
killed, and quantity of weapons at the following correlations 0.691**, 0.504*, 0.560**, 0.610**,
0.654**, and 0.627**, respectively. A moderately strong negative relationship exists between
NVT and percent black (-0.559**), and race of offender (-0.619*).
The number of articles written within the first year since the shooting occurred is
positively correlated with the number of articles written about a given case in general,
correlation is 0.981**. The number of articles written per case is directly correlated with the
Table 3-3. Matrix of cases with exceptional qualities.
Name 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
unknown elementary school X X X
albany high school X X X
alva w dimmitt middle X X X X
appalachian school of law X
buell elementary school X X
burlington high school X X X X
carter g. woodson middle X X X X
columbine high school X X X X X X X X X X
destrehan high school X X
fort gibson middle school X
heath high school
heritage high school X X X X
john batram high school X X X
lake clifton eastern high X X
lake worth middle school X
lew wallace high school X
martin luther king high X X
parker middle school X X
pearl high school X X X
public school 18 elementary X X
ridgewood high school X X X
santana high school X X X X X X X X
shawnee mission south high X X
sumner high school X
thurston high school X X X X X X X
vigor high school X
westside middle school X X X X X X
1 Articles within top 25% of word count.
2 Cases in top 25% of number of articles.
3 Cases in top 25% of UI for first year articles.
4 Number of "full length" articles written on one case is five or more.
5 Cases in top 25% of number of offenders killed.
6 Cases in top 25% of number of offenders.
7 Cases in top 25% of number of victims.
8 Cases in top 25% of number of victims killed.
9 Cases in top 25% of that involve the use of mass-casualty producing weapons.
10 Cases in top 25% of number of weapons used.
11 Towns in top 25% of median family income.
12 Cities in top 25% of white race category.
13 Cities in top 25% of black race category.
14 Cities in top 25% of other race category.
15 Cities in top 25% of total population.
Table 3-4. List of cases from lowest to highest affirmation of NTSV status.
School Displayed Interest Factors Causing Interest Total NVT
columbine high school 3 7 10 N
santana high school 3 5 8 N
thurston high school 2 5 7 N
westside middle school 2 3 5 N
burlington high school 1 3 4 N
heritage high school 2 2 4 N
albany high school 1 2 3 N
alva w dimmitt middle school 1 2 3 N
pearl high school 0 3 3 N
ridgewood high school 2 1 3 N
buell elementary school 0 2 2 N
destrehan high school 0 2 2 N
parker middle school 0 2 2 N
unknown elementary school Boston 0 2 2 N
appalachian school of law 0 1 1 T
carter g. woodson middle school 2 -1 1 T
sumner high school 0 1 1 T
heath high school 0 0 0 T
fort gibson middle school 0 -1 -1 T
lake clifton eastern high school 0 -1 -1 T
lake worth middle school 0 -1 -1 T
lew wallace high school 0 -1 -1 T
martin luther king high school 0 -1 -1 T
public school 18 elementary school 0 -1 -1 T
shawnee mission south high school 0 -1 -1 T
vigor high school 0 -1 -1 T
john batram high school 0 -2 -2 T
Std Deviation 3.04
1. Displayed interest is the number of boxes checked in Table 3-3 columns 1- 4 for each case. This
indicates the actual level of interest that media sources have with regard to a case.
2. Factors causing interest are those factors in Table 3-3 columns 5 14. From the theory one expects
that cases with columns 11 and 12 are more likely to gain special interest from the media. This
column is computed in one of two ways: for those cases with no check in Table 3-3 columns 13 and
14, the sum of checks in columns 5 12 is added. For those cases with checks in Table 3-3 columns
13 or 14, checks in columns 5 12 are added. For those cases with checks in Table 3-3 columns 13 or
14 and no checks in columns 5 12, the checks in blocks 13 and 14 were subtracted. The theory
behind this line of reasoning is that for cases that occur in mostly white populations and those that
occur in more wealthy populations will be likely to gain the unique interest of the media. On the
other hand, in communities where the population is mainly made up of minority groups, the media are
less likely to show a unique interest in the case unless there is some thing sensational about the case
yet still the case must allow for the maintenance of socially acceptable ideologies.
3. The T vs. NT column represents the relative status of the case: is it T or Traditional School
Violence or is it NT or New Trend School Violence. This can only be determined on a relative scale.
number of offenders that are killed in the school shooting, correlation 0.919*. Also, the
number of articles written per case increases directly as the number of victims involved in the
case increases (0.806**), and as the number of victims that are killed increases (0.761**). The
number of articles written per case increases when the race of the victim is either black or
some other minority group (0.776*). The number of articles per case also increases directly
in relation with the quantity of weapons that are reportedly used or planned to be used by the
The number of articles written within the first year of the shooting appears to be
relatively weakly correlated with the median family income of the city where the shooting
occurred (0.395*). The number of articles written within the first year is strongly and directly
correlated with the number of offenders that are killed in the school shooting (0.928*).
Likewise, the number of articles written within the first year increases directly as the number
of victims involved in the case increases (0.745**) and as the number of victims that are
killed increases (0.682**). Cases involving minority victims are likely to have a larger
number of articles written about them within the first year after the incident (0.836*). The
number of articles written within the first year increases directly in relation with the quantity
of weapons that are reportedly used or planned to be used by the offenders (0.712**).
The level of unique interest (UI) of the articles that are written within the first year
appears to be correlated with the percentage of white people in the population of the city
where the school shooting took place in a moderately positive direction (0.467*). The level
of unique interest (UI) of the articles that are written within the first year appears to be
correlated with the percentage of black people in the population of the city where the school
shooting took place in a moderately negative direction (-0.445*). The level of UI of articles
written within the first year is relatively strongly correlated with the race of the offenders in
such a way that if the offender is white news agencies are more likely to display a greater
level of UI (-0.730*).
The level of UI of articles that within the first year represents the average level of UI
for all of the articles published about a given case within the first year of the occurrence. The
level of UI is a broad indicator for type of school violence. A high level of UI indicates that
it is likely that the case is a NTSV case. The lower the level of UI the more likely it is that
the case is a TSV case. This finding indicates that cases with higher levels of UI tend to be
more likely to have happened in communities where there are a higher percentage of white
Americans living in the community. This finding also indicates that the higher the level of
UI for a given case the smaller the percentage of black Americans in the community where
the school shooting took place.
The median family income is higher in those communities where a larger percentage
of the population is white (0.490**) and lower in those communities where a larger
percentage of the population is black (-0.572 ). The correlation between the median family
income of the community where the school shooting took place and the number of offenders
killed in the shooting is 0.495*.
The percentage of white people in the community where the shooting took place is
moderately correlated with the race of the offender in a way such that white offenders are
more likely in communities with larger populations of white people (-0.668*). The percent of
white people in the population is positively and moderately correlated with the SES of the
victim (0.717 ). This means that in communities where a larger percent of the population is
white the victim is more likely to have a high SES. On the other hand, as the percentage
Table 3-5. Correlations of aggregate variables.
Table 3-6. Table of significant correlations for aggregate variables.
Variables Corr. Sign. N Variables Corr. Sign. N
NTSV vs TSV (NVT)/number of articles .724** .000 27
NVT/articles written within 1styr
NVT/unique interest for 1st yr
NVT/median family income
NVT/number of offenders killed
NVT/number of victims
NVT/number of victims killed
NVT/race of offender
NVT/quantity of weapons
number of articles per case/articles
written within 1styr
number of articles per case/number of
number of articles per case/number of
number of articles per case/number of
number of articles per case/race of
number of articles per case/quantity of
articles written within 1st yr/median
articles written within st yr/number of
articles written within 1st yr/number of
articles written within 1st yr/number of
articles written within 1st yr/race of
articles written within st yr/quantity of
unique interest 1st yr/percent white
unique interest 1 styr/percent black
unique interest 1st yr/race of offender
median family income/percent white
median family income/percent black
median family income/number of
percent white/percent black
percent white/race of offender
percent white/SES victim
percent black/race of offender
percent other race/total population of
total population ofcity/SES victim
number of offenders/type of weapon
number of offenders/SES victim
number of offenders killed/number of
number of offenders killed/number of
number of offenders killed/race of
number of offenders killed/quantity of
number of victims/number of victims
number of victims/type of weapon
number of victims/quantity of weapons
number of victims killed/quantity of
race ofoffender/SES of offender
SES victim/SES offender
of black people in the population of the community where the shooting took place
increases the offender is more likely to be a minority (0.694**). The smaller the
population of the city where the shooting took place the lower the SES of the victim (-
0.936 ). White offenders are more likely to have higher levels of SES (-1.00*). On the
other hand, black offenders are more likely to have lower levels of SES. The SES of the
victim is directly correlated with the SES of the offender (0.745*). This indicates that
offenders with higher SES are likely to assault victims with higher SES. Offenders with
lower SES are likely to assault victims with lower SES. This also means that white
offenders are more likely to shoot victims with higher SES and black offenders are more
likely to shoot victims with lower SES.
The number of offenders involved in a given case of school shooting appears to be
relatively strongly correlated with the type of weapon used in a negative direction (-
0.744*). Based upon the coding of the "type of weapon" variable that means that as the
number of offenders involved in an offense increases it is more likely that these offenders
will not use mass-casualty producing weapons. In other words, the larger the number of
offenders the less violent the type of weapons used in the attack. As the number of
offenders involved in an offense increases, the SES of the victim is likely to be less (-
0.745*). Cases involving attacks on poor victims are likely to involved larger number of
There appears to be a strong positive correlation between the number of offenders
that die in an attack and the number of victims involved in the school shooting incident
(0.805**). Similarly, there is a strong positive correlation between the number of
offenders killed and the number of victims killed in any given case (0.802*). The
number of offenders killed in the shooting incident is directly correlated with the number
of weapons used in the shooting in a relatively strong way (0.798*). The number of
offenders killed is highly correlated with the race of the victim. When a larger number of
offenders are killed, minority victims are likely to be involved in the case (0.903**).
The number of victims involved in any given case appears to be correlated in a
strong positive direction with the number of weapons found to be used or planned to be
used in each case (0.866*). In situations where offenders seemed determined to kill a
large number of people or to cause significant damage and thereby they used several
weapons, it is more likely for these cases to involve many victims because that was the
intent of the offenders. Similarly, the actual number of victims that are reported as dead
increases as the number of weapons found to be associated with any given case increases
(0.903*). Another intuitive relation that comes from the data is that there appears to be a
strong correlation between the number of victims involved in a case and the number of
victims that actually die as a result of the shooting (0.942 ). It appears that the greater
the number of victims involved in the case the more deadly the type of weapons used in
the case (0.563*). This means that when the offenders use weapons of mass-casualty
producing capabilities there are more victims produced as a result of the shooting.
DETAILED DISCUSSION AND ANALYSIS OF SPECIFIC INCIDENTS
According to Table 3-4 List of cases from lowest to highest affirmation of NTSV
status, Columbine High School and Thurston High School are two of the highest ranking
schools and John Batram High School and Lew Wallace High School two of the lowest
ranking schools. To keep a more homogeneous group for comparison high schools will
be analyzed here.
A general description of these four cases follows: the number of articles written
about each case is twenty two, two, one, and one corresponding to Columbine, Thurston,
John Batram, and Lew Wallace, respectively. The article covering the Lew Wallace and
John Batram cases are very short compared to the mean length. One of the articles
describing the Thurston case is relatively long while the other is a brief article. Of the
articles covering the Columbine case, there are many "full length" articles and editorials,
and a few short descriptive articles.
4.1 Columbine High School Shooting
The Columbine case is unique in that it ranks the highest on the scale provided by
Table 3-4. The articles that provide information about the Columbine case describe it in
such a way that Columbine appears in the top 25% of the following categories: number of
words per article, number of articles per case, number of offenders killed, number of
victims, number of victims killed, mass-casualty producing weapons, number of
weapons, number of "full length" articles, median family income, and percent white
population of city. These findings lead to the conclusion that it is likely that Columbine
is a NTSV case and that the level of unique interest associated with the case is relatively
high. According to the findings in this study, the actual level of unique interest for the
articles written about Columbine within the first year of the occurrence was 2.2 which is
around the mean UI for all the cases. One would not expect this given that Columbine is
the ideal model for NTSV cases. However, given the apparent inadequacy of using the
level of UI alone as an indicator of whether a case is NTSV or TSV, this finding is
A summary of the findings from the twenty two articles is presented here. The
article coverage is broken down into the following article types: ten of the twenty two
articles are "full length" articles, six of them are editorials, and six of them fall into the
other category which includes an article on poll results, SWAT team tactics, criminal
investigation of the case, response of Catholic Church, and the television and movie
industries' response to Columbine. From the findings in this study, none of the TSV
cases have elicited such a response. This is just one of the multiple ways in which NTSV
cases are treated differently than TSV cases.
One reason for the overall relatively low score on UI for Columbine is that there
are so many articles related to this case. Of the twenty two articles covering the
Columbine case, the "full length" articles contribute an average of 3.6 on the UI scale.
The editorials have an average UI of 1.2. The various other types of articles have an
average of 1.5 UI. The overall UI of 2.2 is significantly affected by the contribution from
the editorials and other articles which do not necessarily provide adequate coverage of
the Columbine case.
The group of articles that falls into the other category provides little insight into the
Columbine case other than the fact that these articles exist. By this, it is meant that with
most other cases there were not a lot of articles written in response to the shooting.
Columbine provided stimulus for a plethora of articles to be written in response to the
shooting. Many of the articles that are written about other cases refer to Columbine at
some point in the text. Columbine was a unique and powerful occurrence that shaped the
way that society reacts to school shootings. Several television companies and a movie
productions agency rescheduled certain movies and television programs so that they
would not appear insensitive to the victims and survivors of the Columbine school
shooting. Pope John Paul II., made a speech in response to the Columbine tragedy letting
the World know that his prayers and the prayers of the Catholic Church would be with
those who were harmed by the shooting. A SWAT team video, which was produced
before the Columbine shooting, to help SWAT teams and other law enforcement agencies
work together to deal with school shootings was in high demand after the Columbine
attack. Several polls were given by news and media agencies in response to the
Columbine shooting in order to determine what America's youths and parents thought
about the attack and how to keep it from occurring again. Lastly, an article briefly
provides information that authorities are looking into the link between two weapons that
were used in the shooting and purchased at a local gun show in Colorado. Of all of these
articles, only the last one would represent anything close to the type of coverage that one
would expect to find in a TSV case.
'By this I mean that TSV cases are not likely to be under special investigation by SWAT teams,
nor are they likely to receive special address from the Pope, nor are news agencies likely to
conduct various polls in response to TSV shootings.
The six editorials can be broken down into three collections of two editorials each.
The first two editorialists provide their opinion about how to solve school violence.
There are no fundamentally moving findings provided by these two editorials. The
second set of two editorials present the views of numerous authors about the Columbine
school shooting in response to other editorials written by other authors. One set of
responses mainly discusses the impact of gun control on America and whether or not gun
control should be the focus of discussion about school shootings and school violence in
general. Another editorial gives a summary of the general theoretical ideas behind this
study and will be paraphrased here: people have become somewhat hardened to urban
school violence; it has been thought of as being related to gangs and drugs. Recently,
new shootings aimed at killing many people have occurred in rural and suburban
communities. The offenders have been mainly middle class kids with no apparent reason
for their rage. It is believed that the reason lies within the American culture ("Looking
for answers in wake of another school shooting," 1999). In one editorial, the author
responses to an editorial that suggests, as this study does, that there are corollary
statements which follow from the statements made in the aftermath of Columbine.
Statements like "this should not happen here" lead to a conclusion that this violent
behavior is expected in other locations. The author of the responding editorial does not
think that these statements have anything to do with race overtly or covertly. He believes
that coverage provided by news organizations is based solely on "circumstance" and
newsworthinesss" of the case under scrutiny (Rickerfor, 1999, p. B6). However, the
author of the responding editorial fails to convincingly show that race does not affect
circumstance and what journalists and editors consider to be newsworthy.
Ishmael Reed made an argument based upon a 1994 statistic that "found that school
children in small U.S. cities, suburbs and rural communities were twice as likely to carry
weapons as students in large cities" (1999, p. 5C). Mr. Reed's argument based upon this
statistic is that Americans should not be surprised to see shootings such as Columbine
and other NTSV shootings. The findings from the 1994 report indicate that these
shootings are a very real possibility, and yet white Americans still act as though NTSV
cases are shocking. The argument being presented by Mr. Reed is a fairly strong case for
gun control laws. While, at the same time, he points out that the critical component of
racism is still a problem in America which permeates every aspect of our daily lives and
yet is not discussed by most media sources. The editorial offers the idea that there is a
general tendency to blame the nation's youth violence problems on minority communities
in inner cities, while perhaps it may be more accurate to take a look in the suburbs and
rural communities in America to determine where the guns are coming from and where
the desire to use them is coming from.
The general tendency of these "full length" articles is to describe, in detail, the
events that happened at Columbine and the response of the survivors, parents, and
officials after the incident. Some particular issues that arise in the set of "full length"
articles are: the Jefferson County Sheriff s office performed an investigation of the
Columbine case in an attempt to prepare instructional material so that other law
enforcement and rescue service personnel will know what to do in the event that a similar
incident occurs (Luzadder, 1999). Another article describes the increased incidence of
parents taking their children to the psychiatrist to be evaluated given the scare that has
occurred in the aftermath of Columbine. Prior to the shooting, parents may have written-
off certain behavior as being part of growing up, but now it is too risky for parents to
simply allow certain behaviors to continue without further evaluation to determine the
cause of the behavior (Jefferson, 1999). A third article describes how the "reluctant
experts" at Columbine are helping other communities, in particular the community in
Santee California with its school shooting. The article describes the information and
advice that counselors provided for the officials at Santana High School. The article also
describes the changes that occurred at Columbine as a result of the shooting:
Columbine doubled the number of campus security guards to four, installed 16
surveillance cameras, instituted the use of identity cards, and equipped doors with
key cards. Two mental health counselors were added to the staff, bringing the
school total to eight. A safe room was established, where distraught teachers or
students could retreat for privacy and quiet. (Carter, 2001, p. A12)
The very simple and yet utterly obvious conclusion from this article is that it is
written as though Columbine is the epitome of school shootings. As though Columbine
is the only school shooting and the experts that have dealt with the Columbine case can
deal with any other case. There is an overtone in the articles that although NTSV is the
only type of violence and the means by which officials deal with this type of case is the
only real means of dealing with any case of school violence. It also appears from the
Luzadder (1999) article that NTSV is the only type of violence that exists. There is a
tendency to ignore the TSV cases that happen in America unless there is something that
generates a unique interest in the case. It is then considered newsworthy and worthy of
being published in non-local newspapers for a larger audience to read.
One article from The Baltimore Sun describes the offenders as part of a gang that
called itself the "Trenchcoat Mafia." This article provides information about the
circumstances that surrounded the case. One interesting factor that came out of this
report was that the two offenders in this case had been "convicted last year for first-
degree criminal trespassing after they broke into a car. They completed a county
program for troublemakers in February" (Morgan, 1999, p. Al). The author asks: If these
children had not been white middle class youths, how would their treatment had been
different? Would they have been able to continue on from that conviction and plan and
prepare for the mass murder of other students and a teacher? The offenders started
planning for the shooting a year or more in advance. How would things have been
different had they been working class youths attempting to plan this shooting after having
been convicted of a serious offense? A parent of one of the surviving students stated that
he was shocked after seeing the house that one of the offenders lived in. The parent
asked, "Who would have thought that a kid who lived in a home like that and had that
kind of opportunity could do something like this?" (Morgan, 1999, p. Al). This is a
sentiment that has been expressed many times and in many ways in this case. There may
be some truth to this sentiment but, at the same time, it still seems to imply that this
behavior is expected of those who do not live in expensive houses and nice
neighborhoods. Another article by O'Brien, Gorov, and Zuckoff (1999) describes that
the offenders had been planning the attack for at least a year and that they planned the
attacks to coincide with Adolf Hitler's birth date. An interesting and important fact that
distinguishes NTSV from TSV is that the offenders wanted to create as much harm and
damage as possible. The police "recovered more than 30 homemade explosives from the
school, as well as two sawed off shotguns, a semiautomatic rifle, and a semiautomatic
handgun, leading them to believe that others had to have been involved" (O'Brien et al.,
1999, p. Al). This is not typical in TSV cases. The TSV offenders usually want to harm
or kill the person that has lead them to perform the act of violence. The TSV offender
does not want to kill as many people as possible and cause as much damage as possible.
One theme that ran in the following two articles is the idea that the offenders were
bullied and this is why they lashed out at others (Yettick, 2002; Kenworthy, 1999). The
Yettick (2002) article also describes the idea that the offenders and the "Trenchcoat
Mafia" bullied other students as well. However, many of those who were killed by the
offenders were football players responsible for bullying and picking on the offenders and
the '"Trenchcoat Mafia." Frank DeAngelis, the principal of Columbine at the time of the
shooting, said of the survivors, "they were deprived of so much of their youth. They saw
so much death. It scares me. I hope it made them stronger" (Yettick, 2002, p. 1B).
While this statement is insightful, it did not find its way into any of the articles related to
TSV that appear in this sample. It seems to be very difficult for white American
journalists to portray TSV in a way that shows this much compassion and understanding,
this much sympathy and respect for the victims and survivors of TSV. Certainly, those
who experience TSV live in communities and experience situations which would allow
them to be deserving of such a statement.2 Additional curiosities that are described in the
article by Kenworthy (1999) include the fact that President Clinton saw fit to address the
Columbine incident in one of his speeches. It is unlikely that the president would address
the country about school violence if Columbine had been a case of TSV. However, it is
understandable that the, now former, president would make some kind of statement given
that Columbine was one of the most violence cases of NTSV and TSV ever. Another
2 For examples of youths in America that have experienced violence and atrocities, see Code of
the Street, by Elijah Anderson. Anderson (1999) provides an explicit and detailed account of the
conflicts that youths in inner city America face on a daily basis.
interesting fact is the number of law enforcement officials that responded to the
Columbine incident in total:
Hundreds of heavily armed police officers, FBI and Federal Bureau of Alcohol,
Tobacco and Firearms agents arrived on the scene, accompanied by dozens of
ambulances and other rescue vehicles. .. [and] an armored personnel carrier. ..
[were] dispatched to help remove the wounded. (Kenworthy, 1999, p. Al)
Again, it is highly unlikely that a TSV case would receive such resources. Not one TSV
case in this study received such resources. At most there would likely be a squad of
police officers and an ambulance or two. There would not likely be so many law
enforcement resources utilized at one scene of violence in a TSV case.
4.2 Thurston High School Shooting
There are two articles that describe the Thurston High School case. There are other
cases that are ranked in the top 25% according to Table 3-4 and that have a larger number
of articles per case that appear in the sample used for this study, but there are few that are
ranked as high as the Thurston case.3 The Thurston case ranks in the 25% percent of
word count, unique interest, number of victims, number of victims killed, mass-casualty
producing weapons, number of weapons, and percent white population of the city. The
sum effect of these findings is that Thurston should be more likely to receive a higher
degree of unique interest and should likely be labeled as a NTSV case.
The first article that describes the Thurston High School shooting is entitled
"School shooting inspires one-act play that allows teens to explore violence" (1999, p.
A2). It describes a play that was written by an Oregon resident who decided to write it
after watching the news about the Thurston High School shooting. The article describes
3 The fact that there are a certain number of articles, in this particular sample, on a given case
does not mean that there are no other articles included in the population of articles that may be
written about the case.
some basic facts about the case such as it occurred on May 20th, 1998 in Springfield,
Oregon. A fifteen year old white male student shot and killed both his mother and father
and then went to school and shot and killed two classmates wounding twenty two others.
It is almost the ideal description of a NTSV case: the offender is usually a white male
who is of a middle class or higher socioeconomic status, the target of the violence tends
to be not only peers but also authority figures, violence is likely related to feelings of
revenge or retribution, types of weapons used are mass-casualty producing and tend to be
excessively violent, number of targets are usually high, and violence seems to be planned,
pre-meditated, and organized. This article scored a 4.0 on the UI scale.
The article describes the opening scene of the play in which the offender is in his
jail cell and is confronted by the ghosts of his victims. These ghosts ask the offender
"why'd you kill me?" ("School shooting inspires," 1999, p. A2). Eventually, in the play,
the offender states why he killed the victims. He did it because he was "wracked by the
pain of his classmates laughing at him" ("School shooting inspires," 1999, p. A2). The
playwright wrote the play in hopes that it would cause young people to face the problems
associated with the "culture of violence" in America. It is interesting that there are no
such plays written about TSV cases that appear in this sample. It seems that there is a
need for such tools to combat violence that takes place in inner city and poor rural
communities throughout America as well.
The second article that describes the Thurston High School shooting case is a more
detailed description of the case. The main theme of this article is the offender's mental
status. His lawyers originally planned to have the offender plead not guilty by reason of
insanity, but ultimately, the offender decided on his own that he would plead guilty to
four counts of murder e.g., his parents and two students. This article indicates a
relatively low score of 2.0 on the UI scale, but this is a result of the inability of this scale
to provide a truly adequate representation of the level of unique interest that the author
appears to be portraying in the case. However, after reading the article, one does realize
how the portrayal of the offender and the victims in this case differs greatly from the
portrayal of the victims and offenders in the TSV cases that follow.
4.3 John Batram High School Shooting
One brief caption explains the situation at one school shooting that happened at
John Batram High School in Philadelphia, Pennsylvania. The caption appears on page
12A in the front page section of The Boston Globe on October 5th, 1999. This caption
had a total score of zero on the variable of UI. The caption simply describes the
photograph which is of an ambulance that came to the school as a result of the shooting.
Apparently, the assistant principal of the high school was shot in the leg when he
confronted the seventeen year old offender about possessing the handgun.
This case is described here because it ranks very low on the matrix in Table 3-4.
The case does not appear in the top 25% on any of the key variables that would indicate
that the case is likely to be classified as a NTSV case. The only variables that this case
appears in the top 25% of is percent black, percent other race, and total population size.
Philadelphia is large and has a high percent of black and other non-white racial groups
when compared to the other cities in the sample.
4.4 Lew Wallace High School Shooting
There is one article in the sample that provides coverage of the Lew Wallace High
School shooting. This article is an editorial written in response to an article that appeared
in the front page section of a newspaper on page 18A. The author of the editorial writes
of the difference in the coverage of the shooting that happened at Lew Wallace High
School and other recent shootings. The author states that there is "no front-page story, no
analysis of how a student was pushed to the edge by dire circumstances, no 'how could
this happen here' commentaries" (Dixon, 2001, p. B8). The author attributes this lack of
interest in the case to the fact that the offender and victims involved in the case are not
white. The implication is that they are black American students that attended an inner
city school in Gary, Indiana and this is why there is differential coverage of this shooting.
On the scale of UI, this editorial received a score of zero.
The Lew Wallace case appears to be off of the scale according to its ranking in
Table 3-4. It does not appear in the top 25% of any of the variables except for the percent
black population in the city. This indicates that the case should a TSV case as opposed to
a NTSV case and there should be little unique interest in this case. Indeed, the evidence
does seem to confirm the theoretical assumptions about the differences between the two
types of school violence.
The author of the editorial describes a case which is classified as TSV in this
project. It gives support to the ideas generated from the theory that TSV is violence in
which the offender may be white or of another racial category and is of a low
socioeconomic status, the target of the violence is generally other students, violence may
be related to gang or drug activities, types of weapons used vary from assault with hands
to assault with weapons other than mass-casualty producing guns, the number of targeted
persons are usually low, and violence seems to happen in more of a spontaneous nature
although there may be planned acts of violence in this cases of TSV.
The goal of this study was to investigate the way U.S. newspapers portrayed
domestic school shootings from 1st May 1997 to 30th April 2002. Using Lexis Nexis
online search utility, major newspapers in the U.S. were analyzed for content involving
school shootings. A total of 100 articles were selected at random from a sampling frame
of 578 articles within the specified time period. The following non-aggregate variables
were coded in this analysis: title, case name, number of words, date of occurrence, date of
publication, section of the newspaper that the article was published in, time difference
between date of publication and date of occurrence, race of offender and victim,
socioeconomic status of offender and victim, number of offenders and victims, number of
offenders and victims killed, type of weapons used, number of weapons used, relation of
victim to offender, and level of unique interest. Each article's qualitative characteristics
were also described. Once the non-aggregate data were collected several of these
variables were used to code for aggregate variables related to each school shooting in this
study: NVT, case name, number of articles per case, number of articles written within the
first year after the occurrence, socioeconomic status of offenders and victims, quantity of
weapons used, type of weapons used, race of victims and offenders, number of offenders
and victims, number of offenders and victims killed, level of unique interest controlling
for time, racial makeup of the city where shooting occurred, median family income of the
city where shooting occurred, and total population for the city where the shooting
occurred.1 The overall focus of this analysis is to determine the nature of news coverage
and portrayal of school shootings.
5.1 Overview of Theories Utilized in this Analysis
The perspectives used to address the news portrayal of school shootings have been
the conflict and structural functional paradigms, with special interest in conflict-labeling
criminology, power-threat hypothesis, media criminology, and critical-media
perspectives. These perspectives view the topic of school shooting as one of a power
struggle between people with lower levels of power and those with higher levels of
power. News agencies in America are in a situation where they have the power to
provide coverage of school shootings. They provide coverage of each case in a way that
will perpetuate social ideals about those parties involved in each shooting. A key
motivating factor for news agencies to continue to perpetuate socially acceptable beliefs
about different cases is that the agencies rely on the business of their customers and the
investments of many individual investors and advertising establishments. There are also
political influences on the type of coverage that news agencies provide (Paletz & Entman,
1981). The critical point to draw from the theory is that there are expected to be
differences in the coverage of school shootings based upon the level of power that the
various parties involved in an incident possess.
5.2 Overview of Important Findings
This section will start with a summary of the findings in the relationship that exists
between the chance that a case is NTSV and other key variables in the analysis. As the
number of articles written per case, articles written within the first year, unique interest
1 The 2000 U.S. Census American FactFinder was used to find the racial makeup of the city
where the shooting took place and the median family income of the city where the occurred.
within the first year, percent white, median family income, number of offenders killed,
number of victims involved and killed, and the quantity of weapons involved in a case
increase there is also a greater likelihood that the case is one involving NTSV offenders.
Cases of NTSV are less likely to happen in communities where a large percent of the
population is black. The NTSV offenders are likely to be white and from the qualitative
data more likely to be male.
Now consider the major results involving the level of sensational characteristics
about a case. As the number of offenders killed, the number of victims involved, the
number of victims killed, and the quantity of weapons used in the case increases the
number of articles covering a case increases. This seems to be due to the fact that when
those factors are high the case is seen as being more newsworthy i.e., more sensational
and dramatic. A finding which appears to be counter to the theoretically expected result
is that number of articles written increased when the victims were likely to be minorities.
This finding, and many others, may be different if one were to control for other factors
such as number of offenders involved and killed, number of victims killed, number of
weapons used, type of weapons used, and median family income.2
A similar finding is that the number of articles written within the first year after the
shooting increases when median family income, number of offenders killed, number of
victims involved and killed, and quantity of weapons used increases. Another relatively
unexpected finding was that the number of articles written within the first year increases
when the victims involved are minorities. Again it would be interesting to determine if
2 One must remember that even though SPSS awards a certain level of significance to these
relationships the relationships may warrant more or less strength based upon the actual number of
cases involved in the relationship. However, these findings do give some idea of the relationships
that exists in the population of articles that offer coverage of school shootings.
this finding changed when holding constant several other key variables. Theory suggests
that cases involving minority offenders are more likely when those cases allow journalists
to perpetuate current socially acceptable ways of thinking about minority members
(Howitt, 1998; Barak, 1994; McQuail, 1992; Paletz & Entman, 1981).
As the percentage of white people in the population increases, news agencies are
more likely to provide coverage within the first year of an incident with a higher level of
UI. Similarly, when the offender is white, the level of UI provided by articles written
within the first year after the shooting is likely to be high. On the other hand, populations
with larger percent of black people are less likely to receive coverage with a high level of
UI. These findings do not deviate from what the theory would lead one to wait for.3
The median family income is larger in communities where a large percent of the
population is white. The median family income is smaller in communities where a large
percent of the population is black. The median family income in communities where a
large number of offenders were killed is likely to be higher. These findings can be
expected based upon the literature.
The percent of white people in the population is correlated with the race of the
offender and the SES of the victim in such a way that as the percent of white people in
the population increases the offender is more likely to be white and the victim is more
likely to be of higher SES. Similarly, white offenders are more likely to have higher
SES. Offenders are likely to attack those in a similar SES category. Therefore, white
3 The percent white and percent black figures were found using the Census American FactFinder
utility and were found for each case in the sample. The previous findings related to the race of
the offenders or victims were found using a world wide web search and the races of all offenders
and victims were not founds for many of the cases. It is important to notice beside the impact this
has on the validity of the finding that there is limited information available on any given case
offenders are more likely to assault white victims and black offenders are more likely to
assault black victims. In a similar way, populations with a large percent of black people
are likely to have black offenders. Black offenders are more likely to have lower SES.
Small cities are more likely to have victims with lower SES.
Cases involving a larger number of offenders are likely to involve the use of less
deadly weapons. When a large number of offenders are involved in a case the victims are
likely to be minorities and have lower SES. The number of offenders killed in a shooting
is directly related to the number of victims involved and killed, and the number of
weapons used. The number of victims killed in a case is directly correlated with the
number of victims involved in a case, which is also correlated with the number of
weapons and the type of weapons used in a shooting. The larger the number of victims
and the number of victims killed the more weapons that are involved and the more
damage those weapons are capable of creating.
A pattern is developing from these results. News agencies are more likely to
provide coverage within the first year after the occurrence and beyond for cases that
involve sensational behavior. Sensational behavior may be classified as behavior that is
not expected within the norms provided by a society. In other words, cases that involve a
large number of offenders and victims, large number of deaths, white offenders of middle
and upper middle SES killing other white people in middle and upper middle SES
communities with large percentage of the population being white is considered to be
which usually varies with the likelihood that a case is NTSV or TSV. Put another way, if the case
is NTSV, there is more information available on the case from various sources on the internet.
The three questions posed at the beginning of this project are the following: Is there
a difference in the type of coverage provided for the two types of school violence? If so,
how does the coverage differ? Finally, why is there a difference in the coverage
For the first question, yes there does appear to be a difference in the type of
coverage that news sources provide for the two types of school shootings. Table 3-4
shows a clear division in the coverage provided on the twenty seven cases included in
this sample. This sample has been taken at random from a larger sampling frame and is
therefore believed to be generalizable to the larger population of articles written on
school shootings in the U.S. during the time frame of the study.
Secondly, the coverage is different in several respects. Many of the results indicate
that NTSV cases were given a great deal of attention by news agencies. For example,
every NTSV case was in the top 25% for at least one of the following variables: word
count, number of articles written overall and within the first year, and number of "full
length" articles written. 4 At the same time, one out of thirteen of the TSV cases were in
the top 25% of the variables listed directly above. The implication of this is that there
exists a divide in the coverage provided for the two types of school shootings. Incidents
of NTSV received most of the attention of news agencies during the time period studied.
Finally, the key question that still remains is: Why is there such a divide in the
coverage of the two types of school shootings? The answer seems to be found in the
division that exists between factors causing interest presented in Table 3-3 and tallied in
Table 3-4. Most of the NTSV cases within the top 25% of variables are likely to be seen
by news producers as being newsworthy and sensational. The cases that provide media
producers with a source of behavior that is not to be expected in the cases involving white
offenders and behavior that is an exemplary instance of activity that is to be expected of
minority offenders will be considered to be newsworthy. From this perspective most of
the offenders of NTSV cases will be white, middle class or upper middle class, and the
targets of their attack will be white, middle and upper SES. Likewise from this
perspective most TSV offenders are likely to be either minority or white, of lower SES,
and targeting others of lower SES who may be white or minority.
The theoretical explanation for the differential coverage of school violence appears
to be supported by this analysis. It is the level of sensational behavior with the
combination of expected behavior that seems to be the determining factor behind the type
of coverage any particular case will receive.
5.3 Critique and Suggestions for Future Research of Differential Newspaper
Future research may include a dataset that will provide more information for a
higher level of analysis. A rigorous analysis of more sophisticated data would better suit
the topic. For example, one could analyze the unexpected findings further to determine
whether they disappear after controlling for other factors as suggested above. In a similar
fashion, one could analyze the many statistically significant findings of this project after
controlling for the many variables to determine if any of these relations disappear.
Additionally, future researchers may search for cases in a different way than this
project has. A suggestion would be to compile a list of cases by school name and
location and from this list pick a random sample of n cases. After compiling a list of
cases to be included in the sample, search for articles on each case instead of picking a
4 See Table 3-3 Matrix of cases with exceptional qualities.
random sample of articles which was the technique used in this study. This would
provide a larger n when running various statistical analyses allowing for regression
techniques to be used assuming that the case is the intended unit of analysis.
Another area of potential research would be to further distinguish the affects of
both sensational components of a case and the expected normal behavior of the case.
They are closely linked, but from the theory one would expect that even though a TSV
case is sensation, it may not receive coverage if it does not allow news producers to
present it in a way that will further promote socially acceptable beliefs about the
offenders and victims of the TSV case. Here the main limitation of this study becomes
painfully clear. The data are not detailed enough to allow for more rigorous quantitative
analysis e.g., linear regression analysis.
Another critical relationship that must be analyzed further involves two key
variables: race and SES. These variables are so intertwined in this and many research
projects that it is difficult to separate the effects of each. In this study, the following
appear to be the case: NTSV offenders are likely to be white males of middle to upper
SES. While, TSV offenders seem to be either white or of a minority racial group but
definitely seem to be of lower SES. There exists a basic idea that NTSV offenders would
be mainly white and TSV offenders mainly black, but this would exclude a significant
number of TSV offenders that are white and non-black minorities. The dynamics
between sensational and expected normal behaviors becomes even more difficult to make
sense of as one takes a closer look at race and SES. Some questions to guide research on
this topic may include: Does differential coverage exist in cases that involve white TSV
offenders and non-white TSV offenders? These offenders, by definition, would be of
lower SES so that is one less variable to consider. Race becomes the prominent
determinant here. The other key variable to consider besides race would be the expected
normal behavior. A fundamental question that would arise is how is the expected normal
behavior for white TSV offenders and non-white TSV offenders different, and how do
those differences deviate from the expected normal behavior of white NTSV offenders?
This study did not consider this topic in that much depth.
This analysis is an initial attempt to determine the nature of news coverage with
regard to school shootings in America. Depending upon future researchers' intent, they
may want to continue analysis of this topic with a critical eye on the media portrayal of
school shootings in the U.S. Future researchers may want to analyze the racial
components of media coverage more thoroughly. There are many different directions for
future study of this topic. A big challenge for any future research is a general lack of data
on the topic of media coverage of school shootings in America. Compiling the correct
data for the correct purpose is especially difficult and time consuming but is the most
critical task that any research faces.
LIST OF REFERENCES
Anderson, E. (1999). Code of the Street: Decency, Violence, and the Moral Life of the
Inner City. New York, NY: W. W. Norton and Company.
Barak, G. (Ed.). (1994). Media, Process, and the Social Construction of Crime: Studies
in Newsmaking Criminology. New York, NY: Garland Publishing, Inc.
Barlow, M. H., Barlow, D. E., & Chiricos, T. G. (1995). Economic Conditions and
Ideologies of Crime in the Media: A Content Analysis of Crime News. Crime and
Delinquency, 4(1), 3-19.
Blalock, H. M. (1967). Toward a Theory of Minority-Group Relations. New York, NY:
John Wiley & Sons, Inc.
Carter, J. (2001, March 7). Santee school shooting; Littleton's reluctant experts try to
help; Aid: officials who went through the columbine horror tell Santee what to
expect. One warns that the trauma has "just begun." Los Angeles Times, p. A12.
Dixon, D. (2001, April 3). Another school shooting. Los Angeles Times, p. B8.
Howitt, D. (1998). Crime, the Media and the Law. New York, NY: John Wiley & Sons.
Jefferson, R. S. (1999, May 7). School shootings put parents on alert, therapist says. St.
Louis Post-Dispatch p. 4.
Kenworthy, T. (1999, April 21). Up to 25 Die in Colorado School Shooting; Two student
gunmen are found dead. The Washington Post, p. Al.
List of school shootings. (1998, May 22). The Buffalo News, p. 4A.
Looking for answers in wake of another school shooting. (1999, April 22). The
Baltimore Sun, p. 19A.
Luzadder, D. (1999, July 11). Country will learn from Columbine Jeffco investigators
create textbook on school shootings so other communities can prepare for such
incidents. Denver Rocky Mountain News, p. 5A.
McQuail, D. (1992). Media Performance: Mass Communication and the Public Interest.
London, UK: Sage Publications.
Morgan, J. (1999, April 22). Dressed in black, fans of Hitler; Students were aloof, gave
Nazi salutes, glorified death; Colorado school shooting. The Baltimore Sun, p. 1A.
O'Brien, E., Gorov, L., & Zuckoff, M. (1999, April 25). School killings plotted in diary;
Conspirators timed massacre to occur on Hitler birth date; Colorado school
shootings. The Boston Globe, p. Al.
Other school shootings. (2000, March 1). The Seattle Times, p. A3.
Paletz, D. L., & Entman, R. M. (1981). Media, Power, Politics. New York, NY: The
Quinney, R. (1970). The Social Reality of Crime. Boston, MA: Little, Brown and
Reed, I. (1999, May 30). The pundits misfire on school shootings; Causes: The media
offer uninformed and often racist reasons for violence by U.S. students, a critic
charges. The Baltimore Sun, p. 5C.
Rickerfor, R. (1999, May 16). School shooting story has no racial undertones. The
Times-Picayune (New Orleans), p. B6.
School shootings. (1999, April 21). St. Petersburg Times, p. 6A.
School shootings. (1999, October 5). The Boston Globe, p. 12A.
School shooting inspires one-act play that allows teens to explore violence. (1999,
February 9). St. Louis Post-Dispatch p. A2.
Tolnay, S. E., Beck, E. M., & Massey, J. L. (1989). Black Lynchings: The Power Threat
Hypothesis Revisited. Social Forces, 67(3), 605-623.
Turk, A. T. (1964). Prospects for Theories of Criminal Behavior. The Journal of
Criminal Law, Criminology and Political Science, 55, December 1964, 454-461.
Yettick, H. (2002, April 20). Echoes of tragedy; Effects of 1999 school shooting are
shrouded yet unmistakable. Rocky Mountain News, p. IB.
John Robert Bennett received his Bachelor of Arts degree from the University of
Florida, Gainesville, Florida, in December of 2001 with a major in sociology. He has
been enrolled in both the undergraduate and graduate sociology programs at U.F. since
January, 2001, pursuing a Master of Arts degree in sociology.
After receiving his M. A. degree in sociology, the author plans to continue at the
University of Florida in pursuit of his Master of Science degree in business
administration with a concentration in management. His main areas of interests are
media, race, and criminology.