Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 287
Deadly Lessons: Understanding Lethal School Violence 9 Lethal School Violence in Statistical Context This chapter presents information about levels and trends in the form of lethal school violence under study, placing them in the con- text of other forms of violence. Such statistical information helps us interpret the information from the cases by revealing the size of the phenomenon under study relative to other forms of violence. It also enables an investigation of the likely causes of the particular form of violence that interests us and, more particularly, whether it seems to move in concert with other forms of violence or in its own independent patterns. EXTENDING THE DEFINITION To do any kind of statistical analysis, it is necessary to develop an operational definition of the phenomenon to be studied. As described in Chapter 1, the operational definition we embraced was “lethal school violence,” which includes the following elements: Lethal violence That took place in schools Was committed by students of the school and Resulted in multiple victimizations. The committee’s application of this definition was fluid enough to include at least one case in which no one died.
OCR for page 288
Deadly Lessons: Understanding Lethal School Violence In fact, it was the idea that the subject of study was a series of incidents defined not only in terms of the amount of carnage, but also in terms of the character of the attacks and the motivation of the offenders that generated a second concept beyond the operational definition of lethal school violence that guided most of our work. After we reviewed the cases and the various literatures that seemed related to them (including the emerging literature on rampage shootings), we began referring to some of the school cases as “school rampage shootings.” In one sense this is hardly controversial. The label “school rampages” seems a reasonable descriptor of events in which a student of a school shot at and killed members of his school community. Yet it is worth noting that the idea of a “rampage shooting” has connotations and meanings that go beyond the operational definition we began with. The idea of a rampage, for example, suggests that an important cause of the event was the agitated, confused mental state of the perpetrator. When we refer to such events as rampages, we are already making claims about the likely causes of these events as well as their consequences. It is also worth noting that this assigns the school shootings to a class of violence (rampages) about which there is both theoretical and empirical information (see, e.g., Fox and Levin, 1998). We could equally well have assigned these incidents to a different specialized class of violence. We could, for example, have characterized the incidents as “school mass murders” (see, e.g., McGee and DeBernardo, 1999). This would have had somewhat different connotations than school rampages. The point is that, as soon as we start using concepts that are more abstract and evocative than a specific listing of the attributes that provide the operational definition of the phenomenon we seek to measure, we are at risk of subtly biasing both our own and our audience’s understanding of the events we are trying to analyze. To examine the choices of how to think about these relationships, the committee used a series of Venn diagrams that reveal the relationship among more or less inclusive ideas. Figure 9-1 shows the relationship between three distinct but partially overlapping sets: Set A is the idea of “lethal school violence in incidents involving more than one victim” (this was the core operational definition for most of our work); Set B is the idea of “lethal school violence” (which is the phrase we often used, but as an analytic matter would include cases in which only one person was killed); and Set C is the idea of “school rampages” (which includes cases in which more than one person was injured, and there was a significant potential for lethal violence as well as those in which people were actually killed). As a practical matter, the violence we investigated included the union of sets A and C. The portion of Set B that involved only single victims (whether fatal or not) was excluded from the analysis. This put the pri
OCR for page 289
Deadly Lessons: Understanding Lethal School Violence FIGURE 9-1 Lethal school violence and school rampages as adjunct sets. mary emphasis on a very small number of cases. Moreover, it turned out that suburban and rural schools were much more significantly represented in these cases than inner-city schools, where many of the incidents of school violence involved single victims. That was not because we thought these cases were unimportant; the committee held the opinion that the special attention given to the relatively few incidents involving multiple victimizations might distort society’s picture of lethal school violence in general. It was, instead, that we had been directed by Congress to give special, but not exclusive, attention to this apparently new form of violence. We also recognized that a significant amount of public and scholarly attention had been devoted to lethal youth violence in urban areas, and that less had been focused on trying to understand this new and unexpected phenomenon. A Venn diagram is also helpful in showing the relationship between the kind of violence that was our primary focus and the other categories of violence with which it was being compared. Figure 9-2 embeds Figure 9-1
OCR for page 290
Deadly Lessons: Understanding Lethal School Violence FIGURE 9-2 School rampages as a component of violence.
OCR for page 291
Deadly Lessons: Understanding Lethal School Violence in the larger frame of more general definitions of violence and more common forms of violence. The largest set is “all violence.” An important subset of all violence relevant to this study is all violence committed by or experienced by youth. Two important subsets of youth violence relevant to our subject are “serious youth violence” on one hand and “all school violence” (regardless of the seriousness of the victimization) on the other. These subsets overlap, but neither is a subset of the other. There is a lot of serious youth violence that takes place outside of school. There is a great deal of school violence that is not serious. Within the category of “serious youth violence,” we can create a subset of “lethal school violence.” Some of this happens in schools, and some happens outside schools. Within the category of “lethal school violence,” we can create two subsets that define lethal incidents in which only one person was killed and distinguish that from other incidents in which many were killed or injured in the same incident. Finally, we can describe a subset of serious (but not lethal) school violence in which many were victimized but no one died. As noted above, the core of the study focused on incidents of lethal school violence in which more than one person was injured (“lethal school violence” for shorthand purposes) and incidents in which many were victimized regardless of whether anyone was killed (“school rampages.”) In the rest of this chapter, we present data on levels and trends in forms of violence organized in terms of these different sets. We look first at levels and trends in the kind of violence that was at the core of our investigation: lethal violence, in schools, by students, resulting in multiple victimizations, and other situations in which multiple victimizations occurred in a way that suggested the fact that no one died was simply a matter of chance. We then look at how this violence fits within the larger patterns of youth and school violence. TRENDS IN LETHAL SCHOOL VIOLENCE The principal dataset available in assessing the level and trends of lethal school violence (with or without multiple victims) was produced and maintained by the National School Safety Center (2001). It is a census of school-associated violent deaths that occurred between 1992 and 2001, based on a systematic collection of newspaper accounts of the incidents. We used this dataset to separate from all lethal violence in schools that portion that meets our core definition. This excludes incidents in which only one person was killed or injured, as well as incidents in which a student started shooting, but no one died. Figure 9-3 presents a simple count of these incidents over the time period from 1992 to 2001. It makes clear that incidents of lethal violence including multiple victims while serious, are indeed very rare events. Overall, there have been 13 of these
OCR for page 292
Deadly Lessons: Understanding Lethal School Violence FIGURE 9-3 U.S. multiple-fatality, student perpetrated incidents, 1992–2001. SOURCE: Data from the National School Safety Center. incidents in this period. In the peak year, 1997, there were only three of these incidents. There are major concerns about these data for our purposes: (1) they go back only to 1992, (2) they focus only on events in which there was at least one fatality (thereby excluding incidents in which many were victimized, but no one killed), and (3) the sample was limited to the United States. Each of these presented a problem for the committee as it sought to put the issues of multiple-victim lethal school violence and school rampages in a broader context. Consequently, we sought to expand and improve on this database. In constructing our own database to complement this one, we relied on the same basic source—namely, newspaper accounts. But we looked in newspapers prior to 1992, and in countries other than the United States. We also had access to the United States Secret Service study that looked at this form of violence over a longer period than 1992–2001. Consistent with our broader operational definition that includes incidents in which many were victimized but none killed, we included incidents that resulted in multiple serious injuries but no fatalities. Figure 9-4 presents the results of that inquiry for the United States: our best estimate of the number of U.S. multiple-victimization, student-perpetrated school violence incidents between 1974 and 2001. Table 9-1 provides some details on these incidents.
OCR for page 293
Deadly Lessons: Understanding Lethal School Violence FIGURE 9-4 Multiple-victimization, student-perpetrated incidents, 1974–2001. Figure 9-4 shows that student-perpetrated school rampages (with or without a fatality) are not entirely new phenomena. There were two such incidents in the 1970s and six in the 1980s. And yet it also seems clear that the frequency of student-perpetrated school rampages resulting in multiple victimizations increased dramatically after 1994. The difference is highlighted in the figure by lines showing the mean number of such incidents per year in the 17-year period from 1974 to 1990 and the 11-year period from 1991 to 2001. The mean number of student-perpetrated ram TABLE 9-1 Multiple-Victimization, Student-Perpetrated School Violence in the United States Date of Incident Name of Offender Location Fatalities Injuries 12/30/1974 Olean, NY Anthony Barbaro 3 11 2/22/1978 Lansing, MI Roger Needham 1 1 3/19/1982 Las Vegas, NV Patrick Lizotte 2a 2 1/20/1983 Manchester, MO David F. Lawler 2b 1 1/21/1985 Goddard, KS James Alan Kearbey 1 3 12/4/1986 Lewistown, MT Kristofer Hans 1 3 3/2/1987 DeKalb, MO Nathan Faris 2b 0 2/11/1988 Pinellas Park, NJ Jason Harless and “companion” 1 2
OCR for page 294
Deadly Lessons: Understanding Lethal School Violence Date of Incident Name of Offender Location Fatalities Injuries 12/16/1988 Virginia Beach, VA Nicholas Elliott 1 1 11/25/1991 Brooklyn, NY Jason Bentley 1 1 2/26/1992 Brooklyn, NY Khalil Sumpter 2 0 5/14/1992 Napa, CA John McMahan 0 2 11/4/1992 Detroit, MI Renard Merkerson, Montrice Coleman 0 6 11/4/1992 Detroit, MI Unknown 0 2 11/19/1992 Chicago, IL Joseph White 1 2 12/4/1992 Great Barrington, MA Wayne Lo 2 4 1/18/1993 Grayson, KY Scott Pennington 2 2/1/1993 Amityville, NY Shem McCoy 1 1 3/18/1993 Harlem, GA Edward Bryant Gillom 1 1 1/23/1995 Redlands, CA John Sirola 1b 1 10/12/1995 Blackville, SC Toby R. Sincino 2b 1 10/30/1995 Richmond, VA Edward Earl Spellman 0 4 11/15/1995 Lynnville, TN Jamie Rouse 2 1 2/2/1996 Moses Lake, WA Barry Loukitas 3 1 2/8/1996 Palo Alto, CA Douglas Bradley 1b 3 3/19/1996 Las Vegas, NV An unidentified 8th grade student 0 2 5/14/1996 Taylorsville, UT Justin Allgood 1b 1 7/26/1996 Los Angeles, CA Yohao Albert Rivas 0 2 2/19/1997 Bethel, AK Evan Ramsey 2 2 10/1/1997 Pearl, MS Luke Woodham 2 7 12/1/1997 West Paducah, KY Michael Carneal 3 5 12/15/1997 Stamps, AR Joseph “Colt” Todd 0 2 3/24/1998 Jonesboro, AR Andrew Golden, Mitchell Johnson 5 10 4/25/1998 Edinboro, PA Andrew Wurst 1 2 5/21/1998 Springfield, OR Kipland Kinkle 2 21 9/29/1998 Miami, FL Felly Petit-Frere 0 3 1/8/1999 Carollton, GA Jeff Miller 2b 0 1/14/1999 New York, NY Camillo Douglas 0 2 4/20/1999 Littleton, CO Eric Harris, Dylan Klebold 15b 23 5/20/1999 Rockdale County, GA Anthony B. “T.J.” Solomon, Jr. 0 6 10/11/1999 Las Vegas, NV Maynor Villanueva 0 2 12/6/1999 Fort Gibson, OK Seth Trickey 0 4 2/22/2001d Portland, OR 14-year-old student 0 3 3/5/2001 Santee, GA Charles Andrew Williams 2 13 3/22/2001 El Cajon, CA Jason Hoffman 0 5c aIncludes offender killed by police. bIncludes offender suicide. cIncludes offender injured by police. dStabbing incident; all others are shootings. SOURCE: List of incidents was compiled using data from the National School Safety Center (2001), the U.S. Secret Service (2000), and a separate media search using online national and local newspaper archives.
OCR for page 295
Deadly Lessons: Understanding Lethal School Violence pages increased from an average of 0.53 incidents per year to an average of 3.27 incidents per year.1 It is important to note that these are very small numbers. It is also important to note that the increase observed in the 1990s could be explained at least in part by a reporting phenomenon. It seems likely that the media would cover fatalities in schools, and particularly fatalities that occurred with multiple victimization, with a high degree of consistency and reliability over the entire period from 1974 to 2001. What we cannot be sure of, however, is whether the media would have covered incidents involving multiple victimizations without a fatality as consistently or reliably over this period. While it seems likely that multiple victimizations in a school setting would be newsworthy throughout this period, we cannot be entirely sure that the media weren’t particularly sensitized to the issue of school rampage shootings in the late 1990s, and therefore began covering these more assiduously (even when they did not involve fatalities) than had previously been true. If the media were sensitized to these events, part of the increase could be accounted for by the increased likelihood of news accounts of such events, not by an increase in the real underlying rate of these events. Still, the difference in the rate of these events is impressive and would easily be rejected as a chance occurrence if the reporting were accurate, even though the numbers are very small. Our media search also uncovered five student-perpetrated school rampages in other countries (Table 9-2). While these results may be biased by the less certain coverage of international events, it seems note TABLE 9-2 International Multiple-Victimization, Student-Perpetrated School Violence Date of Incident Location Name of Offender Fatalities Injuries Method Data Source 5/28/1975 Brampton, Ontario, Canada Michael Slobodian 2a 13 Shooting Media 12/7/1999 Veghel, Netherlands 17-year-old student Father and sister (age 15) were charged as accessories 0 5 Shooting Media 3/16/2000 Rosenhem, Germany 16-year-old student 1 1b Shooting Media 4/20/2000 Gloucester, Ontario, Canada 15-year-old student 0 6b Stabbing Media 3/26/2001 Machakos, Kenya Felix Mambo Ngumbao, Davies Otieno Onyango 67 19 Arson Media aIncludes offender suicide. bIncludes offender self-injury.
OCR for page 296
Deadly Lessons: Understanding Lethal School Violence worthy that only one incident occurred in 1975 and no additional shootings occurred until 1999. The 1999 shooting was followed by three other rampages involving different means of inflicting harm on others (arson, stabbing, and shooting). This suggests that school rampages are not unique to the United States and, since no international school rampages were evident until 1999, rampages in other countries may have been somehow influenced by the U.S. epidemic in the 1990s. One final point: a December 2001 article in the Boston Globe reported that since the April 1999 Columbine tragedy, 12 U.S. school rampage shootings have been discovered and thwarted before they came to fruition. Ideally, we could put these events on Figure 9-4 as a further indication of the trends in time of these school rampage shootings. There are three problems in doing so, however. First, it is quite likely that, given the public concern about the school rampages, the newspapers would be much more likely to report on thwarted incidents in this period than they would have in earlier periods. Second, given efforts to mobilize students to report these events and law enforcement to take them seriously, it is quite likely that the police would find more such events and that they would treat each event as a serious plot that was really to be carried out rather than mere fantasizing by the kids involved. Third, in any case, Figure 9-4 records events that actually occurred. Presumably, for every act that actually occurred, there were some others in which some preparations were made, but for a variety of reasons, the act never occurred. Consequently, we would have to assume that there were even more attempts to be found than completions. What we are observing in the thwarted events, then, are some incidents that might never have occurred even if the police had not found them in time. For all these reasons, it is inappropriate to put these thwarted shootings in the same figure as the other data. Still, the fact that these thwarted events were planned during this period is consistent both with the idea that planning for such events increased in the latter half of the 1990s, and that society and the police got a bit better at learning about and thwarting the events. But the data cannot prove this claim. While the data depicted in Figure 9-4 are weak by scientific standards, they are still important to include in the effort to understand multiple-victim lethal school violence. What they suggest is that school rampage shootings are not a recent phenomenon, nor are they uniquely a U.S. phenomenon. It seems likely that the United States has experienced an epidemic of these incidents in the latter half of the 1990s—that is, an unexpected increase in their number. There may also have been some contagion mechanisms at work—that is, some kind of copycat influence. If the international and thwarted incidents are included in the basic time trend of observed school rampages, then copycat mechanisms seem
OCR for page 297
Deadly Lessons: Understanding Lethal School Violence likely. But there also seems to have been a small previous increase in these incidents in the late 1980s that no one much noticed. The lack of notice may have prevented the escalation of these shootings through the copycat phenomenon. But this is largely speculation, not a scientific claim. It seems unlikely that this phenomenon is either entirely new or entirely unique to the United States. It may have gotten worse recently and— even more speculatively—that may be in part the result of a kind of contagion. But the problem has endemic and international aspects as well as epidemic and U.S. ones. TRENDS IN RELATED FORMS OF VIOLENCE The comparison of school rampage violence with related forms of violence in scale and over time is important for policy as well as scientific reasons. On the policy side, we are interested in its size and trend relative to others to help in making judgments about the relative importance of the different kinds of violence. What proportion of all violence, or all violence involving youths, or all violence that occurs in schools do rampages account for? The question of how fast this component of the violence problem is growing is also of interest. If it is increasing rapidly and other forms are fading, then there is more urgency about efforts to keep it from getting out of hand, even if it is not a major piece of the problem. Comparing school rampages with other forms of violence indicates what priority is appropriate to give to this kind of violence. There are also scientific reasons to be interested in the relationship of school rampage shootings to other forms of violence. One possibility is that all forms of violence spring from the same basic causes that operate at a structural level in society. Another is that one kind of violence tends to cause other kinds of violence somewhat independently of the structural causes. For example, in this view, one might say that the epidemic of youth violence in the inner city from 1985 to 1995 set the stage for or caused the outbreak of school shootings in 1995–2000. We place our estimate of the trends in school rampages against the backdrop of trends in the following other kinds of violence: (1) intentional lethal violence in general, (2) intentional lethal violence among youth, and (3) intentional violence in schools. Figure 9-5 presents the yearly number of all homicide victims in the United States between 1976 and 1999 as measured by the Federal Bureau of Investigation’s Supplementary Homicide Reports (SHR) (Fox, 2001).2 There are two distinct peaks: homicides increased to a peak of 22,953 victims in 1980, decreased through the early to mid-1980s, increased to a second peak of 24,711 victims in 1991, and then decreased to 15,483 victims in 1999. As the figure shows, homicide victims ages 12–18 represent
OCR for page 298
Deadly Lessons: Understanding Lethal School Violence FIGURE 9-5 Homicide victimization in the United States. SOURCE: U.S. Federal Bureau of Investigation (2001). a relatively small portion of the yearly body count. At the peak in 1993, there were 2,634 homicide victims ages 12–18, which represents slightly less than 11 percent of the 24,571 total homicide victims that year. Figure 9-6 presents the yearly counts of homicide victims age 12–18 and homicide offenders age 12–18.3 The number of youth homicide victims increased dramatically between 1985 and 1993, then decreased significantly from 2,476 victims in 1995 to 1,495 victims in 1999. More directly relevant to the focus on rampage shooters, the yearly number of FIGURE 9-6 U.S. homicide victims and offenders ages 12–18. SOURCE: U.S. Federal Bureau of Investigation (2001).
OCR for page 299
Deadly Lessons: Understanding Lethal School Violence homicide offenders age 12–18 followed a similar trajectory. The number peaked in 1993 at 6,285 offenders and decreased significantly to 2,959 offenders in 1999. These patterns of intentional lethal violence in general and intentional lethal violence among youth stand in stark contrast to the epidemic of student-perpetrated school rampage shootings. The school rampages were increasing when broader forms of intentional lethal violence were decreasing. The pattern also contrasts with trends in reported school violence. According to National Crime Victimization Survey estimates, the number of reported violent and serious violent crimes against students decreased between 1992 and 1999 (Table 9-3). Using data collected from media databases, state and local agencies, and police and school officials, Anderson et al. (2001) found that the rate of single-victim student homicides decreased significantly between 1994 and 1999, while the rate of multiple-victim student homicides increased significantly during the same period. Between 1994 and 1999, Anderson et al. (2001) identified 220 school-associated events resulting in 253 violent deaths.4 The vast majority of the events involved single deaths, while only 8.2 percent (18 of 220) involved multiple deaths. While these data do not distinguish school rampage shootings, it is clear that multiple violent deaths in schools represent a small portion of all lethal violence in schools. Student-perpetrated school rampage shootings could be thought of as a subset of multiple killings or mass murders committed by adolescent offenders. Figure 9-7 presents the yearly number of homicide offenders age 12–18 who killed two or more individuals in a single incident. The time series follows the same general trajectory as the trends depicted in Figure 9-6. This is probably because the youth homicide epidemic was TABLE 9-3 Number of Nonfatal Crimes Against Students Ages 12–18 at School per 1,000 Students by Type of Crime, 1992–1999 1992 1993 1994 1995 1996 1997 1998 1999 Violent 48 59 56 50 43 40 43 33 Serious Violent 10 12 13 9 9 8 9 7 Total 144 155 150 135 121 102 101 92 NOTE: Serious violent crimes include rape, sexual assault, and aggravated assault. Violent crimes include serious violent crimes and simple assault. Total crimes include violent crimes and theft. “At school” includes going to and from school. SOURCE: National Crime Victimization Survey, 1992–1999 from P. Kaufman, X. Chen, K. Peter, S. Ruddy, A. Miller, J. Fleury, K. Chandler, M. Planty, and M. Rand. 2001. Indicators of School Crime and Safety, 2001. Washington, DC: U.S. Departments of Education and Justice.
OCR for page 300
Deadly Lessons: Understanding Lethal School Violence FIGURE 9-7 U.S. homicide offenders, ages 12–18, multiple victims. SOURCE: U.S. Federal Bureau of Investigation (2001). driven by gang-related and drug-related violence; there is no reason to believe that a majority of multiple homicides would not share similar circumstances. However, when these types of homicides are parsed from the data, the same pattern is evident. CONCLUSIONS The analyses reported here suggest a number of broad conclusions. First, school rampage shootings, while extremely serious, account for a very small component of all the lethal violence in the United States. Even if we take the smallest subset of violence—lethal violence in schools—we find that school rampages are a small component of this sort of violence. Although it has been growing, it remains small. This point can be emphasized still further by showing deaths in school rampages as a component of all traumatic death for young people and the relative risk of dying from different causes that adolescents face. These comparisons are not meant to minimize the pain and suffering and shock that the school rampage shootings generate. But they underscore the other kinds of risks to which children are exposed in addition to school rampages. Second, the trends in school rampages seem to be somewhat unrelated in time to the other forms of violence. They do not rise and fall together. Nor do they seem to rise and fall in some other relationship to
OCR for page 301
Deadly Lessons: Understanding Lethal School Violence one another, with one rising after the other rises, or one falling while some other component rises. It is quite possible that this form of violence moves independently of other forms of violence, just as the level of domestic homicide or gangland murders seems to move somewhat independently of homicides in general. NOTES 1 The difference between the numbers of events in these two periods would be significant in any probability model constructed to test whether the number of student school rampages increased. However, the issue of statistical significance is not directly relevant here, as we have a census of the events. As a descriptive matter, there is little doubt that the number of student school rampages increased over time. 2 SHR victim data were weighted to match national FBI Uniform Crime Reports estimates of homicide victimization. 3 SHR offender data are limited by missing offender data associated with unsolved homicides. The estimates reported here use an imputation algorithm made available with the SHR computer file that provides estimates for these unsolved homicides based on the characteristics (age, race, and sex) of offenders associated with solved homicides (see Fox, 2001). 4 These events included homicides, suicides, legal intervention, or unintentional firearm-related related death of a student or nonstudent in which the fatal injury occurred (1) on the campus of a public or private elementary or secondary school, (2) while the victim was on the way to or from such a school, or (3) while the victim was attending or traveling to or from an official school-sponsored event (Anderson et al., 2001).
Representative terms from entire chapter: