Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
9America is now and has long been a relatively violent nation (Gurr,1989; NRC, 1993). It has had a particularly notorious history withrespect to lethal violence (Zimring, 1998). In recent years, the gen- eral violence of American society has engulfed the nationâs young (Cook and Laub, 1998). From 1985 to 1994, the United States experienced a historically unprecedented epidemic1 of lethal youth violence that took the lives of young victims, shattered inner-city communities, and ruined the prospects of many young people across the nation (Blumstein, 1995; Moore and Tonry, 1998; Cook and Laub, 1998). The general violence and the fear it causes have also reached into American schools. Serious violence has always been rare in schools, and it remained so even as society weathered the epidemic of youth violence. Recent figures show that between 1994 and 1999, 220 events of school- associated violent deaths occurredâon average, 36 per year. This is less than 0.3 percent of all violent deaths that occurred during that time pe- riod. Young people are far more likely to be killed or seriously injured when they are out of school rather than in it. The school boundary has continued to create a relatively safe haven for the nationâs young. To say that schools have remained relatively safe compared with other social and geographic communities is not to say that current levels of violence and fear in schools are acceptable. Quite the contrary. The prevalence of violent victimizations in schools has more than doubled since 1989: 3 percent of youths reported violent victimizations at school then, compared with 8 percent today. And while it is true that most 1 Introduction
10 DEADLY LESSONS: UNDERSTANDING LETHAL SCHOOL VIOLENCE schools are relatively free of violent crime, some schools experience very high rates, especially those in high-crime areas of U.S. cities (Kaufman et al., 1999). During the past 10 years, a spate of multiple-victim shooting inci- dents in school settings has greatly increased and widened public concern about violence in schools. From 1992 to 2001, 35 incidents occurred in which students showed up at their school or at a school-sponsored event and started shooting at their schoolmates and teachers. These incidents, represented most starkly by the incident at Columbine High School in Littleton, Colorado, left 53 people dead and 144 injured (see Table 9-1). These shootings contributed to a significant increase in homicide rates for students killed in multiple-victim incidents on school grounds between 1992 and 1999 (Anderson et al., 2001). These events shocked the public partly because so many were killed or injured so quickly in a single incident. It was also particularly frighten- ing that in many instances the victims seemed to have been chosen more or less at random. But the intense public concern was also due to the social location of the shootings. Much of the violence occurred in commu- nities that had, until the time these shootings occurred, been spared the kind of lethal youth violence that beset some of their urban neighbors. Moreover, the shootings took place in schoolsâthe place in communities that is supposed to be protective of children. Finally, the fact that these terrible shootings didnât stopâthey kept occurring at an apparently in- creasing rate in a pattern that suggested an emergent epidemicâpushed the level of concern much higher. THE STUDY CHARGE Responding to the high level of public concern, the U.S. Congress asked the National Research Council (NRC) to undertake a detailed study of lethal school violence, giving special attention to these particular events. Specifically, the NRC was asked to convene a committee to âconduct a study regarding antecedents of school violence in urban, suburban, and rural schools, including the incidents of school violence that occurred in Pearl, Mississippi; Paducah, Kentucky; Jonesboro, Arkansas; Springfield, Oregon; Edinboro, Pennsylvania; Fayetteville, Tennessee; Littleton, Colo- rado; and Conyers, Georgia.â Congress also requested that the study should be conducted through the development and analysis of detailed case studies describing the cir- cumstances leading up to the events, what happened in the events, and how the community responded both before and after the event. The goal was to use the case studies to learn, first, about the important causes and consequences of these unexpected lethal shootings and, second, what ac-
11INTRODUCTION tions individuals and institutions could take either to prevent such events from occurring in the first place or to minimize the damage once they begin to unfold. This special mandate from Congress challenged the National Acad- emiesâ National Research Council (NRC) to focus its attention on a very specific, very urgent national problem. Viewed from one perspective, this is not unusual. The NRC is often called on to offer scientific advice to policymakers as they confront urgent, specific problems. The difference here, however, was that the violent incidents that galvanized congres- sional concern seemed to many to be new and unique as well as urgent and specific. Of course, the problem represented by multiple, lethal victimizations in school settings could, on close investigation, turn out to be neither new nor unique. Perhaps there had been episodes like this in the past or in other places that had not been noticed as much as these recent incidents. Perhaps the phenomenon could be easily understood against the back- drop of a larger theory of violence, or of school violence more generally, or as some kind of offshoot of the dramatic increase in youth violence that had occurred earlier in the nationâs inner cities. If, however, on close investigation, the incidents that attracted special congressional attention turned out to be both new and different from what society had seen beforeâproducing different kinds of consequences, caused by different processes, and best prevented through different mechanisms than the other more familiar kinds of violenceâthen the committee charged with characterizing and understanding the problem would find itself in a difficult position. It would be challenged to make sense of a phenomenon for which no established literature existed. Without a strong science base to rely on, the committee would have to abandon its usual procedures of reviewing an extant literature, arbitrat- ing the disputes contained therein, and synthesizing the results. It would instead have to engage in original researchâsomething that is often thought to be done better by individual scholars pursuing their theories rather than a committee trying to reach a shared understanding. To the extent that original research was less than definitive, the committee would enter a realm in which the standards for making claims about causes, consequences, and effective cures were ambiguous. It would be operat- ing in a realm in which its findings had only a little more weight than the opinions of other thoughtful observers. THE COMMITTEEâS APPROACH The task before the committee was to find some satisfactory way to meet these challenges. An important first step was to develop agree-
12 DEADLY LESSONS: UNDERSTANDING LETHAL SCHOOL VIOLENCE ment about the operational definition of the phenomenon under study. An abstract concept does not always point unambiguously to a single operational definition. Consider, for example, the problems that arise when developing an operational definition of a concept such as alcohol- ism. Should the idea include both very high rates of alcohol consump- tion and frequent periods in which a person is highly intoxicated, or is that definition too narrow? Should it be defined only in terms of con- sumption or include the idea that individuals have âlost controlâ of their drinking? The development of an operational definition requires the exercise of some judgment on the part of the investigators. Much is at stake in that judgment. A particular phenomenon can be made to seem large or small in the world. It can come close to ordinary understandings of the word or concept, thus allowing an informed discussion; or it can depart from ordinary meanings of terms, thus introducing confusion and error into the public discourse. It can harden misconceptions about the nature of a problem, or it can help to illuminate. That is why people fight about operational definitions. On one hand, the committeeâs charge invited a very broad defini- tionâone that was almost exactly coterminous with a literature on school violence in general. On the other hand, it pointed to a set of incidents that seemed to have distinctive characteristics. The distinctive characteristics included not only the fact that the incidents occurred in schools and were committed by students, but also that they resulted in multiple deaths and serious injuries in a single incident. Incidents with these characteristics represent only a tiny subset of all instances of school violence: the few cases at the right tail of the distribution of seriousness for such events. While Congress seemed primarily interested in understanding these very serious incidents of school violence, seeing them in a broader context that could help us explain the size and significance of these very frighten- ing events, as well as what the causes of such events could be, seemed equally important. This led us to adopt an approach that focused atten- tion on four dimensions. Serious school violence involves: â¢ incidents of lethal violence â¢ that took place in or were associated with schools â¢ that were committed by students of the school and â¢ that resulted in multiple victimizations in a single incident. We did not have to include âuse of a gunâ in the definition of the events that interested us. It turned out to be true that all but a few of the cases that fit the definition of multiple victimizations involved the use of
13INTRODUCTION guns.2 We could, therefore, add the concept of âshootingsâ to the defini- tion without loss of accuracy. The committee summarized this definition as âlethal school violence including multiple victimizations.â This has the virtue of being short and easy to understand, and it focuses on a problem that is obviously of great public importance. It is also neutral and objective. The words have precise meaning. Furthermore, the definition itself does not suggest much about the causes of the events, and therefore it does not bias attempts to explain why the events occur. Moreover, the facts that one needs to assign particular incidents to this category of violence are relatively easy to ascertain for any given incident. The concept of lethal school violence could include incidents in which only one person was killed or injuredâa phenomenon that is much more common and may be somewhat differently motivated and executed than incidents in which many were killed or injured all at once. It could also exclude such incidents as the Rockdale County case (Conyers, Georgia) that Congress had identified in the legislation. That incident included multiple injuries but no fatalities. The committee didnât think it made sense to exclude this incident, since it was certainly more probable than not that someone could have died in this kind of incident. The committee therefore decided it would be a mistake to apply its operational definitionâlethal school violence including multiple victim- izationsâtoo rigidly. It seemed clear that any incident in which a student walked into a school and started shooting apparently randomly was of potential interest. This moved us away from defining the violence in terms of its consequences (measured in terms of victimization) and focuses instead on the motivations and behavior of the offender. This operational definition of lethal school violence seemed best to reflect congressional intent in requesting the study. The next step was to consider what sources to use to develop an understanding of the incidents. We already knew that we would have cases describing some of these incidents in rich, narrative detailâprovid- ing much of what we would rely on. We also thought it was important to put the cases into a broader perspective. We sought to locate this particular kind of lethal school violence in both the countryâs overall experience with violence and in our theoretical understanding of violence. To do so, we needed to construct a new database to identify all the incidents of violence that met the criteria. This would indicate how large this form of violence was compared with other forms, when this form of violence appeared, and how fast it had been growing. It might even allow us to see whether this kind of violence seemed to move independently of
14 DEADLY LESSONS: UNDERSTANDING LETHAL SCHOOL VIOLENCE or consonant with other forms of violence, thus providing clues about whether its causes were the same as those that caused other forms of violence. We also needed to identify and read the literatures on violence that seemed related to the particular incidents of violence we had been asked to study. This was important not only in itself, but also as a necessary guide to our efforts to interpret the cases. Our theoretical understanding and review of data would help us select cases and construct the templates for gathering information. Our review and understanding of the relevant literature would help guide our interpretation of the cases. At the outset, it was not clear which particular theoretical literature would be relevant, since we did not yet know how to locate the form of violence we had been assigned to study. Again, we knew it was at the tail of some distribution because the incidents were extremely serious and therefore quite rare. But it was by no means clear to what theoretical categories these particular instances of violence should be assigned. In the end, we explored several literatures to help interpret the cases and more generally to help us think about what might be causing the incidents and how they might best be prevented. First, was a very small literature on incidents that looked very much like the ones we had been asked to review (only four studies, and only two of them that met reason- able standards of scientific care and rigor). Second, there are large litera- tures on broad categories of violence, such as violence in general, youth violence, school violence, and the relationship between violence and sui- cide. Third, there were also literatures on some specialized forms of violence that bore some similarity to the incidents under study. These included literatures on mass murders, rampage shootings, and âsuicide by copâ (e.g., incidents in which individuals seemed to shoot in order to provoke a response by the police). Finally, because there might be some contagion effects in the events we were examining, we looked into the literatures that explored the contagiousness of violence and other social events. These are the sources the committee used to understand the incidents of lethal school violence involving multiple victims: the cases that describe six specific incidents; a statistical database constructed for the committeeâs purposes from several existing sources; and several litera- tures that had something more or less directly relevant to say about this phenomenon. Perhaps the most valuable resource we had was the commitment and expertise of the case writers who developed the cases, all senior scholars with a significant amount of experience in qualitative methods and most with substantive knowledge in the area of youth violence. The discus- sions we held were designed to stretch the evidence and knowledge on incidents of lethal school violence as far as we reasonably could, and to
15INTRODUCTION make it available to the nation as it sought to deal with this important, urgent problem. There is almost nothing in this report that meets the usual scientific standard of 95 percent confidence that a statement is true. But there is much that is likely to be true (more probable than not), and some of these things are not obvious. There are also some findings that upset some conventional assumptions about the phenomenon we studied, and there are some pretty clear ideas about where the priorities for future research may lie that could strengthen the science base for understanding and preventing particular kinds of lethal school violence. The report is organized in the following way. Part I presents the cases. They are preceeded by a short analytic section that explains why they were chosen, lays out the template used to collect the information, describes the sources consulted to obtain the information, and describes the procedures followed by the case authors and the committee for the protection of human subjects. Part I also presents a cross-case analysis of the similarities and differences among the specific cases to develop some plausible hypotheses about the causes, consequences, and effective meth- ods of preventing and controlling these incidents. Part II puts the cases into the context of the literature reviewed and uses the cross-case analysis to develop some observations about policy implications and research recommendations. In reading Part II, it is im- portant to keep in mind that the work contained therein reflects sources beyond the cases. It presents statistics about trends in violence and find- ings from the literatures that seemed relevant to the inquiry. We begin with the cases because, even though these are not all that we relied on, they are in many ways the heart of our understanding and the source of some of our most important ideas. It is valuable for readers to enter this field as we didâin an inductive, exploratory way rather than a deductive, hypothesis-confirming way. The cases, presented here as signed, stand-alone pieces by the case investigators, are full of surprising facts, poignant moments, and rich insights. They put a human face on the tragedies that have beset the nation. They fill in the gaps that empirical studies cannot address. Those who are more inclined toward a deductive approach and would like to see what observations we have drawn first and then to test them against the evidence of the cases should feel free to start with Part II and read the cases afterward. NOTES 1Epidemic means an elevated level; it does not necessarily imply a contagious phenom- enon. There could be a contagious mechanism at work, but that requires investigation. 2One case in Portland, Oregon, involved a knife, and an international case involved arson.