In the school gymnasium, the president talked about his "concern" in coming to Springfield. "I was afraid all I could do was to tell you that your country has been thinking about you and praying for you and pulling for you." The president wanted to do more than just offer comfort (which he nevertheless did in his customary and effective role as Consoler-in-Chief). And he did not want simply to assign blame for school violence, although he pointed fingers at: the glorification of violence in popular culture and its desensitizing effect on young minds; easy access to guns; and lax parental supervision.

As a kind of response to the problem of school violence, lists of warning signs have proven extremely popular. During the year surrounding the killing at Thurston High, the American Psychological Association issued "22 Warning Signs" that might indicate a "serious possibility" or "potential" for violence. The National School Safety Center, a nonprofit advocacy group in California, drafted a 20-point "Checklist of Characteristics of Youth Who Have Caused School-Associated Violent Deaths." And The National Center for the Prevention of Crime (NCPC) - popularizers of "McGruff The Crime Dog," who famously helped kids "take a bite out of crime" -- offered a list of "signs that kids are troubled" which might require "action."

A review of the different lists shows loose consensus around a number of warning signs for potential youth/school violence: chronic feelings of isolation or rejection, frequent angry outbursts, social withdrawal or depression, fascination with or possession of weapons, alcohol or drug dependency, history of bullying behavior, and lack of interest in school or poor school performance. Then there are items common to several lists, but not to all, like cruelty to animals and gang affiliation. And then there are some items that appear only on one or another list -- "dresses sloppily," is a "geek or nerd"; "characteristically resorts to name calling, cursing, or abusive language" -- that seem to be only marginally useful as warning signs, and may raise questions about the study's underlying research methodology.

A different type of response to school shootings is coming from law enforcement groups, or from researchers who adopted law enforcement-style approaches. One study, the Classroom Avenger, was conceived as a teen version of the FBI's profile of the "Workplace Avenger." The study's authors hypothesized: "With adult Workplace Avengers the precipitant is discipline and/or termination/rejection by the employer; the motive is vengeance. . . In the case of adolescent 'Classroom Avengers' the precipitant is discipline by parents or authorities and/or rejection by peers or girlfriends; the motive is also vengeance and the action is premeditated shooting spree with parents, fellow students and /or faculty/school administrators as target victims.") Interestingly, the study lists several "exclusionary" characteristics not associated with Classroom Avengers -- "drug/alcohol abuse" and "poor academic performance" -- that almost all of the warning signs checklists associate with high potential for violence. At the same time, the study's list of inclusionary characteristics provided one of the closest fits to Kip Kinkel's case of any reviewed by FRONTLINE.

The FBI itself is preparing a report on "problematic traits" of potential school shooters to be released early in 2000; it is not yet clear whether they, too, will see a connection to their "Workplace Avenger" research. The Secret Service, shedding some of its namesake secrecy, announced the establishment of a "National Threat Assessment Center," where a team of investigators is adapting research on "American assassins" to solve the problem of school violence. A different type of project is being partly funded by the Alcohol, Tobacco, and Firearms Bureau (ATF) and carried out in partnership with a private company, Gavin de Becker, Incorporated -- a self-described "very unusual firm. . .called upon by our government when some zealot shoots an abortion doctor or opens fire on federal employees" to "help predict human behavior" and to"help clients manage both fear and risk." The ATF/de Becker partnership has produced a software program, "Mosaic 2000," which promises to help school administrators detect troubled kids on the verge of violence.

There is no telling for sure how effective any of these responses has been in preventing or reducing incidents of school violence, or how widely or systematically these different approaches have been employed. In the last year, however, a number of warnings have been sounded about the "warning signs" approach itself. Some checklists have been criticized as unhelpfully vague or miseladingly predictive -- a danger that arises mainly when warning signs lists are informally adapted as profiling tools. The National School Safety Center's checklist initially included a scoring component that rated the likelihood for violence, for example; after more careful thought about its potential misuse, the scoring element was dropped. Other checklists and warning signs approaches have been challenged as dangerously likely to identify and label too many students, or as insufficiently based on peer-reviewed research to be considered valid.

The ATF/de Becker "Mosaic 2000" pilot profiling program, which began late in 1999 in some 20 schools nationwide, has drawn some of the heaviest, and most public, criticism. Commenting on "Mosaic 2000," an editorial in the Seattle Times stated:

"One of the more ludicrous and inevitable solutions for preventing violence in schools is a computer program that tabulates risk. . . random violence in schools has fostered a desperation for predictability and control." (12/20/99)

"There's something terribly wrong with the Mosaic-2000 mind-set: that if we can just find the right software program and feed it some surface characteristics of teens, it will assemble for us the three-dimensional understanding of them that so eludes us. Software is no substitute for the real conversations that need to take place, day to day, in American high schools.. . Profiling is a fine technique for FBI manhunts; it is misplaced in American schools." (11/19/99)

Not all of the law enforcement responses to school shootings are profiling attempts, however. The Secret Service response will likely be based on a "threat assessment approach" -- and the methods and underlying principles of threat assessment are fundamentally different from those of profiling. While profiling conceives of dangerousness as a matter of individual disposition not likely to change over time, the "threat assessment" approach focuses on "pathways to violent action" and the specific contexts in which potential for violence is actualized. Threat assessment is not primarily concerned with identifying the immutable character traits of potentially violent actors. Threat assessment considers "the interaction among the potential attacker, past stressful events, a current situation, and the target(s)." Someone with a certain disposition toward violence, or a history of violent behaviors, might be on a pathway to violence, on the threat assessment model, but that person will act violently only under a specific set of circumstances that is hard to predict. [1] Some critics of profiling are more accepting of threat assessment.

So, where do all of these criticisms leave parents or school officials in the aftermath of school violence? The president's dilemma at Thurston High is not so easily argued away, and the facts continue to demand a response: From the summer of 1994 through the end of 1999, there were more than 200 school associated violent deaths, most involving guns. The rate of such incidents is not increasing significantly, but the lethality and number of multi-victim events is rising. And the intense, 24-hour national media coverage of these events contributes to a growing sense of "national crisis" that creates pressure for solutions.

This brings us back to the report that the president ordered on the day of his visit to Springfield, Oregon. The report offered its own list of "warning signs" featuring many of the same items as the other lists. But the report's co-author, Kevin Dwyer, the president of the National Association of School Psychologists and an outspoken critic of a number of checklists and profiling attempts, also made sure to include a thoughtful discussion of "principles" for use of the warning signs, as well as frequent cautions on their limits. "It is inappropriate, and potentially harmful, to use the early warning signs as a checklist against which to match individual children," the report states. "It is important to avoid inappropriately labeling or stigmatizing individual students because they appear to fit a specific profile or set of early warning indicators. It's okay to be worried about a child, but it's not okay to overreact and jump to conclusions."

The cautions are well-taken. Longterm answers to the problems underlying school violence are complex, and may require a number of important actions: rethinking boys' emotional education; taking more seriously boys' depression; understanding the connections between violence and feelings of shame, and injustice; or looking more closely at other crucial areas which we avoid only at our own peril. There are so many ways to lose boys to violence. Checklists of warning signs, behavioral profiles, and threat assessments may be a necessary stop-gap measure, or a comfort to those who may otherwise feel helpless for something constructive to do in the aftermath of a shooting -- and they may even save lives. But even those who write and endorse the lists and profiles will freely admit: this is only a rough draft of a solution to a more complicated problem.

[1]  This review of the differing methodologies relies on "Risk Factors in School Shootings," Stephanie Verlinden et. al, unpublished manuscript (1999), pp. 44-48. For a discussion of the principles of "threat assessment" and the "myths" of profiling, see the 1998 research of Robert A. Fein and Bryan Vossekuil.