Methodology

Over the past two years, students and faculty of the Division of School Psychology at Alfred University followed media accounts, research results, and expert opinion on the causes of lethal school violence and attempts to develop useful assessments of risk. Checklists of risk factors and characteristics of previous offenders result in lists of symptoms and features that these youngsters display. The problem is that these features are common to most students dealing with emotional and behavioral problems, but do not distinguish those at risk for school violence from this much larger group. The danger in profiling from these characteristics is the number of false positives that would occur, and the concomitant unwarranted danger to privacy and civil rights.

So we are left largely with opinion, as informed as it might be, about causes of school shootings. Those that have had the highest visibility include access to firearms, being bullied or mistreated in school, family dysfunction, and various psychological disorders. The first objective for this study was to test these theories and opinions about school shootings with actual students.

Despite the lack of a valid discriminating profile, there are characteristics of events that have occurred frequently with instances of school shootings. School shootings, for the most part, are not impulsive acts. They seemed to be planned and premeditated. As a matter of fact, shooters seem to think about these acts much before they are committed; to generate a plan or scenario for how to go about the shooting; and to frequently tell others about their thoughts and plans. The second objective of this study was then to investigate with students the extent to which these conditions exist in their own schools, with people they know or have heard about. And the third objective was to discover to what extent students harbor these thoughts and feelings themselves - to assess risk from the students own self-reports.

The Division of School Psychology developed the plan for this study at Alfred University. We asked students to tell us their opinions of 16 different possible causes of lethal school violence that have been proposed in the academic, professional, and popular literature. The students were also asked to report their awareness of conditions that have been linked to previous school shootings. For example, we asked students whether they had heard of or knew of other students who had spoken of committing such acts. Finally, we asked the students themselves about their own thoughts, whether they themselves had contemplated committing such acts. The surveys also asked demographic and other descriptive information to facilitate data analysis.

The Division then commissioned Harris Interactive (1) to conduct a national internet survey of public school students in grades 7-12. The sample was selected from Harris Interactive's existing panel of more than seven million students. The 2,017 responses were weighted using propensity score adjustment to reflect a nationally representative sample of junior and senior high school students. Propensity score adjustment uses regular, repetitive, parallel telephone and on-line surveys to adjust for any self-selection bias in on-line surveys. The efficiency and accuracy of the methodology was demonstrated during the 2000 elections, when Harris Interactive predicted national and state election results more accurately than most traditional polls (Harris Interactive was the only major poll to predict a dead heat in the 2000 presidential race). That success played an important role in our decision to use the online survey method. The surveys themselves were anonymous, so we cannot identify individual respondents.

In some of the data analyses, we used three indices developed by Harris to measure students' quality of life, degree of alienation, and media usage. These indices allowed us to explore whether students who scored high or low on these characteristics differed from each other in their responses. The indices were constructed as follows:

Alienation index
Determined by their responses to the questions, "I feel left out at home," "What I think doesn't matter at home," and "What I think doesn't matter at school." Students were classified as having high, medium or low alienation indices.

Quality of life index
Measured by their responses to the questions, "I have a lot of friends," "I get along well with my parents," "I am often bored," "I often feel sad and unhappy," "I have been happy at school this past year," and "I get into trouble a lot. " Students were classified as having high, medium or low quality of life depending upon their responses to those questions.

Media usage
Based on how much time each week they spend watching videos, playing video games or computer games, how much time they spend in chat rooms and how much time they spend looking at websites. Those who spent less than eight hours a week total were classified as low media users; those who spent nine to 17 hours a week were classified as medium media users; and those who spent more than 18 hours a week were classified as high media users.

(1) Harris Interactive (Nasdaq: HPOL) is a worldwide market research, polling and consulting firm. It is best know for The Harris Poll and its pioneering use of the Internet to conduct scientifically accurate market research.