Profiling is Unable to Identify a Child Who May be Prone to Engaging in School Violence – (excellent Secret Service example)

The 2001 Secret Service report titled “Evaluating Risk for Targeted Violence in Schools: Comparing Risk Assessment, Threat Assessment, and Other Approaches” discussed the lack of reliability with endorsing profiling as a means to identify, treat or remove children from schools who are likely to engage in targeted school violence.  As parents, community members and even teachers seek means to “screen” for school shooters, it is important to understand the limitations of screening and profiling, and also the risks of inadvertently stereotyping a group of students and providing a false sense of having screened one’s way to a safe environment.  While intentions are good, resources are limited – and dedicating resources to something that is not statistically sound diverts educators from research-supported activities that decrease school violence, such as promoting school connectedness.  The following is an excerpt from the aforementioned study – I recommend using it as a talking point when responding to suggestions to “screen” for the students prone to targeted school violence.

“Numerous concerns have arisen over the use of prospective profiling to identify and assess the risk students pose for targeted violence in school. First, prospective profiling is not sufficiently sensitive nor specific to identify a child who may be at risk for engaging in targeted school violence, nor for evaluating the child’s likelihood of doing so (Sewell & Mendelsohn, 2000). Prospective profiling to identify students likely to become “school shoot­ers” carries with it considerable risk of false positives; that is, because targeted violence in school is such a rare event, most who “fit” the profile will not engage in acts of targeted school violence (Sewell & Mendelsohn, 2000). In addition, use of prospective profiles would inappro­priately exclude students who do not fit the profile, but who may in fact pose a risk of targeted violence. By way of example, the use of a prospective profile derived from previous assassins would have failed to identify Sarah Jane Moore prior to her assassination attempt on President Ford in San Francisco in 1975. The profile most accepted at that time would have predicted Ford’s attacker to be male, between the ages of 20 and 40, of slight build, born overseas, un­employed, a loner, and someone who suffered from delusions of grandeur or persecution (Weisz & Taylor, 1969). At the time she shot at Ford, Moore was female, in her mid-40s, of stocky build, born in the U.S., employed full-time as an accountant, had been married and had a son, and had no history of delusions.”

Facebooktwittergoogle_plusredditpinterestlinkedinmail

Leave a Comment