The Chicago Public Schools see approximately 30 students killed and many more involved in some kind of school violence each year; that's why they've embarked on a $30 million initiative to lower those numbers, particularly the toll of fatalities. Using a probability model, the district will identify those high school students most likely to be gunshot victims--1,200 have already been spotted--and direct resources (such as a personal advocate, a social worker, and a paid job) towards these pupils in an attempt to reduce the odds of tragedy. Already, the data have revealed significant trends among the most at-risk: They are most likely to be black males, homeless, special-education students, and students at alternative schools. Typically, they are also at least two credits behind in their coursework, are absent for 40 percent or more of the school year, and tend to commit one serious school infraction per year. These stereotypes are not exactly news, but identifying and doing something about the individual students who fit into such categories is innovative and encouraging. The possibilities for using such algorithms to identify students at risk for other things--like dropping out or failing--are endless.
"$30 million plan to save kids," by Rosalind Rossi, Chicago Sun Times, September 4, 2009