Press Release: Predicting Which Soldiers Will Commit Severe, Violent Crimes

October 13, 2015
On October 6, 2015, a report in Psychological Medicine was published online with Ronald Kessler, PhD, McNeil Family Professor of Health Care Policy, as Principal Investigator. The report utilized research funded by the Department of Defense (DoD), and the authors collaborated with the Army Study to Assess Risk and Resilience in Service Members (Army STARRS). The study was part of Dr. Kessler's Behavioral-Based Predictors of Workplace Violence in the Army STARRS and was funded by the DoD's Office of the Assistant Secretary of Defense for Health Affairs, Defense Health Program (OASD/HA). 
 
The data provided administrative details for the 975,057 Regular U.S. Army soldiers on active duty between 2004 and 2009. Based on the analysis of these data, the researchers developed a machine learning model to predict which soldiers would later commit a "severe physical violent crime." The final model used hundreds of potential predictors and identified the top 5 percent of soldiers predicted as having the highest risk. This percentage accounted for 36.2 percent and 33.1 percent for men and women, respectively, of all major physical violent crimes between 2004 and 2009. When applied to 2011 to 2013, that 5 percent accounted for 50.5 percent of major violent crimes. 
 
Dr. Kessler said, “These numbers are striking. They show us that predictive analytic models can pinpoint the soldiers at highest violence risk for preventive interventions. Targeting such interventions might be the best way to bring down the violent crime rate in the Army.”
 
To read the press release, please visit the HMS website.
 
To view the original study, go to the Psychological Medicine website.
 
If you have questions about the information above, please contact David Cameron at the HMS Office of Communications and External Relations