Violence perpetrated by military personnel is a major concern of the U.S. Department of Defense. A new study provides evidence that self-reported data can substantially improve predictions of who is at risk of committing physical or sexual crimes.
The findings were published in BMC Psychiatry.
“I am part of a research group that is studying risk factors for suicide and related outcomes among U.S. Army soldiers,” said Ronald Kessler, the McNeil Family Professor of Health Care Policy at Harvard Medical School and corresponding author of the study.
“One of the unique resources available to researchers working with the Army is access to the vast administrative records collected on each soldier. Our research group has been working for several years to see how well we can pinpoint soldiers at high risk of suicide and related outcomes.”
“But we also realize that administrative data alone have limitations, making us wonder how much we would be able to improve prediction by adding in information collected from self-report surveys of new soldiers.”
The researchers administered a New Soldier Survey to 18,838 men and 2,952 women who were beginning Basic Combat Training. The survey was administered as part of the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS), a multicomponent epidemiological-neurobiological study of suicides and related behavioral health outcomes.
Kessler previously found that a machine learning model of administrative Army data could predict which soldiers would subsequently commit a violent crime. The new study found that including the self-report data resulted in better predictions of physical violence perpetration by men, sexual violence perpetration by men, and sexual violence victimization of women.
“The use of administrative data in conjunction with self-report survey data collected from new soldiers makes it possible to pinpoint a relatively small proportion of soldiers who account for high proportions of several negative outcomes,” Kessler told PsyPost.
“This raises the question whether preventive interventions exist that would be cost-effective to implement that could be administered to these high-risk new soldiers, in an effort to help prevent these bad outcomes and improve chances of having a successful military career and life.”
“The major questions that need to be addressed next involve whether or not interventions exist that would be feasible to implement and sufficiently effective to reduce incidence of the outcomes we studied,” Kessler added.
The study, “Improving risk prediction accuracy for new soldiers in the U.S. Army by adding self-report survey data to administrative data“, was authored by Samantha L. Bernecker, Anthony J. Rosellini, Matthew K. Nock, Wai Tat Chiu, Peter M. Gutierrez, Irving Hwang, Thomas E. Joiner, James A. Naifeh, Nancy A. Sampson, Alan M. Zaslavsky, Murray B. Stein, Robert J. Ursano and Ronald C. Kessler.