Research team to use the science of all data to understand soldier performance

Social characteristics include collective identity, social cohesion, unit dynamics, and a soldier’s commitment to the Army, Army values, and warrior ethos.

Sallie Keller, professor in the Department of Statistics in the College of Science and director of SDAL, and her research team received more than $3 million from the U.S. Army for a grant that spans five years to 2022. This project seeks to apply modern data science techniques to uncover the social characteristics of soldier performance that drives soldiers to successfully meet the challenges of rapid technological change in future wars.

“While the Army has tended to rely on internal factors like training and physical fitness to determine a soldier’s readiness for battle,” Keller explained, “this research focuses on understanding how the social characteristics of soldiers and units influence performance and success.”

The Army possesses vast amounts of administrative data on soldiers’ aptitude tests, demographics, family composition, training, fitness, promotions, and deployment history.

“The Army has yet to integrate these data to create a holistic operating picture,” Keller said. The research team will recategorize these data in a social context to capture Army values and warrior ethos, such as courage, honor, loyalty, and empathy. Additional data, such as that available for U.S. federal statistical agencies, will also be used to understand the influences of the economy and world conflict. Taken together, these repurposed data will be used in statistical models to predict soldier performance and unit readiness.

This is not the first collaboration between the Social and Decision Analytics Laboratory and the U.S. Army Research Institute. The first, now in its third year, is a $1.7 million cooperative agreement to determine the ability to access, evaluate the quality of, and integrate Department of Defense data to support decision-making related to the military population. The research is grounded in the context of a real problem: identifying predictors of attrition within enlisted ranks and understanding the effects of attrition on the Army and how to mitigate them.

“We are committed to building insights from existing data sources and letting identified gaps drive new measurement,” Keller stated. “We are working with the U.S. Army to model the social characteristics of soldiers and unit performance – what is today called people analytics – bringing the power of all data to inform their decision-making process. This is a perfect example of how we are working to fulfill our mission.”