Unit-level variations and peer influences in mental health diagnoses in the U.S. Army

This paper uses U.S. Army personnel data from Defense Enrollment Eligibility Reporting System (DEERS), Defense Manpower Data Center (DMDC), and TRICARE, the U.S. Department of Defense health care system, to examine how unit-level variations in incidents of mental health diagnoses affect the likelihood of an individual developing mental health disorders and the effect peers have on one's mental health outcomes. Both fixed- and random-effect regressions are employed to observe variations across units without unobserved time-invariant differences such as culture or leadership style, as well as effects of time-variant variables such as location and size. At the unit level, we consistently find that having more deployed, female, non-white soldiers, and having at least one soldier who experienced divorce and demotion, are associated with an increase in the likelihood of an individual being diagnosed with a mental disorder, and an increase in the percentage of individuals diagnosed with mental disorders in that unit. At the individual level, we observed that individuals who experienced stressful events are more likely to be diagnosed with mental health problems. Conditional to one's own experience, having peers who currently have or have had stressful events in the past are also associated with a higher likelihood of developing mental health disorders. Separately, mental health diagnoses vary with the unit's geographical location, but further research is needed to determine why this variation exists.