> Using those numbers suggests that 10k of those suicides quoted above were specifically attributable to military service (and 20k would be "baseline civilian r

> Using those numbers suggests that 10k of those suicides quoted above were specifically attributable to military service (and 20k would be "baseline civilian risk of suicide"). This would be the case if probability of joining the military was independent of the risk factors for suicide, which I seriously doubt. Not disagreeing with the point at large, just that the statistical analysis would be even more complicated, at least finding similar demographics between the two populations