National Review editorializes:
Senator Kirsten Gillibrand (D., N.Y.) has made a hallmark issue out of the problem of sexual assault and related misconduct in the U.S. military, and her bill to change the way that crimes are prosecuted throughout the armed forces has drawn support from key Republicans, first Senator Chuck Grassley and now Senators Rand Paul and Ted Cruz. Senator Gillibrand and her associates are doing a service by drawing attention to these detestable crimes, but the solution they have put forth is the wrong one.
The legislation in question involves a good deal more than sexual assault. Senator Gillibrand’s bill, the Military Justice Improvement Act, would identify 37 specifically military crimes, such as disobeying orders or absence without leave, that would remain under the jurisdiction of the offending soldier’s chain of command. Other serious crimes — those punishable by a year’s incarceration or more — would be referred to a new, independent system of military prosecutors. It would, in short, replicate the civilian criminal-justice system. The direct involvement of commanders is part of what makes a court-martial a court-martial, and it is a fundamental part of the U.S. military’s command structure. Upending that structure is a serious thing to contemplate, and it is not at all clear that crime in the U.S. military, lamentable though it is, justifies doing so.
There is a great deal of irresponsible talk about an “epidemic” of sexual assault in the U.S. military, but it is far from clear that there is any such thing. A 2012 troop survey purported to show a 35 percent increase in unwanted sexual contact (USC) since the 2010 survey. (The military being the military, USC is an initialism with a life of its own, and it includes everything from a soldier getting handsy at happy hour to forcible rape, making it a metric of questionable value.) It is worth noting that the 2010 survey showed a 44 percent decrease from 2006. It is possible that USCs really did decline by 44 percent between the 2006 survey and the 2010 survey before spiking up 35 percent afterward; that would mean, among other things, that in these days of “epidemic” sexual misconduct, the rate of offense is lower than it was in 2006. A much more likely explanation is that these surveys, completion of which is not obligatory, suffer from selection bias. Perhaps more important, they also have a very low response rate, which introduces an element of randomness into the results. It is not clear whether such surveys tell us much of anything useful.
Whole thing here.