You’ll be tempted to call this a weird fluke that no one should take seriously. It isn’t. Remember this non-scientific poll from July?

Military troops favor Libertarian candidate Gary Johnson for president over Donald Trump and Hillary Clinton, according to a new survey.

Johnson garnered 38.7 percent of the active duty vote, versus 30.9 for Trump, and 14.1 for Clinton, according to the survey, which was conducted via the popular military personality Doctrine Man…

Among all services except for the Navy, Johnson performed better than Trump and Clinton.

Ludicrous. Surely a scientific survey of service personnel would debunk the idea that also-ran Gary Johnson, who couldn’t muster 15 percent nationally to make the debates, is running at a 35+ percent clip within the military.

Well, here’s your scientific survey, courtesy of the Military Times: Trump 37.6, Johnson 36.5, Clinton 16.3. Among the officer class, Johnson actually leads — and it ain’t close.

Note that Trump finishes third among that group. It’s not surprising that he’d trail Hillary among a better-educated demographic, as he’s been trailing her among college grads in the general population all year. What is surprising is that they’re both trailing Johnson. Johnson also leads both candidates in two service branches, the Navy and the Air Force, and he’s above 35 percent within the Army. The only branch where Trump leads big is the Marine Corps, where he’s above 50 percent while Johnson lands at 26.7 percent. In every single group tested, though, the libertarian dark horse finishes ahead of presidential frontrunner Hillary Clinton.

Another reason not to treat this as a fluke is that Ron Paul did famously well in attracting donations from the military four years ago. There really is a strong faction of libertarian support within the service, it seems, although the “Trump and Clinton are both garbage” factor is doubtless helping Johnson. Even so, 8-10 percent nationally versus 36 percent within the military is a big cultural gap. Any vets out there care to weigh in with theories as to why? One obvious explanation is that servicemen have a stronger incentive than the rest of us to prefer an anti-interventionist president. After Iraq and Libya, you know what you’re getting with Hillary Clinton. You don’t know what you’re getting with Trump, which is a different problem. Some days he sounds like he’s unprepared to enforce American military alliances, other days he sounds like he’s ready to send thousands of troops back into Iraq to smash ISIS and “take the oil.” He’s a tough guy, but tough guys might be more inclined to put American soldiers in harm’s way. Doug Bandow, a fellow at the libertarian Cato Institute, tried to explain Johnson’s strong showing in the July poll:

While service personnel are willing to serve in combat, most do not want to do so absent compelling justifications. And few of the interests involved in Washington’s recent conflicts can be considered serious, let alone vital. A Marine Corps veteran told Egan that service members “realize they’re being utilized for other purposes — nation building and being world’s policeman — and it’s not what they signed up for.”

Does that capture it or is that too pat? And why would there be such a discrepancy between enlisted personnel and officers in their preferences for Trump versus Clinton? Is it really an educational thing? The fact that Trump and Clinton are both deeply dishonorable people makes me think intuitively that that must be helping Johnson among service members, but surely honor isn’t worth 25 points between the military and the overall population. Or is it? Good lord, what a depressing comment on American civilians.