/cdn.vox-cdn.com/uploads/chorus_image/image/46913344/GettyImages-481847564.0.0.jpg)
There was a good deal of controversy over how Fox News winnowed down the GOP field from 17 candidates to 10 for tonight's first presidential debate. Some of that controversy was legitimate — for instance, the network wasn't particularly transparent about its methodology in advance. Other criticisms, though, missed the mark.
One unconvincing criticism was about which polls the network ended up using. Fox News had said it would use the five most recent national polls conducted before 5 pm on Tuesday. Yet when the decision was announced, the network decided to exclude one recent poll from NBC News and the Wall Street Journal — and include a slightly older one from Quinnipiac, as you can see in this image from RealClearPolitics:
Fox News had a perfectly legitimate reason for doing this. (RealClearPolitics, annotation by Vox)
As as Bob Somerby flags, this decision looked very strange to some people, particularly MSBNC host Rachel Maddow, who harshly criticized the decision on her Tuesday program. In particular, Maddow focused on how John Kasich barely made the cut for the primetime debate, and Rick Perry barely missed it — pointing out that in the excluded poll, Kasich and Perry were tied, but in the included poll, Kasich beat Perry by 3 points. "They moved the goal posts in the middle of the game in a way that makes it look like they were trying to justify the lineup of who they wanted on stage," she said, according to Mediaite.
But Fox's decision actually made perfect sense. In its most crucial question, the NBC/Wall Street Journal poll had made the odd decision to only offer 10 candidates as explicit options — and Kasich wasn't one of them, but Perry was:
NBC/WSJ vote wording. Doesn’t list Kasich and others. Very odd. pic.twitter.com/gSyXoXl69Z
— Charles Franklin (@PollsAndVotes) August 4, 2015
That isn't the practice the other polls used. For polls, the options explicitly presented are very important. So for this one, the candidates who were explicitly mentioned — including Perry — had gotten an advantage over those who weren't, like Kasich. (Kasich tied Perry anyway, from respondents who said they preferred one of the other candidates mentioned in a previous question.)
So Fox News decided not to include the NBC/Wall Street Journal — correctly, in my view. The network explained the decision in this release. And for what it's worth, even if the network did use that poll, it wouldn't have made a difference — Kasich still would have topped Perry in the average of the most recent five polls, as most analysts concluded that morning.