Bannon was right.
That was in July; by the end of the campaign, things weren’t quite as bleak as that graph suggests. But polling conducted by the Associated Press and NORC found that suburban women backed Democratic House candidates in 2018 by a 19-point margin. Suburban men backed Republican candidates by a three-point margin. About 6 in 10 college-educated women voted for the Democratic candidate, preferring them to Republicans by a 16-point margin.
This was and is a point of concern for Republicans and Trump. So Trump was clearly eager to offer information to the contrary in a tweet this weekend.
As usual, an out-of-the-blue tweet from Trump can be linked to Fox News. In this case, to an episode of “Fox & Friends” guest-hosted by Katie Pavlich. Her guest was Beverly Hallberg, identified on-screen as part of the Independent Women’s Forum, where she is a visiting fellow.
Her name and affiliation are about the only part of Trump’s tweet that’s true.
Pavlich and Hallberg were discussing a poll conducted not by Hallberg or the IWF but by Zogby Analytics. Zogby found that Trump was “doing very well” with suburban women, with his approval sitting at 50 percent. What’s more, a “significant plurality” of 45 percent backed Trump trying to find funding for a border fence, and 50 percent of the group backed Trump’s declaration of a national emergency to get a barrier built on the border.
Why, just look at this graph!
There are a lot of problems here, beyond how weird that graph is. First of all, while about 50 percent of suburban women agree with the declaration, according to Zogby, slightly more disagree, thanks to the wonders of rounding. About 4 in 10 strongly disagree with the emergency, while 3 in 10 strongly agree. As for the wall funding, support from 45 percent of a population is not the sort of thing that normally gets trumpeted as good news.
More broadly, though, Zogby’s results come from a survey conducted online using a methodology that has been shown to yield results of widely varying quality. What’s more, asking whether respondents agree or disagree can influence more “agree” responses; it’s why The Washington Post and other pollsters instead ask whether people “support or oppose” a policy proposal.
Compare the Zogby results among suburban women with a recent NPR, PBS NewsHour and Marist poll. It asked whether respondents approved or disapproved of Trump’s national emergency declaration. Overall, 36 percent of adults said they approved — while 31 percent of suburban women agreed. Zogby pegs much of its assessment that Trump’s support from suburban women has grown to this idea that they back his moves on border security, but other polling suggests that they broadly oppose those moves.
Whatever. With the Zogby results prominently displayed, Pavlich and Hallberg raced ahead to discuss how this meant that security was “bringing them back” to the GOP after they left the party “a little bit” in 2018, in Pavlich’s words. In that Marist poll, 62 percent of suburban women said that Trump’s emergency declaration made them less likely to support Trump in 2020.
“I think for Democrats,” Hallberg said, “they need to be careful about how they message this issue, because the president seems to be winning on the crisis narrative.”
Again, even the Zogby poll found that 59 percent of respondents disagree with Trump’s emergency declaration.
Notice one number that we haven’t used yet in this article: 70 percent. Trump’s tweet offered that 70 percent of suburban women support border security and the wall, a claim he even put in quotes. But Pavlich didn’t say that, and Hallberg didn’t say that. It didn’t appear in the show’s lower third, either, in the on-screen text that Trump has pulled into tweets as a quote in the past. It doesn’t appear at all in the Zogby article about its poll, either.
So where’d it come from? In a way, a weird way, it doesn’t matter. Everything about Trump’s tweet was wildly off the mark, so what’s a 20 percentage point deviation from already questionably generous poll numbers between friends?
Scott Clement contributed to this report.