I had been wondering if Wendy Davis was going to release one of these.
Sen. Wendy Davis
Democratic gubernatorial nominee Wendy Davis is trailing Republican opponent Greg Abbott by single digits for the first time this year in her campaign’s internal polling, according to a copy of it obtained by the Chronicle.
The Davis campaign’s latest survey, which was conducted last week, shows her taking 38 percent of the vote to Abbott’s 46 percent. A Rasmussen poll released last month also found Davis down by 8 percentage points.
Davis has narrowed Abbott’s lead by nearly two-thirds since January, when her campaign’s internal survey showed him up by 23 points. He led her by 11 in May and 12 in August, according to her campaign’s private polling.
The September survey is the first one Davis’ campaign did since launching a series of TV ads attacking Abbott. The first commercial, which started airing Aug. 8, has drawn the most attention, accusing Abbott of “siding with a corporation over a victim of rape” as a state Supreme Court justice.
A document summarizing the Davis campaign’s internal numbers this year shows the January survey is the only one in which Abbott garnered more than half the vote. The document’s headline reads, ”Davis Chipping Away At Abbott Vote As He Falls Below Critical 50% Mark.”
The Davis campaign’s internal poll suggests a much closer race than a private survey done by Abbott’s people during the last week of August. The Abbott campaign’s internal numbers, which were shared with donors earlier this month, had him beating Davis by 18 points. Most public polls this year have found Abbott leading Davis by double digits.
As with that other poll, which we heard about first from the Ken Paxton campaign, there’s no data or poll questions to examine (not even a polling memo in this case), so I can’t offer any kind of objective analysis. The lack of further information about this poll actually makes it sound a bit more credible to me than the one Abbott touted simply because it makes no claims about subsamples that are hard to reconcile with the overall data. It also fits with my own perceptions about the race, and who doesn’t love a little confirmation bias? Beyond that, take it with the amount of salt with which all internal polls should be taken.
You may wonder why Davis would bother to release an internal poll showing her to be down by eight. Countering that Wilson Perkins poll was surely part of the calculus here. More broadly, this is to try to establish a “she’s gaining momentum” narrative, hopefully to replace the “she’s trailing by double digits” narrative that has pervaded the coverage of the campaign. The Rasmussen poll from last month helped with that, but it’s only one result. With the release of the Abbott internal poll and now the release of another YouGov poll that’s more of the same from their static sample, a different perspective was a good idea. (YouGov also shows Sen. John Cornyn leading David Alameel by 20 points; for a good discussion of how to interpret their overall numbers, see Steve Singiser.)
You may also wonder why her campaign would release a poll showing her at 38 percent. Clearly, the key here was Abbott being under 50, at 46 percent, which was a change from their earlier results and a suggestion that the campaign is having an effect. For what it’s worth, in the large majority of polls that include crosstabs that I’ve seen in recent years, the lion’s share of “undecided” voters come from populations that are generally Dem-friendly, specifically African Americans and Latinos. I presume the same thing is going on in this poll, but of course I don’t have the data, so who really knows.
What would truly be of interest here would be to know how the population Davis is sampling differs from the others we have seen, which I strongly suspect bear a close resemblance to the 2010 electorate. Determining what the electorate will look like is a guessing game that all pollsters must engage in, and when there is a clear failure to accurately predict an election result (think Gallup in 2012), the culprit is often a wrong view of who will be voting. This is one area in which an internal poll can be more accurate than a third party poll, since any viable campaign ought to have a fairly clear view of their voter universe. My favorite example of this was the 2009 Mayor’s race here, where two polls of registered voters gave a lead to Peter Brown (who had been doing a lot of TV advertising) but an internal poll of Annise Parker’s that had a tighter (and in my opinion, more appropriate for that kind of race) voter screen had her leading instead. We know how that turned out. Of course, this can easily turn into an exercise in wishful thinking for a candidate that needs some good news. That’s why election forecasters that use poll averages tend to discount internal polls in some way, since they often represent something like a best case scenario. What makes this particular poll interesting, and why it would no doubt be enlightening to know how its demography compares to other polls, is precisely that there is an unprecedented effort by the Davis campaign and Battleground Texas to boost turnout among less frequent voters. Everyone agrees that will have some effect on the election, but no one knows how much. Knowing what kind of population the Davis team sampled would give us some insight into how they think it’s going. Without that, we’re left with waiting to see the final numbers, as we were before.