Here’s a new poll of four statewide races in Texas, for which there may or may not be any news coverage. The executive summary:
Gravis Marketing, a nonpartisan research firm, conducted a random survey of 602 likely voters across Texas. The poll was conducted from July 3rd through July 7th and has a margin of error of ±4.0%. The totals may not round to 100% because of rounding. The survey was conducted using an online panel of cell phone users and interactive voice responses. The results are weighted by voting demographics. The poll was paid for by Gravis Marketing.
As there is no news story to excerpt, I’ll just go straight to the results:
US Senate: Ted Cruz 51, Beto O’Rourke 42
Governor: Greg Abbott 51, Lupe Valdez 41
Lt. Governor: Dan Patrick 46, Mike Collier 44
Attorney General: Ken Paxton 45, Justin Nelson 41
There are a bunch of approval and issue questions in the polling memo, so feel free to browse through it. I will note two things. One is that Gravis is rated as a C+ pollster by FiveThirtyEight, better than some but worse than many others. Like Quinnipiac, they have no record in Texas prior to this year that I’m aware of. Two, while I haven’t spent any time critiquing subsamples in the polls we’ve seen so far, I have to say that the subsamples in this poll are nuts. Somehow, Gravis found the most ridiculously and unbelievably Republican group of 18-29 year olds and Hispanics I’ve ever seen, as well as the least hostile-to-Democrats Anglos. I have no explanation for this, and to some extent it doesn’t really matter. It is what it is, and what it is is another data point. And that data point brings the Senate poll average to 46.9 for Cruz, and 40.0 for O’Rourke.
I heard about this poll via a campaign email from Mike Collier, who for obvious reasons wanted to tout this result. (The TDP subsequently posted about it.) The low “don’t know/no answer” rate for the Lite Guv and AG questions is suspicious, but maybe that’s a function of their “likely voter” screen. Collier trails Patrick 50.4 to 42.8 among white voters, which is why he is so close in the race despite trailing 43.8 to 36.4 among Hispanics and leading by a mere 57.9 to 34.9 among blacks. Did I make my incredulity about this polls’ subsamples clear enough? You see some wacky stuff sometimes when the subgroups are small, but good Lord. As I’ve said, it’s a data point. Don’t make any more of it than that.