With in-person early voting set to begin in Texas on October 24, the latest University of Texas/Texas Politics Project poll finds Gov. Greg Abbott leading Democratic challenger Beto O’Rourke in the gubernatorial race, 54%-43%, among Texans likely to vote in the 2022 election. While more than half of Republican voters say immigration and border security is the most important issue area informing their vote, Democratic voters’ attention is divided among a list of several issues, topped by abortion.
The poll surveyed 1,200 self-declared registered voters using the internet from October 7-17 and has a margin of error of +/- 2.83 for the full sample. From among this overall sample, likely voters were defined as those respondents who indicated that they have voted in every election in the past 2-3 years; or those respondents who rated their likelihood to vote in the November elections on a 10-point scale as a 9 or a 10. This likely voter screen yielded a pool of 883 likely voters, with a margin of error of +/- 3.3% for the full likely voter sample.
Beyond the two major party candidates, Green Party Candidate Delilah Barrios and the Libertarian Party’s Mark Tippets each earned 1% support while 2% preferred an unspecified “someone else.”
[…]
The results among likely voters found Republican candidates maintaining wide leads in the five other major races for statewide office. In all of the trial ballots, including for governor, undecided, but likely, voters were asked whom they would choose if forced to make a decision. All results for the trial ballots report the results of the initial question combined with this “forced” response. (The poll summary reports the share of voters who expressed no preference in the initial question in each race.)
Lt. Governor. Incumbent Dan Patrick led Democratic challenger Mike Collier, 51%-36%, in their rematch of the 2018 race.
Attorney General. Incumbent Republican Ken Paxton leads Democrat Rochelle Garza 51%-37%.
Comptroller of Public Accounts. Two-term incumbent Republican Glenn Hegar leads Democrat Janet Dudding 47%-35%.
Agriculture Commissioner. Incumbent Sid Miller leads Democrat Susan Hayes 51%-39%.
Land Commissioner. Republican State Senator Dawn Buckingham leads Democrat Jay Kleberg 47%-36%.
The generic ballots for the U.S. House of Representatives and the Texas legislature also revealed continuing advantages for Republican candidates: Republicans lead 53%-44% in the generic ballot for the U.S. House of Representatives, and 53%-42% for the Texas legislature.
This is upsetting mostly because the August poll had Abbott up by only five and had shown a slight but steady drift towards Beto over time. The one caveat here is that the previous polls were of the full registered voters sample, and this is of “likely voters”, which is about three-fourths of the original. It’s not a direct comparison as a result, though of course the pollsters will have done what they think is best to reflect the electorate accurately. If they provided numbers for the full sample in October, I didn’t see them.
The October poll data is here and the August data is here. The underlying atmosphere has not changed in any significant way. Biden’s approval was 40-52 in August and it’s 39-52 in October (the approval numbers are still based on the full sample in each case). Abbott went from 46-44 to 47-44. Dan Patrick and Ken Paxton were actually slightly worse in October, going from 38-37 to 37-39 for Patrick and from 37-38 to 36-39 for Paxton. Either a lot of people changed their minds or that likely voter screen is a big difference maker.
I’ve put my faith in the “the screen is too tight” beliefs before without much success, so I don’t want to go overboard here. If these numbers are accurate, they don’t bode well for Harris County either, suggesting Beto might end up with 52 to 54 percent. At the high end, as I’ve said before, I’d still feel pretty confident about Harris County Dems. Less than that, and I would expect Republicans to win at least some races. Maybe this year is another inflection point, and maybe the dip in the gap between Harris and the state that we saw in 2020 following years of games will not be a one off. No way to know until we start to see some real numbers.
The poll also includes this demographic breakdown in the vote:
White/Anglo: Abbott 64%, O’Rourke 32%
Hispanic: O’Rourke 48%, Abbott 48%
Black: O’Rourke 86%, Abbott 11%
Those are the strongest numbers Beto has had for Black voters in awhile. They’re not great for white voters – compare to the Marist poll, for example, which had Abbott leading Beto by a much smaller 57-37 margin among those voters – and this is another poll that has Beto with no advantage among Hispanic voters; note that was also true in the Marist poll. We saw a great disparity in Hispanic preferences in the 2020 polls, and in the end the ones that showed a smaller lead for Dems were more accurate. I don’t know what else to say here.
I will add that we saw one more poll result released yesterday, from the Democratic AG’s Association (DAGA), which claimed Rochelle Garza was trailing Ken Paxton by two points, 48-46. That linked poll memo is the entire thing – no Beto/Abbott numbers, no Biden approval numbers, no crosstabs, nothing – and it’s basically an internal poll, so maintain a higher level of skepticism for this one. I will note the following from the memo:
The survey was conducted between October 12th-16th using live calls to landlines, SMS text-to-web and live calls to cell phones, and an online panel. The sample includes 879 registered voters and is weighted to reflect a likely 2022 Texas general electorate. The margin of error is +/- 3.24% at a 95% confidence interval.
The results of the survey show that when asked who they’ll vote for as Attorney General and Texas undecided voters are allocated to a candidate, Paxton is only ahead by 2 points, within the margin of error for the survey, landing at 48% Paxton, 46% Garza, with 6% of voters say they’re voting for Libertarian Mark Ash in the AG race.
Another “likely voter” result, though with less detail. They also seemingly pushed the initial non-respondents into picking a side, which I had initially frowned at but I guess if the UT/TPP folks can do it, they can too.
39 days and counting………….