Here’s our first post-convention poll from an outfit that has polled the state at least twice so far this year.
Texas remains a toss-up in the presidential race. But Democrat Joe Biden’s modest – and somewhat startling – lead over President Donald Trump has evaporated in the last two months.
From a 5-point edge in early July, Biden now lags Trump by 2 points among likely Texas voters in a poll released Sunday by The Dallas Morning News and University of Texas at Tyler.
Trump’s lead is 48-46.
That turnaround is sure to gladden the hearts of Republicans, who have no hope of controlling the White House without Texas.
But as Trump has clawed his way back into contention just in time for the post-Labor Day sprint, Sen. John Cornyn has lost ground against Democrat MJ Hegar. His lead now stands at 11 points, down from 13 in early July.
But nearly 3 in 10 voters remain undecided, making for a potentially volatile fall.
“Trump is ahead,” said political scientist Mark Owens, who directed the poll, adding that the fates of the two Republicans at the top of the ticket are closely entwined. “Trump is helped by Cornyn.”
The poll, conducted Aug. 28 to Sept. 2, surveyed 1,176 registered voters. Of those, 901 said they are “extremely likely” to vote in November. The margin of error is plus or minus 2.87 percentage points for the bigger group, and 3.22 points for the subset of likely voters.
The news is better for Biden among registered voters, and since elections hinge on enthusiasm and turnout, it’s worth paying attention to all potential voters and not just those who are already sure to cast ballots.
Biden leads 44-43 with that broader group, though he led by 5 points two months ago.
See here for polling data, which for the first time contains demographic breakdowns of the data. I’ll get to that in a minute. The July poll had Biden up 48-43 among “likely” voters and 46-41 among registered voters. That was easily the biggest lead any poll had shown for Biden, so it seems likely it was a bit of an outlier. The April poll was just RVs and had a tied race, 43-43.
Some fascinating results in the data. Greg Abbott’s approval rating is a solid 54-33, right in line with his July rating. Abbott saw a dip in his approval ratings during the worst of the COVID-19 outbreak in July and August, but there wasn’t that much polling data that included an approval rating for him, so it’s hard to say how much of an effect there really was. Dan Patrick’s approval rating is a much lower 42-45 (he was at 37-37 in July), and Trump has a 40-38 rating, with 22% of respondents, including 22% of Democrats, saying they neither approve nor disapprove. Imagine me sitting here with my mouth hanging open, because that makes no sense at all.
It’s impossible for me to take very seriously a polling result that has both candidates in the 30s and 20s, with 30% or more of the respondents being not sure. Even the “Likely Voter” sample for the Senate race has Cornyn up 39-28, with 28% not knowing. Yet somehow, the result for the question about voting for the “Democratic or Republican candidate for the Texas State House” is 49-48 Dem for RVs and 48-49 for LVs. Seems to me the party preference at this level is going to tell you more about the party preference at the higher levels than anything else.
Interestingly, Biden does better among Likely Voters of color than RVs of color. He’s at 53-26 among Hispanic RVs and 75-10 among Black RVs, but at 58-28 and 87-9 in the LV screening. Make of that what you will. Trump goes from 57-34 among white RVs to 60-35 for white LVs, and 54-41 among white LVs with a college degree, which is better for him than elsewhere in the country, and 68-27 among whites with no college degree. The same thing happens with the “Democratic or Republican candidate for the Texas State House” question – it’s 67-33 Hispanic, 85-15 Black, and 37-56 for Dems among RVs, but 71-28, 90-10, and 36-64 for likely voters. Again, make of this what you will.
There are still many pollsters to be heard from, including UT/Texas Tribune, Quinnipiac, PPP, Fox News, and CBS News. It gets a little trickier tracking the data because now there are more likely voter results, which may or may not also include RV results. We’ll do the best we can. Remember, it’s never one poll result that matters, whether you like that result or not. It’s the aggregate, and as far as that goes, this remains a close race.