New pollster says Romney leads 55-40 in Texas

It would figure that the day after I complain about a lack of polling data in Texas we get a fresh poll result for the state.

Thanks to a lopsided lead among white voters, Romney is leading President Barack Obama 55-40, according to a poll from Wilson Perkins Allen Opinion Research, which carries out surveys for GOP candidates. That’s even higher than John McCain scored against Obama in 2008.

Chris Perkins, who is a partner at the research firm and served as the pollster for U.S. Senate hopeful Ted Cruz in his Republican primary victory over Lt. Gov. David Dewhurst, said he expected a better showing for Obama after the Democratic National Convention.

The three-day telephone poll of likely Texas voters was conducted Sunday through Tuesday, plenty of time for any post-convention glow to reach the electorate.

“I thought the Obama number would be a little bit better,” Perkins said. “It wasn’t there. It’s kind of lining up to what 2008 did — if not better — for McCain.” McCain beat Obama 55-44 in Texas in 2008.

The survey showed Romney with 32 percent of the Hispanic vote, which mirrors national Latino numbers for the Republican candidate in the wake of his party’s convention in Tampa, according to a recent poll. Romney got only 6 percent of the African American vote in Texas, compared with 90 percent who favor Obama.

But Romney’s lead over Obama among white voters in Texas is nothing short of overwhelming — 77-17 percent in the survey — which helps to explain why Republican candidates are maintaining their electoral advantage here even as the minority population explodes.

The full poll data can be found here. This is the first general election poll result from this outfit that I’ve seen, so I don’t know what their overall track record looks like. I will say that I don’t have any particular criticisms of their methodology. Their partisan and racial splits look reasonable – if anything, they slightly oversampled Latinos, at 26% of the total – and their sample voted for McCain over Obama by 55-44, also perfectly reasonable. The difference maker isn’t so much the white vote as it is the change from the 2008 vote. Page 4 of that crosstabs file tells the story. Of 552 McCain voters, 511 were voting for Romney, 20 for Obama, and 21 were neither. Of 442 people who said they voted for Obama in 2008, 380 said they were voting for him again, but 40 were voting for Romney, and 22 were neither. It’s those 40 switchers that depress Obama’s numbers; if he had the same 20 lost voters as Romney had, the result would be 53-42 instead.

So the question is whether this is a fluke or a real thing. To take a genuine stab at answering that we’d need – you guessed it – more data. I can tell you that in that May UT/Texas Trib poll, Romney led Obama among white voters by a 65-24 margin, and in that April PPP poll, he led 61-33 among whites (scroll to page 19), which was the best showing among Republican candidates (Newt Gingrich, Rick Santorum, and Ron Paul were also tested) and was in line with Obama’s 33-62 approval rating among whites. Is it possible that Obama’s support among Anglo voters could have collapsed that much since then? Sure. Is it likely? I’m always suspicious of results that stand out that much from others in the absence of an easily identifiable cause. I don’t reject this out of hand, but I am dubious.

Another way to look at this is to observe that if Romney were leading among white voters by a margin of 65-29, which is in between the UT/TT and PPP results, this race becomes a near dead heat: Romney drops to 48.4%, versus Obama’s 46.9%. Imagine the headline with that result. Even if he wins whites in this sample by 70-24, his lead is a mere 51.2-44.0, which is almost exactly the same margin that PPP found back in April. That would make the remarkably large Latino share of their sample all the more vital. If Republicans really do have to run up the score that much among white voters, you have to wonder how sustainable their edge will be.

What tf this result is indeed accurate? Does that portend doom for downballot Democrats? I’m sure it’s not good for Paul Sadler and the other statewides in that case – I checked, they either didn’t ask about the Senate race or they did not include those results if they did ask – but beyond that it’s not clear. It seems to me that the two most likely parts of the state where Obama might have lost white voters are the rural areas, and some of the suburbs. In the rural areas, Obama generally underperformed the rest of the Democratic ticket in 2008. The easiest way to see this is to scroll through the 2008 Senate results by district. If Obama has fallen further among these voters – there’s only so much more he can go down – then he probably is dragging other Democrats with him, but outside of the statewides and maybe Nick Lampson, there’s hardly anyone else for him to affect. In the suburbs – and here I mostly speak of Collin, Denton, and Williamson Counties – Obama ran ahead of the other Democrats. If he’s lost support here, he’s lost it among people who mostly voted Republican otherwise. One place where that could have an effect is in Wiliamson County’s HD136. In 2008, Diana Maldonado also ran ahead of other Dems as she scored a historic win in HD52 (see page 41 here). As we saw, Obama did better in the new HD136 than other Dems in 2008. It will undoubtedly be to Matt Stillwell’s advantage if Obama hasn’t lost his touch there.

Anyway, as noted it’s one result, and it would be nice for there to be something contemporary to which to compare it. A few other observations from the poll:

– The poll has a breakdown by voting history, and somewhat unexpectedly Obama does better among those who say they have voted in all of the last four elections than the other three subgroups, trailing by a 54-43 margin.

– Going by age groups, Obama does best among those aged 55-64, trailing 52-46. He wins among all women 55 and older, 51-46. He does worst among voters 18-34, losing them 58-35. Color me dubious of that one as well.

– Geographically, Obama wins “Austin” 58-40, loses “San Antonio” 60-38, loses “Houston” 51-42, and loses “Forth Worth” 52-44. I put them in quotes because these are clearly shorthand for the greater metro regions of each – “Houston” accounts for 23% of the sample, which would be a vast overstatement otherwise, as Harris County accounted for less than 15% of the total vote in 2008. “Fort Worth” was 30% of the sample, so this is clearly the entire Metroplex. “Austin” and “San Antonio” were an identical 80 voters each, or a smidge less than 8% of the sample each. Obama carried Bexar County in 2008, so I’ll chalk that up to small sample size weirdness.

Finally, on a tangential note, on the same day this came out I received a campaign email from Paul Sadler announcing that a “new poll” showed a “path to victory” against Ted Cruz. This was a campaign fundraising email, not a press release, so I have no useful numbers to share, but the clear message was that Sadler was competitive among voters who heard his message. Of course, the problem all along is how to get that message out to the voters. You can help by attending our fundraiser on Monday the 24th. It’s big hill to climb but there’s no reason not to try.

Related Posts:

• First pass at analyzing the 2012 results
• Precinct analysis: Comparing to 2012 and 2016
• UT/TT poll: We need more context