The Polling Center: Our Inaugural Election Survey

When we piloted this project using matched random sampling to poll Texans via the Internet last year, our goal was to overcome the growing limitations of phone polling while still getting statistically sound estimates of public opinion in Texas. Both the body of data we've built using this methodology over the last eighteen months, as well as comparisons with some other polling using different methods, tell us we're getting accurate snapshots of the Lone Star state. We'll have more to say about this after we've taken some time to look at all of our results longitudinally, both here and at the Texas Politics website, where we have archived (and will continue to archive) our survey data sets.

The results of the first University of Texas/Texas Tribune Poll, which was in the field from October 20-27 and sampled 800 Texans who identified themselves as registered voters, shows Texas slowly turning their attention to the 2010 elections. Perhaps more to the point, they have become extremely skeptical about the direction of the federal government. Today we’ll focus on the election match ups and what they tell us about the state of play a little less than six months out from the March primaries. (See the full writeup here.)

Republicans appear to be getting more engaged with the marquee primary race for governor between Rick Perry and KBH. Perry leads by 12 points (42-30), just outside the margin of error of 5.14%. More significant than the gap between the candidates, which has swung back and forth over the past nine months, is the drop in the number of undecided voters. At 18%, it has dropped ten points since our June survey. Debra Medina, seeking to carry the mantle of the Tea Partiers, runs a distant third with 7%.

The Democratic race continues to be marked by a lack of engagement and enthusiasm for the top contests. More than half of those planning on voting in the Democratic primary haven't settled on a choice. Kinky Friedman's first place finish with 19% probably testifies to his name recognition in a low salience race. Tom Schieffer is currently pulling 10%, which is good enough for second-place, while no other Democratic contenders break double digits. Are these results about the weakness of these candidates? The lack of interest in what the Democratic Party is offering more generally? The fact that the Republican contest is garnering more media attention? None of these possibilities are good news for Democrats.

Hutchison’s continued occupancy of her Senate seat, without any assured exit plan, means that the non-race race for her seat has yet to draw voters’ attention — or at least their commitment to any of the potential candidates. As in our June Survey, the number of undecided voters is still large. Houston Mayor Bill White appears to have made some headway — he trailed John Sharp and David Dewhurst in June, and now runs neck and neck with Dewhurst at 13% (Sharp garners 10% this time around). All three well are within the margin of error here, and nobody is really making the needle jump. The other candidates all remain in low single figures. The small number of observations in the results for each candidate don’t give us much to hang any reliable candidate comparisons on, but the results tentatively suggest that White is doing better with weak and leaning Democrats and independents, while Sharp and White are splitting the strong Democrats.

The pundits and media mavens who have been saying for months — well, years — that Hutchison would be a stronger general election candidate will find support in some of these numbers. In a series of hypothetical match-ups against the prominent Democratic “front runners,” Hutchison runs stronger than Perry across the board: nine points stronger against Tom Schieffer, five points stronger against Kinky Friedman, and (in a fantasy league grudge re-match) seven points stronger against former Travis County Attorney Ronnie Earle. Both candidates, however, register double-figure advantages in all of these hypothetical head to heads.

Perhaps the most interesting results of these match ups, though, are the items in which we matched both the Governor and the Senator with an unnamed “Democratic nominee.” Hutchison topped a generic Democrat by 11 points (36-25), while Perry edged the unnamed D by only 1 point. To some degree, respondents may well be attempting to convey that they are maintaining an open mind at this point, and in all of these hypothetical races more than a third of the respondents either didn’t know or opted for an unknown third party candidate. But the gap between Perry and Hutchison in the generic match up has to get your attention — particularly when the governor’s approval rating remains at 36%, in the ballpark of his 39% approval in our February poll, his 42% in June, and 38% a year ago. It may be worth noting, however, that those strongly disapproving of Perry jumped 9 points since June, from 17 to 26. His overall disapproval also increased from 32 to 44 — on balance, a negative approval rate in the most recent survey.

Given all this, it is hard not to have the impression that Perry is in the position of at the very least holding his own in a close primary contest with a well-funded opponent, even while his job ratings have been at best mediocre and are currently negative. He appears focused on the immediate task of winning the primary while not worrying too much about his opposition in the general election. Conversely, while the generic ballot shows an electorate willing to entertain a Democratic alternative, none of the D alternatives seem to be fitting the bill.

Speaking of the Democrats: the only good news for Democrats here (and it is marginal) is in the generic legislative match ups. In questions asking which party voters expected to vote for in Congressional and state legislative elections, Republicans enjoyed a nine-point advantage in the Congressional match up, but only a six-point lead in state legislative races. That this passes for good news is commentary on the tough sledding for Democrats, even if it is can be seen as a sign of continuing improvement for their position in legislative races in the last few cycles.

These and other approval numbers in the survey (e.g. Obama — 41% approve-52% disapprove, US Congress — 71% disapprove, including 49% strongly disapprove ), as well as some of the policy results, give us some powerful indicators of just how much the dominant political coalition in Texas, however strained by the pitched Perry-Hutchison contest, has emerged in hostile opposition to the dominant national political coalition. We’ll go into that when we look at some of the policy numbers tomorrow.

Back to methodology in the poll for a beat. There is a lot of interest in our use of the internet for polling. There is a deep discussion of our method in the attached methodology section, and similar discussions for all the polls in the polling section at the Texas Politics website. For those who want to dig still deeper into the underlying statistical methods, the founder of the polling firm we use, Doug Rivers, has been a central figure in developing the matched random sampling methods for use over the internet, and weighed in on some of the issues being discussed in this post at Pollster.com in September. If you’re interested, you can trace the discussion backward and forward from this post, and get a detailed explanation of why the matched random sampling method is different from opt-in polls.

Adam Myers and Jerod Patterson are graduate students in the Government Department at the University of Texas at Austin, and assisted in the analysis on this poll. Since the summer of 2008, they have been research assistants for the UT-Austin Texas Statewide Poll.