Emerson’s weird polls

It’s a poll, so we do the thing.

Joe Biden

A new poll has former Vice President Joe Biden leading Beto O’Rourke in the Texas presidential primary and toppling Donald Trump in a head-to-head showdown.

The survey, conducted by Emerson College for The Dallas Morning News, signals that even with two favorite sons in race, Lone Star State voters want a familiar face as their nominee.

Vermont Sen. Bernie Sanders, the 2016 runner-up to Hillary Clinton for the party’s nomination, was third with 16% and the only other Democrat beating Trump in the general election.

The poll also projects a wide-open Democratic primary race for the Senate seat held by longtime incumbent John Cornyn. At 19%, “someone else” is leading the field, a blow to former Army helicopter pilot MJ Hegar, who’s been campaigning for most of the year.

That “someone else” is leading the entire field is an oddity, but reflects the complexity of the primary race and the conundrum felt by many Democrats.

Hegar was the choice of 10% of those polled, followed by state Sen. Royce West at 8%, former U.S. Rep. Chris Bell at 7% and Houston City Council member Amanda Edwards at 5%. A whopping 51% of respondents were unsure.

West, Bell and Edwards are all relatively new to the race.

“I wouldn’t be surprised to see other people jump into the race,” said Spencer Kimball, the Emerson College polling director. “It’s just that wide open.”

The news is not great for Cornyn, the powerful incumbent who’s held the seat since 2003. Only 37% approved of his job performance, while 31% disapproved. The polls found that 33% of Texans were neutral or had no opinion.

For whatever the reason, the story only includes the head-to-head results in a non-embeddable graphic, so I will reproduce it here:


Candidate   Pct   Trump
=======================
Biden       51%     49%
Bernie      51%     49%
O’Rourke    48%     52%
Buttigieg   48%     52%
Warren      48%     52%
Castro      47%     53%

The poll is of 1,033 registered voters, with a 3% margin of error. They use a combination of automated calls to landlines and an online panel, as described here. You can find the crosstabs here, in a downloadable spreadsheet. They really didn’t want to make this easily to summarize, did they? The head-to-head numbers are very similar to the ones from their April poll, and are not far off from the Quinnipiac poll from June; the UT/Trib poll from June didn’t include two-candidate matchups.

I find the Emerson numbers dicey because I just don’t trust polls where the responses add up to one hundred percent. I guarantee you, there are “don’t know” and “someone else” responses in there, but their questions (scroll down past the disclosure stuff) do not allow for those answers. The crosstabs show that everyone surveyed picked someone, but if you have no choice but to give an answer, I don’t know how much I trust that answer. I’m much more comfortable with a poll that allows for “someone else” and “don’t know”. Emerson has a B+ rating from FiveThirtyEight, but I remain skeptical.

I don’t much care for Spencer Kimball’ analysis of the Senate race, either. MJ Hegar has been in the Senate race for ten weeks, not “most of the year”. She did say she was considering a run for Senate in February, but wasn’t raising any money or doing any campaigning until late April. All the other candidates have gotten in more recently. As I’ve noted before, Beto was still polling in the “majority of people don’t know who he is” area right up to the March 2018 primary. It’s going to take time – and money – for the people to know who the candidates are.

Also, too, the field for Senate is highly unlikely to get much bigger. There’s one potential new candidate out there, though nearly a month after that story I haven’t heard much about her. It’s already later than you think in the cycle, and it’s not going to get any easier to start fundraising and traveling the state to meet interest groups and primary voters. And as I’ve noted before, the fields for all of the Congressional races of interest in 2018 were basically set by this time two years ago. Each of the four top tier candidates entered the race only after some period of weeks or months of speculation, expressions of interest, exploration, and so forth. The only non-candidate out there right now with any association to the race is Cristina Tzintzún Ramirez, and she only gets mentioned occasionally. If the primary field isn’t set, it’s close.

Anyway. I’m still waiting for some head-to-head Senate polling. Even if the candidates are basically unknowns at this point, a “Cornyn versus generic Dem” question still has value. Maybe the Trib will give me that in their October poll. In the meantime, enjoy the results we do have, for whatever they are worth.

Related Posts:

This entry was posted in Election 2020, The making of the President and tagged , , , , , , , , , , , , , , , , , , . Bookmark the permalink.

3 Responses to Emerson’s weird polls

  1. Easter says:

    I found these polls have more questionable results: “They use a combination of automated calls to landlines and an online panel.” Landlines are badly skewed conservative, rural, and older and adding an online panel typically doesn’t correct all of that as well as those respondents being typically being more well-informed and partisan than the voting public. Of course, it is just really, really hard to do polling now.

  2. Pingback: Five for Senate – Off the Kuff

  3. Pingback: Back to the Beto question – Off the Kuff

Comments are closed.