Differing views of likely voters

First we had this.

Rep. Beto O’Rourke

U.S. Sen. Ted Cruz, R-Texas, leads his Democratic challenger Beto O’Rourke by 9 percentage points among likely voters, according to a new poll from Quinnipiac University.

Released Tuesday, the survey found Cruz with 54 percent support and O’Rourke, an El Paso congressman, with 45 percent. Only 1 percent of those polled were undecided.

“The Texas U.S. Senate race between Sen. Ted Cruz and Congressman Beto O’Rourke, and Democratic hopes for an upset win there, have boosted talk of a Senate takeover,” Peter Brown, assistant director of the Quinnipiac University Poll, said in a news release. “These numbers may calm that talk.”

It’s the first time Quinnipiac has released a likely voter survey in the Senate race. Quinnipiac previously polled registered voters three times, finding Cruz ahead by 6 points in August, 11 in May and 3 in April.

Quinnipiac also surveyed the governor’s race in the most recent poll and continued to find a much less competitive contest, with Republican Gov. Greg Abbott leading Democratic opponent Lupe Valdez by 19 points.

I started writing a post about how like everything else this is one result, the first one we’ve had of just likely voters, then I got distracted by all the hot takes about how this means The Race Is All Over And It Was Never Really Close and the shitshow in SD19, so I didn’t get it finished. And then I woke up the next morning and saw this.

U.S. Rep. Beto O’Rourke, D-El Paso, leads Republican incumbent Ted Cruz by 2 percentage points among likely voters, according to an Ipsos online poll released Wednesday in conjunction with Reuters and the University of Virginia. O’Rourke has been closing the gap over the last several months, but this is the first poll that puts him ahead of Cruz.

Forty-seven percent of likely voters told Reuters they would vote for O’Rourke, while 45 percent said they would cast their ballot for Cruz. Three percent said they would vote for “Other,” and 5 percent said “None.” The margin of error on that portion of the poll was 3.5 percentage points.

A Quinnipiac poll released Tuesday put Cruz 9 percentage points ahead of O’Rourke among likely voters. That poll was based on phone interviews, while the Ipsos poll used an online survey. But it’s trying to predict who will show up on Election Day that shifts the numbers, said Ipsos Vice President Chris Jackson.

Ipsos is trying to gauge political enthusiasm on each side, said Jackson. The poll asked respondents to estimate the likelihood that they’d vote in the midterm elections on a scale from one to 10. “More Democrats are registering at the highest part of the scale, at the 10, than the Republicans,” Jackson said. And that’s what’s interesting, he said, because Republicans usually have the momentum advantage in Texas.

“It demonstrates how Democrats are mobilized,” said Jackson. “This election is going to be really competitive and its going be very hard fought.”

[…]

The poll also questioned voters about the Texas gubernatorial election and found that Gov. Greg Abbott leads his Democratic challenger, Lupe Valdez, by 9 percentage points among likely voters.

Well, well, well. Look, this too is just one result and we can’t really judge either of them until we see enough polls to get a feel for where these fit in. That said, this is 1) the first poll result of any kind showing Beto in the lead, 2) the first poll of any kind in at least a decade showing a Democrat with as much as 47% of the vote, and 3) an extremely satisfying quick corrective to all those hot takes from Tuesday.

So what do we make of all this? I think the DMN has it right:

Polling experts have long warned against putting too much weight in any one survey, particularly since different polls can take widely different approaches. Texas is also a special case, since the GOP’s longstanding dominance there means pollsters don’t usually pay it much attention.

Nate Silver, editor-in-chief of Five Thirty Eight, on Wednesday pretty well summed up the dynamic.

“Texas is a tough state to poll (lots of new residents, low turnout among certain voting groups, may be hard to reach Spanish-speaking voters),” he wrote on Twitter. “It’s probably a healthy sign that we’re seeing some disagreement.”

A closer look at the two surveys perhaps further proves the point. Both polls looked at likely voters, which is a key distinction from earlier surveys. The feelings of likely voters are supposed to give a better representation of Election Day results than those of registered voters. But that approach also means some assumptions have been made on who is likely to vote.

The Quinnipiac poll surveyed 807 likely voters earlier this month, tallying a margin of error of plus-minus 4.1 percentage points. It was conducted using live interviews over landlines and cell phones, which many experts say is the best way to approach the task. It had a sample that featured 35 percent Republicans, 26 percent Democrats and 33 percent independents.

The Ipsos poll, done in partnership with Reuters and the University of Virginia, surveyed nearly 1,000 voters earlier this month, garnering an error margin of plus-minus 3.5 percentage points.

It was conducted by way of online surveys. Its sample is also much different, reporting a significantly lower number of independents. So its breakdown is 47 percent Republicans, 43 percent Democrats and nine percent independents.

So live call versus online, and self-reported engagement levels versus whatever formula Quinnipiac used (they did not elaborate on that). Ipsos wound up with a sample that was slightly less Republican, which is consistent with the thesis that Dems are more engaged than usual. Who’s “right” and who’s “wrong” at this point is impossible to tell, though we may get a better feel for that as voting draws nearer. For now, be aware of the differences in methodology and look for any trends that emerge there. In the meantime and even though it’s mixing apples (registered voters) and oranges (likely voters), I’m updating the now-16 poll average to reflect 46.69 for Cruz, and 41.38 for Beto. Until we can say definitively otherwise, this is still a very close race. Washington Monthly and the Current have more.

Related Posts:

This entry was posted in Election 2018 and tagged , , , , , , , , . Bookmark the permalink.

5 Responses to Differing views of likely voters

  1. Manny Barrera says:

    For nearly a year I have been stating that I believe that Beto will win. I still do.

    The first poll mentioned above, Quinnipiac, while is rated highly may have had too many strong Republicans on the poll. It had Trump about equally liked and disliked in Texas. Most all previous polls don’t indicate that Trump is thought that highly in Texas.

    I have nephews that have never voted that are going to vote for the first time because of Beto.

    But if Harris County can turn Blue that will be good enough for me.

    Not on topic but I keep reading by some that the Democrats are trending far left, that is BS from the Republicans and the Media that wants to sell stories. How many so called far left candidates have won and lost? Very few. Compare that to the crazy right wing Tea Party after the Obama election.

  2. Michelle says:

    I agree with your final paragraph completely Manny. People complaining on both sides (too many liberals- they can’t wait and not enough liberals- the party is controlling the votes) need to remember that different districts have different needs.

  3. C.L. says:

    I believe Beto can win as well, as long as he gets more votes than Cruz.

  4. Pingback: Two from PPP (RV): Cruz 48, O’Rourke 45, and Cruz 49, O’Rourke 46 – Off the Kuff

  5. Pingback: What about Neal? – Off the Kuff

Comments are closed.