Hubbell Leads Reynolds in New Iowa Gubernatorial Poll

Emerson College released an e-poll of 1,000 Iowans conducted from September 6-8. It does not provide good news for Republicans.

No candidate polled received a majority support. In the Iowa Gubernatorial poll, neither Fred Hubbell, the Democratic challenger, or Republican Governor Kim Reynolds reach 40 percent support among Iowans.

Hubbell leads Reynolds 35.9 percent to 31.3 percent. 6.9 percent of voters said they planned to vote for someone else. The two additional candidates on the ballot are Libertarian candidate Jake Porter and Gary Seigwarth of the Clean Water Party of Iowa. 

The key number here 25.9 percent of Iowans state they are undecided.

Hubbell had a 40 percent favorable/ 22 percent unfavorable rating among Iowa voters. Reynolds had a 36 percent favorable rating/ 37 percent unfavorable rating.

The margin of error for the gubernatorial poll is +/- 3.2 percent. 

They also polled for the Congressional races with every poll having a margin of error of +/- 6.4 percent with the exception of Iowa’s 4th Congressional District having an MOE of +/- 6.5 percent.  Currently, Congressman Rod Blum (R-Iowa) is the only incumbent who trails, but is within their polling’s margin of error. 

First District:  Abby Finkenauer (D) 42.5 percent, Blum (R) 38.4 percent, Someone Else 7.3 percent, 11.8 percent undecided.  (Sample Size: 250 voters)

Second District: Congressman Dave Loebsack (D) 45.4 percent, Christopher Peters (R) 20.7 percent, Someone Else 5.9 percent, 28 percent undecided (Sample Size: 250)

Third District: Congressman David Young (R) 46.5 percent, Cindy Axne 30.5 percent, Someone Else 7.7 percent, 15.2 percent undecided (Sample Size: 260)

Fourth District: Congressman Steve King (R) 41.4 percent, J.D. Scholten (D) 30.5 percent, Someone Else 12.2 percent, 15.8 percent undecided (Sample Size: 240)

The top lines were interesting.

Voter breakdown by party is interesting. I think it’s appropriately weighted regarding voter registration: 32.3 percent Republicans, 31.4 percent Democrat, and 36.3 percent independents. Turnout, as always, is the big question. In 2016, according to exit polls, independents accounted to 35 percent polled, Republicans 34 percent, and Democrats 31 percent. In 2014, according to exit polling, 37 percent identified as Republican, 32 percent as Democrats, and 31 percent as independents. In presidential election years independents have a greater turnout than they do for midterm elections.

The weighting by age group, I assume, is based upon the latest census data. 18-29 year-olds were 13.9 percent, 30-49 year-olds 28 percent, 50-64 year-olds 20.1 percent, and 65-years-old and older 38 percent. Regarding 2014 and 2016 turnout, more 50-65 year-olds voted than voters 65-years-of-age or older. The 18-29-year-old segment is closer to 2014 turnout (13 percent) than 2016 turnout (18 percent) based on exit polling. I’m doubtful any of this impacts the accuracy of the polling.

The top line data that caught my eye which could impact the accuracy of this poll was the break down of Trump voters to Clinton voters. 39.8 percent voted for Donald Trump while 38.6 percent voted for Hillary Clinton. Iowa was not remotely this close in 2016. 51.8 percent of Iowans voted for Trump while 42.2 percent voted for Clinton. 

Again, turnout could render the Trump to Clinton ration moot, but I think it is worth noting.

Also, 50 percent of their respondents were polled through Interactive Voice Response (IVR) calls to landlines, and 50 percent were surveyed online. Many pollsters have mixed online polling because of the decreasing number of landlines, as well as, difficulty reaching people on mobile phones (most of us screen numbers we don’t recognize). This online polling is not of the unscientific straw poll variety that some websites conduct, but they have their challenges as Scientific America pointed out in 2016:

They typically recruit by advertising on popular websites, so people choose whether to participate, and that means that there might be a built-in bias in their samples. Pollsters don’t exactly know who is missing from the poll, and it’s harder to estimate the reliability of the final poll numbers.

FiveThirtyEight notes that live calling has proven to be the most accurate since live callers can call cell phones where IVR calls can only call landlines. That said, IVR is still more reliable than online polling and hybrid polling. Nate Silver writes in his review of pollsters for 2018:

The clearest trends are that telephone polls — including both live caller and IVR polls — have outperformed online polls in recent elections and that polls using mixed or hybrid methods haven’t performed that well.

The relatively strong performance of IVR polls is surprising, considering that automated polls are not supposed to call cellphones and that more than half of U.S. households are now cellphone-only. It ought to be difficult to conduct a representative survey given that constraint.

We’ve sometimes seen the claim that IVR polls are more accurate because people are more honest about expressing support for “politically incorrect” candidates such as Trump when there isn’t another human being on the other end of the phone. This feeling of greater anonymity would presumably also apply to online polls, however, and online polls have not been very accurate lately (and they tended to underestimate Trump in 2016).

Silver at FiveThirtyEight said Emerson had an average error of 4.1 percent over 51 polls taken of the 2016 presidential election and beyond. They also noted that their polling tends to have a bias toward Democrats of 1.8 percent. These numbers are for polling using IVR methodology, not hybrid polling. With IVR polling they gave Emerson a B+ rating.

Emerson had notable accuracy problems during the 2016 presidential elections, however. 

I have to wonder just how accurate this poll actually is.

That aside, I think it’s accurate to say the gubernatorial race is in a dead heat. Frankly, the Congressional race polling is junk because the sample size is far too small and probably not appropriately weighted.