x
Breaking News
More () »

What went right with polling of the presidential election in Washington state

KING 5 worked with SurveyUSA for the 2020 election polls, which reflected the end results within several points. The pollster explains the methodology.

The KING5/SurveyUSA statewide pre-election poll was on target, and could serve as a model for others around the country.

“We're enormously proud, of course, of the work that we do in the state of Washington,” said Jay Leve, CEO of SurveyUSA, who did a series of polls for KING5.

The polling appears to have exactly matched the final results in multiple state races, including the presidential race, along with the gubernatorial and lieutenant governor contests as well.

Back in October, the KING5 poll showed Joe Biden with a 55-34 lead over Donald Trump. The current margin, as of today, was 58-39.

Gov. Jay Inslee had a 14 point spread in the same poll, and currently leads by 14 points. The poll showed Congressman Denny Heck with a 13 point lead in the race for lieutenant governor, and he is up by 14 now.

Other poll results are holding up. The attorney general, secretary of state, and Referendum 90 poll results all matched the actual results as well, within a couple points.

Leve says his company changed its methodology years ago, as people’s phone habits changed.

“We observed over the last 12 years, which is when we began our transition away from 100% telephone polling,” he said. “We identified I think five, six, seven years ago, a situation where internet polls were consistently ideologically, one sided, they kept favoring the same group.”

Leve said the company made adjustments to correct for that.

“We made an adjustment to say we need to go out of our way to find this particular type of individual who is going to be under counted. Now we did this long before we knew President Trump was going to run for president. So it's not a unique Trump phenomenon. But there are certain types of individuals...  who simply are less likely to participate in a survey because of their lifestyle or whatever," he said. "So we made a concerted effort to include those individuals in our sample. And while I cannot say that we've solved the riddle, what I can say is that we've brought the needle pretty close to back to magnetic north.”

There has been significant attention on polling across the country, which showed Biden with a roughly 8 to 10 point lead on Trump before the election, and one in particular that showed Biden had a 17 point lead with voters in Wisconsin. That state was decided by just a percentage point.

“That was conducted of regular telephones, where you ring the home of a respondent in Wisconsin,” Leve said. "It is in fact a voice phone call. So that means that it is by definition, what we call 'barge in' polling. Barge in polling means it's at the pollster's convenience, it's not at the respondent's convenience.” 

That can lead to skewed results, he says.

“We have to believe that the two or three people who complete the survey out of every hundred numbers we dial are just like the 98 people we didn't reach, either because they slammed down the phone, or they never lifted the phone, they just let it roll to voicemail,” Leve said.

What about the concept, floated by some pundits, that Trump voters quietly show their support and are reluctant to share their views with pollsters? 

“I would much more precisely say it was the defiant Trump voter or the combative Trump voter,” Leve said. “If you have the President of the United States, deriding fake polls over and over again, then among his voters, there is a reluctance to participate in what may they may perceive as a fake process or a doomed process or, you know, that they're somehow going against the President's wishes.”

There is one thing for sure, says Leve.

“There have been some tweets that suggest that polls are some form of voter suppression. And I want to be declarative: that is false.”

Before You Leave, Check This Out