Many UWM students were too young to vote in the 2016 election. Yet, the debacle of the polling results, prematurely calling the win for Sen. Hillary R. Clinton, caused such a stunning uproar throughout the nation when the win ended up going to Donald J. Trump that it is still a topic of discussion that new voters hear about in this year’s election.
The questions among citizens and news media were, “How could the polls be so off in their outcomes?” and “Can the polls be trusted?”
Charles Franklin, nationally recognized Marquette Law School pollster, government scholar and director of the Marquette Law School Poll, does not think voters should worry.
“We’ve done what we know how to do to improve the methodology,” Franklin said.
Franklin said in a PBS Wisconsin “Here and Now” video interview, “Oh, there’s a lot of doubt about it, and that especially blew up after 2016.” Franklin added, “As a political scientist, and a methodologist, I think there are clear reasons to say polling is not broken.”
“I do think there’s reason as a technician to say, polling’s doing OK,” said Franklin. “But I think the idea that polling, especially in 2016, was seen as so overwhelmingly showing Clinton winning – that people had no expectation that Trump would win. And when that happened, the polls took much of the blame for being surprised so much. And I think rightly so. Here in Wisconsin, Michigan and Pennsylvania, there was 134 polls in the fall of 2016. Only one of them in Michigan ever showed Trump ahead, and that was a couple of weeks before the election.”
Eric Wilkinson, visiting political philosophy assistant professor at UW-Milwaukee is a campus expert on polling.
“There are several reasons why the polls in the 2016 election were inaccurate,” said Wilkinson. “Traditional polling methods still widely used in 2016 relied on phoning participants, and this has become increasingly difficult and unreliable since fewer people own home phones, and people with cell phones are unlikely to answer a call from an unrecognized number.”
Franklin added that each candidate appeals to different types of voters, altering polling results.
“Trump reaches a set of voters and mobilizes them to vote, especially in 2016 that hadn’t been mobilized before,” Franklin said. “The Clinton campaign maybe underperformed in voter mobilization by famously not coming to Wisconsin that year for example.”
The distrust in media from Republican voters also swung the polls in favor of Clinton.
“Polls in 2016 also tended to underestimate support for the Republican nominee,” said Wilkinson. “This is partly because conservative voters who voted Republican were more likely to distrust the media in general and were therefore less likely to participate in polling… Additionally, some voters for the Republican nominee have said that they kept their preference to themselves to avoid the judgment of those around them.
“Since 2016, there have been significant changes in how most polls are conducted in response to their inaccuracy in past elections, which may compensate for some of these factors.”
One way the pollsters are increasing awareness in their reliability is their acceptance of room for error.
“I think you have to have a reasonable view of the margin of error in these polls, which is around 3 or 4 percent in most cases,” Franklin said. “Plus, there are other factors that affect accuracy – like how you answered the question and what not. I think we do tolerable well in being in that plus or minus 4 percent on most questions.”
Franklin also added, “So, if you’re thinking about, is the poll accurate to half a percent of the election… that’s an unattainable standard. Is it getting the winner, right? That’s a lot better question to ask.”
With election day in less than a week, Franklin and his cohorts will discover the reliability of their final poll.
“Going into the remainder of this year, I’m certainly hopeful that we’ll be close or under our average error,” said Franklin. “I certainly pray that we don’t have a repeat of the 2016 large error. Now… we’ll just see in November whether it really works out or not.”
Full interview: Charles Franklin on the Marquette Law School Poll’s accuracy | Here & Now: https://youtu.be/CtdTnZbRgZ4?si=Uo4BIgm2X7ZT3Mjb