Berwood Yost talks political polling, midterm elections during Common Hour

By Amanda Leonard || Contributing Writer

Photo courtesy of fandm.edu

Is political polling an accurate way to predict the outcomes of political elections? Were the 2016 Presidential election polls truly accurate? And as the title of this week’s Common Hour said, “Should We Trust What The Polls Are Telling Us About the 2018 Mid-term Elections?”

According to Berwood Yost, director of F&M’s Center for Opinion Research and the featured speaker, “The answer is yes and no.”

Yost first emphasized that the 2016 election was a particularly complex one, with 1/6 of voters having an unfavorable opinion of both candidates and a historically large number of undecided and third-party voters.

Yet, he asserted that the national polls were in fact accurate, and that, “contrary to the beliefs of partisans everywhere, the [national] polls did not favor one party over another.”

He then contrasted this conclusion with that of the state polls. These showed that voters’ preferences shifted throughout the campaign, with 1/7 of Pennsylvania voters making their final choice in the last week and a considerable amount switching in favor of Trump during that period, contradicting the prediction of Clinton’s win.

In understanding this upcoming election year, Yost said, “The basis of understanding voter outcomes lies in voters’ approval of the current president.” With President Trump’s historically low job approval ratings, Democrats will likely represent a larger majority of the pool of voters due to their enthusiasm for change, which gives them a clear advantage. In Pennsylvania, 60% of individuals planning to vote this year are Democrats.

Voter turnout was also discussed as a factor in skewed poll results. Midterm elections generally receive a much lower turnout than presidential elections: according to polls only about half of voters in Pennsylvania are planning to vote. In turn, polling sample sizes are reduced.

In addition, Yost explained how the particular method of polling can impact its results. For example, the Quinnipiac poll collects their data by calling landlines randomly. However, the demographic and the political opinions of those who own landlines and those who do not varies greatly.

A common theme throughout the whole talk was the “polarization” that has clawed its way through the political climate in the last two years. In one of Yost’s polls, a staggering 11% of pollers said that their reason for supporting one candidate was because of their distaste for the other, and this was an open-ended question. Democrats have an advantage on the generic ballot, meaning that many Democrats will be voting solely for democrats. And in regard to ongoing bias in the media, Yost claimed, “There is no doubt that the media outlet that you use affects your opinion on this election.”

To wrap up, Yost admitted that there is significant evidence for an incoming “blue wave.”  Yet there are many inconstant variables in political polling and several outside factors that could make poll results differ slightly, or even largely, from reality.

First-year Amanda Leonard is a Contributing Writer. Her email is aleonar1@fandm.edu.

print

Leave a Reply