After the defeat of their presidential candidate in 2016, Democrats know the perils of poll-induced over confidence.
This time around a majority of party activists and pollsters are hedging their predictions of a Joe Biden victory with a mandatory caveat that incumbent Republican President Donald Trump can still defy the odds and win.
Nevertheless, unless polls can be totally disregarded as a method of predicting elections, the current data shows that it is quite likely that the former vice-president will become the 46th president of the United States.
The polling data published so far was enough for the British magazine The New Statesman to effectively forecast a Biden win in brash language.
A headline at the news outlet read: “Why you should be confident that Joe Biden is going to win the US election.”
The article’s underlying argument was that opinion polls for the 2016 vote were not wrong by that much and that in 17 key states that eventually went to Trump, polls predicted an average Clinton lead of less than a percentage point - way within the margin of error.
In three of these states; Pennsylvania, Michigan and Wisconsin, Trump’s eventual victory was by a margin of less than a percent.
In contrast, Biden enjoys an average poll lead across these 17 states of 5.9 points - significantly higher than anything Clinton enjoyed.
Trump would therefore need to defy the polls by a much larger degree that he did against Clinton to pull off the upset.
According to polls commissioned by the New York Times and published on the last Sunday before the vote, Biden has sizable leads in four key states that went to Trump last time round.
Arizona, Florida, Pennsylvania, and Wisconsin, all went to Trump in 2016, but the Times poll put Biden ahead at six points, three points, six points and 11 points, respectively.
To cut to the point; the polls say that Biden will win.
The seeds of doubt
However, if memories of 2016 were not enough to turn the seed of doubt in the back of Democrat minds into a full grown tree, perhaps polling data out of Iowa might.
A poll commissioned by the Des Moines Register found that support for Trump in the state had surged in recent months.
One poll conducted in September found that Biden and Trump were tied on 47 points each but the October poll found Trump building a formidable lead of 48 points to 41.
According to Selzer and Company, the firm that carried out the polling: “The president is holding demographic groups that he won in Iowa four years ago, and that would give someone a certain level of comfort with their standing.”
If that trend holds across the nation, that spells trouble for Biden and Wednesday morning might bring about more than just flashbacks of the 2016 result.
The prospect of pollsters having to eat another serving of humble pie is exactly what one unorthodox analyst, who correctly predicted the 2016 result, is predicting.
As Politico points out, Robert Cahaly went against the grain and correctly predicted Trump wins in Florida, Michigan, North Carolina, Pennsylvania, and Wisconsin.
Cahaly says that there is a disparity between what people tell pollsters when they are surveyed and what happens when they enter the ballot box.
The phenomenon is known as the ‘shy Trump voter’; essentially a voter who is aware of the potential social consequences of admitting to supporting Trump, given the notoriety surrounding the president, so will not openly admit to backing the president except on the anonymity of the ballot paper.
Cahaly told Politico: “I’m finding that people are very hesitant [to share their preference for Trump], because now it’s not just being called “deplorable.” It’s people getting beat up for wearing the wrong hat, people getting harassed for having a sticker on their car. People just do not want to say anything.”
Pollsters like FiveThirtyEight’s Nate Silver concede that opinion polls may mask the true extent of Trump’s true voter base but disagree on the extent and whether it is because they are ‘shy’.
In an article on his website in September, Silver addressed the concern: “The good news for pollsters is that few, if any, voters are misleading them about their support for Trump. Yes, Morning Consult found small differences in support for Trump via household income and education, with wealthier and college-educated voters less likely to say they support Trump in a live-phone poll, but the shifts weren’t statistically meaningful.”
