Now that the election is over, it’s clear that some pollsters were mistaken with their forecasts that Donald Trump was going to fall a few percentage points short of the White House..
University of South Carolina Dean of the College of Mass Communications and Information Studies Charles Bierbauer told South Carolina Radio Network that there was one poll that had Trump ahead all along, but many ignored it. “The one that people are pointing to is the LA Times/University of Southern California poll, which routinely, regularly, predictably, gave Trump a three or four point lead,” Bierbauer said. “It tended to be constant and therefore a useful marker.”
According to the website Politico, Trump supporters argued that even when some polls showed Trump’s strength, it was ignored or was explained away by the media and analysts.
Bierbauer said that Hillary Clinton could have taken a big hit near the end with all of her perceived baggage. “Hillary Clinton emails and the FBI letters and so you would try to ascertain whether these were making significant moves, incremental moves, whether they were sustained,” said Bierbauer.
Politico reported that Trump’s significant upset victories in Florida, North Carolina and Wisconsin came about when almost every public poll and private polls had Clinton ahead in those states. However, polling in the election’s final week did indicate Trump was pulling closer in the race.
Meanwhile, Winthrop Poll director Scott Huffmon believes national polling was not as far off as critics have suggested. Huffmon noted to WIS-TV that nationwide polls had Clinton ahead within the margin of error and she ended up winning the popular vote by less than one percentage point.
However, polling miscalculated the number of “likely” voters who would turn out in Rust Belt states like Michigan and Wisconsin, switching those historically-Democratic states into Trump victories.
Huffmon also emphasized that most polls use a 95 confidence level of scientific certainty, meaning the results are expected to be within the margin of error 19 times out of 20. He speculated 2016 could have been the odd time out.