MANILA, Philippines — How come some of the major American pollsters got the outcome of the 2020 US presidential election wrong?
According to an expert pollster, this could be attributed to difficulties with telephone polling.
While some polls projected an overwhelming victory for now President-elect Joe Biden, the race was closer than what the polls suggested before the election. Biden eventually won the race with more than 279 electoral votes as of Sunday, defeating President Donald Trump after four days of vote counting.
John Zogby, a renowned pollster, author and founder of Zogby Strategies, said he believed the polls got it wrong because they oversampled Democrat voters against Republicans.
“I saw a number of polls over the months and weeks and days leading up to the election that were clearly oversampling Democrats over Republicans,” Zogby told participants of the US Foreign Press Center Virtual Reporting Tour, which included the Inquirer, in a recent briefing on the analysis of the election.
But he said he did not think “there was anything planned going on” for this development.
Media, university polling
He said some of the polling made by media and university networks oversampled Democrats by 9, 10 and 11 percentage points.
Zogby said his polling firm got the numbers right on the voters who cast their ballot on Election Day itself, the Black turnout, and more men and women voting for Biden.
He explained that most of these polling errors could be “attributed to the fact that we’re seeing the end of telephone polling.” Telephone polling is one of the gold standard methods used by US pollsters.
“There are very few Americans that have landlines anymore, and are schooled now to not answer landlines,” Zogby said.
Likewise, he described as “dismal” the way people responded to their queries on cell phones. Pollsters are having trouble asking many questions to people they end up talking to by phone, he said.
“You just can’t do that when someone is walking down the street or in the men’s room somewhere, and just happens to make the mistake of answering the phone,” he added.
Online sampling
But Zogby said pollsters had a better chance of reaching people and getting better responses through online sampling. Distribution of sampling is also better through online, he said.
“We’re getting better levels of responses by race and by younger people who will not answer a telephone,” Zogby said of online sampling.
He said the polling industry was now “at a crossroads” but added that these were “methodological problems, not ideological problems.’’
Some US pollsters also got wrong the outcome of the 2016 presidential race between Trump and Hillary Clinton.
Doug Schwartz, director and vice president of Quinnipiac University Poll, said this happened because some of these pollsters did not give weight to their samples’ education in some key states, including Wisconsin, Michigan and Pennsylvania.
Briefing journalists on the virtual reporting tour before the Nov. 3 polls, Schwartz said this resulted in the pollsters having overrepresented white voters with college educations and underrepresenting white voters without college educations.
He said it was the white voters without a college education who were key supporters of Trump.
Schwartz said this omission was one of the reasons some of the polls underestimated Trump’s support in the 2016 election.
These polls were now expected to be reliable because their weighing their samples by education, he said.
He said he still trusted major pollsters who use gold standard methods and that included being transparent, using live interviews, and calling by cell phones.
Among the polls he said he trusted were Pew Research Poll, ABC News, Washington Post, NBC News, Wall Street Journal, CNN and Fox News.
“As pollsters, nothing is more important to us than getting it right. Our reputations are on the line,” Schwartz said.