While the polls in the 2016 US Presidential election were not accurate, pollsters have changed certain things for the 2020 elections, in the hopes of a more accurate outcome.

Doug Schwartz, Director of the Quinnipiac University Poll, was speaking with journalists during the US Foreign Press Center's Elections 2020 Virtual Reporting Tour.

"After the 2016 election, pollsters took a look at what happened as a whole and why some state polls underestimated Donald Trump's support. Overall, what they found was that despite the narrative, polls were largely accurate. For example, national polls showed Hillary Clinton winning the popular vote by three percentage points and she won by two. But there were polls in some key states that missed the mark. A major reason was that late deciders in those states wound up supporting Trump. Another reason was that some pollsters underestimated white voters without a college degree," he said.

"The main methodological change that some state pollsters had to make was that since 2016, they've now started to weight by education."

Weighting is a statistical adjustment that pollsters make to ensure that certain demographic groups are accurately represented in the survey based on their percentage of the population.

He said that the problem in 2016 was that in some key states, including Wisconsin and Michigan and Pennsylvania, some of the pollsters there did not weight their samples by education. "And so, what ended up happening is they overrepresented white voters with college educations and underrepresenting white voters without college educations. And the group, white voters without a college education, were a key part of Donald Trump's supporters. So that's why it's one of the reasons why some of those polls underestimated Donald Trump's support in the election. And my understanding is that some of those pollsters are now waiting by education to fix that."He believes that those polls that are now waiting their samples by education, who had not done that previously, will be more reliable."

"In terms of my overall trust, I do trust the major pollsters that use gold standard methods, and I think the public can trust those polls." If a poll is transparent about its methodology, uses live interviewers and call cell phones, industry insiders will typically call it a gold standard poll. "And there's a reason for that. Gold standard polls are typically the most accurate in elections. There are other methodological choices that a poll can make that can contribute to or take away from their accuracy."