Have presidential polls changed since 2016? Pollsters say they’ve learned from mistakes
Sep 28, 2020, 5:16 AM | Updated: 5:10 pm
(Associated Press)
SALT LAKE CITY – Recent national polls say Joe Biden has a lead over incumbent President Donald Trump, with some presidential polls showing the former vice president’s lead is in the double digits. However, with the polls appearing to be so wrong in the last presidential election, can we trust them? Some say pollsters have learned a lot from the mistakes of 2016.
In the weeks before Election Day 2016, the vast majority of polls showed Hillary Clinton had a comfortable lead over Donald Trump, in some cases, she was up by double digits. So, when Donald Trump won the Electoral College, it caught many people off guard.
Some analysts say it wasn’t really the pollsters who were way off as much as it was the pundits on television seeing what they wanted to see. One ABC News analyst was asked, “What would Donald Trump have to do to turn things around?” He responded, “Prayer.”
Pollster Scott Rasmussen says, “The analysis was horrible.”
Still, Rasmussen says there were mistakes that pollsters made, also. For instance, he says there were some states where there wasn’t a lot of polling done, and those states ended up surprising people.
“Everyone expected that they knew how they would turn out,” Rasmussen says. “Why would you poll in Michigan if Michigan always goes for a Democrat?”
Plus, Rasmussen says they’ve improved how they how well they mix the people the poll. In 2016, he says white voters without a college degree were far more likely to vote for Donald Trump, but people with a college degree are far more likely to take a survey than whose who don’t.
Rasmussen’s most recent presidential polls show Biden is ahead by six points, but he projected another poll with more Republican participation, and Biden was still ahead by three.
“Which is exactly where we were in 2016, and Donald Trump could easily win the Electoral College, then,” he says.
There are some big differences in polling numbers if you break down how people vote. For instance, among those who plan to vote in-person, Rasmussen says the president is up by about 20 points.
He says, “Among people who say they’re going to vote by mail, Joe Biden is by about 50 points.”
He believes the real trick is to find likely voters, and that is much easier said than done.
“Anybody who, at this point, claims to know more precisely what the turnout is going to be is either lying to you or fooling themselves. I mean, we just don’t know,” according to Rasmussen.
Pollsters have made a lot of other changes, also. Hinckley Institute of Politics Director Jason Perry says they’re calling more cell phones than home phones and many are transition to online polling because those results may be more accurate. Perry says the more anonymous the poll feels, the more honest the respondent will be.
Perry says, “If someone calls [them] on the phone, they’re more likely to give you what you think you want to hear, even though it’s completely anonymous.”
Also, he says we should never gauge who would win the election based on what we see strictly on social media because those platforms feed you what you want to see.
“If one does not get out of that, one would believe that whatever I’m looking at is what everyone is looking at. That is clearly not the case. It wasn’t in 2016, and it’s even less so now,” Perry says.