Political polls should never be trusted again
By Matthew Gagnon
Prior to Election Day 2016, predictions of an impending Hillary Clinton victory were nearly universal, and those predictions were built on the preponderance of publicly available polling.
In Wisconsin, the RealClearPolitics aggregate polling advantage for Clinton was 6.5 percent. No public poll conducted in Wisconsin in 2016 showed Donald Trump with a lead.
In Michigan, at least 30 consecutive polls conducted by a dozen different pollsters showed Clinton with a comfortable lead of anywhere between 3 to 13 points, with four polls in October showing her up double digits.
In Pennsylvania it was much the same. In July, a Suffolk University poll gave Clinton a 9-point lead on Trump. From the time of that poll until the end of the race, there were at least 43 different polls conducted, with only one showing Trump with a lead.
As we all know, in the end all three states went to Trump.
The polling disaster was more substantial than just these examples. Polling in states like Iowa did show a Trump lead, but his margin was six points lower than it ended up being. This was also repeated in Ohio, Minnesota, North Carolina, and right here in Maine’s Second Congressional District, where polling showed an evenly contested race. It ended up as a 10-point blowout for Trump.
After the failure, pollsters all over the country swore that they were going to be taking a look at what went wrong, and adjusting their methodology to ensure that similar mistakes didn’t happen again. With those adjustments, they were confident that a repeat of the 2016 polling disaster would not happen again.
Well, it did. Polls were so bad this year that David Graham of The Atlantic described the top to bottom failure as a “catastrophe” and said that we are living through a “polling crisis.”
Of course, in some important areas they got the result right, even if the margins were hilariously wrong. In the popular vote, for instance, pollsters had been telling us for weeks that Joe Biden was in for a historic landslide. A poll by The Economist/YouGov showed that Biden was up by 10 points the day before the election, while a Reuters/Ipsos poll released the same day showed Biden up by 7. A day earlier, CNBC/Change Research showed the margin at 10, while Quinnipiac had Biden by 11, with Trump only pulling in 39 percent of the vote.
On Election Day, Trump received roughly 72.2 million votes, pulled in 47.4 percent of the vote, and only lost to Biden by about 3 percentage points in the popular vote.
Here in Maine, polling was particularly heinous. Nearly every single poll conducted in the Second District showed Biden with a comfortable lead on Trump, despite the fact that Trump ended up winning there by more than 7 points. Sen. Susan Collins never led a single poll in her race against Sara Gideon throughout the entire race, and on the eve of the election the three most recent polls had shown her down by 7 points, 4 points and 6 points. She ended up winning by nearly 9 points, which is a shocking double-digit error.
And yet, despite openly wondering whether or not they should even care about them, the media will continue to not only report on polls as though they are accurate representations of public opinion, but will then base their underlying assumptions about a campaign and where the candidates stand on them.
Consider the Collins race. We were told for nearly a year and a half that Gideon was winning the race comfortably, and that Collins was struggling and seemed unable to convince the voters she deserved another term.
This impacts volunteer enthusiasm, political fundraising, and most importantly it influences political media coverage from writers to pundits, all of whom give their supposedly informed take on the realities of political races.
But was it ever true in this case? Given the fact that the polls were almost exactly the same right up to Election Day, it is likely that Collins had a lead the entire time.
The media needs to take a leadership role and stop reporting on public polls, or at the very least stop treating them seriously. But any hopes I have for that will evaporate into dust as soon as political amnesia clouds the minds of reporters and the public.
In the end, neither the media nor the public can help themselves. Polls are irresistible catnip to the politically interested. Pollsters will keep polling, the media keep writing about them, and we will keep believing them.
Matthew Gagnon of Yarmouth is the chief executive officer of the Maine Policy Institute, a free market policy think tank based in Portland. A Hampden native, he previously served as a senior strategist for the Republican Governors Association in Washington, D.C.