Are the polls skewed?
More and more Democrats appear to be worried that they are.
Republican presidential nominee Donald TrumpDonald TrumpNorth Korea conducts potential 6th missile test in a month Kemp leading Perdue in Georgia gubernatorial primary: poll US ranked 27th least corrupt country in the world MORE is within the margin of error of Democrat Hillary ClintonHillary Diane Rodham ClintonThe Armageddon elections to come Poll: Trump leads 2024 Republican field with DeSantis in distant second The politics of 'mind control' MORE in the preponderance of battleground states, and Republicans are suddenly rallying behind him, filling Democrats with anxiety that the vote count could deviate from the polls and produce a Trump victory.
Asked if she bought into the theory that the polls are wrong and Trump might actually be ahead, Democratic strategist Christy Setzer responded: “I do, unfortunately.”
She pointed to Clinton’s decision to advertise in the state of Michigan, which had seemed to be safely out of reach for Trump.
“Look no further than the campaign's travel schedule and ad buys to see where they're worried — battleground states that seemed so out of reach for Trump, Clinton stopped advertising in them...and now are on the airwaves again,” Setzer said.
Partisans on both sides of the aisle are sifting through mountains of late-breaking data, looking for evidence that pollsters are erring in their analyses or underestimating one candidate’s support.
Clinton is the favorite heading into Election Day.
While she holds a small lead nationally, the battleground map that gives her some breathing room.
Trump, on the other hand, must run the table on states Mitt Romney won in 2012 – no easy feat, as Arizona, Georgia and North Carolina look like toss-ups. He must also pull an upset victory in a state where Clinton is favored, such as in Pennsylvania, Michigan, Wisconsin, or Nevada.
Trump argues that polls are too heavily weighted in favor of Democrats.
His concerns have been echoed by some Republicans who say that surveys may be overestimating turnout among minority voters, potentially leading Clinton to underperform on Election Day. They also argue that working class white voters who have never before voted might turn out for Trump, potentially boosting him to a surprise victory.
Most experts, however, believe the Clinton turnout machine will be the great equalizer and that talk of a “silent majority” for Trump is fantasy.
They note that in 2012, the polls underestimated President Obama’s support.
Obama and Romney entered Election Day in 2012 separated by only about 1 point in the national polls. Obama cruised to victory by nearly 4 points.
Some Democrats believe that the polls may actually be skewed against Clinton, arguing that Trump’s rhetoric is sending Hispanics to the polls in record numbers in states like Nevada, Florida and North Carolina – all of which are must-win states for Trump.
“A huge amount of attention has been lavished on Trump's ‘missing white’ supporters,” liberal writer Greg Sargent tweeted Friday. “But missing nonwhite voters could matter more. Irony: The very things Trump does to activate ‘missing white voters’ could actually activate missing *nonwhite* ones.”
The Clinton campaign held a conference call with reporters on Friday claiming it had built an insurmountable early-voting lead in Nevada, Florida and North Carolina based on Hispanics turning out in record numbers for the Democrat.
Polls show Florida and North Carolina are toss-ups, with Trump holding a 2-point advantage in Nevada, according to the RealClearPolitics average.
But many Democrats believe Clinton’s Hispanic support is being underestimated in those states, pointing to a stunning CNN-ORC survey of Nevada released this week that found Trump ahead by 6 points.
That poll attracted scrutiny for not breaking out the number of Hispanics that were surveyed, leading some to suspect they had not polled a large enough sample of Latinos in a state where a more than a quarter of the population is Hispanic.
CNN-ORC does not release cross-tabs for a demographic if the margin of error for the group is greater than 9 points.
Democrats also believe they’re being underestimated in Florida, where more Hispanics have already cast ballots in 2016 than did in the entirety of 2012.
Republicans are skeptical that these are all new voters, arguing instead that the Clinton campaign is merely cannibalizing its Election Day turnout.
“We’ll see Tuesday, but I think there is more evidence that she is having a very difficult time replicating the Obama coalition,” said GOP pollster Robert Blizzard. “She’ll do better with higher educated white voters, but among millenials and minorities — they’ll still vote overwhelming for her, but their composition might be smaller than it was four years ago.”
Meanwhile, Democrats are skeptical about claims that Trump’s support is being underestimated.
There is no evidence that masses of uncounted white voters — either voting for the first time or ashamed to tell pollsters they support Trump — will suddenly materialize on Election Day, Democrats say.
They point to a study from The Upshot’s Nate Cohn that found there has been no surge in white voters since 2012. Rather, the white voters who have joined the voting pool are younger and more likely to support Clinton, the analysis found.
“This year, Mr. Trump’s gains among missing white voters aren’t likely to be even enough to overcome four years of demographic shifts, let alone form the basis of a lasting political coalition,” Cohn wrote
“According to these data, it’s Mrs. Clinton — not Mr. Trump — who stands to gain from a surge of new voters.”
And a Morning Consult survey found that Trump is pulling the same amount of support in live telephone surveys as he is in anonymous internet polls, which would seem to undermine claims made by some Republicans that GOP voters are being underestimated because they’re ashamed to tell pollsters they’ll vote for Trump.
There are some sub-groups where Trump performs better in online polls rather than live surveys — those with a college degree earning more than $50,000 a year, for instance — but it is not meaningful enough to swing the election, the study found.
Still, pollster John Zogby said that if ever there would be a year the polls came out wrong, it would be 2016.
“There’s too much volatility here with undecided voters choosing between two candidates they don’t like,” Zogby said.
That will keep Democrats on edge until votes are counted.
One top Democratic donor acknowledged feeling some queasiness.
“I have always viewed polls to be unreliable as you have to look at overall pathways to 270,” the donor said.
That’s where Democrats are taking solace, believing their superior ground game will ensure turnout is inline with the polls on Election Day.
“This election feels a bit more like 2008 — not in the hope and change excitement, but that Clinton’s ground game and organization was best in class,” the donor said.