War over polls intensifies

Want a tough job? Try being a pollster two weeks before Election Day.

Donald TrumpDonald John TrumpTrump nominates ambassador to Turkey Trump heads to Mar-a-Lago after signing bill to avert shutdown CNN, MSNBC to air ad turned down by Fox over Nazi imagery MORE is ripping pollsters and the media, arguing the surveys are biased against him because many include too many Democrats in their sampling surveys.

Conservative news sites have pounced on Trump’s arguments, pointing to a trio of national polls that show the race between the Republican presidential nominee and Democrat Hillary ClintonHillary Diane Rodham ClintonOvernight Defense: Trump declares border emergency | .6B in military construction funds to be used for wall | Trump believes Obama would have started war with North Korea | Pentagon delivers aid for Venezuelan migrants Sarah Sanders says she was interviewed by Mueller's office Trump: I believe Obama would have gone to war with North Korea MORE is a toss-up.

ADVERTISEMENT

Trump and his supporters argue that mainstream pollsters are under-sampling Republicans to account for a rise in independents, while failing to account for an enthusiasm gap that favors Trump or the new voters he could bring into the fold.

A half-dozen pollsters interviewed by The Hill acknowledged the difficulties of polling the 2016 race.

They blamed unprecedented volatility, two historically unpopular candidates, fast-changing voter behavior patterns and shifting demographics. All of those factors make it tough to figure out exactly who will show up on Election Day.

The pollsters interviewed by The Hill largely dismissed arguments that polls showing a tight race are the accurate surveys, however.

They take issue with the methods those pollsters use and point to 2012, when many Republicans wrongly believed the polls were skewed against Mitt Romney.

“When the candidate starts fighting against the polls as much as their opponent, it means they’re losing,” said Monmouth University pollster Patrick Murray, whose most recent national survey put Clinton’s lead at 12 points.

Still, many pollsters are puzzled by the extreme variations in the surveys.

A new survey from the conservative outlet Rasmussen gives Trump a 1-point lead nationally. A Los Angeles Times/USC poll — one of the most accurate from 2012 — has Clinton ahead by only 2 points. And a survey from IBD/TIPP — which has been among the best for the last three presidential elections — has the candidates locked in a tie.

All three surveys were featured Monday atop the conservative Drudge Report under the banner “Shock Polls.”

The polls showing a tight race are particularly notable, since there are also recent surveys that suggest a landslide win for Clinton.

The Democrat enjoys a 6-point lead in the RealClearPolitics average.

Clinton also leads in most of the battleground states, and the race has tightened in several traditionally red states like Arizona, Texas, Georgia and Alaska, leading some to predict a landslide in favor of the Democrat.

GOP pollster Robert Blizzard dismissed the polls showing a tight race, saying the data is clear: Trump has been unable to grow his support beyond the 40 percent range and is likely headed for defeat on Election Day.

“Throw out the outliers,” Blizzard said, singling out Rasmussen in particular. “There are public pollsters out there that have no idea what they’re doing.”

The pollsters for the surveys that show a tight race have been on the defensive for months and are as eager as anyone to see whose methodological assumptions prove true on Election Day.

“I’m scratching my head just like everyone in America,” said Rasmussen polling analyst Fran Coombs. “Believe me, we’re not putting the thumb on the scale to make Trump look better. It’s a crazy election year and this is what the numbers tell us. We’re comfortable with our methods.”

The L.A. Times/USC survey has drawn scrutiny for consistently being the only survey to find Trump with a lead.

One pollster interviewed by The Hill dismissed that survey as “experimental,” noting that it contacts only the same fixed pool of respondents, rather than reaching out to new voters for each poll.

In addition, the L.A. Times survey is only polling a two-person race, only samples people who have voted before and has faced criticism for how it weights the final results to account for the nation’s demographic and racial mix.

But that same methodology worked in 2012, when it predicted Obama would win by 3.3 percentage points, even when many other polls showed Obama doing worse. The president won reelection by 3.8 points.

“Obviously, the poll’s [2016] results have been an outlier compared with other surveys, but if ever there was a year when the outlier might be right, it’s this year,” L.A. Times Washington bureau chief David Lauter wrote in defense of the poll.

The IBD/TIPP poll has been praised by data guru Nate Silver as among the most accurate for several cycles in a row.

TIPP pollster Reaghavan Mayur told The Hill that he believes the surveys showing Clinton with a double-digit lead are equally deserving of scrutiny.

“In 2008, Obama defeated McCain by 7.2 and that was a blowout election in terms of enthusiasm. Everyone was revved up and to accentuate it, there was the financial tsunami,” Mayur said. “And you’re telling me now that now in 2016 Clinton is up by 12? Does that jibe with anything we’ve seen in the past or common sense?”

“I’ve told myself that I can get a good night’s sleep if I do what I’ve always done,” Mayur continued. “I don’t worry about this. I’ve used the same models over the past four races and have had good showings.”

But some other pollsters say the surveys showing a close race are off precisely because they’re doing the same thing and not accounting for changes in the electorate.

Some believe the white, college-educated voters who were once a Republican stalwart will be less likely to make it to the polls for Trump this go round.

Others say that potential GOP voters are being over-sampled by these three polls, because an increasing number have begun identifying as independents.

And as much enthusiasm as there is for Trump, there seems to be an equal amount of frustration with the candidate among Republicans, potentially suppressing turnout within his own party.

Still, the huge disparity in the polls and the unprecedented volatility has even those pollsters who are in the majority a little bit nervous.

The ABC tracking poll released over the weekend swung 8 points in Clinton’s direction week-over-week. 

And the largest difference between polls in 2012 was around 6 points. This time, it’s 13.

“We really don’t know whose going to vote yet and that makes it more volatile,” said pollster John Zogby. “There are no likely scenario here.”