. WaPo’s Philip Bump Nukes Trafalgar’s Rob Cahaly as ‘Not a Real Pollster’: ‘Laughable’, ‘Absolutely Ridiculous And Statistically Unsound’ - News Times

WaPo’s Philip Bump Nukes Trafalgar’s Rob Cahaly as ‘Not a Real Pollster’: ‘Laughable’, ‘Absolutely Ridiculous And Statistically Unsound’

By News Here - 17:08

Sarah Silbiger/Getty Images, Drew Angerer/Getty Images

As we trudge through the final days of the 2020 election, pollsters’ attempts to read the tea leaves of voter preferences are once again the subject of much speculation, and countless articles and commentary. One pollster with a track record heavily favoring President Donald Trump was the subject of a blistering critique on Mediate founder Dan Abrams’ SiriusXM show Friday, when Philip Bump of the Washington Post slammed him as “not a real pollster,” and shredded his “laughable” methodology as “absolutely ridiculous and statistically unsound.”

Robert Cahaly, the founder of the Trafalgar Group, was the target of Bump’s verbal evisceration. Cahaly is known for being one of the few pollsters to predict Trump would win the 2016 election, based on the “hidden Trump voter” premise. Still, the pollster has faced criticism, and has a C- rating from statistics website FiveThirtyEight.

Trump himself has sneered at the polling throughout this campaign — often in conjunction with his attacks on the “fake news” media.

To introduce the discussion, Abrams played a clip of Cahaly speaking, where the pollster explains what “he’s been doing differently from the other pollsters, and why he thinks he’s going to be right again.”

Cahaly expressed his dislike for “long questionnaires,” because they “miss average people” who “don’t have time to stop, take the phone call at 6 o’clock and talk for 25 minutes.” According to Cahaly, people are unwilling to answer polls that take longer than three minutes, so they keep theirs under that limit.

Cahaly added that they always poll at least 1,000 people in a state, by multiple methods including live calls, text messages, robocalls, emails, online polls, and a longer time frame to complete the poll.

One key factor in the different results from Trafalgar polls is their sampling models. As Cahaly explained, they were targeting “likely voters as we define them” — a group that includes low propensity voters who “have demonstrated the characteristics, as they did in ’16, of participating this year.”

Abrams then asked Bump for his thoughts, and he pulled no punches.

“The word I would use to describe his claims is laughable,” said Bump. “Essentially, everything he said there runs contrary to what we understand about polling.”

Bump rattled off a series of blows, shredding Cahaly’s claims about sample size as “absolutely ridiculous and statistically unsound,” his polling methods as “both unproven and subject to a lot of fluctuation that real pollsters try and avoid,” and the way Trafalgar chooses their sample as “subject to a lot of subjectivity, which I think is part of the reason that he gets the results that he gets.”

Merely “having made a correct guess in one election,” Bump said, “was not necessarily someone you should assume has a strong track record of success.”

Going back to Trump’s attacks on pollsters as biased against him, Bump told Abrams that “the idea that pollsters, broadly… are trying to throw this thing one way or the other is idiotic.”

However, Bump continued, there were some unreliable pollsters, who were willing to “get up there and repeat a factually inaccurate claim from the Trump campaign about what Biden’s view of the virus is, which is just what you heard the guy from Trafalgar doing, right?”

“Trafalgar isn’t even honest about how they do their methodology,” said Bump, criticizing the pollster for not submitting all their data in a way that third parties could review.

“I mean, you heard him hemming and hawing,” Bump continued. “He hasn’t been upfront about that. Real pollsters participate in polling groups, there are professional organizations of pollsters that evaluate work and look and see how people did, and hold themselves accountable.”

“If our pollsters were throwing the race, they would be fired from the Washington Post,” Bump declared, “because our goal is objectivity — no matter what Donald Trump says it is.”

Is Bump’s criticism fair? Reviewing the polls tracked by RealClearPolitics in several key states, Trafalgar is a clear outlier, over and over again.

In North Carolina, polls have had Trump’s Democratic challenger Joe Biden up a few points the entire month of October, with a handful of tie results, Republican pollster Rasmussen twice finding Trump up one point, and Trafalgar finding Trump up 3 in an Oct. 20-22 poll.

A number of states follow similar trends as Nevada, Minnesota, and Wisconsin, where the polls all show Biden ahead, but the narrowest Biden leads are from Trafalgar.

Most of the polls from Florida give the edge to Biden, and again Trafalgar is one of the few standout red lines, saying Trump is up a few points.

Michigan is one of the more extreme examples, with polls all year showing Biden with the lead, and currently an RCP average of +6.5 points. The only pollster finding Trump winning is Trafalgar, showing the president up one or two points four separate times since August.

Listen above, via SiriusXM.



from Mediaite https://ift.tt/35NkqnT

  • Share:

You Might Also Like

0 comments