Undoubtedly, the polling industry will take a hit and face a new round of intense scrutiny because many presidential election and other polls were off.
Clearly, the ABC News/Washington Post poll that had Joe Biden leading President Donald Trump by 17 percentage points in Wisconsin in late October was just wrong. So was the New York Times/Siena College poll that had Biden up 11 points there.
Even if you remember the precaution that polls are snapshots of the period they were taken and don’t predict the future, races just don’t swing by 11 or 17 points in a few days. Biden leads in Wisconsin by less than three quarters of a point as of today.
ABC News & The Washington Post aren’t the only problem pollsters, but one guy who was sure all the other pollsters were wrong has some explaining to do, too — Robert Cahaly, the owner of the Alabama-based Trafalgar Group.
Trafalgar did really well picking Trump as the winner in 2016 in various states, partly by figuring out that many Trump voters didn’t like to admit they favored him.
Cahaly went on Michael Smerconish’s CNN show the week before the election. He boasted about how the election would prove his polling right.
Remembering Cahaly’s polls end within a day or two of the election, here’s the breakdown, as of yesterday at 2:12 p.m.:
FLORIDA: Trafalgar: Trump, 49.4%, Biden, 47.3%. Reality: Trump, 51.2%, Biden, 47.8%.
GEORGIA: Trafalgar: Trump, 49.7%, Biden, 45.4%. Reality: Biden, 49.5%, Trump, 49.3%.
MICHIGAN: Trafalgar: Trump, 48.3%, Biden, 45.8%. Reality: Biden, 50.5%, Trump, 47.9%.
MINNESOTA: Trafalgar: Biden, 48%, Trump, 44.8%. Reality: Biden, 52.5%, Trump, 45.4%.
NEVADA: Trafalgar: Trump, 49.1%, Biden, 48.4%. Reality: Biden, 50.2%, Trump, 47.5%.
NORTH CAROLINA: Trafalgar, Trump, 48.6%, Biden, 46.5%. Reality: Trump, Trump, 50%, Biden, 48.6%.
OHIO: Trafalgar, Trump, 49.2%, Biden, 44.4%. Reality: Trump, 53.2%, Biden, 45.3%.
PENNSYLVANIA: Trafalgar, Trump, 47.8%, Biden, 45.9%. Reality: Biden, 49.7%, Trump, 49%.
WISCONSIN: Trafalgar, Biden, 47.5%, Trump, 47.1%. Reality: Biden, 49.5%, Trump, 48.8%.
So right now, Trafalgar missed in five states and hit in five, though the results in Arizona and Georgia remain really close.
In fairness, he did a lot better than other pollsters like his Wisconsin prediction, for example.
Maybe Cahaly can claim his results fell within the margin of error of his polls, but so can other pollsters.
My guess is Cahaly should have weighted the voters he counted differently from state to state, but that’s just a guess.
I’ll wait until all the results are in before judging the polling, but Cahaly deserves a critical look, too.
Interestingly, his theory that some Trump supporters won’t tell pollsters they planned to vote for him has some smart people thinking he’s right. Pollster Frank Luntz issued a really critical look Saturday on Weekend Edition on National Public Radio. U.S. Sen. Bob Casey told me Friday that Trump’s presence on the ballot appears to influence the way some voters answer polls.
“I think it could be,” he said. “But I’m not a scientist. It will take a lot, lot more analysis, but it just wasn’t happening in ’18 (2018 when he ran for re-election.”
By the Thursday before that election, his polling showed him at 56% support.
“And that’s what I got on Election. I just think Trump on the ballot is a whole different dynamic with regard to polling,” he said.