As they dined at a fancy Des Moines restaurant on the eve of the 2008 Iowa Caucuses, America’s top pundits had dinners disrupted by nabobs from Hillary Clinton’s presidential campaign.
They were furious at a Des Moines Register poll showing Sen. Barack Obama surging ahead and predicting a record turnout. The prediction turned out right and put the Illinois senator on a path to the White House. Hillary finished in third place. The 2004 Register poll had also previously and correctly forecast the fall of Howard Dean and rise of Sen. John Kerry.
Not this year. The poll put Vice President Kamala Harris three points up on Donald Trump in a state Trump handily carried on election day. The President-elect is now suing pollster J. Ann Selzer, the Register, and Gannett for “brazen election interference.”
The litigious Trump is accusing pollster and paper of violating the Iowa Consumer Fraud Act, which lays out penalties for deceptive and dishonest practices. Law professors have told The New York Times that the suit is without merit, but it fits with Trump’s definition of the press as “the enemy of the people.”
“We have to straighten out the press,” he told a news conference at Mar a Lago. It came at the heels of a $15 million settlement with ABC News. Trump has often gone to court. He once sued writer Timothy O’Brien for allegedly underestimating the Trump riches, explaining that he wanted to “make his life miserable.” Such is the incoming President of the United States.
Election polls are proliferating in America. They vary in accuracy. Polltakers face the challenge of dealing with cell phones. They’ve consistently underestimated — “misunderestimated” in George W. Bush-speak — Trump’s support. Several 2016 surveys put Hillary Clinton’s chances of winning above 90 percent. Surprisingly, all seven “battleground” states in this year’s election went for Trump.
Pollsters have had mixed records for years, going back to the 1936 Literary Digest survey that predicted election of President Alf Landon. More famously, there was George Gallup’s prediction of President Thomas E. Dewey’s 1948 victory over Harry S. Truman. To use the old comeback, mistakes were made.
In this Washington, we’ve seen polls that were good, bad, and ugly. Just look at two years ago. A pair of Republican pollsters, Trafalgar and Moore Information, had long-shot challenger Tiffany Smiley closing to within two points of Sen. Patty Murray. (Moore also had Dino Rossi ahead in 2010.) On election day Murray won by double digits. A Democratic survey, Public Policy Polling (PPP), working for the Northwest Progressive Institute, got the margin correct.
Why do polls matter? They help determine where the money goes, but not always accurately. The 2010 Moore surveys, showing a tight race, prompted the U.S. Chamber of Commerce to put nearly $1 million into Rossi’s campaign. The Democratic Congressional Campaign Committee, in 2022, snubbed long-shot congressional hopeful Marie Gluesenkamp-Perez in Southwest Washington, who was doen in some polls. Once again, PPP got it right, detecting the last-minute trend that carried MGP to the country’s biggest congressional upset.
So why have polls been wrong about Trump? They’ve failed to detect his strength among folks who rarely vote. And a substantial number of voters, mindful of negative coverage, kept secret their intention to vote for “the Donald.” Blue collar voters defected from the Democrats. The high tide of support for Harris among “moderate” Republican women failed to materialize.Â
The pollsters failed to detect these trends. We pundits, myself included, failed to listen. Trump proved to be a master at harnessing resentment, and representing the darker side of the American dream.
I should have paid attention to my friend Mike, a longstanding nonpolitical pal. He was once a Young Democrat but fell away bigtime. He was part of the middle class that did not keep pace, making a living selling used cars and later time-share condominiums. He never went to college, became an evangelical, grew mad at government, and discovered Donald Trump.
But Trump connected with just enough of the Mikes out there whom Democrats had disconnected with, to narrowly win back the presidency.  And the consequences for pollsters getting it wrong can become consequential. Iowa’s Ms. Selzer, a polltaker renowned for her accuracy, has retired on the heels of a rare missed call.
The sad truth is we are in for rule by oligarchs, Iowa included, whatever the polls had promised.Â
Despite plenty of reliable information, the press continued to inundate us with the latest opinion polls–a form of information with inherent unreliability. That is, with no information in the press report about how the poll sample was selected, how the questions themselves were phrased and without any check on likely incentive to conceal true intentions, poll results can have little meaning to voters.
Poll coverage should be minimized by the press. They may have value for donors making strategic decisions about fund placement, but to us general voters polls are– like astrological predictions–just fantasy diversions.