Pollster Ann Selzer: We’ll look at data to try to understand Iowa Poll miss
What a big miss for The Des Moines Register/Mediacom Iowa Poll. It followed an unprecedented worldwide fascination with an outlier poll showing Kamala Harris leading Donald Trump by 3 percentage points.
The final poll followed a surprising September poll showing the vice president had closed Trump’s lead over Biden in a June poll by 14 points. So, the October poll appeared the next step in an upward progression for Harris. Except that turned out not to be true.
My philosophy in public opinion research is to take my best shot at revealing the truth of a future event, in this case Election Day. Without fear or favor, we used the same method as the final poll this year to show a healthy Trump lead in both 2020 and 2016. Those turned out to capture the mood of the electorate reasonably well, though both took fire from Iowans who doubted the findings could be true.
My inbox and my voice mail have been full of questions the last few days — those who wanted to know why I “manipulated” the data to show a false Harris lead, and those who wondered if the data were too good to be true.
In response to a critique that I “manipulated” the data, or had been paid (by some anonymous source, presumably on the Democratic side), or that I was exercising psyops or some sort of voter suppression: I told more than one news outlet that the findings from this last poll could actually energize and activate Republican voters who thought they would likely coast to victory. Maybe that’s what happened.
The team at Selzer & Company has begun a review to raise any plausible question of what happened between Thursday night the previous week, when we finished interviewing, and when the votes were tallied on Tuesday night. That work has begun, but it will be awhile before it is complete.
More: Pollster J. Ann Selzer: 'I’ll be reviewing data' after Iowa Poll misses big Trump win
In 2004, the final Iowa Poll had John Kerry with a small 3-point lead over President George W. Bush. In the end, Bush won Iowa by less than a percentage point. I had the good fortune to run into former Gov. Terry Branstad, who gave me a masterclass on why things move late. He credited an enormous rally in Sioux City, closed to the press, with activating a turnout strategy he thought led to a substantial widening of the Bush winning margin in the 5th Congressional District in western Iowa. He also suspected that rally was the cause of Sen. Tom Daschle getting drummed out of office in South Dakota.
That story stays with me now. But, what other stories are out there to explain the miss? We’ll be looking at turnout rates at the polls and comparing them to our demographic mix. We’ll be looking at what amount to tea leaves in our poll about the trending story about Black and Latino men’s growing alignment with Trump and his policies. We’ll be looking at how the late deciders fell, if we can figure that out without a traditional exit poll for Iowa.
At an anecdotal level, I had a sense of the late shift by listening to ordinary voters on a platform called 2WAY, created and helmed by Mark Halperin. The ethos of the platform is peace, love and understanding, and giving others the presumption of grace. It is an outlet that specializes in civil discussions of voting and why people choose the candidates they do.
The last few days featured proportionately more women who had been consistent Democratic voters opting for Trump. Absent was a counterpart of consistent Republican voters opting for Harris, the kind of cohort led nationally by Liz Cheney. A few months ago, one woman still making up her mind said something that has stuck with me. Trump disgusted her, but Harris scared her. I doubt I can find data that can reveal how common these thoughts were.
But I will go looking, both to confirm or deny the underlying theory.
J. Ann Selzer heads Selzer & Co. of West Des Moines. She first worked with the Iowa Poll in 1987 as a Des Moines Register staffer, then started her own firm and has conducted the Iowa Poll on a contract basis since 1997.
This article originally appeared on Des Moines Register: Ann Selzer: We’ll look at data to try to understand Iowa Poll miss