A side issue that arose in this recent election was the Presidential poll by Iowa’s Selzer & Co, a firm treated with enormous respect in the polling industry and rated a 2.8 out of 3, at the time, by FiveThirtyEight’s rating of pollsters. Collected and released late in the voting cycle, the poll gave Harris a three point lead in the state of Iowa, a shocker in that Iowa had strongly favored Mr Trump in his previous two Presidential runs, which were in 2016 and 2020. Of course, I was tracking Senate races; the Selzer poll was noted only in the context of whether or not it would effect Senate races (“no”). But I was curious as to the Iowa results.
In the end, Vice President Harris did not win Iowa’s six Electoral College votes. The margin? 13.2 points.
That’s not even close. Off by 16 points.
What went wrong? WaPo’s Philip Bump has an analysis, based on Selzer’s comments, which might be best summarized as dated methods. The ways in which we may be contacted are becoming more and more specialized, anonymized, and, for the purposes of statistics, self-selecting, meaning we have some idea of who’s calling, even if the caller’s name is not made available. In the old days, you whippersnappers, the phone rang and you either let it ring or you picked up without knowledge, generally, of who was calling. That’s not true today. The majority of citizens don’t want to answer a poll, leaving only a few that do (self-selecting) or don’t pay much attention to caller’s ID (those curious about the world, perhaps) or refuse to pay for it where it costs (cheapskates).
It’s a recognized problem in the industry, and it’s a problem because it skews the raw results, meaning corrections built on unverifiable assumptions must be applied.
The Des Moines Register had a discussion of its own:
Selzer’s review has taken the form of testing plausible theories against available data. To date, no likely single culprit has emerged to explain the wide disparity. But I wanted to walk you through what has been looked at and what relevant data shows.
Following the Register’s long practice, we already released the poll questionnaire. For transparency, we’re also releasing the poll’s full demographics, crosstabs and weighted and unweighted data, as well as a technical explanation from Selzer detailing her review.
The balance of the analysis comes to Nothing yet, still looking! I did notice they did not address the issue of poll respondents simply lying, but then that would be a tough reason to address. If true, though, to the extent such a supposition can be proven, it would be very disappointing concerning the moral character of Iowans. I’d rather not think so.
Selzer has also announced she’s leaving the business:
Public opinion polling has been my life’s work. I collected my first research data as a freshman in college, if you don’t count a neighborhood poll I did at age 5. I’ve always been fascinated with what a person could learn from a scientific sample of a meaningful universe.
Beyond election polls, my favorite projects were helping clients learn something they did not know to help them evaluate options for their companies, institutions or causes. That work may well continue, but I knew a few years ago that the election polling part of my career was headed to a close.
Over a year ago I advised the Register I would not renew when my 2024 contract expired with the latest election poll as I transition to other ventures and opportunities. [Des Moines Register]
Taking her at her word, it appears the polling game is becoming more and more difficult, not due to the usual factor of increased competition, but because the data they are mining, if I may use a precious metal metaphor, is becoming more and more difficult to obtain, and thus interpret.
For those of us who like polls as a way to learn how the nation is leaning, this is a horror-filled message.
Finally, Mr Trump is suing Selzer & Co:
But Trump, apparently still smarting from having to endure a day of people wondering if he might lose Iowa, sees himself as the one who had to deal with dire consequences. On Monday, his lawyers filed a lawsuit in Iowa alleging that the poll’s release was an example of “brazen election interference.”
Selzer would “have the public believe it was merely a coincidence that one of the worst polling misses of her career came just days before the most consequential election in memory,” the lawsuit claims. It later adds that the poll “was no ‘miss’ but rather an attempt to influence the outcome of the 2024 Presidential Election.” [WaPo]
Mr Trump will, or should, be laughed out of court. Or he’ll endanger a bunch of bad Republican pollsters. I think Selzer should ask how much money Trump has via the press, and then comment how much fun it’ll be to strip that money out of Trump’s operations. Once that’s finished, have his lawyers disbarred.
