The devil didn't poll down in Georgia
There's a familiar storyline missing from the next big election.
Hello! Welcome to Margin of Error, a newsletter from me about the polls and the way they are covered.
If you’re new here, thanks for signing up! If you find it interesting, I’d be grateful if you would encourage others to sign up.
If someone sent this your way or you found this post through Twitter or other channels, make sure to subscribe below.
For both new and longtime subscribers, an apology for the lapse in editions. I’ve written partial and full posts over the past six weeks, but when I was writing and/or finished with them, they didn’t feel quite right. My friend and former colleague Myles Udland addressed this on his good newsletter about a similar writer’s block: “I simply wasn’t all that interested in what I’d written. So why would anyone else be?”
To be quite honest, it feels somewhat difficult, even in this edition, to write about polls. A sizable faction of the United States Congress has for the last month been engaging in an overt attempt to subvert the will of a majority of voters in the presidential election. Every other storyline, at least on the political front, feels like it doesn’t really matter.
Now that I’ve gotten that off my chest, I hope you enjoy this edition. Because we’re focusing on the Georgia runoff elections, and they do matter.
By now, we mostly know the story of the 2020 general elections and the polls. The polls missed results in the key states of 2016, which led to a lot of focus on how pollsters were going to fix it. That, combined with the compelling story they were telling and the insatiable appetite of an increasingly interested public, made polls an even more ripe area for coverage in 2020.
This time, though the polls correctly predicted the winner, they also exhibited a lot of the same mistakes as 2016—sometimes even worse.
The first big election of 2021 (and, really, the last election of the 2020 cycle) comes Tuesday in Georgia. It is hard to overstate the consequences of the two runoff elections, which feature Democrats Jon Ossoff and Raphael Warnock facing off against incumbent Republicans David Perdue and Kelly Loeffler, respectively. A lot of observers would probably disagree, but even though liberals seem increasingly frustrated with him, a Chuck Schumer-led Senate makes a huge difference for a Joe Biden presidency.
Outside of the aforementioned plot to subvert democracy, the political world has been focused on the Georgia elections over much of the past month. Yet there has been a glaring void from the conversation: No one talks about the polls anymore.
They just aren’t really a feature of the coverage of these runoffs. To be sure, the most closely watched presidential election in quite some time, with juicy subplots in the Senate and other down-ballot races, is different from two Senate runoff elections.
Yet it hardly feels like polls are part of the conversation. There are few pollsters (if any) previewing their results hours or days ahead of time, or building excitement for poll releases that dominate the Twitter conversation for the day. FiveThirtyEight has poll averages for both races on its homepage, but they’re both kind of sad and tiny, a fitting testament to the shriveled impact of polling on the conversation. There haven’t even been a slew of frantic campaign emails trying to raise money off a bad poll.
The Georgia runoffs present an interesting case study for the future of polling. Does it mean that polls have peaked in their impact? Is it a signal that pollsters are moving more cautiously and trying to figure out where they went wrong in November? Or is it evidence of recognition that no matter what pollsters do, polling in the era of Donald Trump will always be missing a certain level of support?
The limits of the old methods
As of Sunday, there have been 22 public polls conducted of either or both Georgia runoff races. Not too many of them have been by well-known or otherwise respected pollsters. The most frequent pollster of the race is Trafalgar, the firm that found a much closer race than others and then prematurely beat its chest before all votes were counted.
Tom Jensen, who runs the Democratic firm Public Policy Polling, said PPP has conducted private polls for clients that haven’t released the results, a fairly standard practice. He told me a few factors are all converging to create the current dearth of polling:
The runoffs do not have a particularly unique storyline. It’s baked in that they will be close, something that, so far, the limited Georgia polls are confirming. (As of Sunday, both Democrats led by about 2 points on average.)
The campaigns have been taking place during the Thanksgiving, Christmas, and broader holiday seasons, a particularly challenging time for pollsters to reach people (thus also making it more expensive).
Jensen said it’s likely some more well-known firms, especially those aligned with universities, used up budgets for the year. (A few university pollsters either declined or didn’t respond to requests for comment for this newsletter.)
Jensen said there could be some “skittishness” about polling—even though, as he pointed out, Georgia was one of the few states in which public polling ended up being pretty accurate. (The final FiveThirtyEight average found Biden leading by 1.2 points in Georgia, where he ended up winning by about 0.3 points.)
The relative lack of quality polling is feeding into a media environment that has also been more cautious in drawing conclusions from the data. Multiple journalists have told me that their outlets have deemphasized or even outright discouraged including mentions of polls (and polling averages) in stories.
You see where this cycle is going.
Emerson College in Boston, which has an A-minus rating from FiveThirtyEight, has been one of the few quality pollsters engaging in the Georgia runoffs. (It found both Republican candidates slightly ahead but well within the margin of error.)
Spencer Kimball, the director of Emerson College Polling and assistant professor at the college, echoed Jensen’s sentiments. He told me it was especially difficult to pick the right time to poll the races, given the holidays and the continued importance of early voting.
“I am sure budgetary concerns are an issue for pollsters, but also the news media might feel a bit burned by the pollsters and will focus on other aspects of their campaign coverage,” he said via email.
But Kimball told me that while Emerson’s review of its 2020 performance is still in process, he is looking at the runoffs as he does every election: a chance to “test new methods and refine old ones.”
My first thought after the polls’ relative miss in 2020 was that it could be good for the industry. I thought that it could follow a similar pattern as the aftermath of 2016, as general interest would only build as pollsters continued to try to figure out the puzzle. It’s too early to say definitively, but the Georgia runoffs have shown the limits of the old methods.
Thanks for reading Margin of Error. If you have any tips, comments, or insights about polling, email me at bplogiurato@gmail.com, or find me on Twitter @BrettLoGiurato.
If you liked what you read, share it with a friend and make sure to subscribe.