What Do Polls Of The Georgia Senate Runoffs Actually Tell Us?

There's little data ― and a lot of room for caution ― about public opinion on the two elections.
People sign in to vote in the Senate runoff elections at the Lawrenceville Road United Methodist Church in Tucker, Georgia, o
People sign in to vote in the Senate runoff elections at the Lawrenceville Road United Methodist Church in Tucker, Georgia, on the morning of Jan. 5, 2021.

For a high-stakes, closely watched set of campaigns, the Georgia Senate runoffs have been the subject of relatively little publicly released polling. After election polls in November again substantially underestimated the GOP’s performance across a number of key states, many pollsters are still drilling down on what, exactly, went wrong last year.

Expensive, live-interview phone surveys, the type conducted largely by academic and media sponsors, have been close to non-existent in Georgia (exit polling, which includes phone calls to mail-in voters, has been conducted in the state). The campaigns themselves have also deemphasized their use of polls to gauge the state of the horse race between candidates, although survey data remains a tool for testing campaign messages.

What polling there is points to a pair of close contests. FiveThirtyEight’s aggregates show less than 2 percentage points of separation between Democrat Jon Ossoff and Republican David Perdue, and an almost equally tight race between Democrat Raphael Warnock and Republican Kelly Loeffler. With margins that slim, the takeaway that the competitions appear deadlocked matters more than which candidate happens to have the apparent edge.

Given last year’s misses, it’s understandable if people wonder whether even that sort of broad characterization is meaningful. One point in favor of the latest Georgia numbers: The polling error in 2020 appeared highly regional, with surveys missing widely in the Upper Midwest and faring better in some other places. An average of presidential polling in Georgia last year accurately suggested that the state would see a close contest.

On the other hand, accurately polling a special election is an uphill battle even at the best of times, because it’s especially challenging to make assumptions about who will turn out to vote. That difficulty is compounded this year by the ongoing pandemic, the rise in early voting, and the timing of the election, which falls just after the holidays.

“Should we just expect the Georgia polls to be right? I think that would be a little bit of a mistake,” Democratic pollster Nick Gourevitch told Politico. “Everybody fundamentally understands that it’s going to become an issue of partisan turnout. And anybody who tells you they know exactly what’s going to happen in terms of partisan turnout in a special election with two senators to decide control of the Senate in a post-Trump era when he’s not on the [ballot] — nobody knows the answer to that question. It’s a completely unique situation.”

Taking it at face value, the polling we have shows two very close races that could both plausibly end in narrow wins for either side. If either of the candidates defies expectations by netting a runaway victory, it’ll raise another set of questions about election polling error.