What Polls Showing Democrats Ahead Really Mean | Opinion

Whether it's Rep. Tim Ryan (D-Ohio) running even with Republican Senate nominee J.D. Vance, Wisconsin Lt. Gov. Mandela Barnes sporting a healthy lead over incumbent GOP Sen. Ron Johnson, or Democratic governors maintaining advantages in the Michigan, Wisconsin and Pennsylvania governor races, Democrats are polling better than they have all year in the Blue Wall states that fell to Donald Trump in 2016. But after what seems like a decade of underperforming poll-based expectations, no one seems to believe it.

Democratic presidential nominee Hillary Clinton led Donald Trump by 6.5 points in Wisconsin polling averages and lost the state by less than 1 percentage point. In Michigan, Trump erased a 3.4-point average lead on election day, and in Pennsylvania it was 1.9.

Outcomes in 2018, however, did not actually support the idea that polling in these states is hopelessly broken. Democratic Senate incumbents in Wisconsin, Pennsylvania, and Michigan won by margins similar to polling averages, and Democrats won their gubernatorial races in these three states by roughly the anticipated margins.

Biden Holds a Rally
Supporters cheer at a Democratic National Committee rally at Richard Montgomery High School on Aug. 25, 2022, in Rockville, Maryland. Drew Angerer/Getty Images

The big misses in 2018 were in the Ohio, Florida, and Iowa governor's races, where Democrats held small leads in final polling averages only to lose, and in a handful of Senate races, including Missouri, Tennessee, and Indiana, where polls predicted competitive races but Democrats got beaten like drums. And in 2020, presidential polling in Pennsylvania was spot on and close to the mark in Michigan. But once again, the polls dramatically underestimated Trump in Florida, Wisconsin, and Ohio, as well as nationally.

What to make of all this? For many analysts, like the University of Chicago's Nicholas Marchio, "polling as we know it seems to be broken." Yet there is a very clear pattern in the polling, and it's not "every Democratic lead is a mirage."

I compared final election results for the 2016 presidential election, 2018 Senate and gubernatorial races, and the 2020 presidential election against the final Real Clear Politics averages in Wisconsin, Michigan, Pennsylvania, Ohio, and Florida, where analysts like Sean Trende are most skeptical of the polling. Across those 20 races, the Democrat generally got within a point or so of their final polling number, overperforming by an average of just 0.8 points.

It's the GOP's numbers that have been way off. Republican candidates in these races gained 4.1 points on average from their final polling numbers. The lower the GOP candidate was polling and the greater number of undecided voters, the greater the effect tended to be. In 2018, for example, incumbent Sen. Sherrod Brown had a huge lead over GOP nominee Jim Renacci in Ohio, but Renacci gained over 8 points from his 38.6 percent average to make it much closer.

There are three big lessons to draw from this history. First, polling in Pennsylvania has been fine. People seem to forget that polls correctly suggested very close races in 2016 and 2020 and close races are what we got. Second, you can take the Democratic number in polling averages to the bank, but victory is unlikely if a Democrat is leading but averaging in the low-to-mid-40s. And third, expect the Republican candidate to do better than averages would suggest. Polling undershot the final results for Republicans in every single one of these 20 races.

Where does that leave Democrats in 2022? Well, Tim Ryan leads J.D. Vance in Ohio 44.8 to 43.8 in the FiveThirtyEight average (Real Clear Politics curiously omits a series of public polls of this race). If those were the numbers on election eve, I wouldn't put one dollar on Ryan, who would probably lose to Vance handily. There have only been two post-primary polls of the Wisconsin Senate race, but Democrat Mandela Barnes has been over 50 percent in both of them. If that trend continues, Ron Johnson is in much deeper trouble than GOP strategists seem to realize.

In Pennsylvania, GOP quack and viral grocery shopper Mehmet Oz is a goner unless these numbers turn around soon. John Fetterman is at 47.8 in the RCP average and 49.3 at FiveThirtyEight. Ditto for the Michigan race where Democrat Gary Peters is well over 50% with a huge lead over Republican John James. And with Democrat Val Demings polling at just 42.8 percent against incumbent Republican Marco Rubio in Florida, Democrats should cut their losses here unless they see those numbers turn around dramatically.

Democrats also look like the favorites in the Wisconsin, Michigan, and Pennsylvania gubernatorial races, whereas both Mike DeWine in Ohio and Ron DeSantis in Florida should start thinking about their second terms, which are overwhelmingly likely.

The bottom line: there is no reason to dismiss out of hand all polling showing Democrats in good position to hold the Senate and to hang onto swing state governor seats that might prove critical in 2024. Polls may never be as accurate as they were in 2008 and 2012, but analysts who disbelieve every Democratic lead in the Midwest and Florida will be in for a big surprise on election night.

David Faris is an Associate Professor of Political Science at Roosevelt University and the author of It's Time to Fight Dirty: How Democrats Can Build a Lasting Majority in American Politics. His writing has appeared in The Week, The Washington Post, The New Republic, Washington Monthly and more. You can find him on Twitter @davidmfaris.

The views expressed in this article are the writer's own.