Making sense of 2020 presidential election polling By Luke Perry
Polling was a major topic following the 2016 election in light of an upset victory by Donald Trump. Four years later questions remain about Trump’s prospects for reelection and the role of polling in predicting what happens this fall.
Polling is a picture taken at a particular moment of time. The image is not as clear as reality itself. Seeing Paris firsthand is different than looking at a picture, though a picture provides a good sense of what it is like.
Polling is not easy. Social Scientists are mindful of the complexity of human thoughts and actions. Every poll includes a margin of error, acknowledging upfront that outcomes are likely to range, typically 3 to 5 points.
Current polling for the presidential election involves national polls and polls from swing states. Recent Quinnipiac University polling had Joe Biden leading Donald Trump nationally by 10 points among likely voters.
Nearly 60 percent believed the country was worse off now than in 2016, a key consideration this cycle, while the president had a 43 percent approval rating, 53 percent disapproval.
Analysts also consider the aggregate results of many polls conducted over time. This month Biden has an average lead of 7 points over Trump among all national polls.
Where people vote; however, is as important as who they vote for. This month Biden leads in aggregate polling in every major swing state: Michigan (+4.2) Wisconsin (+6.8), Pennsylvania (+4.3), Florida (+1.6), and North Carolina (+0.6).
These numbers can change and it is important to note that several swing states are very close. Even though Biden is currently up in Florida and North Carolina, for example, it is far from certain he will win.
Moreover, just because probabilities suggest one candidate is the favorite to win does not necessarily mean that will happen. In 2016, for instance, political analysts thought it would very difficult for Donald Trump to win with just 45 percent national support.
National polling was right. Trump finished with exactly 45 percent of the vote, but won by securing important victories in particular states, notably Michigan, Wisconsin, Pennsylvania, Florida and North Carolina.
State polling in Michigan and Wisconsin was off because of subtle changes in turnout that favored Trump, including lower African-American turnout and higher turnout for white men without college degrees. Polling was accurate in the other 48 states.
Scientific polling is reliable and insightful, but also fluid and imperfect. Polling is not as bad as critics think, nor as certain as some like to believe.
Current polling favors Biden, for now. The first debate on September 29 is the next major test for each campaign.
Luke Perry (@PolSciLukePerry) is Professor of Political Science at Utica College and Director of the Utica College Center of Public Affairs and Election Research.