After the 2016 U.S. presidential election, pollsters shook their heads and tried to figure out how they went wrong. In case you forgot, everyone went to bed expecting a Hillary Clinton presidency. We woke up with Donald Trump.
This year, all the major pollsters said they adjusted their models and learned from their mistakes.
Yet here we are, a few days after another presidential election, once more scratching our heads. Why were the polls off base once again?
“No two polling failures are quite alike and I think that held in this year,” said W. Joseph Campbell, a professor of communication at American University and author of “Lost in a Gallup: Polling Failure in U.S. Presidential Elections.”
“2020 is not the same kind of polling surprise as it was in 2016.”
The truth is, as mail-in votes continue to be counted, there is growing evidence the polls were not that far off. While many pollsters predicted a blue wave, there has been instead a blue stream. Vice President Joe Biden has a solid lead in electoral votes with 290. Democrats have not taken back the Senate as anticipated although they picked up one seat. They are barely clinging to their majority in the House, but they still hold the upper hand.
Georgia is currently swinging towards Biden. This would be the first time it has chosen a Democrat for president since 1992. It also looks as if both Georgia’s senatorial races will head for run offs.
On the other hand, in two heavily contested races, both Republican Senators Mitch McConnell (Kentucky) and Lindsay Graham (S.C.) held on to their seats.
Other factors that made this election unusual and polling more unreliable:
- An avalanche of mail-in votes. Many Republican state legislatures mandated the votes could not be counted until election day, which delayed results. Generally, because of the pandemic, Democrats preferred to mail their ballots while Republicans went to the polls. This accounted for the late turnaround in some states.
- Unprecedented voter turnout. Pollsters often use voter files to select their samples. The logic is to include people who have voted before and those who are eligible to vote. In 2016, for example, far more working-class, rural white men turned out and voted for Trump than was expected based on their previous track records. This year had the highest voter turnout since 1900, which muddied the results. Many people were first-time voters.
Polling itself has new challenges. Landlines, for example, once the lifeblood of pollsters, are vanishing. And those who have them tend to be older, skewing the demographic. Just finding respondents is becoming more difficult.
Other factors that influence polls:
- Voter preference can change between the poll and the election. There is a time lag between voters telling the pollster which candidates they will vote for and their voting behavior on election day. People may read new information or the candidate might do something to change their mind.
- Voters provide socially desirable responses in polls. Poll respondents often provide answers that they believe show them in a positive light to the interviewer. In 2016, many voters were reluctant to admit support for Trump because of his erratic behavior, even though they agreed with him on issues. This behavior may have repeated this year.
- Intentional misleading. Some participants that do agree to be interviewed may not trust the questioner and may provide misleading responses that cause pollsters to arrive at the wrong conclusion.
In the end, you have to ask yourself, why do the polls matter at all? This isn’t a horse race – no one is betting on the winner in advance. All that really matters is the actual vote.