What are the polls really telling us this year?

The BDN Opinion section operates independently and does not set news policy or contribute to reporting or editing stories elsewhere in the newspaper or on bangordailynews.com

Matthew Gagnon of Yarmouth is executive director of the Maine Policy Institute, a free-market policy think tank based in Portland. The Hampden native previously served as senior strategist for the Republican Governors Association in Washington, DC

One of the stories – some would say a scandal – of the 2016 and 2020 presidential campaigns was how bad the polls were, particularly around Donald Trump.

In state after state, particularly in swing states, polls underestimated Trump’s expected performance against Hillary Clinton and Joe Biden compared to his actual performance on Election Day. This polling error in 2016 ultimately led to wildly inaccurate predictions of an almost guaranteed victory for Clinton. In 2020, polls showed Joe Biden with a comfortable lead in most swing states, yet he only won the election’s three tipping point states by 0.23 percent (Georgia), 0.31 percent (Arizona), and 0.63 percent (Wisconsin).

There is something very wrong with the polls, but interestingly it seems to be particularly pronounced when Trump is on the ballot and less so when he is not on the ballot. In the 2018 and 2022 midterm elections, for example, the poll numbers were halfway decent. There are many theories about what caused the so-called “Trump effect” in polls, with “non-responsive bias” — the historically low response rate from Trump supporters that pollsters are trying to achieve — being the most compelling explanation.

And yet this particular problem was obvious and much discussed after the failure in 2016, causing most pollsters to re-evaluate the way they conducted their polls in an attempt to get it right. After doing well in 2018, it looked like they might have figured it out. But then 2020 happened and it was obvious that all their efforts were in vain.

This has left pollsters wondering once again what they can do to correct the mistakes and get it right this time. This has led them to change the way they collect data, weight survey responses differently, and try to use new tactics to survey specific groups.

But did these changes fix polling? Or will the surveys be wrong again in 2024?

If you look at the current swing state polls, this is an important question. If they don’t have the polls fixed, Trump should win a relatively easy election in two weeks. If they actually sort things out, this will be another very close election.

Let’s see what’s going on.

As of October 23, the RealClearPolitics aggregate polling average in the state of Pennsylvania shows Trump leading the state by 0.8 percent. He has been leading in six of the last seven polls in the Keystone State and in seven of the last 10 polls. Still, a 0.8 percent lead would suggest it is very close and we could expect a repeat of 2020 results.

However, if you look back in time, as of October 23, 2016, Hillary Clinton was ahead of Trump by 5.6 percent in the RealClearPolitics average. In 2020, Joe Biden led Trump by 5.1 percent. Trump significantly outperformed the polls in both years, winning Pennsylvania in 2016 and losing by just 1.17 percent in 2020.

If the Pennsylvania polls have been corrected, then all we see is an accurate reflection of public opinion in a state that has been remarkably close in the last two presidential election cycles. However, if there continues to be some kind of “Trump effect” that causes the polls to be somewhat in favor of the Democrats, then there will be a shift towards Trump.

The same story shows itself in every single swing state.

At this point in 2016, Clinton was ahead by 11 points in Michigan, according to the RealClearPolitics average, while Biden was ahead by 8.1 percent in 2020. Trump currently has a lead of 1.2 percent. In Wisconsin, Clinton was up 6.2 percent and Biden was up 5.4 percent, while Trump is now ahead by 0.4 percent. Arizona saw an increase of 1.5 percent for Clinton and 2.4 percent for Biden and is now showing an increase of 1.8 percent for Trump. In Nevada, both Clinton and Biden rose 5.2 percent, while Trump was now up 0.9 percent.

You can also see this in North Carolina, where both Clinton and Biden were leading, but Trump is up 0.4 today. This also applies to Georgia, where Trump is currently up 2.5 percent.

Trump currently has a lead in every single swing state, which is usually quite narrow. The question remains: Did the pollsters “correct” the polls and is what we see now accurate? Or is there still a problem causing pollsters to underestimate Trump’s performance?

The answer to this question – and the impact it has on the outcome – will only be known after Election Day.

Less than two weeks left.

Election Notice: As of Wednesday, October 30th, the BDN will no longer accept letters and columns related to the November 5th election. Not all submissions can be published.