Kerby Anderson

Before Thanksgiving and Christmas push the election from our minds, we need to remind ourselves how wildly inaccurate polling organizations were in 2020. We should recognize that political polling has always been an inaccurate science.

Kristen Soltis Anderson reminds us that in 1936 Literary Digest declared Alf Landon would be the next president. The Gallup organization that year got it right only to later be embarrassed by being on the wrong side of the “Dewey Defeats Truman” upset.

Four years ago, the pollsters did not predict that Donald Trump would defeat Hillary Clinton. We were assured that they caught their mistakes and would be more accurate this time. A few hours into election night, it was obvious that these pollsters were wrong not only about the presidential race, but the races for the US Senate along with the races for state legislatures.

Back in 2012, the polls were off because pollsters didn’t call enough cell phones. Then in 2016, they were off because they missed voters without college degrees. This time, the critics said they were missing the “shy Trump voters.” But in a previous commentary, I mentioned that most pollsters denied this group even existed. And the argument might work for the presidential race but can’t really explain why there was no huge blue wave for the House, Senate, and state legislatures.

Another argument being used is the Trump factor. He is a polarizing figure in American politics. That may explain why the presidential polls were accurate in Maine and South Carolina, but it fails to explain why the polls were so far off in the Senate contests in those states.

While the pollsters are trying to figure all of this out, I have a recommendation for you. Don’t trust the polls. Perhaps they will fix their survey techniques over time, but we don’t have to listen to them until they do.