The Polls Are Still Getting It Wrong

The Polls Are Still Getting It Wrong

( – In 2016, the world was shocked when Donald Trump defeated Hillary Clinton. Virtually no one in the media saw it coming. No one except for those who were paying close attention to what the polls were actually saying and accurately reporting what the Trump campaign was doing.

Was it intentional? Were pollsters attempting to create a self-fulfilling prophecy to elect Clinton? Is it happening again as Biden has unrealistic polling leads over Trump?

It’s almost impossible to know if pollsters were trying to rig the 2016 election in Clinton’s favor. There was little question that pollsters oversampled Democrats, undersampled Republicans, and had other significant structural problems in 2016. Biden holds double-digit leads in some polls; are pollsters getting it wrong again?

According to some in the industry, they are — and they haven’t fixed the underlying problems from the previous election that led them to get it wrong.

Polls Now Hide Their Segmentations

In polling, segmentations are vital to understanding how pollsters arrive at their conclusions. In political polling, segmentations include political party affiliation, race, gender, and educational level. In 2016, they significantly oversampled Democrats, which led to a false sense of security that Clinton was the odds-on-favorite to win the election.

Logic would dictate if Trump has a 94% approval rating in the Republican Party today, he ought to be polling higher against Biden than he is. Additionally, if it’s true that excitement for Biden is lower among Democrats, how would he poll so high and Trump so little?

There are a couple of factors at play. Party affiliation is one factor. In the most recent polls included in the Real Clear Politics average, Biden has a 9.5 point advantage. However, by linking to each survey and looking at each one’s sampling, they are not telling us their political affiliation sample sizes any longer.


It makes one wonder and question their credibility. If pollsters are oversampling Democrats, again, then it helps explain why the polls are skewing so hard for Biden.

Elections Are Not Won on National Popularity

Another factor is that these are national polls. Presidential elections are not won on national popularity. They are won state-by-state in the Electoral College. Campaigns know this very well and poll by states but do not release their public findings. It explains why Trump’s 2016 campaign worked hard in the rust belt states that hadn’t voted for a Republican since Ronald Reagan in 1986.

No comprehensive polls are being conducted at the state level that is not campaign driven right now.

Pollsters Talking to Registered Voters, Not Likely Voters

There is a drastic difference between a registered voter and likely voter. A registered voter is someone who is just that, a registered voter. However, it doesn’t mean they vote because they are registered.

In presidential years, a 60% turnout is considered a high number. Likely voters are segmented by those who voted in the most recent elections and show a high probability that they will vote in the next election.

Pollsters have access to publicly available voter files in every state. Yet, current questioners are talking to registered voters and, therefore, skewing a potential turnout that is more indicative of how the vote could go for one candidate or another.

Pollsters Not Reaching Lower-Income Whites

Finally, and profoundly important, pollsters are not accounting for differences in education or people telling them the truth. In recent history, lower-income whites have shifted away from the Democratic Party towards Republicans, while upscale whites have gone over to the Democrats.

There are considerably more lower-income whites than upper income, which is significant. If pollsters are not reaching lower-income whites, the poll results will be skewed.

The combination of white lower-class voters not being accounted for and people not telling pollsters the truth is pointing polls to significantly lean in favor of Democrats.

The bottom line is that polls are opening themselves up to failure once again and the public not being able to trust them. This leads to other questions.

Do they have an intentional bias against Trump? Are they purposefully skewing the results?

It’s not clear, but their tactics leave open the door that it’s possible.

By Don Purdum, Freelance Contributor

Copyright 2020,