Are polling numbers reliable in elections?

Polling isn’t always the most accurate way to measure the Presidential Election, and the numbers can also be incorrect.

Politico claims the national surveys from the 2020 Presidential Election were the least accurate in 40 years.

Dr. Mike O’Neil joined “Arizona Horizon” to discuss the problem of the polls and the upcoming Presidential Election.

One of the problems of polling is that people aren’t willing to participate in the polls, which is affecting the accuracy. 

“One of the most important things is there’s a long term deterioration in people’s willingness to participate in polls. In the 70’s, in my dissertation research, I did a poll and got an 88% response rate. 88% of the people consented to be interviewed. We’ve been on a long, historical, multi-decade downturn in those kinds of numbers of 1 or 2% a year. Today’s polls are not to the same extent built on a foundation of mathematics,” said Dr. O’Neil. 

“We all have been harassed today with sales calls for awhile, and just the volume of interruptions [means] you’re just sick to death of the interruptions, and people to a very great extent got fed up, whereas in the old days, we would call somebody and say, ‘We’re doing a poll,’ and in comparative terms, it was easy. But now people are just fed up,” said Dr. O’Neil. 

In 2020, the polling seasons didn’t come out as great as people thought it would, and it was worse in 2016. 

Polls try to stay as neutral as possible for predicting the candidate. They try to pick the one who will win, and they want to be correct. They lean away from picking one side. 

“Everybody wants to be right. Even if they have a D or an R leaning, they would rather be right than pick one side” said Dr. O’Neil.