Voters question the reliability of the polling system
Aug. 25, 2022
Since the 2016 election, the reliability of polling has been severely scrutinized and brought into question.
Michael J. O’Neil, PhD of O’Neil Associates Inc. discusses polling issues for the governor’s race and addresses the accuracy and response rate of recent polls.
“I’ve been doing this for forty years and we used to get over 80% of the people that we called to respond, if we were persistent and if we were professional,” O’Neil said.
That number, according to O’Neil, has been dropping one or two percent for the past forty years due to the lack of trust in institutions, journalists and the government.
While many people now use cellphones instead of landline phones, O’Neil said that this does not play a huge role in the fact that survey response rates are decreasing.
“It was a lot easier when you had landline telephones that were geographically based. It’s a tricky element, but it’s not an insurmountable one,” said O’Neil.
The real problem that exists is the public’s unwillingness to respond to polls. Now, it’s virtually impossible to reach the same level of accuracy that poll surveys once achieved years ago due to the shortage of people replying.
Concerning the GOP primary race in Arizona, recent polls have demonstrated varying results with Kari Lake ahead in some and Karrin Taylor Robson taking the lead in others. Conducting an accurate and representative poll is expensive and hard to manage with so many people refusing to participate, which is a big reason why many polls are showing different results.
To create a precise and accurate poll, O’Neil recommends starting out with a group of representative people and continuously pursuing them. This cannot happen overnight, but many news media outlets oftentimes want immediate results.
O’Neil also touches on the problems with margin of error and how it can be misconstrued due to other errors in the poll, such as the wording of questions or whether or not people are replying truthfully.