As this election season drives forward in the mad dash to May (and then November), please keep one thing in mind when you read article after article on so called "public opinion polls."
Take a look at who they are sampling.
For example, the Elon Poll garnered hundreds of media hits over the weekend and into this week on their supposed election poll. But a closer look at their sample reveals, they survey North Carolina households, not necessarily registered voters or much less likely voters. Thus, their "poll" of supposed voter opinion factors in the opinions of a large percentage of people whose opinions do not matter cause they aren’t registered and/or don’t vote.
In 2004, 85% of the voting age population was registered to vote and only 64% of those voters actually voted. Thus, only roughly half of the voting age population voted in the last Presidential election.
Therefore, since the Elon Poll surveys random households, we can guestimate that their survey is about half right. Or to put it another way, they could be equally right or wrong in their survey results. A 50% margin of error is not good.
Yet, the press continues to report this poll as a reflection of voters’ intent, which is completely misleading. (See this AP story and very misleading headline that ran in tens of media outlets).
Many criticize phone or automated polls like the ones conducted by Democratic firm Public Policy Polling as inaccurate based solely on the technology they use. Say what you want about their technology and the "science" of weighting responses, but at least they are sampling the right audience. I’ll believe the results of their poll over Elon’s any day of the week.
Our monthly DecisionMaker poll is set to be released on Wednesday, but keep in mind that we sample likely General election voters. And as such, I’ll be the first to admit that our sample is not an accurate reflection of Primary voters and questions regarding the primary ballot should not be taken as such. While it may be more accurate than Elon’s since we actually sample voters, it is not as accurate as a poll of likely Primary voters.
The focus of our poll is on issues, not the horserace, and how those issues will play out in the General election. So when you see results for the GOP or Democratic gubernatorial primary, please remember that it is a sample of General election voters. We put those questions in there mainly because the media is obsessed with the horserace and not issues and frankly, we like the media attention.
I will say, however, that I think our General election polling matchups are pretty spot on.
We’d love to have you out for the poll lunch and release if you are available. 12 noon on Wednesday at the Clarion in downtown Raleigh. For more information or to sign up, do so here.
Leave a Comment