It’s now en vogue to criticize someone else’s polling as biased. It’s the easiest way discredit a pollster’s results, especially when you don’t like them. We’ve done it here, here and here. Civitas has done it here, here and here. Under the Dome here; Progressive Pulse here; and Capital Beat here. There nothing wrong with a strong critique of polling questions. In part, that’s why this blog exists. But sometimes I think we may go too far, and end up missing the story.
Claiming bias is also a good way to avoid having to analyze polling results. Just cast the results aside as biased and you’re done with it. The N&O fell in that trap yesterday when discussing our recent Raleigh survey concerning Dorothea Dix Park. Ryan Beckwith makes the observation:
A follow-up question asked if voters would support leasing a smaller section of the park, but it's tainted by wording of the "historic core" that is "crucial to the success of a future destination park."
Fair enough. But he doesn’t report the results of the question, and most importantly misses the fact that while he considers the wording tainted, it is still true. Sometimes the truth is biased and that’s not a bad thing. Bias is also a matter of opinion. Some may consider a question biased, while others disagree. Some polling questions are also intentionally biased as a scientific way to measure the strength of an argument.
I tried to avoid simply claiming bias when discussing the most recent poll released by the NC Association of Realtors. The poll was biased, no doubt. But instead of claiming bias and walking away, I tried to point out instances when the question wording was not just bias, but factually inaccurate or misleading.
I didn’t further analyze the results of that poll on this blog, but there are lessons to be learned from the results. And in counties that will now face a transfer tax referendum, I suggest people look at these poll results for insight into how their pro-and-anti-transfer tax campaigns should be run.
Even in the face of bias we shouldn’t be scared to analyze, there is still something to be learned. Factual inaccuracy, however, is a different story.