Biased Poll? No Weh: Top Republican Demographer Validates Poll

On February 21 ("Shooting the Messenger") I wrote that Allen Weh accused me and my students of deliberately rigging a poll we took, showing Domenici to be the favored candidate at this early stage in the Republican primary for governor, with a very high undecided factor. Not content when these accusations made the front page of the Las Cruces Sun News, Weh's campaign then circulated an email to Republicans repeating accusations we had "grossly inflated" the results and adding another, that we might have "misused university resources" to the benefit of one candidate.

Up to that point I had decided to let the grapes sour where they might, but these new accusations of serious impropriety were so over the top I wrote him a letter suggesting he should put specific allegations of misusing public funds in writing and send them to my superiors for action. I also sent all of the 444 interview forms and the spread sheet we used to compile the poll to Senator Rod Adair, from Roswell, asking him to review our work for any signs of bias. Adair is one of the top two or three political demographers in the state, widely respected for his analytic capabilities, he is Republican, and he has not endorsed anyone for governor. Here is his answer:

"I have reviewed the full documented record of the poll of the Republican
gubernatorial field conducted by a New Mexico State University political
science class under the direction of professor Jose Z. Garcia.

The internal documentation of the mechanics of the poll is as solid and
complete as can be found for any survey of opinion, and is open for peer
review or public examination. The methodology for the poll both for sampling
and demographic weighting is without error. The poll is random, scientific
and valid by every statistical measure. I can see no evidence of any intent
to bias the poll in favor of any candidate."

Rod Adair
New Mexico Demographic Research
Sent from my BlackBerry Smartphone provided by Alltel

I received this email from Adair on Monday and sent it on to Mr. Weh. Subsequent polls are completely consistent with our findings. I have yet to hear a peep from the Weh campaign. Some of my students have wondered out loud how Weh could jump to conclusions, repeating outlandish accusations without a shred of evidence. I keep hammering at them to internalize the concepts of "validity" and "reliablity," explaining why you need to be careful in research. The politics of polling puts pressure on the pollster to defend the poll. In this case we did quite well. Perhaps the Weh campaign learned something too: you can try to shoot the messenger, but that won't make the grapes taste one bit sweeter.