From a source close to the Perry campaign comes this e-mail about the methodology used by Survey USA in determining that the governor's approval rating has slipped by 9 points since the previous poll. (See "Poll Axed," August 22, below). This will not be entirely new to dedicated readers, both of you, who will remember that I previously criticized the methodology of both the Survey USA and Zogby/Wall Street Journal polls along similar but far less sophisticated lines.
"Survey USA uses auto-dialers whereby they make calls into Texas until an arbitrary number of calls have been answered. This is obviously flawed and will not allow an accurate representative sample of the state. Using this flawed system, Survey USA apparently believes 50% of all Texans live in East Texas and the state is evenly divided according to party affiliation (37%-34%). Both of which we know are not true. Another example is that, according to the sample used in the poll, Survey USA believes that 18-34 year olds are just as likely to vote as Texans 35-54 years old and MORE likely to vote than folks over 55. This is simply not true. Also, this is a poll of adults, not likely voters or even registered voters, so their demographics are messed up. They also believe that the majority of Texans are pro-choice when polls consistently say Texas is solidly pro-life. Also, this poll swings wildly month to month, which is a telltale sign of problems with methodology. For these reasons, not many media outlets use this poll anymore."
My response: I agree with everything said here, but writing about bad polls is better than not writing about polls at all. The life of a blogger is hard and unappreciated, and I am grateful for anything to fill the space.
While I am giving the Perry folks their say, I should correct an error I made earlier in saying that Perry, along with Tom DeLay, could have avoided the mess in Congressional District 22 by calling a special election. That, of course, is wrong; the special election winner would have served only to fill out DeLay's term. In a previous post, I had said that if the special election had been called, a court might have allowed the winner, if a Republican, to replace DeLay on the ballot. While that can't be considered an error, it sure doesn't seem very likely, in retrospect. I withdraw the comment.
- 1 week