“Things are seldom what they seem. Skim milk masquerades as cream.” –Gilbert and Sullivan A reputable GOP consultant who is not engaged with either gubernatorial campaign has concluded after looking at the crosstabs of the Rasmussen poll that the poll’s conclusion that Hutchison is leading by Perry by 2 points is unreliable. I have seen the crosstabs, and they are weird. The only question that matters is who are you for? This question has two columns for respondents. One column is for “Republicans.” The second is for “Other.” Other? Who is “other?” It appears that Rasmussen did two telephone surveys on September 16. One was to 790 “likely” Republican primary voters. The other was an unknown number of calls to an undefined group of recipients. I put likely in quotes because I hear from pollsters that affirmative responses to a question about whether a respondent will vote in a primary is notoriously unreliable. Most people, if asked, will say that they intend to vote. Here is the result among likely Republican voters: Perry 44% Hutchison 39% Medina 2% Not sure 15% Now, here is the result among “Other” voters Perry 28% Hutchison 41% Medina 5% Not sure 25% This is more or less what you would expect from Democrats and independents — almost a 50% edge for Hutchison and a lot of undecideds, since they don’t know what the final lineup of candidates is going to look like. Oh, here’s something else from the crosstabs: “Other” voters approve of Obama by 2 to 1. Presumably, the 40-38 edge for Hutchison is the result of weighting the result of the “Other” voters in comparison to the likely primary voters. I can’t begin to come up with an explanation for why Rasmussen would produce a poll that is essentially worthless. This is entertainment, not mathematics. My consultant source says Rasmussen wants people to subscribe to his service. But as soon as any political pro looked at this, they would know it is statistically unreliable. The only defense of Rasmussen I can think of is that the pollsters believe that there will be an influx of non-Republicans (and Republicans who do not typically vote in primaries) into the Republican primary and is making an allowance for this. A number of commenters to this blog have criticized the Rasmussen poll, although their criticism has not been accompanied by much, if any, analysis. My response has been that Rasmussen is well regarded in the polling world and I am not going to look behind the result of polls. I think that is the right course to take generally. But in this case, I have to say that Hutchison’s lead is a mirage and that Perry has a five point lead. That means his ten point lead in July has been cut in half. Here’s a question: The Perry folks had to know that the Rasmussen poll numbers were substandard practice. Why didn’t they question them? Do they want Hutchison to think that she is in the lead? Did they not want to show concern? This poll was easy to disparage, but they did not send out a response to the media — at least, I know I didn’t get one.