Here’s the full polling data for that KHOU/KUHF poll. Of greatest interest to me is the bit where they note that they asked people whether they’d voted in the November election; about 75% of them said Yes. Given that turnout for this election was about 20%, I find it impossible to believe that so many respondents in a random sample would have actually done so. Hell, even if they thought the question referred to last November’s Presidential election, this would be too high. But even if you believed them completely, why wouldn’t you then say “thanks for your time” to the No respondents, hang up on them, and then go to the next name on the list till you find another self-reported voter? What possible value is there in asking these non-voters for whom they plan to vote in the runoff? I don’t get it.
Rant aside, the rest of the sample seems reasonable enough. In fact, the Anglo/African-American/Latino ratio is very close to that of the Parker poll, which surely sampled more-likely voters than this one did. I’d say that lends credence to the hypothesis that lower turnout benefits Parker, while higher turnout benefits Locke. Anyway, it’s all there, including the Controller and At Large Council races, so check it out. Link via Robert Dahnke.
UPDATE: As sure as mosquitoes follow the rain in Houston, campaign press releases about polls follow polls. The Locke campaign has sent one out touting the fact that two out of three recent polls – theirs and KHOU’s – show him trailing by a modest amount. Of course, they resorted to the old “statistical dead heat” silliness, about which I will direct you here and here. By my calculation, there is about a 77% probability that Parker is actually leading the race, given the KHOU poll. Given that three different polls have her in the lead, I’d say the odds are greater than that.