What is there to say about the third Chronicle poll, which shows Senator Kay Bailey Hutchison winning by a 61-27 margin over Barbara Radnofsky? Well, that poll was done by Zogby, which as it happens just released its latest batch of Zogby/WSJ Interactive polls. That poll pegs it as Hutchison 55, Radnofsky 36. Zogby says his interactive polls are valid, so it'd be nice to see him explain the differences between the two.

That same Interactive poll, by the way, has the Governor's race as Perry 36.7, Bell 28.5, Strayhorn 15, Friedman 14. Like I said, it'd be nice to have Zogby explain why one set of results is so different from the other.

As with the CD22 poll, 90% of the respondents said they were "very likely" to vote, with the rest claiming to be "somewhat likely". Either this means they also talked to a bunch of non-voters, whose results they then excluded, or it's as much BS as the CD22 sample was. Whatever the case, taking the respondent's word for it is highly suspect. You could screen by past voting history instead, for instance.

One interesting thing from the crosstabs, which are in popup windows that I can't link to: The respondents were asked whether the US was going in the right direction or the wrong direction. By a 47.9-43.7 margin, they said "wrong direction". Hutchison gets nearly 80% of the "right direction" voters, but Radnofsky wins the "wrong direction" folks by a slim margin, 41.6 to 40.9 (the Lib candidate gets 8.9% of these, with the rest being "other" or "not sure"). I didn't see where they asked the same question for Texas, but they have data on that. People are more sanguine about Texas, with 52% percent saying "right direction" and 40% saying "wrong direction". KBH gets 75% of these "right direction"ers, with Randofsky taking 45% of the "wrong"ers (to 39% for KBH). It'd be useful to know which of these groups is really more likely to vote, but alas, we don't.