Yes, the polls show Hillary reasserting a slight lead here. But I've now heard former Texas Rep. Martin Frost, a one-time DCCC chairman who knows Texas demographics about as well as anyone, make the following case a few times:
1.) Don't overlook early voting, which has been significant here. (More than one million people voted early in Texas, versus only 800,000 who voted here overall in 2004.) Unlike the February 5 states, the demographics of early voting seem to favor Obama. That means a chunk of his now-vanished lead in the polls has been locked in.
2.) Frost argues that decades of attrition have reduced the Texas Democratic Party to a kind of liberal rump-party. If that's true, then the Democratic primary electorate could be more favorable to Obama than a lot of us have anticipated. True, the primary here is open, which means more than just the liberal rump will be voting. But, even so, you have to figure the non-Democrats who cross-over will tend to favor Obama. Hillary hasn't been attracting many Republicans and Independents.
Last week, I thought Obama's performance among white voters was unrealistically high in several polls. But Frost's insight suggests those polls may have been accurate. If Obama gets within 5-7 points of Hillary among whites, which he did in the latest SurveyUSA poll, for example, I think he'll win the primary. (The polls showing him trailing in recent days typically put him down 20 among whites.)