Polling for Texas: Invitation to Disaster

Pollster does a masterful job of highlighting the different demographic makeups of various Texas polls this go-round. Add to the mix, Ron Brownstein's column on some overall trends in Democratic Party makeup. From the vantage point of Harris County in particular, it seems entirely believable that Hispanics will be swamped a bit by a boost in African-American turnout, which is just exacerbated a bit by the boost in turnout among the multi-degreed/professional class.

That seems like bad news for Survey USA (which has had a generally good rep this season), Rasmussen, and Insider Advantage. And it might mean that ARG is worth looking more closely at despite their sometimes awful misses. I'm not sure what to translate that into if you want to dismiss the ones with the odd samples - it's just too easy to dismiss bad polls and accept good polls in this state based on things like this.

For what it's worth, ARG shows an early Obama peak followed by a more recent Hillary comeback. Granted, nobody's writing that story. All we get from the press is that it's their seeming obligation to note that Obama has a double-digit win streak and Hillary doesn't. Nevermind if it may or may not be relevant to the race at hand.

Of course, caveats abound ... it's a big state and there's no simple, easy way to identify if there's a pro-Hillary boost in South Texas (or anywhere else) that compensates for the Obamania we see in the urban counties. Is anyone writing about whatever sentiment may or may not be prevailing in East Texas? There's a number of viewpoints to this campaign that the massive media onslaught (welcome though it always is) has simply bypassed.

All that to say this: hurry up and toss the ball so we can all jump. One of these days, though, someone ought to figure out how to do a reliable job of polling statewide in Texas.