Karl-Thomas Musselman analyzes polling out of Texas on a Senate District by Senate District basis (which is a little imprecise) and comes to a startling conclusion. Even with a 49%-41% state-wide lead, Hillary Clinton is poised to come out of Texas with a net loss of six delegates. At least, she stands to lose six delegates in the primary portion of the Texas contest. Texas will also conduct a caucus on the same day as the primary, and that will be factored into the final allocation of delegates. Musselman’s conclusion is based on delegate splits. In regions where Clinton is polling ahead, there are very few districts with an odd number of delegates available, and the opposite is true in regions where Obama is leading.
Musselman makes a few other interesting observations. There is an interesting split between voters that are planning to vote early (Obama 46% Clinton 42%) and those that plan to wait until election day to cast their vote (Clinton 51% Obama 40%). This odd distinction can probably be explained by differential enthusiasm. Obama voters are more motivated, even if there are less of them. Yet, that greater motivation that has served Obama so well in caucus states could wind up hurting Obama in Texas’ caucus. Why?
The caucuses begin at 7:15pm, after the primary vote is over. People that voted early may be less inclined to show up than people that voted the same day. (To learn more about the complex system in Texas, go here and here).
Overall, Clinton’s problem is that she needs to get 62.5% of the vote in her areas to get extra delegates, while Obama has a lot of opportunities to pick up delegates with a bare plurality of the vote in his areas (primarily Austin, Dallas/Ft. Worth, and Houston). I don’t know if Clinton can do better in the caucus portion, but there are some troubling signs. The Obama campaign is more motivated and has a better ground game.
Obama is going to build an initial lead in early voting, unlike in California where early voting strongly favored Clinton. And as long as he can keep it close in Hillary’s areas, he will come away with a delegate lead in the primary portion. Recent history suggests that he’ll do better in the caucus portion, too. Since this is the lay of the land with Obama polling behind by 8 points, imagine the situation if he narrows the gap?
Hillary might win Texas, but her chances of significantly narrowing Obama’s lead in pledged delegates is not good. Perhaps Ohio will offer more fertile ground.
yes, yes, the dual TX prima/caucus is the reason I’ve said Texas will surprise. As noted the math does not favor Clinton.
And There’s time for Obama to close the 8% gap.
Time = advantage, Obama.
New Rasmussen puts her up 54%-38%.
Thanks BooMan. that’s not a typo – but who to believe?
I read a different poll over at TPM:
ain’t over ’til it’s over….And the fat lady sings…”something is happening here”
who to believe? Maybe this is also a typo?
We have another poll, today Feb. 15 2008, from ARG: Texas Primary Preference seems in line with that IVR poll.
Obama 48%
Clinton 42%
Hmmmm. I’ll try to ignore.
One of the things that gets frequently overlooked by the lay public (and talking heads, too) in polling data is methodology — particularly the distinction of “registered” (voters, Dems) and “likely” voters.
Another source of polling variability, which interacts with the likely v. registered sampling categories, is the “open” and “closed” primary. Open primaries are difficult to forecast largely because independents can participate in either party’s primary. This also compounds yet another large source of variability — turnout.
There are other problems, too, such as wording differences and data collection techniques (e.g., automated polling, etc.).
I doubt people are interested in all this. Given the high inherent variability, a good rule of thumb is to roughly average out the different polls, as well as expecting low predictivity.
Fascinating. Bizarre, but fascinating. Is there some part of Texas’ delegate apportionment process that involves drinking shots of tequila lined up at the bar at two in the morning?
Texas. It’s a whole other country.
I always hate to get my hopes up about Texas because … well, it’s Texas. But this certainly looks promising.
I think the John Lewis endorsement might be the final straw. This is a lot bigger than the media is letting on.
And you have to love the subtle racism in this NYT headline:
Tilting, is he? Interesting choice of words. I might have used a more obvious verb, perhaps “switches” or “changes.” But the Times sees as Lewis tilting. Do they mean “leaning?” No. Lewis has stated that he’s changed his position; he’s not “leaning” in a new direction. This leaves us with only one other possibility:
Yes, friends, Lewis is doing Sancho Panza to Barak Obama’s Don Quixote.
And, according to the Times, we, too, are tilting at windmills.
The liberal media strikes again.
p.s. Has image embedding been disabled for this website? I tried like five different image-hosting services and none of the HTML tags would work. Please forgive my ignorance if this an old question.
It’s hard to say what Lewis means. SEE: http://www.cbsnews.com/blogs/2008/02/15/politics/horserace/entry3837215.shtml
That ARG poll is interesting since the ENTIRE primary season they have been wrong and ALWAYS over estimated the ClintonS lead.
That poll with Obama behind 6 is a good sign!
I dunno. I have this strong sense that whoever wins Wisconsin wins it all. No rational basis, just my sense of how things are working.
Ye old trendlines are not good. Ohio actually looks better for Clinton. I suppose they are fighting for Texas to show two “wins” so they can sputter on.