Texas, Ohio, Rhode Island and Vermont vote this Tuesday, March 4 to allot 370 pledged delegates in the latest round of the Democratic primary tussle. Tandem wins by Barack Obama in the big states of Texas and Ohio would end Hillary Clinton's effective chances of capturing the nomination and cause her withdrawal from the race. Twin victories for Clinton would revive her cause and extend the campaign at least seven weeks to the April 22 contest in Pennsylvania. A split decision might leave the door open to either scenario, but Obama would very much retain the upper hand. What lies in prospect?
In Texas Obama has caught up among likely voters and now leads an assortment of polls by an average of 1.2%. That's a razor-thin margin but the trend is clearly for
Obama. Clinton had an average 10-point edge two weeks ago. The delegate apportionment in Texas also favors Obama, since it's based on the Democratic vote in 2004. The African-American districts, where Obama is strongest, turned out in much greater numbers than the Hispanic districts where Clinton is most popular. That means he'll gain extra delegates in his base areas. There is also a caucus procedure that takes place after the balloting on Tuesday to choose about one-third of the total delegates. Don't ask me why; I have no idea for the reasons behind such an unusual system. Whatever the cause, Obama has an army of 155,000 volunteers in Texas and has demonstrated superiority in the caucus format the entire primary season. He's outspent Clinton on media buys about $10 million to $5 million too. Both candidates will spend all day here Monday. With all this considered I feel Obama will garner at least 104 delegates to 89 for Clinton in Tuesday's big prize, even if his percentage win is only 1%.
Ohio has 141 pledged delegates at stake. Clinton is stronger in this heavily white working class state, and leads the poll averages by 5.9%. That includes a Columbus Dispatch survey that has Clinton up by 16. That survey polled registered voters rather than likely voters, an important distinction, and is so far off all the other results I'm inclined to throw it out. The averages of the other six polls have Clinton up by 4.2%. Obama has narrowed an earlier 16-point deficit here, so he is probably closing to within this margin now. Obama has spent about $5.3 million on media buys in Ohio to Clinton's $3 million. Both candidates are stumping in Ohio all day Sunday and will leave things to their surrogates tomorrow. Clinton will probably hold on to win narrowly in the Buckeye State. A charitable delegate outcome for her would give her 74 to Obama's 67.
The two smaller states look like easier picks. Rhode Island gets 21 delegates, and an average of three polls gives Clinton an average of a 10.7% lead in the Ocean State. If that holds give her 12 delegates to Obama's 9. Vermont appears strong for Obama. Two polls give him an average lead of 19%. He should get 9 delegates to 6 for Clinton.
The upshot of the four states looks to be a close win for Obama in Texas and a strong one in Vermont. Clinton should get a squeaker in Ohio and a more comfortable victory in Rhode Island. My estimate, which is pretty conservative, projects Obama to add 189 delegates to his total and Clinton to add 181 to hers. This looks like a pretty even split, and so it is. Clinton's problem is that she already trails by 155 pledged delegates and needs to make up ground, not fall farther behind. After March 4 there are only 611 pledged delegates left in 11 states and 2 U.S. territories. If Obama takes 189 Tuesday, Clinton would need to win 63% of these remaining pledged delegates to catch up. Given the Democratic Party's proportional distribution system it is highly improbable she can achieve that. Unless she rather strongly exceeds the performance I've outlined in this post you can expect to start hearing more frequent calls from Democrats for Clinton to suspend her campaign for the good of party unity beginning Wednesday.
2 comments:
Given the gross inaccuracy of the polls so far this election season, averaging X number of polls proves what? Is it your theory that you can improve their individual accuracy by averaging?
Good question, Ed. Yes, that is the theory. Standard survey theory holds that the bigger the sample the more accurate the poll is likely to be; that's how they come up with that "margin of error" figure that's associated with each poll. Going one better than this is to average the different whole polls too, to try to average out methodological variances. That's the practice made popular by Real Clear Politics, the current "in thing" in political polling.
As for accuracy this year, can you think of a single state after the second one (New Hampshire on the Democratic side) where the polls did not correctly predict the winner? I can't. I accept your point that we can't take what they say as settled fact, especially after they got egg on their faces in New Hampshire. But after that they have proven pretty reliable once again.
Post a Comment