Let me take the results of the ABC/Washington Post polls in Texas and Ohio and the terrific graphic in Sunday's print Post and try to estimate the delegate count in the primaries if the vote splits the way it does in the poll. Statewide, the poll showed Clinton ahead by the statistically insignificant margin of 48 to 47 percent.
Texas allocates 127 delegates by proportional representation within the 31 state Senate districts (with a 15 percent threshold, which no one but Hillary Clinton and Barack Obama will meet). I have Clinton coming out on top in the delegate count by 64 to 62.
That's a pretty dicey count. The poll has Clinton ahead 55 to 38 percent in East Texas, but the four districts entirely in that region have four delegates each, and that split, just barely, leaves the delegate count in each at 2-2 rather than 3-1 Clinton, which she would get with a couple of extra points. I have Obama leading in the two districts with the most delegates, 5-2 in the 13th (Houston blacks) and 6-2 in the 14th (Austin liberals). The count in the latter could be less if Austin Latinos vote heavily for Clinton. I have several four-delegate South Texas Hispanic districts (19th, 20th, 21st) going 3-1 Clinton on the basis of the poll's 61-to-35-percent South Texas split; but if she falls a couple of points short, the districts split 2-2.
The bottom line: If the popular vote in Texas splits about evenly, either candidate could end up with maybe a 10-delegate edge if she or he got lucky in her or his percentages in certain districts, but it's likely to get much closer than that. A close popular vote Clinton win gets her no significant distance closer to pledged delegate parity. And of course Obama looks likely to be better organized to score in the caucuses that will be held later that night. Clinton needs to improve on her current standing in Texas to get her closer to the nomination. And a popular vote loss here looks like curtains for her campaign.
Ohio allocates 49 statewide delegates by proportional representation and 92 delegates by proportional representation within congressional districts. The WaPo poll shows her ahead statewide by 50 to 43 percent. That gives Clinton a 46-43 delegate lead.
In the race for district delegates, I show her ahead 50-44. The poll has her about even in the southwest (Cincinnati and Dayton media markets) and northwest (Toledo media market). All but one of the districts there have even numbers of delegates; I gave Clinton a 3-2 advantage in the 3rd District (Dayton). It could easily go the other way. In the central area (Columbus media market and southeast), which the poll scores as even, only two districts have odd numbers of delegates; I gave those to Clinton (12th Columbus, 18th rural), but with no great conviction on the 12th, which has lots of black and university voters.
The poll gives Clinton a 51 to 39 percent advantage in the northeast, the most Democratic part of Ohio, but doesn't make clear whether that includes Cuyahoga County (Cleveland), which is 28 percent black. I gave Obama only a 5-3 edge in the east side black-majority 11th District (where Rep. Stephanie Tubbs Jones has supported Clinton); it could easily go 6-2. The three other northeast districts with odd numbers of delegates are unlikely to net Clinton more than a one-delegate advantage each.
The bottom line is that a 7-point advantage in the poll (50-43) seems likely to give Clinton a 6 percentage point advantage in delegates (76-67). That's nine delegates, a long way from overcoming Obama's current lead of 99 delegates (1,374 to 1,275) in realclearpolitics.com. The Democrats' proportional representation rules, combined with determining so many delegates by congressional districts, make it hugely difficult for any candidate who is behind in a close delegate race to forge out into a lead. Nobody seems to have thought through the consequences of having allocated so many districts an even number of delegates: In a close race, voters in those districts have almost no chance of casting a decisive vote. Perhaps some trial lawyer will bring a suit on behalf of Democratic voters in districts with an even number of delegates, arguing that they have effectively been shut out of the presidential nominating selection process.
I await the outcome with interest. I have long argued that papers presenting government officials with policy options should have an even number of options. Otherwise there is a middle option, the second of three or the third of five, which is obviously the Goldilocks option, the best of all possible worlds. Providing policymakers with an even number of options, two or four or six, requires them to make a real decision. In the Democratic delegate selection process, voters in districts with an even number of delegates are effectively disenfranchised. My proposal would be to require all districts to have an odd number of delegates, so that every Democratic voter would know that her or his vote counts.