Maybe you noticed that John Edwards finished second in the Iowa district delegate vote count, but got one fewer national delegate than Hillary Clinton. Perhaps you were surprised that Hillary Clinton got more votes in New Hampshire, but the same number of delegates as Barack Obama. Surely you've read some Obama claims that he "really" won Nevada in delegates 13-12, even though Clinton got more votes. It's enough to make me flashback to the Electoral College map. So, what's going on here, and how does it relate to New Jersey? This is the answer as I understand it.
In Jersey, we've been told there is "a" primary coming up, but there are actually twenty primaries happening on February 5. Each of these primaries occur in a "Delegate District," which is two of the familiar "Legislative Districts" combined. For example, I'm in Delegate District Two, which is LD3 (Sweeney's) and LD4 (Madden's) combined. Each district elects three or four delegates to the national convention. Now the LDs have (roughly) equal population, but the districts that have fewer Democratic voters only get three delegates. Each of these are truly separate primaries, in the sense that if Hillary Clinton turns out a crushing victory in District 20, it has the same effect in my district as her votes in Los Angeles: None whatsoever. There's a total of 70 delegates at stake in these twenty elections.
Note that at the end of the day, outcomes like a 5% win or a 10% win in votes may not be reflected at the district level. In a three delegate district, I think there are realistically only two possible outcomes: Either Clinton-Obama-Edwards tie with one each, or one of the them gets two, one gets one, and the other gets nothing. A candidate gets nothing if "he" fails to get 15% of the votes (I don't think that will happen to "she"). A four delegate district, however, could realistically split 2-2-0 or 3-1-0 or 2-1-1. As the district results are added up, it's possible that just as in the early states, a small lead by one candidate is not reflected in the delegate totals. Obama's counterintuitive Nevada result was powered by a small victory in an odd delegate district while splitting even-delegate districts by not losing by too many votes.
The average of the last five NJ polls gives Clinton 44%, Obama 29%, and Edwards 10%. This state vote, of course, may vary widely in the individual districts (did you know Cape May is different than Hudson County?), but it seems to me that the other 17% (assuming they show up and vote for one of the big three) could thrust Edwards into viability, Obama into an effective tie in delegates, or Clinton into a big win. My candidate, Obama, clearly needs to get close enough to tie the four-delegate districts, thereby blunting the practical effect of a Clinton victory in the delegate counts, and better yet eke out some wins in the three-delegate "Republican" districts. Edwards is in a tough spot (pending, of course, the results in South Carolina), and his supporters face a difficult choice of trying to get to 15% viability with the risk of not getting any influence on the delegates, or casting a vote that may count between the other two but doesn't reflect their first preference. (If you disagree, tell me in the comments.)
Ah, but I said twenty-one primaries. Another 23 at-large and 14 "party leader" delegates are allocated based on the state-wide results. So here, we all count equally, and the current polls suggest a Clinton advantage, but the delegates will be divided up based on the votes. Once again, the 15% minimum applies. Another twenty delegates are not selected by voters (for example, Senator Menendez gets to be a delegate, and he's already decided on Clinton), but someone can explain the full process.
New Jersey, of course, will be just one of twenty-two states voting on February 5th, and by Democratic Party rules, none of the them are "winner take all." For an idea of how the district-based and state-based voting might turn out, look at poblano's Delegate Projections at dailykos. I sometimes notice online comments -- typically by Brits or Canadians -- mocking this process as undemocratic, but remember that in their countries voters get the same say in their national leader as we do in deciding whether Nancy Pelosi, Steny Hoyer, Rush Holt, or some other Democrat is Speaker of the House. (That influence, if you haven't noticed, is none.) Our bizarre process balances the developing national tradition of allowing voters some say in the Presidential candidate, and the longer tradition of leaving it up to party leaders.
