Three Reasons Why O'Donnell won in Delaware

I realize that tomorrow morning many Delaware Republicans are going to wake up with a headache. Although they're not Blue Dog Democrats, the continued erosion of moderate Republicans against Tea Party candidates is an interesting phenomena that mirrors left-challenges to conservative Democrats. As an Delaware-adoptee, here's my immediate reaction to this result...

Reasons Christine O'Donnell won:
1. Closed Republican Primary
There were 182,796 people eligible to vote in Delaware's Republican primary today, yet 57,582 actually did. I suspect that the percentage of leaning Republican identifiers (i.e., those not 'registered' with the Republican Party, simply those moderates who support Castle) WOULD have voted in the Delaware Republican primary if the election was open. The people who did come out: the Tea Partiers, who feel intensely about their candidate.

2. The Electorate
Given that the people who usually vote in Republican primaries are the most politically participatory, it might not be surprising that more Republicans describe themselves as being more conservative. Christine O'Donnell won because she earned the support of the Delaware Republicans most conservative, plus maybe a few moderates. Mike Castle had to get out the vote among more traditional Republicans who may be a little less likely to vote. That could be the difference between 3,500 votes sometimes.

3. The Geography
A few interesting observations from Congressional Quarterly:

Delaware is deeply divided ideologically with its two southern counties Kent and Sussex greatly more conservative than New Castle. 55% of the state's registered Republicans are located in New Castle, but only 50% of the votes in the last comparable Republican primary were cast in New Castle. If Tuesday's vote mirrors past elections, 15% of the rest of it should come from Kent and between 30-35% from Sussex.

Castle couldn't overcome O'Donnell's early lead. Why? Because the early vote was more likely coming in from New Castle County (50%) of the electorate, who were voting for O'Donnell marginally over Castle. When the more rural, small town, conservative Kent and Sussex county votes came in later - it could only buttress O'Donnell's lead.

In short, for Castle to have pulled it off, he would have needed a better GOTV operation in New Castle County. In all likelihood, it's where he's MOST popular, and could make the easiest sell. I think polling this race would have been enormously difficult, because pollsters simply didn't know the electorate. If Castle's ground game was better, PPP would have got their projection completely off. Fortunately for them, for O'Donnell, and in all seriousness, for the Democrats, they're weren't.