Monday, February 11, 2008

Caucus Effect, Believe It!

Obama won again on Sunday in Maine, keeping his streak of impressive caucus victories alive.

Via Kevin Drum, comes this from Brendan Nyhan, which seems to conclude that the caucus effect may actually be mathematically significant.

Brendan examines a 5 conventional wisdom statements about Barack. 1) Caucuses are good for Barack, 2) Barack does well in states with especially high or low black populations, 3) Barack doesn't do well among Hispanics, 4) Barack does not do well in large population states, and 5) Barack does not do as well in high population states.

He finds that the first and second have some significant statistical support. The rest, while the evidence does support them, the correlation is much much worse. In fact, in the update he says he "wouldn't put too much stock in the results of any of these hypothesis tests because (a) hypothesis testing is riddled with epistemological problems and (b) it's difficult to achieve significance in small samples." So take it all with a grain of salt.

Also, this doesn't attempt to answer WHY he has the caucus advantage, just that it seems it exists.

2 comments:

bebedoc said...

I have seen commentary that the caucus effect may be due to his wider margins among higher income brackets and college grads, who are more likely to have the job flexibility to attend. I think it's also due to his enthusiastic grassroots campaign. Maybe because I'm young, but I have yet to meet anyone really "fired up" about Hillary

IMUnaware said...

Yeah. I am sure that some people are "fired up" for Hillary, but no one I know. I'd be totally happy with her being the Democratic candidate, but I'm not excited about her the way I am about Obama.

Grandma voted for Hillary because "all my life I've wanted to vote for a woman for President." I hear that. :-D