On Wed, 9 Jan 2008, Doug Henwood wrote:
> Also, it is very hard to predict likely voters in these nonpartisan
> primaries. Partisan affiliation is an important predictor in general
> elections, but where Dem-leaning independents can vote for McCain, all
> bets are off.
Is there any way to read off if that last factor played a significant role? Because I was wondering if Obama might have been a victim of weird feedback loop -- i.e., independents who wanted the winners to be Obama and McCain, reading that Obama was ahead by double digits and McCain was in doubt, decided they'd vote for McCain. That would have been a rational thing to do (assuming you are this odd kind of double independent creature who thinks like this). And it could theoretically have created a large drain of Obama votes.
I assume someone else must have thought of this first and dismissed it for a good reason, but I haven't run into it yet.
Michael