Implications of Gallup's Dem Defections

Two interesting points on the Gallup survey I linked to earlier, which finds that 28% of Clinton supporters would defect to McCain if Obama were the nominee versus 19% of Obama supporters who would do the same if Clinton were the nominee.

1) From Pew Research via Brian Schaffner via Mark Blumenthal, this type of threatend defection is not uncommon. In 2000, 14% of McCain backers said that they would vote for Gore rather than Bush. Two points here. First, 14% is a long way off from 28%. Second, and this is a point that Schaffner makes, Bush had over 7 months to recover from the diletarious effects fo the primary campaign. How long the Democratic nominee (likely Barak Obama) has to recover has a lot to do with when Clinton chooses to make an exit. If it is not until August at the convention, leaving Obama only 3 months to convince the base that they should come out to the polls, we may see a very different result.

2) More convincingly, Blumenthal points out that:

the Gallup analysis focuses solely on self-identified Democrats that say they vote in primaries. It does not cover to the ability of the two Democrats to attract independent or cross-over support from those who say they do not vote in Democratic primaries.

The implication of this is that while the Gallup poll shows a large difference in defections between Clinton and Obama to McCain (28% vs. 19%), the polling questions and methodology do not take into account the independents who would defect from Obama to McCain if Clinton were the nominee. In fact, Blumenthal goes on to note

Pew's Scott Keeter reported
that roughly equal numbers of voters are Obama-not-Clinton or
Clinton-not-Obama in matchups against John McCain. It would be
interesting to replicate those calculations using the Gallup Daily
data, although the fact that Obama gets 44% and Clinton 45% against
McCain suggests that the rough parity in these defector/cross-over
groups persists.

In fact, according to a new poll by NBC/WSJ, Obama may have an easier time bringing in Democrats than the Gallup poll suggests.

Because among Obama voters, Clinton has a net-negative personal rating
(35-43) while Clinton voters have a net-positive view of Obama (50-29).
Taken together, this appears to be evidence that Obama, intially,
should have the easier time uniting the party than Clinton.