Quinnipiac polls the big three, uncovering both Democrats' usual weaknesses

Apart from today's fascinating primary polls, there are a number of other surveys that are worth looking at closely, starting with Quinnipiac's latest release of their polls from the general election "Big Three," Ohio, Florida and Pennsylvania. Quinnipiac brings mixed results for Democrats and finds Hillary Clinton polling better than Barack Obama, surprisingly so given that most recent polls have shown the latter outpacing her.

• In Florida, Clinton beats John McCain 44% to 42% but the Republican leads comfortably against Obama, 46% to 37%.
• In Ohio, Clinton distances McCain 48% to 39%. Obama edges out the GOP candidate 43% to 42%.
• In Pennsylvania, finally, Clinton leads 48% to 40% while Obama is ahead 43% to 39%.


• The first major difference between Clinton and Obama come from the level of Democrats who cross-over to vote for McCain. If Clinton is the nominee, 9% (Florida), 8% (Ohio) and 11% (Pennsylvania) of Democrats vote for McCain. If Obama is the nominee, those numbers go up to 20%, 19% and 17%.
  

(Update, in response to comments: The issue of cross-over Democrats is not the same as that of the polls that have recently showed that Obama and Clinton primary supporters would not support the other candidate in the general. Those polls show that roughly equal a proportion now say they would abandon either candidate and those who would go to McCain have generally been found to be independents voting in the Democratic primary, not registered Democrats. In these general election polls, however, the cross-over does not affect both candidates equally and these voters are registered Democrats, not independents).

• Second, Clinton is performing better among the white vote. She gets 38% of whites in Florida, versus 27% for Obama. In fact, McCain beats Obama 2:1 among whites in that state (54% to 27%). In Ohio, Clinton wins white women by 11%, while Obama loses that group by 9%.
• Also noteworthy is that Obama and Clinton come in with roughly the same numbers among men, and that all the difference Clinton creates is among women. Enough to say Obama's problem are white women Democrat?
• Obama, meanwhile, does slightly better than Clinton among Republicans and independents. In Pennsylvania, he leads McCain by 9% among indies while Clinton ties. The margin among independents is the same in Florida -- which is also significant given that Obama runs 11% worse overall.
• As for trendlines, both Democrats have improved in Ohio and Pennsylvania since the last poll in February (when McCain led both Dems in Ohio), but only Clinton improved in Florida.

As always, it would make little sense to draw overarching conslusions about which candidate is most electable based on these polls, except perhaps in the case of Florida as a consensus seems to be developing that Florida is a state in which Obama would struggle more than Clinton, just as there is a consensus developing that Clinton would struggle much more in Oregon, Colorado and Washington.

The general election polls that are coming in do not tell us much about who will win in November, but they are very instructive about what states we can expect to be competitive and, very importantly, what are the weakest points of both Democrats. And here we have been getting a fairly consistent result: Clinton significantly outpaces Obama among registered Democrats despite the fact that she usually is weaker among African-Americans (Quinnipiac did not release the breakdown of the black vote); she also tends to perform better among women. Her advantage among whites in these polls, however, is not something we are accustomed to; given other polls that show Obama putting the Wright controversy behind him, we should wait for other surveys to confirm these results (especially in Florida) to draw conclusions.

Despite his apparent problem with white Democratic women (remember New Hampshire...), Obama outpaces Clinton with equal consistency among independents, usually men (though not particularly in this poll) and gets more cross-overs from Republicans. Depending on which groups a poll has weighted more heavily (what proportion of Dems? of independents?), Clinton and Obama alternatively look stronger.

It is thus significant when one polling group comes out with surveys showing consistently better results for one of the candidates (Quinnipiac today for Clinton, Rasmussen earlier this week for Obama) for that confirms that the main difference between these polls is the turnout model the polling group projects. But pay attention to the internal numbers, for that is where the electability debate will play out: If Clinton is in any position to argue her way to the convention and the electability debate is actually debated in August, will she have improved her position among black voters? Will Obama remain weaker among registered Democrat? And so on.