PA is getting harder and harder to read, especially for poll junkies.
Here's the good news, part I, for Obama supporters:
Survey USA shows Clinton up 50-44. Why is this good news? With 2-3 exceptions throughout the ENTIRE primary season, SUSA has underreported Obama's support--and sometimes massively so.
Here's the bad news, part I:
SUSA, if you're not great at adding up to 100, is showing 6% Undecided/other. Why is this bad news? For this we turn to Zogby, who says the Undecided sample at this point largely is white Catholic voters, who, according to Zogby, are more likely to vote for Clinton or not to vote at all. Quinnipiac puts the spread at just over 2-1 for Clinton among all Catholics.
Here's the good news, part II:
The Undecided sample is undecided, still. This last group of mostly white Catholic voters, who Clinton ought to have in her hip pocket already, aren't yet convinced.
Here's the bad news part II:
The more recent the polling period, the larger the gap, almost across the board for Zogby (Sunday's one-day numbers, which, however, have a massive margin of error), Suffolk, SUSA, SV, and Quinnipiac. Friday-Sunday polls show less gap than Saturday-Sunday polls, which show less gap than than Zogby's one-day numbers for Sunday.
Here's the good news, part III:
The Suffolk Poll, which shows a 10-point Clinton Edge, polled ONLY 8% of its sample from the under-35 electorate. SUSA in its poll last week in PA captured 19% of its sample from the under 35s. (I can't locate the breakdown by age for their poll from this weekend, nor breakdowns from the other polls.) My guess is that more than 19% of the Tuesday's electorate will be under 35. If Obama matches the turnout of that group in PA's neighbors Ohio, NY, and Maryland, he'll make that group somewhere around 24-25% of the electorate. With land-line only calling, it's certainly possible that most polls are underepresenting Obama support.
Here's the bad news, part III:
Zogby, however, over the last two or three primaries, doesn't appear to be one of the pollsters under-representing Obama's support.
In the end, this will be about two main questions:
1. Will the undecideds indeed vote for Clinton 2-1, or not vote, or vote as if Clinton is the incumbent and break for Obama 2-1?
2. How many under-35s will vote?
PPP will put out a poll today sometime, which might or might not help sort this out.
http://www.realclearpolitics.com/epolls/2008/latestpolls/index.htmlhttp://www.realclearpolitics.com/RCP_PDF/Suffolk_PADemsMarginals_4_20_08.pdfhttp://www.quinnipiac.edu/x1327.xml?ReleaseID=1171