Election Preview: Tale of Three States

Here is a Thursday update to my election analysis, using the most recent RCP information. Assuming BO overpolling 1 pt.

Pennsylvania – BO 52.3 and JM 42.8; Undecideds 4.9; BO will overpoll by 1 pt.; Undecideds, less 2.0, 2.9 will break 5/7 for JM and 2/7 for BO – based on 2008 Democratic Primary Performance. Prediction given these assumptions: JM – 45.8 BO 52.2

Ohio – BO 49.2 and JM 43.4; Undecideds 7.4; BO will overpoll by 1 pt.; Undecideds, less 2.0, 5.4 will break 4/7 for JM and 1/7 for BO, remainder to 3rd parties – based on 2008 Democratic Primary Performance. Prediction given these assumptions: JM – 47.4 BO – 48.9

Florida – BO 48.5 and JM 45.0; Undecideds 6.5; BO will overpoll by 1 pt.;Undecideds, less 2.0, 4.5 will break 2/1 for JM (same as prior assumption – given screwed up Democratic primary in Florida). Prediction given these assumptions: BO 49.0 JM 49.0

Based on this review – to win the presidency, John McCain needs to overpoll by at least 6.2 pts. in Pennsylvania in order to win, by at least 4.75 pts. in Ohio, and by at least 4.0 pts. in Florida to win.

Can this happen? Yes. Is it likely? I don’t know. The problem with many of these polls are the statistical samples being used and the assumptions made concerning democrat versus republican turnout. Many, if not most, of the polls upon which the RCP is based have made certain depressed assumptions regarding republican turnout — as we have seen at least anecdotally from early voting in areas throughout the country – republican turnout appears to be substantial (as Michael Barone notes regarding his comparison 2006/2008 preliminary comparisons). If republican turnout is higher than used for weighting on these polls, then automatically McCain’s polling numbers would be adjusted up, thereby requiring less of a leap to make victory possible.

How is this best demonstrated? I don’t know if I can point to specific factual information – I can provide anecdotal information from my own backyard.

In the days before the 2004 election, John Zogby had Kerry up in Michigan by 6 pts. – he won by 3.5. I would submit that this was predicated largely on assumptions made concerning republican turnout, which was underestimated. Zogby has Michigan down by an even larger amount this race and I can safely say that Obama will not win Michigan by a larger margin than Gore won in 2000, which was approximately 5.1 pts. If we were to compare this result with current polling projections, this would suggest that some or many erroneous assumptions are being made concerning republican turnout.

I think it’s important to remember that this election boils down to a few wildcards – some unknowables right now – will there be a Bradley effect? (I personally hope not, but I don’t know if that’s likely or not – it’s unknown) will there be a substantial democrat defection particularly from democrat women who would have voted for Hillary? what is the Palin effect — simply good for turnout of conservative base or is it something more?

We shall see soon enough. More update on Monday.