If you buy the notion that the 29th's suburbs might hold the key to this election, what are the numbers needed to drive change?
My model says 5,5,2 and 1.
If 5% of those who voted for Kuhl stay home in 2006, and 5% of those folks vote for Massa instead of Kuhl, and if those trends are magnified by 2 additional percent in Monroe, and 1 percent in Ontario, then Massa squeaks through. Without the input of the 'burbs, it's 51/49 Kuhl, all other things equal.
Let's look at those numbers and the assumptions underlying this model.
The first assumption is that Republicans are dissatisfied, and that they'll express their dissatisfaction in two ways: by voting for Kuhl's opponent, and by staying home. Is 10% Republican dissatisfaction a reasonable number? I dunno, but remember that it's 10% of Kuhl's vote in 2004, not 10% of Republicans - some of those '04 Kuhl voters might be Independents or Democrats.
The second assumption is that suburbs are more likely to change, and that change will be 4% more in Monroe, and 2% in Ontario. Overall, the model assumes that 14% of voters in Monroe, and 12% in Ontario, will either stay home or switch alliance. Fourteen percent is pretty fickle. The Ontario number is smaller than Monroe because Ontario includes suburbs (e.g., Victor) and rural communities -- it isn't a "pure" suburban community.
Finally, we need to deal with the '04 third-party candidates. The model assumes that the Conservative vote will break 2/3 Kuhl and 1/3 Massa. This might be a bit controversial, but I think some hard-core conservatives who aren't satisfied with the Bush administration might want to cast a protest vote. The model also assumes that the Independence party votes will split 50/50: they're independent, after all.
Now this isn't a "real" model -- it's more like "fun with numbers". A real model would, at a minimum, drill deeper than the county level, address differential turnout by party registration, and have a lot more knobs to adjust the underlying assumptions. But I hope this at least captures the basics of one of the more interesting dynamics of this race.