We’re five days away from Election 2012 and the race is stabilizing. Among many things to consider, we have various national polls on the popular vote and, the projected number of Electoral College votes for Barack Obama and Mitt Romney. The national percentages are close, and it is clear that the victor will hardly walk away with a national mandate to govern — a proposition that poses its own challenges after the election. For now, it’s interesting to consider some of the polling out there,both nationally and in the swing states.
Consider the following two polling sites, Real Clear Politics (which highlights many different independent polls and aggregates them), and also the DeSart and Holbrook political science model election forecast out of Utah Valley University. Real Clear Politics (RCP) has the national race tied, in terms of poll average, at 47.4% for the two candidates. Their electoral map puts the following states into the toss-up category: Colorado, Florida, Iowa, Michigan, Nevada, New Hampshire, North Carolina, Ohio, Pennsylvania, Virginia, and Wisconsin. Looking at these states individually, poll averages put either Obama or Romney ahead by a given percentage, ranging from just 0.6% Obama in Colorado to 3.7% for Obama in Wisconsin. Romney is up 1.7% in Florida, and up by 3.8% in North Carolina. So, again within swing states, there is a variation, but these are tight races less than a week to go.
Over at the DeSart and Holbrook forecast, we can look at those swing states with a different model. For those of you curious about how this model is constructed, I’ll point you to this blog post over at my former professor’s Presidential Power blog, a source for nonpartisan political analysis. Numbers-wise, they project the national two party vote percentages will be 51.44% for Obama to 48.56% for Romney with an 86.84% projected Obama win probability, which is very different fromthe RCP projection, which is again, a polling average. Finally, DeSart and Holbrook have Obama with 281 Electoral votes to Romney’s 257 votes.
Now, look at their state win probabilities from November 1. Ohio, New Hampshire, Iowa, Michigan, Nevada, Pennsylvania, and Wisconsin are all in the Obama column with Ohio being the closest (or most "competitive") state, with still a 69.6% probability win for the president. Compare that with Romney’s column which has Colorado, Florida, and Virginia, but with very close probability wins, hovering just above the 50% mark. These probabilities suggest that this is Obama’s race to lose, although it is hardly a given.
Even if Obama were to lose Colorado, Florida, and Virginia, which again are in Romney’s column just barely, he still stands a good chance at winning the key states of Ohio and Pennsylvania, in addition to Iowa, Michigan, Nevada and Wisconsin. Notice also that the RCP average for Obama winning in Colorado is just 0.6% along with Virginia at just 0.5% for Romney, and Florida at 1.2% for Romney. Those average percentages aren’t all that different from the DeSart and Holbrook model suggestion that those three states are all in Romney’s column, albeit just by a small win probability. Obama can lose those states, so long as he takes other key swing states and still surpass the 270 vote threshold to win. Romney on the other hand, could win those three competitive swing states and still not make the magic number. For him, it will be a tougher road to victory.