Thursday, November 8, 2012

Wrong, but not by much



The Romney campaign conceded Florida today.

That makes the final Electoral College tally 332-206. I predicted it would be 303-235. Florida, the only state that I did not guess correctly, accounts for the difference.

When all of the ballots are counted the difference between the two candidates in Florida will probably end up being about five or six tenths of a percentage point. Thus, I came pretty darn close to hitting the electoral count on the nose.

On the other hand, it is also true that the 332-206 electoral count was fifth on my list of most-likely scenarios. The reason it was so far down the list was my incorrect assessment that it was more likely that Romney would outperform the polling averages in combinations involving Virginia and Colorado than that Obama would outperform the polling averages in Florida.

Florida was the biggest surprise of the election for me.

In addition to Florida I was a little bit surprised by the strength of Obama’s showing in Virginia and Colorado and mildly surprised at Romney’s stronger than expected showing in Ohio, which Obama won by just under two percentage points.

Other than that, the election played out pretty much as I had anticipated.

While the Obama victory was no surprise at all, I was a bit taken aback by how many seemingly rational people seemed to be so genuinely surprised that Obama won; including, apparently, Romney himself.

Did Romney, Ed Gillespie and the rest of the crew really think their campaign plane was going to land on Fantasy Island?

As I stated in my blog post the day before the election; “Predicting the outcome of the 2012 presidential race boils down to this; for Mitt Romney to be elected president the state poll trends and averages will have to be wrong, and not just in one state, but in several.”

That was, if I do say so myself, a spot-on analysis. And, while I would love to bask in the glory of my brilliance, the fact of the matter is that any kid capable of earning a C in a middle school civics class could do the same thing. It’s common sense, not rocket science or political science.

In the last several presidential elections these aggregated poll averages have proven to be very accurate predictors of the state by state outcomes, especially in elections where there is no strong third party candidate. Obviously, that reputation has been burnished by the 2012 election.

Thus, the election results are an unabashed landslide victory for those who make predictions based on aggregated polling data; especially Nate Silver of the New York Times FiveThirtyEight blog.

Not only did Silver correctly call the presidential election he also correctly called every senate contest except for Montana and North Dakota, both states where the polling data he had to work with was thin.

While it would be foolish to put a lot of stock in a single poll it is equally foolish to assume that the aggregated mountain of polling data available in the weeks leading up to the election was wrong.

Thus, with a heavy heart and great empathy, I have to break the disappointing news to Mr. Roarke and Tatoo that Mitt Romney’s campaign plane will not be landing on Fantasy Island.

No comments:

Post a Comment