Media Electoral Predictions and the Horse-race

It’s November 5th, and you know what that means–that’s right, Election Day 2012 is almost upon us. As the race draws to a close, I thought it would be valuable to examine electoral maps published by five major news outlets: ABC, CBS, CNN, Fox News, and NBC. I have compiled their state-by-state ratings below:

Now, none of these networks predicted a clear winner (at least in their Electoral College maps). President Obama had a plurality of electoral votes in each of the maps I examined, but did not have a majority (270 or greater) in any of them. I could have rated some sources as predicting a victory for a certain candidate–Fox News featured two articles with their map predicting a Romney win, with one even predicting a landslide. But since none of the maps produced a clear victor, I rate them all as “toss ups.”

The anomaly of this group is clearly Fox News–they have rated several more states as toss ups, despite recent polling data indicating otherwise (such as Michigan and Pennsylvania). Fox’s outlook actually featured three different maps. The scenario you see here is the primary one on their page, whose data is taken from Real Clear Politics. Karl Rove’s map was also featured, and it was laughable: It rated Oregon and Minnesota as toss ups, in addition to those already listed above.

I find that major news outlets (such as those above) are generally worthless when it comes to electoral predictions. Yes, this claim is probably a bit shocking, given that I spent time compiling the given information and writing this blog post. But ask yourself: Besides presenting the news, what do the media exist to do? The answer: Make money. And what kind of news sells? Well, anything exciting, frankly. The media have a big incentive to make elections appear as close as possible, even if polling data indicates otherwise. This is what we call the horse-race–the media’s portrayal of elections (especially presidential ones) as a never-ending race in which each candidate is vying for an advantageous position. One is always “ahead” while the other is always “behind.” This coverage is evident in the current campaign. Just look at the chart above–the fewest number of toss up states were on both the CBS and NBC maps, but even they had seven. Making elections appear closer than they really are is economically beneficial for news outlets, so they do just that–make elections look like one long (and close) horse-race.

The most reliable election predictions I have found are at FiveThirtyEight, run by statistician Nate Silver. Silver is perhaps best known for correctly predicting 49 of 50 states in the 2008 presidential election and correctly predicting every Senate election in 2008–even Al Franken’s upset of incumbent Norm Coleman. He also predicted the Republican takeover of the House in 2010 (he was only eight seats off–not bad, considering that there are 435 total). Needless to say, Silver is one of the top people in his field, and is widely acclaimed for his accuracy in predicting elections. His map looks much different than those above:

As you can see, this map features no toss ups. Instead, each state is rated based on the probability of a candidate’s victory in that state. This model is much more realistic than the “Obama-Romney-Toss up” model, simply because each state must have a winner–there can be no toss ups once the votes have been tallied. Sure, polls can be statistically tied (i.e. within the margin of error), but Silver’s method–which takes adjusted polling averages and state fundamentals into account–is proven to be incredibly accurate. The accuracy of Silver’s predictions stems from his background–he’s a statistician, not a journalist. Yes, his blog was picked up by the New York Times, but that doesn’t mean he suddenly has an incentive to make the election appear too close to call. Find me a model published by a major news network that gives Obama an 87 percent chance of winning Ohio, and then I’ll retract everything I’ve written here and admit that I’ve been listening to the wrong people.

Given that Election Day is so close, my message might be too late for this election cycle. But even in the future, the media will continue to make races appear close–even when they’re not. If you’d really like an accurate idea of how future elections will turn out, keep FiveThirtyEight in mind. And if you still go to majors news outlets for electoral predictions, remember–take what they say with a grain of salt (or maybe even the whole shaker).


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s