Media and Prediction Markets Don't Reflect Reality

If you've ignored forecasting sites such as this and FiveThirtyEight but have only listened/read what the media has been saying, you would believe the Republicans are all but certain to gain the majority in the Senate after the November elections.

No doubt, it's an attractive story for the news organizations to tout: President Obama is unpopular and there will be a Republican wave akin to 2010 because the public is so fed up. But that's not really what is going on.  

Yes, President Obama is unpopular.  His approval rating stands at about 42% (about 4% higher than Bush's approval at this same point in his Presidency).  But Congress and both parties are also deeply unpopular, and that is why Republicans aren't actually running away with this election.  

Democrats were always going to lose a few seats in the Senate.  In 2008, a banner year for Democrats, the party won in very conservative states such as Alaska, Arkansas, North Carolina, Louisiana, South Dakota, Montana, and West Virginia.  Those candidates were largely helped by huge turnout from Obama voters.  Those same voters are less likely to turn out in a midterm election, and many are dissatisfied with the President.  

So why does the media continue to claim that Republicans will almost assuredly re-take the Senate?

As I mentioned above, part of the reason is because it's a favorable storyline.  Conflict and comebacks are always good stories that Americans, in particular, respond well to. Therefore, the more it is reported, the higher the ratings.

A second reason is due to an oversimplification of the polls.  Many reporters and news anchors don't understand that a reported four point lead in the polls still yields only about a 75% chance of victory.  Now, that is certainly good betting odds, but many of the races that could decide the fate of majority status in the Senate are much closer than four points. However, the media choose to cherry pick polls that work in its favor, leading to an overall simplification.  

One can see the effect this has in betting markets such as iPredict.  The New Zealand prediction website currently has almost a 92% odds for the GOP to re-take the Senate. That is wholly incongruous with reality.  As of yesterday, Battleground270's forecast gave the GOP a slight advantage with a 63.3% probability.  Nate Silver's FiveThirtyEight blog gives the GOP a slightly lower chance at 62%.  The New York Times LEO calculator puts their odds at 69%.  

The only forecaster to line up with the betting market and media is the Washington Post, which puts the GOP chances at 93%.  We've already discussed why the Washington Post's predictions are higher (and unfounded), here.

Should Democrats be worried? Yes.  Should Republicans begin popping the champagne? Absolutely not.  There is still two more weeks until the election, a lifetime in politics.  There are many races which haven't been polled often or in awhile, which makes our model even more uncertain.  Stay tuned to find out the real odds, and stop listening to the talking heads.