Can Political Scientists Predict Presidential Elections?

Are political scientists soothsayers who already know who will be president in 2013?  As New York Times politics blogger Nate Silver observes, presidential elections forecasting has proliferated, with many political scientists developing statistical models that can be used to predict outcomes.  The models vary, but tend to emphasize the effects of the economy--and to ignore the attributes of candidates, like Mitt Romney's purported inability to connect with voters.  Silver, however, argues that the track record of these models is dismal; they are regularly wildly off in predicting vote totals.  John Sides defends the political science models, noting that they have correctly predicted the winner 85% of the time.

Rate this posting:

{[['']]}