Greg Marx had two nice posts at the Columbia Journalism Review, in which he shows just how hard it is for reporters to resist interpreting these off-year elections as harbingers of Something Significant.
To buttress that point, via Brendan Nyhan comes Alan Abramowitz’s analysis of the predictive power of the VA and NJ elections, “What Happens in VA and NJ, Stays in VA and NJ”:
bq. However, the results of the previous year’s gubernatorial elections in Virginia and New Jersey did not predict the results of the midterm elections. Not only is the estimated coefficient for the Virginia/New Jersey election variable small and statistically insignificant, but it is in the wrong direction: the better Republicans did in Virginia and New Jersey, the worse they did in the subsequent midterm election.