Brendan Nyhan brushes aside a good chunk of the post-Super Tuesday analysis by claiming that primary data mean next to nothing when it comes to predicting what will happen in November. So ignore all that stuff about high turnouts in the Democratic primaries being a good sign for the party, and all the exit polls showing that Clinton or McCain have "problems" with certain demographic groups that will haunt them in a general election. The pundits know nothing.
Unfortunately for a lot of bloggers, he makes a pretty good case.