Part of the reason pundits and journalists don't much like the analysis of political scientists about elections is that political science can be kind of boring. Political science tells us, for instance, that structural factors are much more important than the actual conduct of campaigns. If you want to know what's going to happen in an election, just use a few variables, particularly economic ones like real income growth or unemployment, and you can predict with a fair degree of accuracy what the outcome will be.
From the pundit's perspective, that removes all the interesting stuff. It leaves you no reason to talk about the spectacle of politics, the inept candidates and shocking ads and hilarious gaffes. It doesn't allow room for the armchair strategizing ("What the Democrats ought to do is...") that is the pundit's stock in trade. The political scientist responds, well, sorry about that, the truth is the truth. While the political scientists may not be able to tell you what's going to happen in any particular race, they make a convincing case that there's a fundamental trend, and individual races will be minor influences around that trend, pulling it only marginally one way or the other. For instance, Jon Chait points to one study using a model with just three variables -- the fact of a midterm election, the ruling party's "exposure" (how many seats they have to defend), and personal income growth -- that predicts that the Democrats will lose 45 seats and therefore lose the House. You can take that as a baseline and conclude that more or fewer losses than that can be attributed to strategies and candidates. But the fact is that just based on the makeup of Congress and the state of the economy, Democrats are almost sure to lose control of the House no matter how skillful they might be.
But the pundits and journalists are still focused on all the interesting stuff like witchcraft and Aqua Buddha, and they're the ones who will determine how we understand this election. That understanding will then influence the decisions political actors make in response. Assuming the Democrats lose the House, the prevailing explanation will probably be some combination of "Obama overreached" and "Americans are mad at big government," both of which are wrong. But pundits and journalists (unlike most Americans) understand ideology and think in ideological terms, and so are prone to attach an ideological explanation to elections.
Of course, it's in Republicans' interest to characterize the results as a mandate for their agenda. And if they do take the House, Republicans are going to be getting a lot of airtime, so they'll have all the opportunity they could ever want to advance their narrative. When Democrats respond, "No, it was the economy," their boring explanation will be dismissed with a wave of the hand.
-- Paul Waldman