Primaries and Generals

What can we learn about the likely success of presidential candidates in general elections from their performance in primaries and caucuses? My answer as a political scientist is: very little. Neither Obama's success in red states with independent voters, nor Clinton's success with working-class whites in some big blue states, will necessarily carry over to the general election. The constituencies of both candidates are going to look quite different when they are matched up against McCain. Moreover, even in this remarkably competitive nomination season, only about 27 million Americans have voted. In the general election, participation is likely to be roughly four times that, and there's every reason to believe that general election voters will be quite different from primary voters.

I'm happy to see that a journalist, Jeff Greenfield, agrees, though I would add to his analysis that voting in the general election is likely to be largely driven by factors--the state of the economy and the status of the war in Iraq--that none of the candidates, or their campaign managers, can control.  Political science research on presidential campaigns tends to be fairly fatalistic; it suggests the choice of Obama versus Hillary will have much less impact on the fortunes of the Democrats this fall than supporters of either candidate seem to assume. Here's a good example of how political scientists think about presidential elections. According to Doug Hibbing's model, John McCain can expect to win 46-47% of the vote in the general election. Hibbing acknowledges the influence of more idiosyncratic matters--a tough nomination fight might have repercussions in the general election, for example--but like most political scientists who study elections, Hibbing emphasizes larger structural factors. If there's a full-blown recession in the third quarter of 2008, McCain and the whole Republican Party are going to be swimming upstream against a very heavy current.