Mark Schmitt notes a few instances of districts featuring vulnerable Republicans where the "wrong" Democratic candidate prevailed in the primaries against an establishment-backed moderate, only to have the establishment write the seat off (at least temporarily), and then the Democrats won the seat anyway. "Is there a lesson here? It's not a big sample size, but it suggests that in a district where a Republican was vulnerable to defeat, a plain-spoken progressive could do it at least as easily as a focus-grouped moderate. Perhaps even better."

Maybe that's the lesson. I'm inclined, however, to see a different lesson. Consider once again Carol Shea-Porter. Mark observes that she "won a four-way primary, defeating a veteran state legislator who had the support of the DCCC, got a campaign visit from Tom Daschle, and out-raised Shea-Porter 10 to 1." My guess would be that the real lesson here is that a candidate who manages to win a four-way primary against, among others, a veteran state legislator who had the support of the DCCC, got a campaign visit from Tom Daschle, and out-raised Shea-Porter 10 to 1 probably just had strengths as a candidate that weren't obviously there on paper. As everyone knows, actual issues and policy views have only a limited impact on voting behavior -- there are a lot of intangible factors in play, and primaries put those intangibles to the test.

One of the oddities of 2004 was that because Dean and Gephardt focused so much of their fire on (successfully!) bringing each other down, and then John Edwards waged a "nice guy" campaign aimed at securing the Vice Presidency, Kerry emerged victorious without really being tested. It's better, I think, to have real races insofar as their are real disagreements between the candidates. In a way, this is especially true for more moderate candidates who'll have a better chance at getting credit for their moderation if, like Bill Clinton, they actually succeed in facing-down alternatives and securing a mandate for re-positioning the party.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.