Let’s be clear about this: Bill Halter and Joe Sestak, in broad terms, represent challenges from the left, and their success is fueled by the energy and intensity of liberal activists. ~Greg Sargent

This strikes me as common sense, but as E.J. Dionne points out common sense apparently has nothing to do with voters’ preferences:

Also, there is this fascinating bit of data reported by Tom Schaller at fivethirtyeight.com. He cites a blog post by Public Policy Polling’s Tom Jensen reporting that in the last poll taken in PA-12, Sestak led Specter by nine points, 44-35. “Among liberal voters in the district Sestak and Specter are actually tied at 44% each,” writes Jensen. “Sestak’s lead comes largely thanks to a 49-21 advantage among conservatives within the party [bold mine-DL] who make up 26% of the primary vote in this district, almost equal to the 25% who identify as liberals.” This goes to show the kind of branding problems that can occur when one switches parties.

This finding is bizarre, but perhaps there is a way to reconcile the two claims. In fact, Sargent is correct that Sestak and Halter are challenging their respective incumbents from the left and derive important support from progressive activists. They have based their candidacies on being more progressive representatives of their states. That said, it is necessary to distinguish between the ideology of activists and the preferences of the electorate, and it is also necessary to distinguish the ideological leanings of the candidate from that of the people who end up voting for him for whatever reason.

It is important to emphasize how relatively unimportant ideological identifications are in determining intra-party voting. For example, Ron Paul was obviously to the right of all of his competitors in the presidential primaries on almost every issue (and I would argue that his foreign policy views also put him far to the right of the 2008 field), but he received a disproportionately large part of his support from self-described moderates because they tended to be weaker Republicans and therefore more likely to be responsive to a dissident candidate. On paper, it makes no sense that moderate and conservative Democrats would prefer Sestak to Specter while liberals are evenly split between them, but then it doesn’t make rational sense that John McCain won the antiwar and anti-Bush vote in New Hampshire in 2008.

What many of these voters probably have in common is weaker partisan attachment. Specter has the backing of the state party machine, and so there will be liberals who are also strong partisans who rally around the candidate the party machine supports. What those McCain voters apparently perceived in McCain was someone whom they identified as being at odds with Bush, mainly because of their rivalry in 2000, and they bought into his reputation as someone who broke ranks with his party, so they may have assumed that the way to express their opposition to Bush and the party leadership was to support someone who was, in fact, virtually identical to Bush on every major policy question of the day. Perversely, the anti-Bush and antiwar Republicans with weak attachment to the party supported the party establishment’s original preferred candidate because they had the superficial, basically false impression that McCain was not a reliable partisan and Bush ally. It seems likely that Sestak, who is running as the real, progressive Democrat against the phony newcomer, will receive support from weaker partisans while the opportunistic “centrist” with establishment backing will receive a surprisingly large amount of support from the left because those voters are also strong partisans.

My guess is that Sestak’s support among moderate and conservative Democrats in Pennsylvania matches up very closely with the disapproval these voters have for Obama and the Washington Democrats with whom Specter is identified. This just underlines why many strongly ideological voters are often frustrated with the direction of their parties: many strongly ideological voters support candidates endorsed by the party establishment because they are also strong partisans and tend to conflate what is good for their party with what is good for their political views. Bizarrely, the people more likely to oppose the Congressional Democratic leadership’s priorities are going to help put a reliable supporter of those priorities into the Senate and help throw out an opportunist who would probably turn against the Democratic leadership at the drop of a hat if he thought it would serve his self-interest. Realizing this actually makes the result of today’s voting in Pennsylvania even harder to interpret and weave into a larger narrative.