To add to what Phil wrote below, I think there’s a very simple question to ask when evaluating whether to support a flawed candidate: What’s at stake in the election? That should be followed by a second, more complicated question: What is the likeliest outcome of a particular candidate’s election?
In Massachusetts, I reluctantly voted for Bill Weld, Paul Cellucci, and the pre-conservative Mitt Romney because I agreed with them on the issues that were actually in play during the election, thought they would have little impact on the issues where I disagreed with them, and found the viable alternative worse. I would have voted for Rudy Giuliani in the New York mayoral races of 1989, 1993, and 1997. But I voted against Weld when he ran for Senate, voted against Romney in the Republican presidential primary, and could not have pulled the lever for Giuliani even in the general election.
Why? For me, the stakes changed. New York City was going to remain pro-choice whether Giuliani lost or won. But it didn’t necessarily have to remain mired in high rates of crime, taxes, and welfare dependency. Yet Giuliani as the GOP presidential nominee might have had an impact on whether the Republicans remained a pro-life party. (I wasn’t wild about his foreign policy either, but that’s another argument for another day.)
What impact will electing Dede Scozzafava have? Best case scenario, not much: she’ll be another minority vote for John Boenher for speaker (for now, anyway) in a chamber where the minority seldom has much influence. Worst case scenario, she will create confusion about the Republican brand, frequently vote with the Democrats anyway, and help move the GOP to the left. In a district where there is a better alternative with some realistic chance of success, then party disloyalty seems like a no-brainer.