The numbers Nick cites on the relative liberalism of the new Democrats are interesting, but strike me as methodologically weak. These measures that evaluate how often politicians break with their party, rather than on what issues they break with their party, are problematic.
All issues, after all, are not created equal. Iraq is not the same as federal holidays is not the same as college loans. So it's telling that Joe Lieberman, who votes for every Iraq extension imaginable and is currently trying to whip up a war with Iran, is the 9th most liberal member of Congress, when most liberals would currently call him a neoconservative. Were we in a massive fight of universal health care right now and a given Democratic politician was voting right on everything but that and Iraq, I also couldn't consider them a liberal no matter how many times they made the right call on conservation issues. A better metric, probably, would be to take a survey measuring what liberals think were the 15 key votes of the last year and see how various politicians came down. The salience of the votes they broke ranks on are at least as important, if not more important, than the quantity.
