One new poll of Iowa Democrats, by InsiderAdvantage, came out yesterday, and another, by Washington Post/ABC, came out late last night. The former created a big buzz among Edwards supporters, because it showed him up 4 over Clinton and up 6 over Obama among likely Caucus-goers. The latter had Obama up 4 over Clinton, and 13 over Edwards. InsiderAdvantage hasn't done a previous Iowa poll, so there are no trend-lines to look at; the WaPo/ABC poll shows small trends towards Obama and Clinton and a small drop for Edwards since last month.

The top line aside, the two polls agree on some things, most notably John Edwards' increasingly strong performance in second-choice preferences, and the now-familiar Obama dominance among younger and highly-educated voters. As Chris Bowers at OpenLeft points out in an excellent analysis, the InsiderAdvantage poll stipulates an extremely low turnout among voters under 45, even by Iowa standards, making its numbers for Obama especially suspect (the WaPo/ABC poll isn't accompanied by age breakouts).

So: while campaigns can be expected to spin any given favorable poll as indicating a "surge" or a "comeback" or whatnot, the real deal remains very hard to measure. Typically, "likely voter" screens become more accurate as the actual event approaches, but the legendary difficulty of determining likely participation in the Iowa Caucuses makes even that prediction perilous. About the only statements that can be made with any degree of confidence based on a variety of recent polling are that John Edwards will probably do better than his first-preference polling suggests, and that young voter turnout will probably be decisive, one way or another, for Obama.