... poll, and the trendlines of pooled polls tend to lag the trendline of the underlying population by half the width of the pooling time window.

There is a distinct difference for Senator Clinton in finishing second and finishing third in Iowa ... finishing third in Iowa will result in her taking a substantial hit in New Hampshire, and with the short time frame, there is time for exactly one cycle of post-IA polling and poll results before the NH primary.

That's why my scenario one was, suppose that one (of six possible) configurations, followed by Senator Clinton finishing third in NH behind a surging Edwards chasing a rising Obama.

As far as a 15% gap ... if that's the gap, that's just at the edge of the "Iowa bump" that can be received in NH from a strong victory in Iowa. But in a three way race, he only needs an eight point bump with an eight point drop by Senator Clinton to cross into second place, and with his "third, two lengths back" position all of 2007, a second place finish would be positive news for Edwards ... obviously, though, winning NH would be massive for Obama.

However, that's NH ... where there has been a lot of political engagement for months. For the balance of the country,

Her 20% lead in all the other early states (bar South Carolina) gives her a huge cushion against possible "bandwagon" and other defections.

... if she finishes third in both Iowa and New Hampshire, her 20% leads vanish everywhere except in the Northeastern Feb 5 states.

The Iowa caucuses is when everyone else starts tuning in, and if it upsets people's vague impression that she's the front-runner, and that is then confirmed in NH, all of her campaign messaging from 2007 is flushed down the dunny.

OTOH, if she wins IA, she wins NH, and she has the nomination in the bag, so while she is the least likely to win the IA caucuses, she definitely has the clearest run to the nomination still.

I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.