John Edwards did a very responsible thing by dropping out of the contest for the Democratic party nomination. He had little chance of winning the nomination, but held the support of a non-trivial fraction of Democratic voters (see the NYT Dem Election Guide for fractions that go up to 17% in NH and SC). Given the tight race between Senators Clinton and Obama, it may very well be the case that where the Edwards supporters go will be a decisive force in the contest.
In this article praising Edwards for his "unabashed populism" and "campaign based on ideas", Paul Krugman explicitly articulates this question:
And so Mr. Edwards won the arguments but not the political war.
Where will Edwards supporters go now? The truth is that nobody knows.
Here is one crude analysis of where they'd go:
Look at Jan 2008 Google Trends for "clinton obama -edwards" -- (I think) this means volume of queries that mention Clinton and Obama but not Edwards -- this could be taken as a measure of how many people are comparing Clinton and Obama, without mentioning Edwards. Now look at the Trends graph for "obama edwards -clinton" (queries that mention Obama and Edwards, possibly comparing the two) and for "clinton edwards -obama". The interesting fact is that when you look at all three graphs or just the latter two graphs (for a closer view) you'll notice that there are a lot more queries that "compare" Obama and Edwards than there are queries that "compare" Clinton and Edwards. This "implies" that a lot more of Edwards supporters will likely vote for Obama than they would for Clinton.
So you heard it here first - be sure to pat my back after Super Tuesday :-)
Of course, Edwards supporters might be Edwards supporters for a variety of reasons (can't stand a woman being President, can't stand an African-American being President, don't like Clinton, like his anti-poverty policies, much better haircut than either of the other candidates, etc.), and this "analysis" doesn't take into account any of these -- it's just a crude interpretation of relative query volumes. However, to the extent that search queries are a reflection of the society's thinking at large, this just might be a sound basis to predict an Obama victory.
ps: as I type this post on Blogger, its built-in spell-checker stupidly underlines the word "Obama" (but not "Clinton" or "Edwards") -- little does it know...