The Iowa caucus, which sounds the starting pistol in the Republican nomination race, gets underway tonight. Yet as voting fast approaches, there is still no clear frontrunner. Polls show that some two out of five voters in Iowa are still undecided.

There is no clear consensus among the pundits either, who variously predict that either Mitt Romney -- currently topping most national polls -- Ron Paul, or Rick Santorum could win in the state. These three candidates are almost evenly tied.

Richard Cohen [2] at the Washington Post (who predicts that Romney will be victorious) notes that none of the other candidates have emerged from Iowa with their campaign in-tact:

The Iowa caucus has turned out to be a demolition derby for Republicans. With the exception of Romney and Santorum, they all have been damaged. Perry showed he couldn't debate (or talk), Bachmann had trouble with the truth, Gingrich acts like R2-D2 with a short circuit and Paul has been soiled by the ugly newsletters his foundation published in the past. Santorum emerges undented, (al dente?) but that could be because until too late he was not considered worth denting. Aside from him, though, only Romney came out of Iowa as he came in -- boring, but inevitable. He wins because everyone else loses.

Romney is cultivating this sense of inevitability around his campaign, seeking to give the impression that the party is coalescing around him. A big win in Iowa would give this tactic a significant boost, given that it has thus far looked like a tight race. On the other hand, if he falls into third place, he may have to do some explaining, although this will not necessarily spell disaster for the rest of his campaign.

What happens in Iowa does not necessarily reflect the eventual national outcome -- it is an oft-quoted fact that Mick Huckabee won in Iowa in 2008, although the nomination eventually went to John McCain. Indeed, since 1972, only three non-incumbent candidates have won the Iowa caucuses and went on to win the presidency -- Carter, George W. Bush, and President Obama. It is easy to make arguments for why this largely agricultural state does not reflect the US as a whole; yet it does represent the first test of the voting public, and a reasonable indication of the viability of a candidate's campaign.

For this reason, it can be almost more important who does badly than who does well. All of the second-tier candidates have insisted they will continue with their campaigns regardless of what happens in Iowa, but it is not unheard of for low polling candidates to drop out of the race.

Paul, who has stood for presidency twice before, will be particularly affected by this. In the past, he has been held back by the perception that he simply does not have sufficiently wide appeal to take the fight to the Democrats. A more organised campaign this time has worked to broaden his support base outside libertarians and students, and a win in Iowa could provide a counter-argument to those who maintain he is not a viable candidate.

Quite apart from what Iowa means for individual candidates, the level of voter turnout in this swing state -- important in the general election -- should give some indication about the strength of partisan feeling. As Michael Shear [3] notes at the New York Times Caucus blog:

Fourteen months after a tidal wave of Republican energy helped sweep many Democrats out of Congress, the Iowa results will provide a hint about whether that intensity of purpose remains.

If 140,000 or 150,000 voters show up to the caucuses, that would be a good sign for Republicans (who have said for months that they have succeeded in adding to the rolls of registered Republicans). If fewer people show up than last time, it may suggest that the excitement of 2010 has faded a bit.

In a race so far characterised by uncertainty and swift rises to the top of the polls, matched in speed only by falls from grace, Iowa will give the first reliable test of public opinion. All eyes on the results.