New polls from Nevada and South Carolina show states ready to split

With one more full day of campaigning for Republicans in South Carolina and for Democrats and Romney in Nevada, we get a new batch of polling from both states with results that are inconclusive at best.

Nevada: Let's start out west, because we only have one poll from NV and its results are not yet fully known. But the Las Vegas Sun has already leaked the major story: Romney and Clinton are both up by significant margins. Romney is leading the field by 15% and Clinton by 9%. The survey was conducted by Mason-Dixon, a reliable pollster. There is no question that such a decisive victory in Nevada would be a divine surprise for Clinton and, to a lesser extent, for Romney (whose victory here has become expected); and it would allow Clinton to ease the pressure of what looks like a probable defeat in South Carolina.

But keep in mind that the only two other polls released this week (Research 2000 and ARG) had a toss-up in the Democratic race and a tighter Republican race. The reason for this is evident, and the Sun admits it: "The problem is that no one — and I mean, no one — has any idea what the turnout will be." The Nevada caucuses have never played such a role, turnout in 2004 was only 9,000 voters, and absolutely no pollster has any clue how to project turnout for Saturday. I will of course update with the actual numbers tomorrow morning -- and we should also get Zogby's take on Nevada then.

South Carolina: This is a primary, and is much easier to poll. And get ready, because there was a deluge of surveys in the past few hours, so much so that I am going to have to separate out Republicans and Democrats. Since the GOP primary is only in two days, let's start with Republicans. And it is obvious that polls disagree very much on whether McCain has a significant lead, though they all put him at the top and agree on a basic order:

• Two polls out this morning from Zogby and from PPP had McCain up 7% and 8%. And they are supported by the ARG survey that has McCain up 33% to Huckabee's 23%. Romney is close behind at 20%, with Thompson at 13%.


• But three other polls have much closer races. The most important is surely the Mason-Dixon survey in which McCain leads 27% to 25%, with 15% for Romney, 13% for Huckabee, 6% for Paul and 5% for Giuliani. Evangelicals favor Huckabee 33% to 20%, but McCain leads 2:1 among non-evangelicals.
• Second, the SUSA poll (conducted on the 16th) has McCain at 29%, Huckabee at 26%, Romney and Thompson at 17%, Paul at 5% and Giuliani at 3%. The previous SUSA poll was done on the 7th, so post-Iowa but pre-New Hampshire, and it had Huckabee leading McCain 36% to 17%.
• Finally, Rasmussen came out with the tightest poll of them all, showing McCain and Huckabee tied at 24%, with Romney at 18%, Thompson at 16%, 5% for Paul and 3% for Giuliani.

Just like Iowa, it all hinges on evangelical turnout for Huckabee. Evangelicals came out in higher than expected numbers in the caucuses, and propelled Huck to a 9% lead despite the fact that he was under-organized. If Huckabee manages to make that happen once more, he could very well take the lead Saturday night. If he fails, however, his candidacy will be in serious peril, for he cannot succeed in an electorate of which about half is evangelical, can he win anywhere?

Notice also that all four polls agree with the basic order: McCain - Huckabee - Romney - Thompson - Paul - Giuliani, setting up an honorable result for Mitt and yet another shocker for Giuliani. For the candidate who had the national lead until a month ago, getting in under 4% in 3 of the first 4 contests has got to be humiliating. Naturally, this order could be completely shuffled up, as all the polls also agree that most of the margins are well within the margin of error.

And now on to the Democratic race, where four polls were released this afternoon. All give a healthy lead to Obama who has consolidated the black vote and looks unlikely to look back in the next 9 days. Which makes a Clinton victory in Nevada that much more important:

• Rasmussen gives Obama a 44% to 31% lead, which is a significant improvement from the 5% lead he had a few days ago (that poll looks like an outlier as Obama was up double-digits in Rasmussen surveys for a few weeks now).
• Mason-Dixon agrees with the storyline, giving Obama a 40% to 31% lead, with 9% for Edwards.
• Insider Advantage has it outside the margin of error as well, 41% to 33%, with 13% undecideds. Among black voters, Obama leads by a crushing 61% to 17%. Remember when Clinton led the black vote just a few weeks ago?!
• And ARG pits the race a bit more close, 44% to 38% with 9% for Edwards.

Clinton has entirely lost the black vote, but Obama trails widely among the white electorate. And to put it in frankly, Clinton's problem is very simple: John Edwards is taking a part of the white vote that might go to her otherwise. The electorate here is divided along racial lines, and Obama's disadvantage among the white vote is being offset by the fact that Clinton and Edwards are dividing up that vote. The most clear indication yet that race is playing an actual role in determining results -- just as Clinton relied on the female vote in New Hampshire to press ahead.