Might we have seen the end of Mitt? A new Newsweek poll suggests we might have.

The most dramatic result to come out of the poll, which is based on telephone interviews with 1,408 registered Iowa voters on Dec. 5 and 6, is Huckabee's emergence from the shadows of the GOP race into the front runner's spot in just two months. The ordained Southern Baptist minister now leads Romney by a two-to-one margin, 39 percent to 17 percent, among likely GOP caucus-goers. In the last NEWSWEEK survey, conducted Sept. 26-27, Huckabee polled a mere 6 percent to Romney's 25 percent, which then led the field.

Huckabee has also opened up a wide margin over the next three leading candidates, who all show signs of fading in Iowa: Rudy Giuliani, who dropped from 15 percent in the last survey to 9 percent in the current one; Fred Thompson, who fell from 16 percent to 10 percent; and John McCain, who slipped from 7 percent to 6 percent. "You rarely see anything like [Huckabee's surge]," says Larry Hugick, who directed the polling for Princeton Survey Research Associates. Hugick added that the reason has as much to do with a leeriness of the other candidates among Republican voters as Huckabee's folksy success on the stump. "He's filling a vacuum," Hugick said. "Nobody on the Republican side was getting strong support."

This survey was in the field just as coverage of Mitt Romney's impending speech on his faith blanketed the news media -- and it shows. According to the Newsweek survey, Huckabee holds a massive 47 percent to 14 percent lead over Romney among Republican Evangelicals voters in Iowa, a spread that contrasts clearly with the 24 percent to 24 percent tie between the two candidates among non-Evangelicals. Nearly half of Republicans (46 percent) say "at least some Iowa Republican voters will not consider supporting Romney because of his Mormon faith," and 16 percent admit to personally being less likely to support Romney because he is a Mormon.

On the Democratic side, Newsweek sees less change -- though perhaps a two-way rather than three-way contest.

Unlike the GOP race, standings in the Democratic campaign have not changed dramatically since the September NEWSWEEK poll in Iowa. However, Barack Obama has gained some ground, moving to within a point of Hillary Clinton among all Democratic voters (29 percent vs. 30 percent), with John Edwards in third place at 21 percent. Among those most likely to attend the caucuses, Obama has moved substantially ahead of Clinton, 35 percent to 29 percent, while Edwards falls back a bit, to 18 percent. Obama also gets more support from those who say they will "probably" attend a Democratic caucus (40 percent vs. 27 percent for Clinton). While the Illinois senator's lead among Democratic caucus-goers in this poll is not large enough to be statistically significant, things seem to be trending his way, Hugick said. "It's evolving into a two-person race, with Edwards hanging on," he said.

Now for the important caveats from this poll. The margin of error for likely Republican voters is plus or minus 7 percentage points (and plus or minus 5 percentage points for all Republican voters). What's more, based on Newsweek's screen, more than 400,000 Iowans would be considered likely Republican caucus goers. But comparison, current predictions estimate GOP turnout on January 3 will be in the range of 75,000 to 95,000, meaning that Newsweek's screen may be way too loose on the Republican side (perhaps by a factor of 5 or more).

The looseness of the voter screen on the Democratic screen isn't much less potentially problematic. According to my back of the napkin math, it looks like Newsweek would consider about 575,000 Iowans to be likely Democratic caucus goers, about 3.8 times larger than an expected turnout of 150,000, which seems to be more in line with other predictions. Even leaving the poll's screening methods aside, the margin of error among likely Democratic caucus goers is plus or minus 6 percentage points, meaning that the gap between Hillary Clinton, Barack Obama and even John Edwards is not statistically significant.

On top of all of this, polls of this nature (basically all public polling) only has such margins of error at a 95 percent confidence level, meaning that one in 20 polls is expected to be an outlier. I'm not saying that this Newsweek poll is an outlier, only that one would be well served by taking some caution (i.e. wait too see other polling showing similar results) before putting too much of a stake in these numbers, particularly the ones on the Republican side of the ledger (the Democratic numbers don't diverge from the average of other polling to near the same extent, or much at all, in fact).