Eight polls have been released for the Iowa caucuses within the past week. For the Democrats, four of those polls show Hillary Clinton ahead, two show Barack Obama in the lead, one has John Edwards on top, and one shows a tie between Edwards and Obama with Clinton a point behind.

Among Republicans, four of the last seven polls show Mike Huckabee ahead while three show Mitt Romney in the lead. Four show Fred Thompson in third place while three award that spot to John McCain. But, five of the seven show McCain and Thompson within three points of each other.

So, as Rasmussen Reports first noted several weeks ago, those who want pre-caucus polls to tell them who will in Iowa are sure to be sorely disappointed. Consider the results from four years ago. Five firms released polls in the final week of the campaign and showed John Kerry’s support ranging from 21% to 26%. Kerry was slightly ahead of Dean in three polls while Dean had a statistically meaningless lead in two others.

The end result? Kerry outperformed the polls by double digits and ended up with 37.6% of the delegates. He more than doubled the 18.0% support picked up by Dean. John Edwards came in second even though four of the five polled showed Dean narrowly on top of Edwards.

Still, while the 2004 polls didn’t nail the final result, they did provide some important clues as to what might happen. Most notably, on a collective basis, they showed upward trends down the stretch for both Kerry and Edwards while also showing a sharp decline in support for Dean. The trends were right, if not the precise numbers. Still, it’s worth noting that the 2004 polls correctly identified the top three finishers by correctly highlighting a drop in support for Dick Gephardt.

The data from 2004 also highlights the biggest single problem in polling caucus results—determining who will actually show up and participate. The entrance polls in that year provided a very accurate picture of the final results (that poll found Kerry with 35%, Edwards with 26%, Dean with 18%, and Gephardt with 11%). The entrance poll was able to produce an accurate result because it was dealing with the one thing earlier telephone polls could not determine, the actual sample of caucus participants.

Looking now at the data for 2008, we see similar challenges. The final Des Moines Register poll of the Democratic race showed Barack Obama opening a wider lead. That could very well be accurate. But, the poll also showed that 60% of those included in the poll would be attending their first caucus. That’s an astounding figure and the bulk of those first timers are coming out to support Barack Obama. If they show up, they will highlight Obama’s appeal to a wider audience. If they don’t Obama’s advantage will disappear.

The Des Moines Register poll also shows that just 55% of the Democratic caucus goers will actually be Democrats. Four years ago, that figure was 80%. Why does that matter? Because Clinton leads among Democrats while Obama leads among Independents. An excellent column by the Register’s David Yepsen addresses this issue and notes that “If pollsters adjust the party identifications in 2008 to look like they did in 2004, Clinton could beat Obama 31 percent to 29 percent.” So, depending upon the turnout model, the same poll can show different caucus victors.

As noted above, while the Register shows Obama with a growing lead, other polls present different leaders. It is likely that adjusting the turnout model for each of those polls would yield different results as well.

A similar dynamic is in play on the Republican side of the aisle. As noted in a separate Rasmussen Reports article on the caucus, the ultimate GOP winner in Iowa may be determined by something that polls can’t measure—the organizational strength of the evangelical church leaders supporting Huckabee. Romney has built a more traditional Iowa caucus organization.

In his article, Yepsen also points out another challenge to pre-caucus polling-- a large number of caucus goers make their final decision in the closing days of the campaign (21% of all 2004 caucus goers said they decided in the final three days). Add to this the many people will have yet to finally decide if they will even show up days and it becomes clear that no poll can reliably project the winner until the night of the caucus itself.

What the polling has provided is a good overview of the race to date. Among Democrats, it’s been a close three-way competition for months. Gradually, all the other contenders have faded away. Among Republicans, the Huckaboom surprised everyone but it remains to be seen whether Huckabee can hang on for a victory or if Romney will have the final comeback of this particular campaign. The final answer will be found in the turnout, not the polls.

The confusion over who is likely to win in Iowa can easily be seen in the Rasmussen Markets. They currently suggest that Obama has a % chance of winning in Iowa while Clinton is given a % chance, and Edwards a % chance. For Republicans, the markets imply a % chance for Huckabee and a % chance for Romney. This data is from a prediction market not a poll. If you disagree with these percentages, enter the market and add your opinion to the collective wisdom. You can sign up and participate in the market at no cost.

With the Iowa caucuses rapidly approaching, Fred Thompson recorded a fifteen-minute video message as his closing argument for that state’s voters. Other recent videos have been released by John McCain, Hillary Clinton, Mike Huckabee, Barack Obama, Mitt Romney, John Edwards, and Rudy Giuliani.