[Author's note: This is a hell of a long post...apologies beforehand!]
I have long espoused the views that polls in the Democratic presidential primary are not as important as the media (and the blogosphere, to some degree) makes it out to be. Nevertheless, we now stand barely 6 months from the Iowa caucuses (unless they are moved up earlier), so I figured it would be a good time to take a look at what the national polls are telling us about the race for the Democratic presidential nomination at this point in time. It was around this time in the 2004 election cycle that Howard Dean began his meteoric ascent in the national polls, which then remained largely unchanged until John Kerry's victory in Iowa. That being said, the dynamics around the race in 2008 are quite a bit different. I'll summarize the differences to keep in mind, then turn to discussing the national polling that has been done to date - and what it means for the candidates so far.
2008 != 2004
The first - and most important - factor to keep in mind is that this election will not be like the last. Why? The biggest factor, first and foremost, is that George W. Bush will not be on the ballot. Due to this, I think there will be a much less likely scenario where you have a "Dated Dean, Married Kerry" situation that happened. Because Democrats wanted to dispose of Bush last time around as much as possible, it's been posited that primary voters ended up going for the more 'electable' Kerry instead of either Dean or Dick Gephardt, who were the presumed favorites in Iowa (Dean had the money, Gephardt had the close labor ties and geographical proximity). That's why I think it's less likely that you'll have a similar surge in the Iowa polls, like Kerry did in 2004, in the last couple of weeks in the race.
The next important factor? George W. Bush will be on the ballot. By this, I am referring to the political environment surrounding the 2008 election. Although neither he nor Dick Cheney will be on the ballot, the circumstances that America faces are a direct consequence of their mishandling of the country. And instead of facing Bush directly when his approval was in the mid- to high-40s, he is now sinking into the mid-20s - with nothing to stop his continued descent towards Nixonian levels. The war in Iraq is deeply unpopular, and that has infected the public's view of just about everything regarding the White House and the Republican Party. What does that mean for Democratic primary voters? I think it's also likely that there will be less focus on the 'electability' argument as a result. The general sentiment around the country - whether it's the generic ballot polls, most of the head-to-head matchups, the online trading futures, you name it - is that the Democratic nominee will be elected in November 2008. Therefore, it will reinforce the point I made above; voters will be less likely to vote based on this notion of who is the most electable.
Now turning an eye towards internal party dynamics, I think the biggest difference within the Democratic field this time around is that there are much bigger names. In 2004, Dean was able to catapult himself into the spotlight because none of the other candidates - Kerry, John Edwards, Joe Lieberman, Bob Graham, and so forth - were well-known to the general public. This time around, three Democratic candidates - Hillary Clinton, Barack Obama, and Edwards - are pretty well-known by the voters, with another prospective candidate (Al Gore) getting a lot of media attention. I think this is the reason why the poll numbers (as you will see below) have been rather static. With so many big players on the stage already, it's hard for a Bill Richardson or a Chris Dodd to break out of the low single digits nationally (although Richardson has been rising slowly in Iowa). It's why, despite 258,000 donors to his campaign (a staggering amount by anyone's measure), Obama has been unable to make steady advances in the polls against Clinton.
Analyzing the Polls
The national polls that I included below are all from data coming from PollingReport.com. In terms of prominent polls that are missing, I was unable to obtain Rasmussen's numbers, as they appear to be behind a subscription-only firewall at their website. That being said, you can view their numbers here at Wikipedia. As another point, I have only included polling data that has ending dates after February 10 (when the last of the 'Big Three', Obama, officially entered the race) and polls that have at least 3 observations between then and the present.
One problem with ARG's methodology is that it includes likely Democratic primary and caucus voters. Does this focus strictly on registered Democrats, or does it include Democratic-leaning independents? A little clarity would be nice. Nevertheless, this poll doesn't show promising trends for either Edwards or Obama. Edwards dropped off quite a bit after nearly catching up to Obama, who appears to have steadily dropped off since officially announcing. Meanwhile, Clinton continues to steadily climb in the polls, commanding a 20-point lead over Obama as of the latest poll.
The NBC News/Wall Street Journal poll shows Clinton climbing back up to a double-digit lead after having just a 5% lead over Obama, well within the margin of error. Again, the glaring error in this poll is that it includes non-Democrats who said they would vote in a Democratic primary. As this varies state by state, it's probably not the best idea to include such factors that make it statistically more difficult to model. I would suspect that Obama's bounce in the April numbers could be attributed to a post-1Q fundraising boost. Nevertheless, it seems that Clinton jumped back up, possibly aided by her strong performances in the 3 debates to date.
Quinnipiac only has numbers that included Gore, so I had no choice but to graph the numbers as given. As you'll see later on, I think that what we're beginning to see is the marginalization (or demotion) of Edwards as a top-tier candidate. Gore's numbers have continued to rise in recent months - winning the Academy Award for "An Inconvenient Truth", releasing his new book, "The Assault on Reason", and his just-completed Live Earth concert series have been giving the former vice president plenty of media attention. Edwards, meanwhile, has only made notable news for the mix-up regarding his haircuts - everyone else in this poll rises except for Edwards in the latest numbers. To boot, Quinnipiac also polled Democratic voters, which presumably includes only registered Democrats. It might have been better to apply a likely voter model.
These numbers from CBS, which are defined as 'Democratic primary voters', are probably a decent reflection of where most primary sentiment is nationwide (although the Big Three were the only named choices) - and it's a boon for Clinton and highly negative for Edwards. Clinton has steadily risen in the polls, likely a product of the debates; Edwards has fallen since April, probably because of the haircut news and possibly because voters haven't been particularly impressed by his debate performances (he's gotten better, but he has only stood out for calling out Clinton and Obama in the second debate over Iraq). Obama has seemingly found a plateau (or floor, perhaps?) at 24%, but this illustrates that the number of donors and the money raised doesn't move sentiment.
The Fox News polls, which I'm generally disinclined to favor - it is Fox News, after all, and the results are from all registered Democrats - but it seems in line with what other polls are seeing - Clinton is steadily increasing her margins, while Gore continues to overtake Edwards despite being undeclared. One additional note is that when Gore is not added as a choice, Hillary Clinton benefits the most. That may surprise many of us in the netroots, where potential Gore supporters are more inclined towards Obama or Edwards. But it's a trend that is playing out in other polls, meaning that much of Gore's support in these polls may not be 'grassroots' support. It may be a result of high name recognition amongst low-information voters, or it could be that there is a lot of 'institutional' support for Gore, so to speak.
CNN's poll is interesting in that some of Clinton's potential rise in the polls is instead given to Gore when he is included in the poll. Nevertheless, while the numbers differ slightly, it's evident that Clinton still maintains a healthy double-digit lead over the other candidates. In addition, Obama doesn't show any real upward trend in his numbers since declaring. It's not clear if the polls are simply missing what appears to be his enormous grassroots support, but it's not showing up in the national numbers. That being said, this poll should be taken a little more lightly than others due to the inclusion of Democratic-leaning independents.
Cook, in conjunction with RT Strategies, includes independents in this poll. It could explain why the lead between Clinton and Obama is lessened, but I don't think it is a big contributing factor. While there is more variability in these numbers, it shows the same trends: with Gore in the race, he overtakes Edwards; without Gore, the top three candidates are all separated by a tangible percentage of voters.
Gallup, compiling data for USA Today, by far has done the most polling for the primaries to date. They are also the only poll to show Obama actually overtaking Clinton at one point (although it is only when Gore is included), although the lead is statistically insignificant. In addition, this poll once again includes leaners, which is disappointing. On to the numbers show a fuller picture: Clinton's lead, contrary to most of the other polls except for Cook, seems to be trending downwards. Obama has trended upwards slightly, but took a hit in the last poll numbers. Edwards has been steady, if not trending down slightly. With the exception of a few dates in April, Gore has led Edwards; again, with Gore out, Clinton benefits the most. I would be inclined to think the June 3rd results were somewhat of an outlier. But USA Today has other interesting numbers to peruse: the results of a direct Clinton vs. Obama matchup.
This seems to confirm the general trends shown above: since Obama's entry on February 10th, he has trended upwards, while Clinton has come down. True, she still leads Obama by 11 points, but that is a far cry from the 29-point lead she had just as Obama declared his candidacy. What is there to take away from this? It's possible that the anti-Clinton forces would benefit by coalescing behind either Obama or Edwards (who wasn't polled in a similar fashion by Gallup) and focusing their attention on the front-runner instead of each other, as seems to be the case online.
Conclusions
- Hillary Clinton remains in a strong position nationally. While I still think that whoever wins Iowa may be able to steamroll, the presence of an effectively 'national primary' on February 5th is undeniable - and makes these numbers somewhat relevant. This bodes well for Clinton, who has been able to maintain her lead despite being outraised in primary dollars by Obama. While some polls show her trending either up or down, she still maintains a healthy lead over her closest competitors.
- Barack Obama must find a way to expand his support. While he has made a splash with his entry into the race, along with raising ungodly amounts of money through an amazing base of support, Obama has barely been able to make the kind of permanent dent in Clinton's edge that he needs. I'd be interested in knowing how many of his donations came from people who are registered Democrats, as it's clear that he can claim support from independents and Republicans. It could just be the Clinton machine being more efficient than Obama's, as he hadn't been planning to run for president until a few months before he announced. Nevertheless, Obama has not seen the numbers on the national scale (or the local polls, either) that he should be seeing.
- John Edwards is being marginalized and is on the verge of becoming a second-tier candidate. It'd be interesting to measure the effect of the haircut snafu - however unfair it was to the candidate (and I thought it was), it seems to have had a chilling effect on his poll numbers (and perhaps his 2Q fundraising as well). He continues to lead in Iowa, but barely - both Clinton and Obama are breathing down his neck, and if he cannot maintain his lead in Iowa and win there, his campaign is dead. Regardless, his declining poll numbers - sometimes into the realm of single digits - is worrisome.
- Al Gore's potential candidacy will have a material effect on this race. If Gore does not enter this race, it will be an immense boost to Clinton's campaign - somewhat ironic, considering that Gore and Hillary Clinton never much liked one another. It is somewhat paradoxical, but having another 'anti-Clinton' candidate like Gore in the race would actually help make it more competitive. I also suspect that Gore's support in these polls is understated, as the average primary voter will remember what happened in 2000 - and then recall that Gore has been right on every single issue since then. Will he or won't he? That's a question that we will not have a definitive answer for until after the Nobel Prize winners are selected. At this point, though, only Gore's entry would have an earth-shattering impact on the primary.
(also cross-posted at MyDD)