Yesterday, I posted a chart of the
national primary preference polling since January 2003. Those polls are great fun for trendwatching, as I said at the time, but it's the state-by-state contests which will actually determine the nominee.
As such, I follow up today with charts of the primary and caucus polling in Iowa, New Hampshire, and South Carolina since January 2003.
But wait. Surely Arizona and Oklahoma and the other Feb. 3 states are important, too! And you're absolutely right, they are. But there's a problem: I have three polls apiece for AZ and OK (with incomplete information on two out of three OK polls) and none or nearly none for the rest. Charting them yields no useful information, so I haven't.
As before, these charts stretch the hell out of the page, and as such, are posted behind the extended content link. Once again, monumental kudos to Adam in Mass for hosting these.
The first point to note is that because the polls have shifted less violently (compare the
raw national polls) in these three states, charts of the raw polling are still more or less legible. However, as before, the more meaningful information is derived from the moving averages (see my original post for the explanation of why) and as such I'll base my analysis on those. One note: because there is less data to work with, I use a three-poll, and not a five-poll, moving average on the state polls. This has the effect of flattening out spikes slightly less.
Some people had problems with the charts displaying properly yesterday; if the image is illegible, there are links below the charts.
Iowa
(raw polling, moving averages)
I remarked yesterday that the most remarkable feature of Dean's Teflon is that his polls never experience any sort of sustained drop. You'll notice this same trend in Iowa. His polls are either tracking upward or flat; he's never dipped all year.
This is extraordinarily bad news for Gephardt -- you'll notice that he's spent the entire year boxed in a band roughly between 20 and 25 points with a brief break out to a high of about 27. This seems to suggest that he has a strong core of support (well, we all knew that) but has done remarkably poorly at reaching outside of it. You'll note towards the end that Dean is pulling thoroughly away, and thus Gephardt appears to have broken the cardinal rule of the poll-watching anti-Dean: never let Dean pass you. There are an enormous number of variables still to intervene, not least the fact that what matters in a caucus is not the size of your support but your ability to get it there on the night; on this basis, there's no way Gephardt is toast yet. But for the moment there does seem to be a tenuous Dean edge in Iowa.
It's very important to note that these polls predate both the Gore endorsement and the capture of Saddam. We won't see how that's affected the race until the first polls are taken next year.
Kerry appears to be solidly entrenched in third, hugging the 15% line. Edwards made a spirited bid for it but never came within striking distance and is now slumping back down again.
Everybody else is utterly irrelevant.
The reason, by the way, that some of the lines are longer than others is that the most recent poll is the 8 December SurveyUSA poll; and SurveyUSA doesn't report percentages for the minor candidates, which in this case includes Lieberman and Clark as well as the usual three.
New Hampshire
(raw polling, moving averages)
So many polls, so little we didn't know already. Dean is walking away with New Hampshire unless he's found in bed with the proverbial dead girl or live boy, and even then he'll still beat Kerry. The New Hampshire polls are notable for one thing: for about thirty days in September-October, the voters handed Dean his first measurable trough in polling. I still can't figure out what the hell caused this; but he recovered handily and is now well on target to break 50% in the eventual primary and, if he gets a bounce off Iowa, possibly even 60%.
No, the real interest is down near the delegate allocation threshold, with the Kerry-Clark fight for third. Notice that late in the year the two lines were almost exactly mirroring each other, suggesting that Kerry and Clark were regularly swapping supporters. However, it looks like Kerry has now rallied and cemented his hold on the second spot -- although it remains to be seen what effect the capture of Saddam will have. I've argued elsewhere that a strong but not too strong Kerry is exactly what Dean wants: a Kerry second in NH is Pyrrhic, as Kerry will be squeezed to death between Dean and Clark if he hangs on through the Feb. 3 primaries; however, an expectations-defying Clark second is worth several points in momentum going into Feb. 3, and those few points could decide whether Arizona, in particular, goes for Dean or Clark.
Below that, there's the fight for the moral victory of fourth place. When the overall numbers, never mind the gaps, fall inside the MoE, it's dangerous to speculate, but Lieberman does seem to have succeeded in grabbing a decent hold of fourth. There's a possibility, albeit not a big one, that he might overtake Clark. Gephardt and Edwards are fighting it out for fifth, and everybody else is, as always, blipsville.
It's worth noting that it's possible to get delegates in New Hampshire with only about 7% of the statewide vote, because most of them are allocated by congressional district; if you poll well in District 1 and lousily in District 2, for example, you'll get a share of District 1's delegates. Thus, Dean, Kerry, and Clark will almost certainly walk away with some delegates. It's about a fifty-fifty chance, at the moment, whether Lieberman gets any.
South Carolina
(raw polling, moving averages)
SC is a hopeless tangle, moving averages or no moving averages. Two polls now have shown Dean with a slight lead, enough of one that he has pulled ahead by a nose in the moving averages, but going just on the polling there are no less than six candidates with a realistic shot at winning -- and embarassingly enough for the Senator from Massachusetts, he shares the 'no-hopers' zone with CMB and Kucinich.
However -- and it's something of which the other candidates should definitely take note -- we have the same phenomenon here as we did in Iowa and the national polls: Dean's polls are always trending up, never down. However, this time there's a second candidate doing the same thing: Clark. A lot will depend on the events of Iowa and New Hampshire, but at the moment a genuinely competitive Dean-Clark two-horse race in SC does not look out of the realm of all possibility. The Dean people have finally woken up to the fact that SC is winnable for them, and are pouring in resources.
The same proviso to New Hampshire applies to South Carolina; with six congressional districts, virtually everybody of the major six is likely to take some delegates home if they're still in the race.
So what does all this actually mean? I don't know. Dean has a slight edge in Iowa, a commanding lead in New Hampshire, and South Carolina's anybody's. I'll make this prediction: if Dean wins Iowa, New Hampshire, Arizona and South Carolina, game over. If someone else wins South Carolina, Dean's favoured but it will take a while longer. If someone takes South Carolina and Arizona away from him, we have a race.
Some time in the New Year: Bush re-elect.