Expertinent is a regular Stumper column featuring interviews with experts on the news of the day.
By now, you've likely seen yesterday's most important numbers: 58 and 41. The first was Barack Obama's share of the vote in the crucial Wisconsin primary; the second, Hillary Clinton's. That 17-point spread immediately transformed Obama, a one-time insurgent candidate, into the presumptive frontrunner for the Democratic nomination. As every pundit pointed out, Obama had beaten Clinton in a state where she had no clear excuse for defeat, setting the stage for showdowns in Texas and Ohio on March 4 and leaving little leeway for further losses.Are those losses inevitable? Does Obama's Wisconsin win mean victory in Ohio and Texas? Or can Clinton battle back? Seeking answers, I decided to call Charles Franklin, a poly-sci professor at the University of Wisconsin and co-founder of Pollster.com. Using last night's exit polling as a guide, Franklin mapped out where Clinton is losing ground--and the challenges she faces going forward. Excerpts:
What was surprising about the results in Wisconsin--especially considering where the race stood a month ago, or even a week ago?
The real effect is we went from polls that showed a small Obama lead to this 17, 18 point ultimate win, and how deep that win is among demographic categories. On the Clinton side, it's how her strongest groups were groups that she managed to only barely win. That, clearly, is the message of yesterday's primary. Obama ran even or even won among Clinton's key supporters--women, middle-aged people, union members, Catholics and core Democrats--while running up huge margins of 20 to 70 points in his strongest demographics: black voters, young voters, etc. If Clinton is going to build a winning coalition in the upcoming states, she's going to have to do a lot better with her "base." Sure, she did okay with them in Wisconsin--but only because she didn't lose them by double digits [like she did with Obama's strongest groups].
Do those shifts--in effect, Obama's poaching of core Clinton supporters--have to do with something specific about the character of the state? Is Wisconsin different in some essential sense from earlier states?
Actually, these differences are largely in line with what we saw in Virginia and Maryland. They seem to be on course with a trajectory of Obama improving across the demographic groups from earlier in the process to Super Tuesday and then in these post-Super Tuesday states, where that improvement has continued. They're consistent with a long-run, rising trajectory for Obama. It's been pretty broad, actually, in terms of the groups Obama has cut into. It's not that Obama is only winning African-Americans or only winning people under 35. This advantage he's been gaining has really been across a whole lot of groups.
Would that cut against the Clinton camp's charge that some of these 10 states that he's won were outliers?
I think it would. You don't look at the states post-Super Tuesday that we have exit polling for and see them stand out as exceptionally different. There is variation in the racial composition, from what turned out to be about 9 percent black last night here in Wisconsin to substantially more than that in Virginia and Maryland. But if you look at other demographics, Wisconsin's not that different from Ohio, for example. And the bottom line is that we've seen these same patterns across states, not just in one or two.
Let's talk specifically about this trend of broad demographic improvement you've been seeing in Obama's numbers.
Sure. Take the white vote, for example. We've seen trends of Obama's share of the white vote going from 24 percent in South Carolina to 31-44 percent on Super Tuesday to now running very close, neck-and-neck with Clinton among whites. As long as Obama does that well among white voters--he actually ended up winning whites by six points here in Wisconsin--then the racial divide that we heard so much about back in January is either effectively a net zero or a very small advantage for Obama.
What about the gender gap?
The gender gap is one of the more interesting and telling ones. Clinton managed to win women by three percent, but she lost men by two-to-one, 66 to 32 percent. Clearly, her campaign has only managed to marginally gain an advantage among women, doing just better than barely breaking even. But whether it's her campaign that has alienated men, or it's Obama's campaign that's attracted them, or whether this also has something to do with latent levels of sexism--men being reluctant to vote for a women--nevertheless the bottom line is that for a group that's nearly half of the Democratic electorate, Clinton has been doing stunningly poorly among male voters. Looking ahead, if she's going to get her campaign back on track, this is a group where she desperately needs to cut down Obama's advantage.
Obama is essentially neutralizing what her campaign thought would be their silver bullet: women.
Right. If it's a three-point advantage for Clinton among women, that's pretty small. And that despite the fact that female turnout, at least as a share of the electorate, went up in Wisconsin from 52 to 57 percent from 2004 to 2008. So something mobilized more women as a portion of the electorate, but Clinton certainly did not win a lion's share of women--only a very small margin. Ultimately if Ohio and Texas are going to be Clinton's firewall, the firewall is constructed out of the bricks of the individual demographic groups. And that means she's going to have to be running far better among many of those groups in those two states than she's been doing the post-Super Tuesday exit polls.
The Clinton campaign would argue that demographics are not destiny in this case--that Ohio and Texas are going to vote in a very different way than Wisconsin. Do you see any reason to believe them?
I would agree with the campaign to the extent that there are some armchair analysts who take the vote shares in demographic groups for one state and then multiply them through the the demographic contest of another state and then use that to make some guess of the vote. The trouble is, states differ a good bit from one another. They move up and down across all groups in their support for one candidate or the other. So that misses the issue of whether Texas is just intrinsically friendlier to Clinton across groups.
Having said that, though, are there any lessons in the Wisconsin vote for the upcoming contest in Ohio? Is Ohio intrinsically friendlier to Clinton? I mean, she leads in the polls by 15 or 20 percent.
The differences between Ohio and Wisconsin are modest. Some of them favor Clinton--maybe more favor Clinton than Obama. For Clinton, for example, Ohio is about six percent more older voters, 60 and older. That's the one age group that Clinton won, and won it big--she won it here by 13 points. So that's a little bit of an advantage. Ohio is about eight points fewer college graduates, and again that's a group--non-college grads--in which Clinton held her loss down. On the party side, there's another advantage. Democrats are about ten percentage points bigger within the Democratic primary in Ohio, and Republicans and Independents are five points lower each in Ohio than they are here. And among Democrats, Clinton lost--but only by about three points. Union members, too--they make up nine points more in Ohio than in Wisconsin.
But the demographics could cut the other way--especially when you factor in Obama's gains in Wisconsin. For example, union members broke dead even between the two here, and non-union members who live in union households went nine points for Obama. Ohio is eight points more African American, so that should help him; seven points less Catholic, which also should help him. Neither advantage is gigantic. After Wisconsin, Ohio is sort of a draw in demographic terms. The real question for me in Ohio is what's that split among white voters. Will the break for Obama like they did in Wisconsin? That's something we really don't know yet. We need polling data--or an election--to find out.
So do you think these demographic differences account for the fact that Obama was leading in the Wisconsin polls before the primary but Clinton is still way ahead in Ohio? Or is it just unfamiliarity--the fact that the campaigns are just starting to make their local appeals.
I think that it has more to do with unfamiliarity. We think that the campaign has been around for a year, but there are an awful lot of voters that are just beginning to realize it's coming up. My niece asked me at Christmas: "Oh yeah, who's running for president this year?"
The point being that there are plenty of people like your niece in Ohio and Texas.
Absolutely. It's important to realize that normal people don't pay the kind of attention to politics that we do. It means there's still a lot of learning to go on in Texas and Ohio.
Is that good or bad for Clinton?
In one sense, that's good for both of the campaigns. They can legitimately think that they have a little more control over their fates in these upcoming primaries. But it also means that these upcoming primaries are occurring in an environment where the story has been set by Super Tuesday and especially these ten post-Super Tuesday Obama wins. So that's probably pretty bad news for Clinton. Even if she's been leading in the polls, that should raise some concerns in her campaign.
I wrote the other day about Gallup's latest national tracking numbers, which gave Obama advantages among women, Latinos and lower-income Democrats--all the traditional Clinton core supporters. Is that an effect of the national narrative--and, if so, is it something that will trickle down into Texas and Ohio?
It's certainly part of the national narrative. Campaign people and especially self-interested campaign people like to point out that national polls don't mean anything when you go to the state polls. But in fact, that's empirically false in one sense. If you look at the trends over all the states where we have polling over the last 14 months or more, and you compare those trends in the individual states with the trends that you see in the national polling, all of them move in a close parallel pattern. Some states have higher or lower levels of support for a particular candidate, but when it comes to the dynamics it's pretty much the case that all states are moving in similar directions. When we take, say, the Gallup data, it's a reasonable estimate based on the statistics--if not some vague political sixth sense--that if it's moving in Obama's direction nationally it's moving in his direction in Ohio and Texas. That doesn't mean he's necessarily ahead in those states. But the trends should be in his direction based on the track record of all the states over the last fourteen months--something upwards of 1,000 polls.
I know this is perilous, but do you have any hunches or predictions--particularly from a statistical point of view?
From a numbers perspective, it's looking hard for Clinton to manage big enough wins in Texas and Ohio to substantially improve her delegate total. Maybe she'll do that. But if she does, she'll do it by changing substantially the demographic profile we've been talking about. Part of that is the question of what message can she use that hasn't been used effectively before--a message that changes the views of Barack Obama among men or among whites or among union members. So far, that message has not been ready at hand. We saw her try to find that message in Wisconsin with the plagiarism charge, but I don't think there's any evidence in these exit polls that it was successful. Where is that criticism of Obama that ends up being really telling with the groups he's been winning by double digits, and that could bring him down to single digits or even give her an advantage? We're waiting to see that.