4.07.2008

Poll roundup: Clinton loses her grip on Pennsylvania

For the first three weeks of the Pennsylvania campaign, Hillary Clinton not only maintained but even slightly expended her lead. But since then, it has been Ohio and Texas all over again: Poll after poll show the race tightening. And though there has only been one survey for now showing Obama with the lead, ARG released one this morning showing a tie between the two Democrats:

  • Clinton and Obama both come in with 45%. Clinton's 16% edge among white voters is offset by Obama's 89% among back voters.
  • ARG's previous poll released late March showed Clinton ahead 51% to 39%, a quite dramatic drop within two weeks.
Needless to say Clinton needs to win Pennsylvania to even stay in the race until the next context, let alone have any chance of taking the nomination. An Obama victory in the Keystone state would at once undermine many of Hillary's central arguments, in particular that Obama is too weak among key groups (registered Democrats, blue-collar voters) to be competitive in the general and that the late campaign is favoring Clinton with voters coming to realize they prefer her.

Do not write off Clinton's chances in Pennsylvania, of course. In the run-up to Ohio, I explained that the fact that we were talking about who would win that contest already meant Clinton would lose since she didn't need to just win Ohio but get a double-digit victory. But when all was said and done Hillary pulled out a 10% win. The trouble for her is that she needs even more than that in Pennsylvania to truly make her points, so if polls confirm that there now is a serious contest for the win rather than for the margin it is hard to see how Clinton can recover.

Meanwhile, two North Carolina polls suggest that the state's giving itself to Obama by a massive margin (Rasmussen suggested 23% yesterday) is not a foregone conclusion:

  • McClatchy shows Obama leading comfortably, 45% to 36%. 20% of Clinton supporters (shall we say "only 20%" at this point?) said they would vote for McCain if their candidate did not get the nomination.
  • Meanwhile, a Raleigh Observer poll was also realeased showing Obama ahead 32% to 26% -- as you can see undecideds were really not push by the pollster here.
North Carolina's primary seems to be the type of contest Clinton needs in Pennsylvania, one in which what is at stake is the margin without much suspense about the winner. Polls are showing an Obama victory anywhere from high-single digits to the 20% range. Even if Clinton is alive to fight on May 6th and even if she gets satisfying results in Indiana, she needs to make sure to cut her losses in North Carolina -- and she particularly needs to focus on winning the blue-collar white vote, the loss of which would undercut her central electability argument.

Finally, Rasmussen released two general election polls yesterday showing mixed results for both parties -- but mainly underscoring John McCain's strength:

  • In Iowa, one of the rare states in which there is a consistent electability difference between the Democrats, Obama edges out McCain 46% to 42%, but the Republican crushes Clinton 51% to 36%.
  • In Tennessee, however, it is Clinton who looks stronger, though both Democrats are trounced by McCain: Clinton is led 52% to 38% and Obama 58% to 31%.
Iowa will likely be one of the most important battlegrounds of the fall campaign after it handed George Bush a very narrow victory in 2004; Tennessee is less important in the logic of a general election campaign given that it did not even win for Al Gore in 2000. Let's note, however, that past polls showed that Tennessee could be in play: A survey in late February showed Clinton trailing McCain by only 4%, while Obama was crushed by 16%; an even earlier poll back in November showed tighter margins for both Democrats.

Labels: , , ,

10 Comments:

  • Great article and exactly right. Clinton needs to win PA by a larger margin than she did in Ohio. So a 15% plus victory which just a week or two ago seemed possible is looking unlikely. She has been able to focus on PA and PA is older, less eductaed and whiter than even Ohio which reinforces her demographic advantages. In this time the press have been harder on Obama than in the past and the Wright controversy blew up. So she really should win by 15%+ let alone that she needs that to make good gains on the pledged delgate (and popular vote) count.

    Clinton did well in OH and TX because for the first time she ran the 3am ad (in TX only) and that made the national news. It will be interesting to see if anything of similar impact could be run 48-96 hours before polling. If not then she probably won`t get a late bounce that the polls miss.

    By Anonymous Anonymous, At 07 April, 2008 11:49  

  • article was all B.S. Clinton is dead woman walking, it's over! Unless she can win ALL THE REMAINING STATES BY AT LEAST 20%!!!!.Than she only ties Obama!!! Don't you get it??? It's over!!

    By Anonymous Anonymous, At 07 April, 2008 12:40  

  • Rolling poll average, PA
    Clinton vs. Obama
    Polls only from 04/01-04/06
    9 Pollsters, one poll per pollster:
    ARG, Insider Advantage, Muhlenberg, Morning Call, PPP (D), Rasmussen, SUSA, Strategic Vision (R), Quinnipiac

    Averages:

    Clinton: 47.7%
    Obama: 41.4%
    Undecideds: 10.9%
    Margin: Clinton + 6.3% (2.8 outside the MoE)

    Averages for the entire month of March (some pollsters took more than one poll):

    Clinton: 51.8%
    Obama: 36.6%
    Undecideds: 11.6%
    Margin: Clinton: + 15.2

    Margin difference between March and April thus far:

    Clinton - 8.9%
    Clinton is losing ground, and very fast.

    To catch up and surpass Obama in the PDs, she must win at least 65% (that is at least a 30% margin) in every single one of the last ten contests, without exception. Her current margin of + 6.3% is 25.7% away from what she needs. A resounding 55%-45% landslide may look impressive, but won’t do the trick in order to take the PD lead.

    What a 55%-45% landslide could do, if 80% of 4,000,000 registered democrats in PA go vote, would give her a plurality of about 400,000 votes and would cut deep into Obama’s PV lead.

    If the polls continue at this rate, then she will win PA with way under a 10% margin.

    By Blogger Statistikhengst, At 07 April, 2008 12:44  

  • As I've written previously, these recent polls show that we can't take any one of them as definitive. Sure, I look at them too to try to get a hazy picture of where things are going. But they should be examined collectively, not as the last one overruling all the ones that came before them. The polling methodologies vary too much to reach hard conclusions based on the last poll alone (sequential polls from the same pollster might be more valid in showing trends, yet even those may not be predictive of what will happen next).

    That being said, it does seem that PA is tightening. An objective look at the numbers would indicate that the nomination is practically decided anyway, so the question now is timing.

    At present, some superdelegates may be hesitant to commit because they don't want to be accused of taking the nomination away from the voters. But when it becomes (even more) evident as to who the pledged delegate leader will be, I'd bet that some of them will say the outcome is sufficiently clear that it's not worth dragging out the process anymore at the expense of the party. PA may not be that time, even if Clinton wins by less than 10%; NC and IN could be the trigger, and I'd guess that OR and KT on May 20 would definitely get most remaining superdelegates of the fence because there would be so few pledged delegates to go. So I see Obama getting enough delegates for the nomination by early June or so.

    (Excuse me while I prepare to eat some serious crow, just in case.)

    By Blogger dsimon, At 07 April, 2008 13:22  

  • I would say if the superdelegates are going to step in before the primaries finish it will be either after PA (say if Clinton wins <10%) or after NC and IN (Obama wins NC >15% and IN is very close). I don`t see them steeping in at KY and OR because the primaries have only 2 more weeks after those two states so not point here. NC and IN is still one month before the end - a month is valuable, 2 weeks less so.

    By Anonymous Anonymous, At 07 April, 2008 14:39  

  • Superdelegates may step in before OR and KT, but there might not be enough of them to secure the nomination total without more pledged delegates. So the process may have to play out a little longer for the numbers to add up.

    By Blogger dsimon, At 07 April, 2008 15:22  

  • I wonder if she wins, but only by 4% or 5%, will she drop out so she can go out on a high before being blown out in North Carolina?

    By Anonymous Anonymous, At 07 April, 2008 18:59  

  • I wonder if she wins, but only by 4% or 5%, will she drop out so she can go out on a high before being blown out in North Carolina?

    I don't think so. I doubt she can go out on any win in PA, even if it's close, because her supporters will see it as a win and will still see the delegate race as close, and they'd be upset if she dropped out at that point. I'd guess she'd want to keep her core supporters happy, so on we go.

    I think it will drag on until Obama gets the delegates he needs (without MI and FL, whose fates will be decided later by how big the delegate spread turns out). But I expect that will happen by early June at the latest.

    That's my armchair analysis, anyway.

    By Blogger dsimon, At 07 April, 2008 19:37  

  • my figures at Anon 12:40 are off and I cede to Mark for the correct #'s.

    By Anonymous Anonymous, At 08 April, 2008 07:31  

  • lol.

    Hi anon, no problem, my numbers have not always been correct either.

    Poll average correction. The big SUSA poll came in a number of hours ago, replacing the last SUSA numbers.

    So, the new averages for 04/08

    Clinton: 48.2%
    Obama: 41.6%
    Und: 10.2%
    Margin: Clinton + 6.6%

    SUSA shows a huge lead for her (18 points), but the other polls from yesterday and today do not show this. All nine mixed together, and the margin average for Clinton has improved today by 0.6%

    I would say, don't uncork that Champagne quite yet.

    By Blogger Statistikhengst, At 08 April, 2008 18:35  

Post a Comment

Subscribe to Post Comments [Atom]



<< Home