FiveThirtyEight.com: Politics Done Right: Obama's bounce in state polling

Wednesday, June 18, 2008

Obama's bounce in state polling

Below is a just-the-numbers list of all states that have been polled since the conclusion of the Democratic primaries that were also polled by that same polling firm in May; positive numbers indicate an Obama lead.

State               May      June     Change
AR  Rasmussen       -24       -9      +15
KY  SurveyUSA       -24      -12      +12
OH  Quinnipiac       -4       +6      +10
KS  Rasmussen       -19      -10      +9
FL  Quinnipiac       -4       +4      +8
NY  Siena           +11      +18      +7
WA  Rasmussen       +11      +18      +7
PA  Quinnipaic       +6      +12      +6
WI  Rasmussen        -4       +2      +6
IA  Rasmussen        +2       +7      +5
VA  Rasmussen        -3       +1      +4
MI  Rasmussen        -1       +3      +4
WA  SurveyUSA       +14*     +17      +3
NC  Rasmussen        -3       -2      +1
NC  Civitas          -5       -4      +1
MN  Rasmussen       +15      +13      -2
MN  SurveyUSA        +5       +1      -4
OR  Rasmussen       +14       +8      -6
=========================================
AVERAGE            -0.7     +4.7    +5.4

* Average of two May surveys.

99 comments

Anonymous said...: New Quinnipiac polls have Obama up in Fla, Penn and Ohio...it's a nice way to start the summer.; June 18, 2008 8:19 AM
Anonymous said...: wow; June 18, 2008 8:20 AM
Mike Plugh said...: Great work Nate. I don't know what the ceiling is for Obama in certain states, but it feels like this is just the beginning of a rather significant shift. It's prone to the ups and downs of the long campaign season, but (to use a baseball metaphor) it's like Obama's taken a 5 run lead early and has a stacked bullpen waiting in the wings. McCain is looking a bit like the Padres offense at this point.; June 18, 2008 8:20 AM
Jessica said...: drudge calls McCain

UNDERDOG!; June 18, 2008 8:23 AM
Falsehood said...: Now, if we only knew how much of this is "bounce," and how much of this support is his to keep.

Nate, how does that list of changes correlate with your coefficients in determining the size of a state's poll adjustment?

Also, are you displaying the actual polls in your data in the right column, or are all of those past sumber subject to change with the super tracker?; June 18, 2008 8:23 AM
Nate said...: Clearly, Obama's biggest improvement seems to be in Appalachia. That described Arkansas and Kentucky, and the Quinnipiac poll also showed him with much better numbers in Southern Ohio. That's why the model seems to think that West Virginia might be a swing state. Obama trailed by 9 there in a poll conducted just before the primary campaign ended, and if he gets an AR or KY type bounce there, it suddenly starts to look like a toss-up.; June 18, 2008 8:32 AM
Anonymous said...: Looks like the states with the biggest bounces are nearly all states where Clinton won the primary.; June 18, 2008 8:35 AM
Nate said...: Yes, I'm now using Clinton's vote share in the primaries (with a modifier for caucus states) as a variable to derive the state-by-state polling trends.; June 18, 2008 8:36 AM
NoDak said...: The Appalachia improvement demonstrates how toxic Bill Clinton's race-baiting campaigning style for his wife was there. Obama should keep Bill Clinton far from Appalachia if he uses him at all.; June 18, 2008 8:36 AM
DU said...: drudge calls McCain

UNDERDOG!

Pfff. Being a loser isn't enough to make one an underdog. You also have to be NOT part of the last 15 years' ruling party and heir apparent to the Chimperor.; June 18, 2008 8:36 AM
Christopher G said...: Nate, is there a reason the pacific seems to have disappeared from the state-by-state projections on the left side of the page?; June 18, 2008 8:41 AM
David said...: Nate-- in your Scenario Analysis, could you add a couple more "for fun" scenarios-- the probability of either Obama or McCain winning all 538 electoral votes? I've been a little surprised the last couple of days by the height of the bars on the very ends of the electoral vote distribution, but since there's no legend I'm not sure how many simulations that represents.; June 18, 2008 8:46 AM
Anonymous said...: That swing to Obama is so consistent across the states that it's scary.

What we need is some more randomness. Bring in some ARG polls, please!; June 18, 2008 8:47 AM
Anonymous said...: Wow is right... I'm Praying to God that it Holds and increases ...; June 18, 2008 9:00 AM
Anonymous said...: Anon @ 8:47: IIRC, I think the weight applied to ARG polls is so low that a standard 500-person poll that came out today would be weighted below 0.05. I'm still not 100% sure on how ratings translate into weights, though.; June 18, 2008 9:07 AM
Aaron said...: "Nate, is there a reason the pacific seems to have disappeared from the state-by-state projections on the left side of the page?"

Major earthquake. Oh the humanity.; June 18, 2008 9:09 AM
Anonymous said...: Couldn't find this in the FAQ.
Do you calculate Popular Vote and Electoral Vote numbers as an average of all the simulations?; June 18, 2008 9:11 AM
Anonymous said...: Couldn't find this in the FAQ.
Do you calculate Popular Vote and Electoral Vote numbers as an average of all the simulations?; June 18, 2008 9:11 AM
Juris said...: Sinking the Pacific Coast into the ocean? That was Lex Luthor's big plan. Superman!! We need you!!; June 18, 2008 9:23 AM
Anonymous said...: Can you explain to me why you have Obama up +2.8 points in Indiana when the sum of your numbers listed would put McCain up by +5.9 points? I mean, ok, new model, national polling, whatever, but... An 8.7 point adjustment, really? In Indiana? I'm just not seeing it. And would it be possible to add a line showing the impact of whatever adjustments you're making, because apparently that isn't all included in the regression.; June 18, 2008 9:31 AM
Matthew H said...: First off, I just can't believe your calclulations for Nevada, W.V., and Arizona. Something has to be broken here. When every survey in NV has shown it going for McCain, including one done last week, it seems wrong to predict it for Obama. It brings your methods into question.

By the way, how did you decide to divvy up the states this way on the left? I would have put Missouri into North Central and West Virginia into the Rust Belt, which in both cases fits better into how they vote and how they think. That leaves Highlands as Arkansas, Kentucky, Oklahoma, and Tennessee, The 'Southern' states not on a coast (yes, I know Kentucky fought for the North. Most folks in Kentucky don't seem to, though).; June 18, 2008 9:34 AM
Alex said...: Matthew, the most recent poll in NV is from a pollster that doesn't seem to have a history there, so we don't have as much information from it as we think. Wait until the Rasmussen/SUSA polls come out for NV--I imagine they are going to show a significant bounce, like the other state polling.; June 18, 2008 9:40 AM
Anonymous said...: PPP poll from Virginia -

Obama 47
McCain 45

Confirming Rasmussen's recent numbers.; June 18, 2008 9:50 AM
Anonymous said...: Great numbers from Virginia.
With the hard work from the liberal groups and the Obama army for registered new voters, Virginia will be blue in November.; June 18, 2008 10:10 AM
Frank from Germany said...: N

Pollsters and Poll Compilations

Other Resources

Labels

Blog Archive

Wonks

Wednesday, June 18, 2008

Obama's bounce in state polling

99 comments