Wednesday, July 9, 2008

The Electoral College Map (7/9/08) [Update]

This morning I was like a kid on Christmas morning (Well, maybe not that excited, but...). I was all prepared for a rather boring glance back at the electoral college map we put out on Sunday and a brief discussion of the everyone's favorite game: who's that VP. And then I saw that Zogby International had state-level, head-to-head polls out in 34 states with some fairly interesting results:

New Polls (July 6-9)*
AlabamaZogby Interactive+11
ArizonaZogby Interactive+3
ArkansasZogby Interactive+2
CaliforniaZogby Interactive+20
ColoradoZogby Interactive+2
ConnecticutZogby Interactive+16
FloridaZogby Interactive+4
GeorgiaZogby Interactive+6
IllinoisZogby Interactive+20
IndianaZogby Interactive+1
IowaZogby Interactive+4
KentuckyZogby Interactive+5
LouisianaZogby Interactive+7
MarylandZogby Interactive+24
MassachusettsZogby Interactive+25
MichiganZogby Interactive+14
MinnesotaZogby Interactive+16
MissouriZogby Interactive+2
NevadaZogby Interactive 0
New HampshireZogby Interactive+3
New JerseyZogby Interactive+13
New MexicoZogby Interactive+16
New YorkZogby Interactive+21
North CarolinaZogby Interactive+9
OhioZogby Interactive+5
OklahomaZogby Interactive+5
OregonZogby Interactive+16
PennsylvaniaZogby Interactive+10
South CarolinaZogby Interactive+1
TennesseeZogby Interactive+5
TexasZogby Interactive+3
VirginiaZogby Interactive+5
WashingtonZogby Interactive+13
WisconsinZogby Interactive+10
*All polls from Zogby International. Follow link and click state for poll data.

The real surprises were the deep red states that have Obama in the lead: Arizona, Arkansas and South Carolina (Yeah, I know South Carolina isn't deep red--really isn't now--but I just can't shake recent and even not so recent history in the Palmetto state.). Arizona is the shocker. I can't imagine that the Grand Canyon state will go against its senior senator in November, but the fact that it is following neighbors, Colorado, Nevada and New Mexico into potentially competitive status, can't come as welcome news to the campaign of the presumptive Republican nominee.

On the blue side of the ledger, things remain pretty much the same. Obama is still running strong in the northeast (New Hampshire being the only possible exception. The Granite state's numbers jumped after Obama's clinch but have come back down to earth a bit since. Is that a bounce?) and through that swath of states from Wisconsin over through Michigan and Ohio over to Pennsylvania.

Despite all the new polling, there really isn't that much in the way of change. Ohio, back on June 18 was the last state to shift from one side to the other (red to blue), but since then the map and the average behind have simply been fine-tuning the positions of both candidates in each of the states. We have a good amount of information in most of the states now so that even the polls that appear to be outliers are quickly absorbed without altering the fundamental shape of the race. South Carolina is one of those exceptions. There are only three polls out now from the Palmetto state, but each has been in the single digits. One double digit poll (likely in favor of McCain) would shift things substantially with that limited amount of polling. In this case however, the third poll is one that shows Obama ahead in South Carolina by one point. That techinically isn't an outlier since the one and only poll that we had prior to last month came out in late February and pegged McCain's lead at just three points. What the new poll does do is pull South Carolina back into the toss up category it had been in until last month's Rasmussen poll.

Changes (July 6-9)
New MexicoToss Up ObamaObama lean
North CarolinaMcCain leanToss Up McCain
South CarolinaMcCain leanToss Up McCain

Beyond South Carolina, though, there were only two other states that switched categories. Both New Mexico and North Carolina moved toward Obama. In each, unlike South Carolina, there has already been a fair amount of polling, the most recent of which have trended toward the Illinois senator. Even with that said, both Zogby margins were a bit beyond what what has been witnessed in the Old North state and the Land of Enchantment. Each serves as an point of punctuation, if not peak, in the recent polling in both states for the Obama campaign.
[Click Map to Enlarge]

All that really happened then, was that the electoral vote numbers on both sides were tweaked without shifting any over to the opposition. Obama slid a handful into the lean category while McCain lost some leaners to toss up status on his side. The Arizona senator's strong and lean electoral votes add up to 176; one more than the number Obama currently holds solidly. And with Obama moving to the center--with a good amount of press coverage but no real damage to his campaign--that cannot be good news for the McCain folks. While that doesn't necessarily bode well for McCain, the fact remains that he only trails in this accounting of the electoral college by 58 electoral votes with 13 states and 140 electoral votes up for grabs.

The Watch List*
Alaskafrom McCain leanto Toss Up McCain
Arizonafrom Strong McCainto McCain lean
Floridafrom McCain leanto Toss Up McCain
Minnesotafrom Strong Obamato Obama lean
Mississippifrom McCain leanto Strong McCain
Nevadafrom Toss Up Obamato Toss Up McCain
North Carolinafrom Toss Up McCain
to McCain lean
Ohiofrom Toss Up Obamato Toss Up McCain
Oregonfrom Obama leanto Strong Obama
Virginiafrom Toss Up McCainto Toss Up Obama
Washingtonfrom Strong Obamato Obama lean
*Weighted Average within a fraction of a point of changing categories.

Of those 13 states, only four are currently on the Watch List (states that could potentially move given future polling). And all have been trending toward Obama since May with the possible expecption of Nevada. The Silver state just looks close. With North Carolina and Virginia already "giving" McCain 28 of his 240 total electoral votes, that leaves just 25 electoral votes in Nevada and Ohio as the one most immediately up for grabs. That would only cut the margin in the map above by less than half; obviously not enough to put McCain in the White House.

The list did lose Massachusetts, Missouri, New Mexico and Texas while adding Arizona, Oregon and Virginia. Of those Missouri, Virginia and possibly New Mexico are the only competitive ones. The eleven above though are the ones to watch as we head into the weekend.

...and hopefully some increased polling (now that July 4th is over).

[Update]: I should also have mentioned that Bob Barr did really well in these Zogby polls, getting anywhere from 2 to 10% in each. I'll have to look at those averages across all 34 states, but if the former Georgia congressman were to continue pulling 8% in places like Colorado or Iowa, it would make McCain's job of reaching 270 that much more difficult.

Recent Posts:
Polling Alert

Jesse Helms and the Current American Political Climate

The Electoral College Map (7/6/08)


Jack said...

Saw these polls earlier on another site. I'd love to believe them, but I can't. AZ, AR, KY, NM. NC, OK, SC, TN, and TX seem extremely odd, and LA and MO seem a bit suspicious too. The aforementioned site also was quite skeptical about the value of these polls.

Anonymous said...

I'd assume you are most likely referring to Nate's post on FiveThirtyEight this morning. I agree with Nate that Zogby's methodology can certainly be questioned and that, furthermore, some of these numbers are beyond what we've seen in recent polling.

As I mentioned in the post, though, none of these polls really shook things up all that much. I can't agree with you more that the polls in the states you mentioned look suspicious. They do. Some, though, are likely aberrations, but all the same, didn't severely alter our weighted averages. The exceptions are states like South Carolina, where polling to date has been sparse. It is pink on the map, but won't stay that way.

Anonymous said...

Much as I'd like to believe them, Zogby's numbers mean nothing. This was an interactive poll, meaning anyone who wants to sign up to be included in these polls can be. By definition, then, they're not random but heavily skewed toward those who go to the Zogby site and sign up.

Jack said...

I was indeed referring to the FiveThirtyEight post.

About the Barr results, I'm rather suspicious of the relatively high numbers we've seen for him in some polls, largely because of recent election history. Some of it might be explained by some sort of Ron Paul effect.

However, I have been particularly surprised by polls that show Ralph Nader with decent numbers. There was one national poll, I believe, that had Nader getting 4% of the vote. That would be about ten times what he got in 2004. I don't think Nader has become ten times as popular in the last four years.

Is it possible that people are more likely to state that they would vote for a third party candidate in polls, especially early on, but would be afraid to waste their vote come election day?

Unknown said...

Anonymous @9:25:

As I understand it, it's not quite as bad as you make it out. You do have to sign up, and that most assuredly introduces bias. But you sign up generically. Only a small fraction of people actually get polled on the Presidential election. In other words, it's a random sample of a non-random population.

That's true to some extent of all polls. After all, people who use CallerID to screen calls from people they don't know aren't in the population. For Rasmussen, people without a land line aren't in the population.

I agree that these Zogby Interactive polls have very fishy results, particularly in the states with the smaller sample sizes, and I'm basically ignoring them until I see confirmation. Some day this methodology may work fairly well, but we're not there yet.

Robert said...

I can't believe that you all are dissing Zogby. In 2000, he was the only one who predicted a Gore win in the popular vote but a much closer vote in the Electoral College (sorry Anonymous for bringing that up). He was also more consistently right in the Democratic primaries this year than any of the other pollsters. I put more credibility in Zogby than in any other pollster.

Unknown said...

Robert--There's a difference between Zogby's phone polls and "Zogby Interactive" (these internet polls).

Anonymous said...

We have a good discussion going here with a couple of different branches. The intent of the post was obviously the electoral college, but the Zogby issues are certainly worth discussing.

Does the Zogby data differ from the other data we have? Yes. Again though, I'm not terribly worried about it. Extra weight is only given to the most recent poll and in most cases the Zogby polls come sequentially before the most recent polls. That means that, in most cases, this data has been pushed into the "all the rest" pile. In other words, they have been absorbed by the average. Yes, outliers skew an average, but none of the states (with the possible exception of South Carolina) were pulled, in my estimation, out of whack. Then again, the only other polls we have out of the Palmetto state were also single digit margin. In McCain's favor, granted, but still single digit margins.

Still, the blue states seem about right, while the red states don't. That is potentially problematic. The deeply red states weren't shifted any and states like North Carolina and Missouri continue to hover right around the line between Toss Up and Lean where they are expected to be. The other states remain where they are expected to be given the information we had prior to the Zogby data being released.

Anonymous said...

The other issue is the support Barr is getting. By any measure of central tendency, Barr receives 6% across these 34 states. That is a touch high given what I have seen thus far.

Eh, you know what, let's push this into its own post. Barr deserves a bit more discussion.