Friday, October 24, 2008

The Electoral College Map (10/24/08)

All that polling Thursday and what do we have to show for it? A lot of blue, that's what. Thursday was not (I repeat, NOT) a good day polling-wise for the McCain campaign. You know it is bad when Montana is blue in a poll this late in the game and Indiana is giving Obama ten points in another. Now, these are individual polls, so as is FHQ's custom, we don't like to overreact to anything other than a clear trend. I'm hesitant to jump on board with the assertion at FiveThirtyEight yesterday that these polls are manufactured to a certain extent to play into the pollsters' perceptions of where each state is in a national context. [And I'm sure Nate was saying that just to get suckers like me to link to the post.] Some other outlets for electoral college analyses have those states and others, like Ohio, shifting in a big way toward Obama. In all three cases, there were consequential shifts for those three states, but by our measure, there was no line-crossing action. However, each of the three either moved on or off the Watch List (see below).

New Polls (Oct. 23)
StatePollMargin
Arkansas
University of Arkansas
+15
California
PPIC
+23
Florida
Quinnipiac
+5
Florida
St. Pete Times/Miami Herald
+7
Georgia
Rasmussen
+5
Illinois
Big Ten/Univ. of Wisc.
+29
Indiana
Big Ten/Univ. of Wisc.
+10
Indiana
Survey USA
+4
Iowa
Big Ten/Univ. of Wisc.
+13
Kansas
Survey USA
+12
Louisiana
Rasmussen
+16
Maine
Critical Insights
+21
Michigan
Big Ten/Univ. of Wisc.
+22
MinnesotaNational Journal
+10
MinnesotaBig Ten/Univ. of Wisc.
+19
MinnesotaRasmussen
+15
Montana
Montana State Univ.
+4
Ohio
Quinnipiac
+14
Ohio
Big Ten/Univ. of Wisc.
+12
Pennsylvania
National Journal
+10
Pennsylvania
Quinnipiac
+13
Pennsylvania
Big Ten/Univ. of Wisc.
+11
PennsylvaniaSurvey USA
+12
TexasRasmussen
+10
WashingtonRasmussen
+11
Wisconsin
National Journal
+13
Wisconsin
Big Ten/Univ. of Wisc.
+13

There was also a lot made of the BigTen polls that came out of states in which the BigTen athletics conference schools are in. This was the set of polls that had that 10 point Obama margin in Indiana. That one along with the Ohio poll seemed extreme -- especially given how close this set of polls was just a month ago -- but I'm in agreement with Jack on this one. Other than those two, though, the rest just don't seem to be extreme outliers. And with Friday's Insider Advantage poll of Ohio also showing a double digit margin for Obama, the BigTen and Quinnipiac polls don't appear to be all that out of whack. On election day they may prove to have been off, but for this moment in time, that doesn't seem to be the case. First of all, we'll, as I often suggest, need more information to confirm or disprove the margins in these polls. [Strategic Vision, for example, disagrees with these three polls showing just a three point Obama McCain lead in the Buckeye state.] But we also see that our averages for these states did not cross into stronger categories for Obama. And, again, that underlines the power of including past polls in our averages because it tamps down the day to day volatility that polls like these trigger.

Changes (Oct. 23)
StateBeforeAfter
Pennsylvania
Obama lean
Strong Obama

Now that the FHQ electoral college methodology public service announcement is over, we can focus on the actual data. So what changed after all those polls, volatile or otherwise? Well, not too much on the map. Pennsylvania moved into the strong Obama category and the Keystone state's shift is the culmination of several weeks worth of polling that has pushed the state's average closer and closer to the double digit mark. Pennsylvania's position as the state for the McCain folks to target seems a bit misplaced in this context. And the reported -- and apparently falsely so -- move on the part of the McCain campaign to pull out of Colorado (and Iowa and New Mexico) really looks strange in light of Pennsylvania's current polling. As I said when Florida turned blue last week, it means something when a state moves here -- typically that it is a lasting move that will be difficult to reverse.

...especially with just eleven days left in the campaign.
[Click Map to Enlarge]

And while Pennsylvania is the only move of the day, it does increase the proportion of his electoral votes in that category to over 70%. [For the record, McCain's strong states make up 76% of the electoral vote share he has on the map above.] If Obama is safe in states with 238 electoral votes then the Illinois senator needs but 32 more to cross the 270 threshold. And the four Obama lean state do that without even factoring in which toss up states may or may not favor Obama.

The Electoral College Spectrum*
HI-4
(7)**
ME-4
(157)
NH-4
(264/278)
ND-3
(160)
LA-9
(67)
VT-3
(10)
NJ-15
(172)
CO-9***
(273/274)
WV-5
(157)
KY-8
(58)
IL-21
(31)
WA-11
(183)
VA-13
(286/265)
GA-15
(152)
KS-6
(50)
RI-4
(35)
OR-7
(190)
OH-20
(306/252)
SD-3
(137)
TN-11
(44)
MA-12
(47)
IA-7
(197)
NV-5
(311/232)
MS-6
(134)
NE-5
(33)
MD-10
(57)
MN-10
(207)
FL-27
(338/227)
TX-34
(128)
AL-9
(28)
NY-31
(88)
WI-10
(217)
MO-11
(349/200)
AK-3
(94)
WY-3
(19)
DE-3
(91)
PA-21
(238)
IN-11
(360/189)
SC-8
(91)
ID-4
(16)
CT-7
(98)
NM-5
(243)
NC-15
(375/178)
AZ-10
(83)
OK-7
(12)
CA-55
(153)
MI-17
(260)
MT-3
(163)
AR-6
(73)
UT-5
(5)
*Follow the link for a detailed explanation on how to read the Electoral College Spectrum.
**The numbers in the parentheses refer to the number of electoral votes a candidate would have if he won all the states ranked prior to that state. If, for example, McCain won all the states up to and including Colorado (all Obama's toss up states plus Colorado), he would have 274 electoral votes. Both candidates numbers are only totaled through their rival's toss up states. In those cases, Obama's number is on the left and McCain's is on the right in italics.

***
Colorado is the state where Obama crosses (or McCain would cross) the 270 electoral vote threshold to win the presidential election. That line is referred to as the victory line. It is currently favoring Obama, thus the blue text in that cell.

But since we're on the topic of toss up states, we may as well discuss them. The list of them remains the same, but some of the ordering on the Electoral College Spectrum (above) has changed. On the weight of a couple of (super-) favorable polls in Ohio, Obama's average lead in the Buckeye state is now above the one point mark. We discussed Nevada nearly making that distinction yesterday and Virginia has already passed that point, but now Ohio has joined them. The Buckeye state is now off the Watch List and is moving even more toward Obama.

Indiana was also a fairly big mover on Thursday. [Just to note, there is a difference in what we're talking about here in terms of the word move. In the Pennsylvania discussion, move meant changing categories, but in this context it refers to the repositioning on the Spectrum's rankings. Pennsylvania changed categories, but didn't actually move on the Spectrum. Indiana, however, didn't change categories but did shift in the rankings.] Sure, it just jumped North Carolina, but it went from being on the Watch List for a potential move into the McCain lean category to very nearly being on the list as a possibility to turn blue (It is only one one-thousandth of a point away from that distinction.). In our averages, that's a pretty big shift. But why did it jump North Carolina? The Tar Heel state has had a bunch of pro-Obama polls lately and has not really gone anywhere. Well, we spoke about how well our model did earlier, and here's where we talk about one of the drawbacks. The discrepancy in the number of polls between Indiana and North Carolina means that a state with fewer polls (Indiana) is more susceptible to bigger shifts given new and decidedly different polling information. The flip side is that North Carolina is being polled quite heavily right now and that if the current trajectory continues, the Tar Heel state will continue to move closer and closer to a tie (or to turning blue).

If Indiana has fewer polls than North Carolina, Montana has fewer polls than Indiana. The four point margin the Montana State poll gave Obama in the Treasure state pulled Montana onto the Watch List for a potential switch into the toss up McCain category. And though it was a pretty big shift within the average, Montana didn't change categories, but it did move from the upper end of the lean category to the lower end in a hurry.

The Watch List*
StateSwitch
Coloradofrom Obama lean
to Toss Up Obama
Floridafrom Toss Up Obama
to Toss Up McCain
Michigan
from Obama lean
to Strong Obama
Missourifrom Toss Up McCain
to Toss Up Obama
Montanafrom McCain lean
to Toss Up McCain
Nevadafrom Toss Up Obama
to Toss Up McCain
New Mexicofrom Obama lean
to Strong Obama
Virginiafrom Toss Up Obama
to Obama lean
*Weighted Average within a fraction of a point of changing categories.

The striking thing is that Montana is really the lone red state represented on the Watch List. Every other state is operating within the toss up McCain to Strong Obama categories. Sure, there are three states and 52 electoral votes that could shift into that McCain toss up category, but even that looks tough for McCain given how the momentum in the race is moving. And even though there is something of a mixed message from the polling out so far today, it doesn't really seem to be shifting wholesale toward McCain, at least not in a way that is going to move enough electoral votes to push him over 270.


Recent Posts:
While You Wait for the New Map, Here's a...Map

The Electoral College Map (10/23/08)

What the Bradley Effect Might Look Like

13 comments:

  1. Fair enough. Real life has to take priority. We (and I believe I can speak for all the other regular readers in this) appreciate what you do for this site.

    ReplyDelete
  2. Wow.

    On October 24, at one minute after midnight, let it be noted that academia, where Josh and I both make our livings, was just referred to as "real life."

    The apocalypse is nigh. :D

    ReplyDelete
  3. No, Josh, the GOP (Strategic Vision) poll in Ohio today shows MCCAIN with a 3 point lead in Ohio!!!

    ReplyDelete
  4. Just for fun, now that Pennsylvania is a lovely dark blue, and it seems to be the big targeted state for the clueless McCain campaign, what would it take for it to shift into his column? Two polls showing him up 100-0?

    Maybe he's relying on Literary Digest polls?

    ReplyDelete
  5. Dneedle,
    Good catch. It has been fixed. I just glanced at that poll as I was finishing up (and getting ready to head out the door to catch the bus) and apparently didn't look closely enough.

    If anything, this lends some credence to what I referenced FiveThirtyEight as saying yesterday. Strategic Vision is a Republican-leaning firm after all. But they have been in line with most of the other polling recently.

    ...until this flurry of double digit Obama leads.

    We now have a divide in Ohio between the Rasmussens, Mason-Dixons and Strategic Visions on the one hand and the Quinnipiacs, BigTens, etc. on the other. And that's just in the last week. Interesting.

    ReplyDelete
  6. Jack,
    The 100-0, single-shot poll scenario you described wouldn't even get Pennsylvania out of the Obama lean category for McCain. With as many polls as Pennsylvania has, it is going to be really difficult to sway by FHQ's measure.

    ReplyDelete
  7. Josh:

    The only striking number I caught clicking on the link of the Strategic Vision poll was that in that sample Obama had only a net 8 point favorable rating. Still, that isn't really THAT striking. You would know far more than I would, but Strategic Vision seems like a serious outlook. The poll is an outlier, but I don't think you can just dismiss it, like you could that 1 point McCain national deficit showing a huge evangelical percentage, from yesterday.

    ReplyDelete
  8. dneedle,

    Check out Jay Cost's post this evening.

    http://www.realclearpolitics.com/horseraceblog/2008/10/a_note_on_the_polls_1.html

    ReplyDelete
  9. I like how once McCain said he was pulling out of Michigan, pollsters basically stopped polling it. Not a single "discretionary" poll since Oct. 12. (I'm not counting the Big Ten poll, since they "have" to poll Michigan, or polls by in-state groups.)

    Wisconsin and Iowa, just to name two examples, continue to be polled.

    That throws a weird wrench into the methodology here. As long as a state is polled at a constant rate, whether high or low, it should eventually respond to changes. But once Michigan changed, it stopped being polled much. So the old, close race is "locked in" without new polls coming in to change the average.

    In other words, Michigan is not being polled much because it is believed to be solid Obama. But we can't confirm that it's solid Obama because it's not being polled much. Sigh.

    ReplyDelete
  10. Sorry for double-posting, but I just came up with what I think may be the explanation for a lot of what seems to be irrational behavior by the McCain campaign in their choice of states to allocate resources to.

    Think, for a moment, what is the most likely victory scenario for McCain in terms of electoral votes. Just move up the Electoral College Spectrum until you get to 270?

    No. The problem is that some of those states have already voted in large numbers. For McCain to win, he has to have the state of the race change nationally. But if that change occurs late (say, something equivalent to the Bin Laden tape of 2004 the weekend before the election), then its effect is tempered in early voting states.

    Here's the map of states with early voting.

    McCain's best chance of winning is to get some of the states on Obama's side of the partisan line that don't have early voting.

    Those are:

    MN, WI, VA, MI, DE, PA, NY, CT, RI, MA, and NH.

    Start by throwing out the deepest of the blue states. Now we've got, in reverse order of their position on the spectrum,

    VA, NH, MI, PA, WI, and MN.

    Looks an awful lot like where McCain is concentrating his resources, doesn't it?

    The one exception there is Michigan. And if they're actually thinking things through (which I admit doesn't seem to be the case), then in the event of one more game-changer they can sweep into Michigan and try to launch a surprise attack of sorts. McCain has enough resources to do that in one largish state, and in the mean time he can save money and resources by leaving it off the list.

    The other mystery is why he seems so fixated on Iowa. That I can't understand.

    But at least this explains why the interest in Colorado, New Mexico, Nevada, and even Ohio seems a little tepid. McCain needs to defend his states up through Florida, early voting or no (and thus we do see plenty of activity in North Carolina and Florida). That brings him to 227. Assume a major game-changer in the last week. He now adds the non-early-voting states of Virginia, New Hampshire, and Pennsylvania. He's at 265. McCain now wins if he can somehow pick off Nevada or Ohio (early voting, but pretty close), or Michigan (no early voting, but would need a last minute blitz).

    It's a long shot, but that's the point. To win, McCain has to assume something crazy goes his way, and then be ready to capitalize on it if it does. Going after Pennsylvania, rather than the western states, is the best way to do that.

    So the football analogy is no longer the hail Mary pass. Now it's down by 16 with under a minute to go and no time outs. If you can score a touchdown, the plan is clearly to go for a two point conversion. Not because that's the high percentage play--in fact, it's likely to make the defeat worse in point terms. But because if lightning strikes (in the analogy, recovering an onside kick), then at least you're poised to take advantage of it.

    That's hypothesis one. Hypothesis two is that they're just clueless.

    ReplyDelete
  11. Scott,

    Save yourself the trouble of trying to understand the McCain campaign.

    I would like to think that the vast majority of early voters are people who are 100% certain they support their candidate. While it's obviously an advantage to have them in the bag, it can't possibly be such an advantage that it's worth giving up a state that's polling relatively close in favor of a state in which you're 15 points down.

    Also, the Iowa thing, as well as so much more, supports Hypothesis Two.

    ReplyDelete
  12. Good, Scott. I'm glad you brought this up.

    ...and beat me to it in the process. This point actually came up in our weekly discussion group yesterday and I was going to post something about it today.

    But now I'll just move this over to its own post. It deserves it's own spot.

    Jack, I'll move your comment over there as well.

    ReplyDelete