Wednesday, April 2, 2008

The Electoral College Maps (4/2/08)

While last week's post focused on the Clinton-Obama match up through the lens of the electoral college, one thing was neglected: John McCain. Yes, he even gets a pass here...apparently. Well, he did until now. Obama's competitiveness in some traditional red states based on these early head-to-head polls was interesting (alarming to some), but McCain is also competitive in some states that have been blue in recent elections (see also CQ's look at recent state polls in NJ, MI and WA). In the match up with Clinton, McCain is ahead in Oregon and Washington and against Obama he is running close in Massachusetts. Massachusetts, Oregon and Washington haven't voted with the GOP presidential nominee since 1984, and while Oregon and Washington have had close statewide elections recently, that hasn't translated to the presidential level.

Massachusetts, on the other hand, is a surprising inclusion on the list. And given that Obama is being backed by both senators and the governor of the state that is something of an anomaly. What is driving this is one poll (of the three conducted in Massachusetts thus far) has McCain and Obama tied. The other two have Obama with comfortable seven point leads.

Well, what about New Jersey (CQ does make a good point.)? The last time the Garden state voted for the GOP nominee was when the first George Bush was elected in 1988. However, coming down the stretch in 2004 the race between Kerry and Bush was fairly tight; tight enough to make the GOP consider expending more resources there (Those of you with access to Lexis Nexis out there can check out this Boston Globe article from mid-October 2004.). New Jersey then, wasn't as eye-catching as Massachusetts (or some of the other red states in which Obama is competitive).

Well, what about the map update?

Twelve states had new polls this week
(CA, CT, FL, MI, MO, NJ, OH, OR, PA, VA, WA & WI), and there are some shifts. The overall conclusion though, is still the same (right down to the electoral count in the McCain-Clinton case). Here are the changes:

NJ: Lean Clinton to toss up.
PA: Obama and McCain are now tied (a change from toss up leaning McCain)
VA: McCain Lean to Strong McCain in the match up with Clinton; Toss up to McCain Lean against Obama
WA: Lean McCain to Toss up (still favoring McCain)
WI: Toss up to McCain Lean (v. Clinton) and Obama Lean to Toss up (still favoring Obama)
Clinton then, makes no gains on McCain in the electoral vote count (still 314-224) but the toss up total increased by 16 electoral votes in her direction. Among the big three, she takes Pennsylvania and Ohio, but loses Florida.
***CORRECTION: Wisconsin should be an "Obama lean" state (one poll was mistakenly omitted from the analysis.)

The McCain-Obama map offers a new distinction: a statistical tie in Pennsylvania. And the outcome of this hypothetical election would come down to which way the state broke. Obama holds a 260-257 lead in electoral votes. The 21 in Pennsylvania would put either candidate over the top. In this scenario then, Pennsylvania is the new Ohio is the new Florida.

A note on the methodology:
Some people have taken issue with me using the average of the polls since Super Tuesday; claiming that that gives Obama an advantage. The reason I made Super Tuesday the cut off, was that that was the point at which the race entered the current deadlock phase. Both Clinton and Obama were on "pretty" equal footing from that point forward. The averages help to absorb the shock of any one event on the polls (Bosnia, Wright, etc.). If the most recent polls were used as opposed to these averages the results don't significantly change. Clinton would gain Florida and Nevada but lose Minnesota and New Jersey while tying in New Mexico. Obama would lose Michigan, New Hampshire, New Jersey and Wisconsin to McCain, but gain both Pennsylvania and Ohio with Colorado as a tie.

The other argument is more damning. "It is too early." I can't argue with that. It is, but this does give us a fun glimpse into each of the states and further, the direction they are leaning seven months out. One thing is certain, as we add polls, the picture will get clearer...provided the Democrats actually make a decision. As such, I'll keep updating this weekly (every Wednesday) so we can all keep tabs on the progression.

How much difference does it make?: a set of companion maps.

And an update for 4/9/08

Update for 4/16/08

Update for 4/23/08

Update for 4/30/08

Weighted Averages 4/30/08

Weighted Averages 5/7/08

Update for 5/14/08 (weighted)

Update for 5/21/08 (weighted)

New Maps? (5/25/08)

Update for 5/28/08 (weighted)

Update for 6/3/08 (weighted)


Robert said...

I think this analysis is very useful. Yes, it is too early, but it gives an indication of where we are at this point. It does help sort out some of the claims made by all three candidates. Events will shape the race down the line, but the voters in PA, NC and IN will not have the luxury of knowing what those events will be in October. They need to make their decisions in the next few weeks with the information that is available.

Brian Dell said...

The Dems cannot win if they lose Michigan, Ohio, Penn, Florida and let McCain in the door in NJ and Mass.

In none of your maps have you assumed that Obama will lose any of these, other than Florida, which you assume Clinton will lose as well.

You've fundamentally begged the question here when you take that approach. Obviously, Obama is the choice if the Great Lakes and New England are not going McCain, since Obama opens up the west which Clinton does not.

But the fact is that Clinton was consistently polling better than Obama vs McCain in mid-March in the key swing states. Obama would take Colorado while Clinton would lose it but Colorado is simply not enough. You're kidding yourself if you think Obama can really turn Texas and North Carolina into purple states.

It's true that mid-March was an unusually low point for Obama vs McCain since it was the height of the Wright contoversy. But at the same time the end of March is arguably unusually low for Clinton because her Bosnia claims were in the media. Both controversies will blow over in a couple weeks (or both will come back). One has to look at deeper arguments and on that count it seems clear that the Senator from New York is identified with by larger demographic groups in the northeast.

Michael Barone on US News defends his contention that Hillary could win the Dem popular vote despite the fact it is now April and not mid-March. If that happens, every impartial person will have to admit that Hillary's vote is more favourably distributed on the College. Obama's Dem vote run-up in deeply red states like Utah etc simply makes an enormous blow-out there to McCain a major blow-out.

Brian Dell said...

I meant to say SOUTH Carolina, not "North Carolina", is not seriously considered a purple state, even with Obama as the nominee, despite what your map suggests (although NC isn't purple either...)

Anonymous said...

NC actually *is* purple, despite what the map suggests. NC is a purple state on the ground, with an overwhelming number of state and local offices held by Dems. Our House party is split: 7 Dems, 6 Repubs, with at least one likely Dem pick up in the fall. We currently have 2 Repub Senators, but Liddy Dole's favorables are under 50% and there are strong efforts underway to oust her in the fall, which would once again yield a purple Senatorial delegation.

And we haven't held our primary yet. Campaigning here has just begun, but Obama had already shown a "leans Repub" versus "strong Repub" polling trend (as opposed to Clinton v. McCain matchups). I suspect that *after* NC's primary, the VA-NC-SC corridor will all be colored "toss up" on your map.

Robert said...


What Josh's analysis suggests that Clinton does better in traditional Democrat states while Obama brings more states into play. Clinton and her surrogates are arguing that she has a better opportunity to capture the Electoral College versus McCain than Obama does. She is working on the same strategy that did not work well for Gore or Kerry. Obama's strategy is more in line with Howard Dean's 50-state strategy. Although there are many limitations in Josh's analysis, it does quash Clinton's claim that she has a better claim to the Electoral College than Obama. Also, his analysis takes into account all polls back to Super Tuesday. Thus it takes into consideration both high points and low points for Clinton and for Obama. I would be very careful to say any candidate can't win this year. The old rules don't seem to apply. If they did we would be speculating on how well Clinton would do against someone other than McCain. Much will change in the polls between now and November based on external events beyond the control of the candidates.

Josh Putnam said...

I'm not assuming anything here other than the idea that these state head-to-head polls will reflect general election votes in those states. It is an exercise, the validity of which we can all talk at length about, but an exercise nonetheless.

I don't assume that any one candidate has won or lost a state. If they are ahead, however tenuously, they get the electoral votes for that state. I don't assume that Obama wins some toss ups and Clinton doesn't. It all depends on how the average comes out.

Now something that we can discuss and may have to be revisited over time is the line drawn between the "lean" states and the toss up states. 4.99 points may be a bit high or low at this point for determining a toss up. It just happens to be the number I started with. It may be instructive to move that line at some point, but we've only been working with this data for a week, so the trends have yet to emerge and create a potentially more natural line between those two categories.

Josh Putnam said...

Good to have some Tarheels (whether you're for Carolina or not) stopping by. I'm a native North Carolinian and will be keeping my eye on that Dole race this fall. That seat is the more stable one. Helms held it from 1972-2002 and passed the torch to Dole then. The other seat is the carousel seat, switching parties every six years since the 70's (I think. Let's see: Burr, Edwards, Faircloth, Sanford, East, and that is where my memory is deserting me.).

Of course, the governor's seat is up for grabs as well. It is a good year for high stakes politics in the Old North State.