2012 Electoral College Predictions

[JoeTheJuggler]

Ah. . here's where the discrepancies arise:

All the percentages I listed (except where specifically called out as popular vote) are 538's predicted "chance of victory in this state".

I thought you were citing poll numbers.

Not that I think these other analyses are wrong, but it's certainly more a question of analysis than simply pointing out that Romney would have to get a 15% change in votes (compared to how he's now polling) to win those couple of states.

 

[remirol]

Without drilling into Nate's figures more than I have (and probably more than he makes readily available, of course) I wouldn't know why Ohio is rated as 'harder to move' for Romney, though historical polling this election cycle might be a good indicator, as well as a possible Ryan effect for Wisconsin specifically.

Nevermind. I can actually tell you at least one probable reason why: looking at the latest Quinnipiac poll (which was linked from the Ohio summary on the 538 site), two sets of figures jump out:

Q: Favorable opinion of Romney?
FL: 45% Yes, 42% No, 13% ??
OH: 39% Yes, 45% No, 15% ??WI: 44% Yes, 43% No, 13% ??

Q: Favorable opinion of Obama?
FL: 50% Yes, 45% No, 5% ??
OH: 51% Yes, 45% No, 5% ??
WI: 51% Yes, 45% No, 4% ??

Simply put: Ohioans don't like Romney. That's gonna make it hard for him to move the numbers in that state. Plus, he's actually lost ground compared to the prior poll in late July (see the link), where he was 40% Yes, 43% No, and 18% were undecided.

This is, of course, only a single poll and it's one that showed favorable results to Obama. But it might provide a bit of insight.

 

[Beerina]

Obama 47.2% Romney 48.6%

EC Obama 272, Romney 266

There are structural reasons it will be hard for Obama to lose.

Didn't California, pissed at 2000, pass a law that their electors shall be bound to go for the national popular vote winner?


So Romney would win in your scenario. Jesus would that be sweet.

Oh holy (this week in unnecessary censorship) god, watching 57 million people screaming as their "good idea" explodes in their face, exactly as they were warned it might. Oh please, Lord. Oh, please!

 

[JoeTheJuggler]

Simply put: Ohioans don't like Romney. That's gonna make it hard for him to move the numbers in that state. Plus, he's actually lost ground compared to the prior poll in late July (see the link), where he was 40% Yes, 43% No, and 18% were undecided.

This is, of course, only a single poll and it's one that showed favorable results to Obama. But it might provide a bit of insight.

OTOH, I can well imagine people with an unfavorable opinion of Romney nevertheless voting for him.


Still, I'd say it's Obama's election to lose at this point.

 

[JoeTheJuggler]

Didn't California, pissed at 2000, pass a law that their electors shall be bound to go for the national popular vote winner?

I think you're talking about this:

http://en.wikipedia.org/wiki/National_Popular_Vote_Interstate_Compact

It would only go into effect if the law is passed in enough other states so as to guarantee a majority of votes in the electoral college.

And frankly, I doubt it would pass a constitutional challenge. (State legislatures can't effectively amend the U.S. Constitution without going through the normal process of ratifying a constitutional amendment.)

ETA: If they want to argue that it's a compact among the states and not effectively an amendment of the U.S. Constitution, they run afoul of things forbidden the states in Article I Section 10 Clause 3. States may not enter compacts with other states. At the very least, this is something that would have to be passed by Congress and signed into law by the POTUS.

 

[mikedenk]

Remirol or Joe (or anybody), what are your opinions on Romney's chances in Michigan?

A few of the factors off the top of my head that might be making the polls so close:

- Romney's history in the state
- The decrease in union influence
- An actual drop in population in urban centers like Detroit

 

[remirol]

Remirol or Joe (or anybody), what are your opinions on Romney's chances in Michigan?

Near-zero. Wasn't it Romney who came through and peed all over the state during the primaries, telling them they had to "toughen up" in a speech? (I actually don't remember and I'm lazy today)

A few of the factors off the top of my head that might be making the polls so close:

I see +3.6 in raw average, +5.5 in adjusted average, +9.1 if you throw out the polls entirely and go by historical factors, and +6.4 for his "if it were held today" model. None of these are particularly close when you're rolling into the tail end of August and need to make up a big pile of ground.

(from 538)

 

[BenBurch]

Remirol or Joe (or anybody), what are your opinions on Romney's chances in Michigan?

A few of the factors off the top of my head that might be making the polls so close:

- Romney's history in the state
- The decrease in union influence
- An actual drop in population in urban centers like Detroit

Obama saving what was left of the auto industry will get him Michigan.

BTW, I had the occasion to drive past the Chrysler plant in Belvidere, IL late Saturday night, and it was simply awesome to see the parking lot full, all the lights lit and a train pulling out on the UP line full of product. (Or at least I assume those carriers were full of cars. Can't see into them easily any more.)

 

[mikedenk]

Maybe a better phrasing would be "relatively close".

I asked because there are stories in the news today about a poll from Mitchell Research that shows a tie in the state. I'm not very familiar with Mitchell Research's methodology or reputation - this poll was of likely voters. Their previous poll from a month ago had Obama with a 5-point lead.

The RCP average for the state is 2.0 points for Obama. Obama won in a blowout in 2008; Bush lost the state by about 5 points (2000) and about 3.5 points (2004).

A Romney suprise in Michigan would obviously be huge for his electoral chances. If we assume that Romney takes his "must-wins" in Florida and Ohio, plus his very likely win in North Carolina, Michigan would put him on the brink (269 EVs).

 

[remirol]

Maybe a better phrasing would be "relatively close".

Part of why I added that particular smiley was directly because of the graphic I linked, showing Michigan as 90% likely to go in favor of Obama.

A Romney suprise in Michigan would obviously be huge for his electoral chances. If we assume that Romney takes his "must-wins" in Florida and Ohio, plus his very likely win in North Carolina, Michigan would put him on the brink (269 EVs).

He's in a dead heat in Florida and well behind in Ohio. I'm not certain what the point you're trying to make here is, unless it's that any possible Romney path to victory must necessarily include him overcoming some highly unlikely odds. It's certainly possible to _assume_ a whole bunch of things; for example, if Obama eats a live baby on network TV we can assume that the polls will shift significantly towards Romney. But I see little point in making assumptions that don't at least somewhat jiv(b)e with current knowledge.

FiveThirtyEight has proven itself over the last 4 years (starting with the 2008 election, where Nate's model got 49 of 50 states correct) as being an extremely reliable predictor in a number of areas. There's a reason I take numbers from it, and not from RCP or any individual poll -- they are, simply put, more likely to be right than anything else out there.

 

[JoeTheJuggler]

He's in a dead heat in Florida and well behind in Ohio.

And I wonder if the Florida polls have yet reflected the selection of Ryan as running mate.

 

[JoeTheJuggler]

FiveThirtyEight has proven itself over the last 4 years (starting with the 2008 election, where Nate's model got 49 of 50 states correct) as being an extremely reliable predictor in a number of areas. There's a reason I take numbers from it, and not from RCP or any individual poll -- they are, simply put, more likely to be right than anything else out there.

That may be so, but in itself isn't really an argument in support of their numbers. As I pointed out, Missouri went to the winner in presidential elections for more than half a century, but then last time it didn't.

I remember when Clinton was running for reelection someone had a very simple model that would have been accurate in all modern elections up to and including that one, which then failed in 2000. (IIRC, it was simply that the incumbent or the incumbent VP wins unless a few economic keys are triggered.)

 

[Dumb All Over]

Electoral College: Romney 320, Obama 218
Popular vote: Romney 52.9, Obama 47.1

Romney 52.9 percent, 47.1 for Obama.

EC Romney 320 vs 218 Obama

Great minds... yada yada.

 

[mikedenk]

I like to mull through different EC scenarios. My "point" is merely discussion of these scenarios... and that's about it. Any confusion is unintended.

The assumption was made for the sake of argument - an Obama victory in Ohio or especially Florida would pretty much hand him the election.

 

[LarianLeQuella]

Another site that I like for tracking polling and such is http://electoral-vote.com/ You can even get a graph of the EC totals with the safe numbers, and the toss up states. The also have a set of data that excludes Rassmusen.

Great minds... yada yada.

That would be funny if you didn't actually believe it!

EDIT: Oh, I am supposed to throw in a prediction (as useless as that is...)...

EV:
Obama: 320
Romney: 218

Popular:
Obama: 49.2%
Romney: 48.6%

Florida will manage to be a cluster again because they are Florida, and could screw up a free lunch.

 

[remirol]

I like to mull through different EC scenarios. My "point" is merely discussion of these scenarios... and that's about it. Any confusion is unintended.

No worries. That was the impression I was sort-of getting but wanted to make sure.

The assumption was made for the sake of argument - an Obama victory in Ohio or especially Florida would pretty much hand him the election.

Florida for sure just because of the numbers (even if one of the other states I listed as 'lock' flips, Florida will more than outweigh it), and Ohio... well, it doesn't look good for Romney if he can't pick up Ohio's votes, agreed.

 

[MaGZ]

Romney 284
Obama 254

 

[Magyar]

I have heard this before about NH. It vexes me as NH has been a traditionally Republican State and Romney is a "Local Boy".

Aside from Kerry, most "local boys" win.
http://www.salon.com/2011/10/28/just_like_dukakis_romney_is_taking_the_fun_out_of_new_hampshire/


Repubicans had a 111 year run in control of the General Court (4th Largest in the English Speaking World), and regained control in 2010.http://en.wikipedia.org/wiki/New_Hampshire_General_Court

NH being a swing state is very new in terms of politics.

Slight derail - There was an interesting story on NPR yesterday about NH and it's economy.
Short version is that basically NH has been a parasite state. IT has been able to maintain it's tax free status at the expense of MA where all the wealth and jobs are created (using tax dollars) and then sucking off the rich who can afford the commute.



So it's jsut another example of how non of the policies of the right actually really work.

 

[remirol]

I asked because there are stories in the news today about a poll from Mitchell Research that shows a tie in the state. I'm not very familiar with Mitchell Research's methodology or reputation - this poll was of likely voters. Their previous poll from a month ago had Obama with a 5-point lead.

The RCP average for the state is 2.0 points for Obama. Obama won in a blowout in 2008; Bush lost the state by about 5 points (2000) and about 3.5 points (2004).

This explains a lot, as usual: http://fivethirtyeight.blogs.nytimes.com/2012/08/28/aug-27-michigan-isnt-a-tossup/

Short version: Mitchell himself is biased.

 

[Baylor]

My prediction: