PDA

View Full Version : 2012 US Presidential Election Thread



Miroir Noir
06-07-2012, 11:59 AM
Five months out. Nate Silver's first election forecast is out today (http://fivethirtyeight.blogs.nytimes.com/2012/06/07/election-forecast-obama-begins-with-tenuous-advantage/).

SlowMotionApocalypse
06-07-2012, 12:06 PM
In the last few weeks I have been getting a nagging feeling that Mitt Romney might actually win.

TomAz
06-07-2012, 12:13 PM
Eat more fiber.

mountmccabe
06-07-2012, 01:28 PM
Ahh, sabermetrics applied to presidential elections. Part of me wonders why this sort of thing wasn't already the standard but there are still many reports on nationwide polls that purport to be meaningful.

TomAz
06-07-2012, 02:23 PM
Yeah that sort of thing is interesting and all but one shouldn't read too much significance into the results. It's all poll-based and we all know polls can be flawed (young people underrepresented, etc). No matter how scientific their polling methodology, though, when you split data that fine you run into statistical credibility issues. Also the types of information they are supplying has to be based on extrapolations and judgment-based assumptions. All of which means this is an interesting way to form a framework to think about issues, but the numbers themselves are probably pretty squishy.

summerkid
06-07-2012, 09:24 PM
Yup, the best recent example is the Wisconsin recall election. CNN was reporting that exit polling data showed Walker and Barrett as being 50-50 and Walker ended up winning by 6 points.

Miroir Noir
06-07-2012, 09:56 PM
That was a little different though, because that was just the first wave of exits. The same thing happened with a bunch of exit polls in the Republican primaries where Drudge and others would leak big Romney wins only to see Santorum close the gap in the later wave of exits.

mountmccabe
06-08-2012, 04:05 AM
Yeah that sort of thing is interesting and all but one shouldn't read too much significance into the results. It's all poll-based and we all know polls can be flawed (young people underrepresented, etc). No matter how scientific their polling methodology, though, when you split data that fine you run into statistical credibility issues. Also the types of information they are supplying has to be based on extrapolations and judgment-based assumptions. All of which means this is an interesting way to form a framework to think about issues, but the numbers themselves are probably pretty squishy.

I don't understand any of what you're getting at. Of course current results aren't binding and no one should treat this as a closed race.

Also is the problem that it is all polls or that it includes extrapolations and judgment?

And how is polling a thousand people in an attempt to get a balanced sample from the entire country have less statistical credibility issues than polling a thousand people from a single state? That's the comparison I mean to highlight: the former is almost meaningless while the latter points to a vote that will actually decide a thing.

mountmccabe
06-08-2012, 04:32 AM
I mean, the state-by-state polls are going to happen. Since that data is there it makes sense to aggregate them state-by-state and watch for trends. And those state-by-state results to build a predicted result from the electoral college. The innovation I see is looking at the 538 rather than the 120 million or whatever.

I haven't had a chance to study all the newer models and methods but getting 49 states plus DC right in 2008 does not, to me, suggest squishiness.

gaypalmsprings
06-08-2012, 04:43 AM
http://www.seagrasswatch.org/seagrass_images/SW_election_cartoon_2001.jpg

Miroir Noir
08-11-2012, 08:33 PM
Here's the required pre-veep pick reading on Paul Ryan, from Jonathan Chait (http://nymag.com/news/features/paul-ryan-2012-5/) and Ryan Lizza (http://www.newyorker.com/reporting/2012/08/06/120806fa_fact_lizza).

As a liberal, I love this pick. It should force the Republicans to run on the specifics of what they are actually planning on doing if they win the election, most of which happens to be incredibly unpopular when explained to voters. At the same time, I have a bit of fear: Ryan is not a Palinesque fuckup. This guy is a total operator, and even if I dont agree that he's the policy guru and "serious" thinker that the press all too happily report him to be, he is unquestionably a dynamic figure with a very real agenda to unmake the New Deal and Great Society. You don't get elected to Congress at age 28, become the chair of the most important House committee at age 42, and convince basically everyone in your party to pronounce their undying fealty to whatever you want to do unless you are an immense political talent.

obzen
08-11-2012, 08:40 PM
You secretly jock Paul Ryan, bitch.

mountmccabe
08-12-2012, 08:15 AM
It's not a bad pick. Romney's not going to win this unless things change fairly significantly. Ryan will shake things up more than some safer choices would. I guess we'll see how much and where things fall.

Miroir Noir
09-13-2012, 03:08 PM
Crushing blows (http://firstread.nbcnews.com/_news/2012/09/13/13848313-polls-obama-holds-the-edge-in-florida-ohio-and-virginia?lite)

mountmccabe
09-14-2012, 06:17 AM
The crazy thing is that Obama does not even need FL, VA or OH to win.

Obama is also leading in CO, IA, NH, WI, NV (and other more states Romney has less of a chance in) that puts him at 272.

There's almost two months to go and plenty can happen but, wow.

bug on your lip
09-14-2012, 07:37 AM
Romney is still leading in the 'Entitled White Guy" vote

but has been declining with Women & Latino voters.
None of this should be a surprise. He doesn't have the demographics to win