If you can read this, either the style sheet didn't load or you have an older browser that doesn't support style sheets. Try clearing your browser cache and refreshing the page.

(APS Physics)   The sports nerds at FiveThirtyEight just got some physics geek competition. Politics tab trumps sports and geek   (physics.aps.org) divider line 25
    More: Interesting, FiveThirtyEight, physics, voter model, home counties, Mount Rainier, Physical Review Letters  
•       •       •

1790 clicks; posted to Politics » on 18 Apr 2014 at 7:10 PM (31 weeks ago)   |  Favorite    |   share:  Share on Twitter share via Email Share on Facebook   more»



25 Comments   (+0 »)
   
View Voting Results: Smartest and Funniest
 
2014-04-18 02:30:38 PM  
Since TFA's provided link to the underlying paper seems broken at the moment, another source.
 
2014-04-18 04:02:14 PM  
So no need to have anymore elections. Carry on citizens.
 
2014-04-18 07:46:14 PM  

IgG4: So no need to have anymore elections. Carry on citizens.


I'm blanking on it, but there is a great Asimov short story where we develop a computer model strong enough that we can predict the election from one voter who places his/her selection.
 
2014-04-18 08:17:07 PM  
Hari Seldon sought for questioning.

/Obama is the Mule
 
2014-04-18 08:28:18 PM  

IgG4: So no need to have anymore elections. Carry on citizens.


We can't trust that responsibility to rubes who might vote Republican.
 
2014-04-18 08:36:22 PM  
FTFA:  The model uses population and commuting data for every US county and assumes that voters may change their political preferences based on interactions with others at home or at work.

I see a flaw in their model.
 
2014-04-18 08:40:40 PM  
"What is exciting [about the result] is the idea that one can be very quantitative about the statistical behavior of people," says Jean-Philippe Bouchaud, a physicist at Capital Fund Management in Paris.

*blink blink*
Um... Have you never heard of psychology, or sociology, or anthropology before? Sure they've only been around a couple of hundred years - relative newcomers compared to physics or biology - but they've been very quantitative about the statistical behavior of people before you jumped into the fray. Don't get me wrong, I'm sure your models are very thorough, but don't be so naive as to suggest that you're pioneering the concept of measuring people's behavior.
 
2014-04-18 08:58:29 PM  

Psychohazard: FTFA:  The model uses population and commuting data for every US county and assumes that voters may change their political preferences based on interactions with others at home or at work.

I see a flaw in their model.


It happens. I've seen senior citizens who have been voting Democrat since Roosevelt suddenly start voting for Republicans because they hate the n*bong. I've also seen middle-aged people and seniors who have been voting Republican since Goldwater angrily abandon that party over women's rights or taxes or the war or Medicare or religious extremism or the racial hatred in that party or any of the other things on the list of reasons that there are literally not enough electrons to transmit. The ultra-orthodox Jews of Brooklyn are rapidly switching from Democrat to Republican largely over gay marriage and an increasing intolerance in Democratic ranks for herpetic rabbis sucking off infants. Asian-Americans, previously a very winnable demographic for Republicans, have shifted toward Democrats almost alarmingly quickly over economic growth reasons and immigration policy. Latino preferences have also shifted toward (before 2004) and rapidly away from (everything after) Republicans based on the party's tolerance for brown folks. And in Kentucky, "Well, thank God, I believe I'm going to be a Democrat."

Evangelicals are rock-solid, it's true. And most Millennials will never vote for a Republican in their lives (although the white males seem to occasionally feel the siren call of Libertarianism) after the crap we've seen pulled. But there are plenty of persuadable groups; it's just that if you push demographics and groups too far in a single direction (which the model seems to show) you can reduce the probability of changing those preferences, due at least partly to network effects.
 
2014-04-18 10:12:26 PM  
It isn't hard to create a model that fits past data; you just tweak the variables until it works. A model only becomes useful when it shows it can predict.
 
2014-04-18 10:33:28 PM  

captainktainer: And in Kentucky, "Well, thank God, I believe I'm going to be a Democrat."


Fair enough.  And I imagine that linked article makes about 2% of Republicans (erroneously) clutch for their wallets and wish for a plague.
 
2014-04-18 11:10:11 PM  
First, assume a spherical voter...
 
2014-04-19 01:12:00 AM  
For all that it isn't a fantastic model for politics, it's a thousand times better than 538's interpretation of climate science. Seriously, I don't know what Nate was thinking. He looks like a drunk undergrad who paid the corner hobo to write his final project. It's embarrassing.
 
2014-04-19 01:21:58 AM  

captainktainer: Psychohazard: FTFA:  The model uses population and commuting data for every US county and assumes that voters may change their political preferences based on interactions with others at home or at work.

I see a flaw in their model.

It happens. I've seen senior citizens who have been voting Democrat since Roosevelt suddenly start voting for Republicans because they hate the n*bong. I've also seen middle-aged people and seniors who have been voting Republican since Goldwater angrily abandon that party over women's rights or taxes or the war or Medicare or religious extremism or the racial hatred in that party or any of the other things on the list of reasons that there are literally not enough electrons to transmit. The ultra-orthodox Jews of Brooklyn are rapidly switching from Democrat to Republican largely over gay marriage and an increasing intolerance in Democratic ranks for herpetic rabbis sucking off infants. Asian-Americans, previously a very winnable demographic for Republicans, have shifted toward Democrats almost alarmingly quickly over economic growth reasons and immigration policy. Latino preferences have also shifted toward (before 2004) and rapidly away from (everything after) Republicans based on the party's tolerance for brown folks. And in Kentucky, "Well, thank God, I believe I'm going to be a Democrat."

Evangelicals are rock-solid, it's true. And most Millennials will never vote for a Republican in their lives (although the white males seem to occasionally feel the siren call of Libertarianism) after the crap we've seen pulled. But there are plenty of persuadable groups; it's just that if you push demographics and groups too far in a single direction (which the model seems to show) you can reduce the probability of changing those preferences, due at least partly to network effects.


Speaking as a political scientist who specializes in behavior, yep. While some segments are getting more polarized, others are very much in flux. And as you alluded to, it seems to be a generational trend. The only cohort of Latinos who still vote Republican are Cuban expats who left when Castro came to power. Even their kids and grandkids are at least split 50/50 or leaning democratic.
 
2014-04-19 02:40:12 AM  
With the randomness term tuned to about 3% of the other effects, the simulation yielded a map whose correlations and voter fraction fluctuations remained stable over time and matched those of the US presidential election results between 1980 and 2012.

Speaking of 538, Nate Silver made a note to caution against models that tout how well they've made predictions of things that have already happened. That could be a signal of overfitting the model. Have they actually used this model to predict a future election yet?
 
2014-04-19 03:02:11 AM  

uknesvuinng: assume a spherical voter...


For many voters this is not an unreasonable assumption.
 
2014-04-19 08:05:21 AM  

GoldSpider: IgG4: So no need to have anymore elections. Carry on citizens.

We can't trust that responsibility to rubes who might vote Republican.


It's funny because Republicans are actively trying to disenfranchise voters.
When you can find a Demonrat pushing "Voter ID", we'll talk.
 
2014-04-19 08:28:15 AM  

GoldSpider: IgG4: So no need to have anymore elections. Carry on citizens.

We can't trust that responsibility to rubes who might vote Republican.


Well, it turns out depressions are bad for almost everyone so we should try and avoid those. Removing republicans is one of the more surefire ways to avoid stupidly large boom bust cycles.
 
2014-04-19 09:39:30 AM  

Shakin_Haitian: GoldSpider: IgG4: So no need to have anymore elections. Carry on citizens.

We can't trust that responsibility to rubes who might vote Republican.

Well, it turns out depressions are bad for almost everyone so we should try and avoid those. Removing republicans is one of the more surefire ways to avoid stupidly large boom bust cycles.


What do you propose we do to make it harder for these people to get elected?
 
2014-04-19 10:28:12 AM  

TheManofPA: IgG4: So no need to have anymore elections. Carry on citizens.

I'm blanking on it, but there is a great Asimov short story where we develop a computer model strong enough that we can predict the election from one voter who places his/her selection.


http://en.m.wikipedia.org/wiki/Franchise_(short_story)
 
2014-04-19 10:44:21 AM  

GoldSpider: Shakin_Haitian: GoldSpider: IgG4: So no need to have anymore elections. Carry on citizens.

We can't trust that responsibility to rubes who might vote Republican.

Well, it turns out depressions are bad for almost everyone so we should try and avoid those. Removing republicans is one of the more surefire ways to avoid stupidly large boom bust cycles.

What do you propose we do to make it harder for these people to get elected?


Let people vote?
 
2014-04-19 10:50:35 AM  

Wolf_Blitzer: GoldSpider: Shakin_Haitian: GoldSpider: IgG4: So no need to have anymore elections. Carry on citizens.

We can't trust that responsibility to rubes who might vote Republican.

Well, it turns out depressions are bad for almost everyone so we should try and avoid those. Removing republicans is one of the more surefire ways to avoid stupidly large boom bust cycles.

What do you propose we do to make it harder for these people to get elected?

Let people vote?


I would go up a notch to mandatory voting.
 
2014-04-19 10:51:56 AM  
Mandatory voting could be opted out of by showing up and filling out a form of course.
 
2014-04-19 12:20:53 PM  

captainktainer: I've also seen middle-aged people and seniors who have been voting Republican since Goldwater angrily abandon that party over


what Bush did after 9/11. not just the damn patriot act and the many abuses that brought but the war in iraq. i did vote for goldwater and even bush in 2000 but god lord was that a mistake.
 
2014-04-19 12:59:01 PM  

Wolf_Blitzer: GoldSpider: Shakin_Haitian: GoldSpider: IgG4: So no need to have anymore elections. Carry on citizens.

We can't trust that responsibility to rubes who might vote Republican.

Well, it turns out depressions are bad for almost everyone so we should try and avoid those. Removing republicans is one of the more surefire ways to avoid stupidly large boom bust cycles.

What do you propose we do to make it harder for these people to get elected?

Let people vote?


img2.wikia.nocookie.net
 
2014-04-19 02:01:30 PM  

Russ Feingold's Brass Balls: Speaking as a political scientist who specializes in behavior, yep. While some segments are getting more polarized, others are very much in flux. And as you alluded to, it seems to be a generational trend. The only cohort of Latinos who still vote Republican are Cuban expats who left when Castro came to power. Even their kids and grandkids are at least split 50/50 or leaning democratic.


Talk about snatching defeat from the jaws of victory... the Republicans were making shockingly strong moves into the African-American community during the Bush years. The gay marriage laws combined with the national insecurity after 9/11 was a great wedge opportunity for the GOP. I knew lots of African-Americans who were quite happy to have the government discriminate against gays. It was mainly based on religion and boisterous machismo, two languages Republicans speak well. Then the Democrats nominated Obama and the true Republican values on race were brought back into the spotlight. The ironic thing about their using outlawing gay marriage as a successful wedge during the Bush years is that it set up the conditions for the current strong backlash against it, bringing marriage equality into the mainstream decades before it likely would have been otherwise.
 
Displayed 25 of 25 comments

View Voting Results: Smartest and Funniest


This thread is closed to new comments.

Continue Farking
Submit a Link »






Report