Some polls now have Romney ahead.

[mshan]

"Perhaps because it was wrong before then."

Great model, eh?

 

[Charles Kozierok]

Time to stop spewing the bullshit. The model has been around before August and they haven't kept anything to themselves.

http://spot.colorado.edu/~mcnownr/working_papers/presidential_voting_mscpt.pdf

I'm not "spewing bullshit". I'd heard of this model before and spent some time searching on Google using date filters. Could not find a single reference to it before August of this year. Believe what you want, but I doubt you can find such references either.

Doesn't matter anyway. The model is inherently one-dimensional and its predictions silly.

 

[xBiffx]

I'm not "spewing bullshit". I'd heard of this model before and spent some time searching on Google using date filters. Could not find a single reference to it before August of this year. Believe what you want, but I doubt you can find such references either.

Doesn't matter anyway. The model is inherently one-dimensional and its predictions silly.

Took me 30 seconds using the same Google filters. I take it back, you aren't spewing bullshit, you just suck at Google. Sorry for that.

 

[xBiffx]

Great model, eh?

Considering the elections it has been correct on, yes. So even if you count back to 1972 and both elections were incorrect its still 80% accurate. Still right much more often right that wrong. It just sounds better to say 100% back to 1980. After all, this is a better way to advertise your Political Science department.

 

[Charles Kozierok]

Took me 30 seconds using the same Google filters. I take it back, you aren't spewing bullshit, you just suck at Google. Sorry for that.

Yeah, I suck at it. I've only been using it for 13 years, though; I'll get better.

I entered "colorado economic model" into Google and searched with the range of 1/1/2008 to 12/31/2008. Not a single news article about it anywhere.

Wonder why that is?

 

[jpeyton]

Why Romney will never win Ohio: http://www.youtube.com/watch?feature=player_embedded&v=a3ZqP2JQqLM#!

His 2008 "Let Detroit Go Bankrupt" editorial cost him the 2012 election.

 

[jackstar7]

Considering the elections it has been correct on, yes. So even if you count back to 1972 and both elections were incorrect its still 80% accurate. Still right much more often right that wrong. It just sounds better to say 100% back to 1980. After all, this is a better way to advertise your Political Science department.

If it's settled science for you, are you putting any money down? Seems like an easy lock on some free money for you.

 

[buckshot24]

Are you comparing popular vote to EC again?

Where would you get that idea?

 

[PokerGuy]

Why Romney will never win Ohio: http://www.youtube.com/watch?feature=player_embedded&v=a3ZqP2JQqLM#!

His 2008 "Let Detroit Go Bankrupt" editorial cost him the 2012 election.

I agree that Romney won't win Ohio, but the detroit bankruptcy line of attack doesn't make any sense. Last I checked, those companies DID go bankrupt. Why would someone be mad about one guy saying he'd let them go bankrupt, when the other guy in fact DID let them go bankrupt?? What am I missing?

 

[buckshot24]

I think you are confused and just throwing numbers around. Can you elaborate on the bold part? They were wrong by 5.8%, with a bias towards the Republican. This has to do with their models.With the link you provided (did you read the text?), they tend to assume there are more Republican than really are and their polling methods also favor Republicans (ie. old white people sitting by phones). You then go on to compare the actual error rate (ie. how far they were from the real results) with the margin of error to individual polls, which is a function solely of how many participants are used (margin goes down with more people).

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

I would say you are reading the data wrong and seriously don't understand it.

I'm pretty sure that they are looking at actual results vs poll predictions in both. the article talks about the republican weighting as a reason for Rasmussen's error being in favor of Republicans.

 

[Farang]

Considering the elections it has been correct on, yes. So even if you count back to 1972 and both elections were incorrect its still 80% accurate. Still right much more often right that wrong. It just sounds better to say 100% back to 1980. After all, this is a better way to advertise your Political Science department.

Banging my head against the wall here--THE MODEL IS USELESS. It is retroactively applied. It's like finding a coincidence that occurred every time before an election and saying it is a predictive model. I.e. I'm sure someone, somewhere, wore boxers every day a Republican was elected, and briefs the day a Democrat was elected.

 

[Charles Kozierok]

I agree that Romney won't win Ohio, but the detroit bankruptcy line of attack doesn't make any sense. Last I checked, those companies DID go bankrupt. Why would someone be mad about one guy saying he'd let them go bankrupt, when the other guy in fact DID let them go bankrupt?? What am I missing?

What you're missing is that Obama's plan involved the government providing necessary funds to keep the companies going at a time when no bank would have provided any money, which Romney opposed. Romney is now pretending that there was no banking crisis in 2008/2009 and that therefore the bankruptcy would have proceeded without liquidation, which is nonsense.

 

[Farang]

Lot of political nerds in here, think you'll enjoy this: http://www.theatlantic.com/politics...-game-that-could-put-him-over-the-top/264031/

 

[xBiffx]

If it's settled science for you, are you putting any money down? Seems like an easy lock on some free money for you.

Nope. This was simply to point out that there are descenting views to those media outlets who have already called the election as was pointed out a couple pages back.

 

[xBiffx]

Banging my head against the wall here--THE MODEL IS USELESS. It is retroactively applied. It's like finding a coincidence that occurred every time before an election and saying it is a predictive model. I.e. I'm sure someone, somewhere, wore boxers every day a Republican was elected, and briefs the day a Democrat was elected.

Continued usage of the words prediction and forecast seem to dictate otherwise. Perhaps every news outlet has used them incorrectly.

 

[Farang]

Continued usage of the words prediction and forecast seem to dictate otherwise. Perhaps every news outlet has used them incorrectly.

Yes! They did! It's a coincidence model, not a prediction model. When this thing became a big headline a couple months back it frustrated me greatly how poor the reporting was. These guys were made out to be election forecasting wizards. I think the way they framed their findings was deliberately misleading.

 

[xBiffx]

Yes! They did! It's a coincidence model, not a prediction model. When this thing became a big headline a couple months back it frustrated me greatly how poor the reporting was. These guys were made out to be election forecasting wizards. I think the way they framed their findings was deliberately misleading.

If this is the case, I don't see how the professors are to blame, they didn't run the stories about it. Also, you've got yourself a pretty giant mess up on the part of most of the media.

 

[Todd33]

I'm pretty sure that they are looking at actual results vs poll predictions in both. the article talks about the republican weighting as a reason for Rasmussen's error being in favor of Republicans.

I see. In that case you are still using the data wrong.

http://fivethirtyeight.blogs.nytime...rate-quinnipiac-surveyusa-performed-strongly/

That average (110 polls) 5.8% error with a 3.8% is compared to other pollsters with the same races, the best getting 3.3% error with no bias. This isn't a subjective thing, you can be good and bad at this and they are currently pretty bad.

Moreover, Rasmussen’s polls were quite biased, overestimating the standing of the Republican candidate by almost 4 points on average. In just 12 cases, Rasmussen’s polls overestimated the margin for the Democrat by 3 or more points. But it did so for the Republican candidate in 55 cases — that is, in more than half of the polls that it issued.

Of course I suspect right leaning people love them and thus defend them. Rasmussen tells them what they want to hear, not the truth.

 

[Charles Kozierok]

I'll start paying attention to Rasmussen -- other than relative trend -- when he gets his head out of his ass and addresses the major problems with his methodology.

 

[buckshot24]

I see. In that case you are still using the data wrong.

http://fivethirtyeight.blogs.nytime...rate-quinnipiac-surveyusa-performed-strongly/

That average (110 polls) 5.8% error with a 3.8% is compared to other pollsters with the same races, the best getting 3.3% error with no bias. This isn't a subjective thing, you can be good and bad at this and they are currently pretty bad.

Rasmussen was off in 2010. I'll accept that. I won't argue that. In fact I said that. Historically, however, he wasn't that bad. He was basically average. Silver also points out that it is easier to be wrong in state and district polls and 2010 was all state and district results. In presidential elections its a different story. According to Silver's database its "harder" (error rates are over doubled as compared to district races) to be wrong in national polls. 2008 Rasmussen wasn't off in his final national poll by much. You should compare apples with apples.

Of course I suspect right leaning people love them and thus defend them. Rasmussen tells them what they want to hear, not the truth.

I don't take Rasmussen as the gospel truth, don't get me wrong. Dismissing them as totally biased in a presidential election isn't supported by evidence.

I posted a video earlier where Silver predicted that the Republicans would pick up 45-50 seats on 10-18-2010. His results were biased towards Democrats by 15-18 seats and that includes Rasmussen's skewed data. Why cut Silver slack and Rasmussen none?

 

[lotus503]

Rasmussen was off in 2010. I'll accept that. I won't argue that. In fact I said that. Historically, however, he wasn't that bad. He was basically average. Silver also points out that it is easier to be wrong in state and district polls and 2010 was all state and district results. In presidential elections its a different story. According to Silver's database its "harder" (error rates are over doubled as compared to district races) to be wrong in national polls. 2008 Rasmussen wasn't off in his final national poll by much. You should compare apples with apples.

I don't take Rasmussen as the gospel truth, don't get me wrong. Dismissing them as totally biased in a presidential election isn't supported by evidence.

I posted a video earlier where Silver predicted that the Republicans would pick up 45-50 seats on 10-18-2010. His results were biased towards Democrats by 15-18 seats and that includes Rasmussen's skewed data. Why cut Silver slack and Rasmussen none?

Nate's a sabermetrician, thats why

 

[Charles Kozierok]

Rasmussen's tracking poll uses only robocalls -- no cell phones. That is a major reason why he is so right-leaning in his results. More people switch away from landlines every year, but he doesn't change his methodology, which is why he is getting worse over time. I see no reason to believe that he'll be more accurate than in 2010 when he hasn't changed his methods.

 

[buckshot24]

Rasmussen's tracking poll uses only robocalls -- no cell phones. That is a major reason why he is so right-leaning in his results. More people switch away from landlines every year, but he doesn't change his methodology, which is why he is getting worse over time. I see no reason to believe that he'll be more accurate than in 2010 when he hasn't changed his methods.

I think SurveyUSA use robo calls and Nate likes them.

 

[buckshot24]

Nate's a sabermetrician, thats why

What does he say about the Tigers in the World Series?

 

[Todd33]

Rasmussen was off in 2010. I'll accept that. I won't argue that. In fact I said that. Historically, however, he wasn't that bad. He was basically average. Silver also points out that it is easier to be wrong in state and district polls and 2010 was all state and district results. In presidential elections its a different story. According to Silver's database its "harder" (error rates are over doubled as compared to district races) to be wrong in national polls. 2008 Rasmussen wasn't off in his final national poll by much. You should compare apples with apples.

I don't take Rasmussen as the gospel truth, don't get me wrong. Dismissing them as totally biased in a presidential election isn't supported by evidence.

I posted a video earlier where Silver predicted that the Republicans would pick up 45-50 seats on 10-18-2010. His results were biased towards Democrats by 15-18 seats and that includes Rasmussen's skewed data. Why cut Silver slack and Rasmussen none?

Read the link, that 5.8% vs 3.3% were for the exact same 110 polls. Apples for apples. Silver is not a pollster, he is a statistician - stop conflating things. He aggregates the polls. Rasmussen is constantly Republican biased when compared to the aggregate this year, so they have not shown any improvement since last cycle. Gallop has the same problem, but they are falling in line now - they don't want to be cause with bogus numbers when the election hits.