I guess Romney is really surging in Ohio.

[Charles Kozierok]

You have two measurement methods. One give you an answer of 40 with and error of 4%. The other gives you an answer of 60 with an error of 5%.

That doesn't happen with properly-designed polls. The 95% confidence intervals usually overlap.

If one poll has an interval that doesn't overlap, it will generally be dismissed as an outlier.

Can you find any pair of Ohio polls that don't have overlapping intervals?

 

[OrByte]

See:





So I am increasing the uncertainty of my age by adding in more dating methods.

lol! nice.

I wish I can do that with my checking account 🙂

 

[DrewSG3]

This was posted in another thread by someone.

http://www.realclearpolitics.com/epolls/2008/president/oh/ohio_mccain_vs_obama-400.html

Yep. Ramussen had it as a tie, Obama won 52-47

 

[Charles Kozierok]

He was also the only pollster to call Florida for McCain.

 

[woolfe9999]

Rasmussen was not the most accurate in 2008. Both of those sources you cite are outdated and invalid. They were compiled in early November, before all ballots had been tallied.

The "Report Card" was based on a 6.5 Obama margin. Obama, in fact, won by 7.2 percent. Notice that it shows Rasmussen being off by 0.5 percent. Rasmussen was, in fact, off by 1.2 percent. The Report Card is fatally flawed.

The Fordham "Initial Report" was based on a 6.15 margin. Again, Obama won by 7.2 percent. Fordham later released a complete analysis based on the final official popular vote outcome. Eight pollsters were found to be more accurate than Rasmussen.

http://www.fordham.edu/images/acade...campaign_/2008 poll accuracy panagopoulos.pdf

Nice work. I didn't know this until you posted it. It's right there in wiki:

http://en.wikipedia.org/wiki/Rasmussen_Reports#2008

It's amazing that Rasmussen's supposed top accuracy in the 2008 PV polling is still being publicized even this year. For example,

http://www.newsmax.com/Newsfront/pew-rasmussen-most-accurate/2012/09/11/id/451377
http://www.breitbart.com/Big-Govern...back-rasmussen-most-accurate-pollster-in-2008

Not only was Rasmussen off by 1.2%, but of course it was in McCain's direction. As was its state by state polling that year. As was its Congressional polling a heavy R lean in 2010.

This would be why people like Silver are treating Rasmussen as having a 2 point R lean house effect.

- wolf

 

[xBiffx]

That doesn't happen with properly-designed polls. The 95% confidence intervals usually overlap.

If one poll has an interval that doesn't overlap, it will generally be dismissed as an outlier.

Can you find any pair of Ohio polls that don't have overlapping intervals?

So it doesn't apply in this situation because the numbers are closer together? Ok. Doesn't mean I was wrong, it just doesn't apply here apparently.

 

[Doc Savage Fan]

He did fine on the overall number. His state polling was abysmal.

He nailed Ohio in 2010.

http://www.rasmussenreports.com/pub..._rasmussen_poll_results_2010_senate_elections

 

[Doc Savage Fan]

dp

 

[Charles Kozierok]

He nailed Ohio in 2010.

Can't find specific Ohio numbers. But overall he sucked in 2010 as well.

So it doesn't apply in this situation because the numbers are closer together? Ok. Doesn't mean I was wrong, it just doesn't apply here apparently.

No, it's more like it doesn't apply to any situation because the situation you described doesn't happen. If two polls are off by that much, one or the other is bogus (or both) or the data cannot be properly sampled.

 

[Doc Savage Fan]

Can't find specific Ohio numbers. But overall he sucked in 2010 as well.

I linked 2010 numbers that clearly showed the specific results for Ohio. Did you read my link? /confused

 

[Charles Kozierok]

I linked 2010 numbers that clearly showed the specific results for Ohio. Did you read my link? /confused

I was confused by that link also. I think you messed something up there. 🙂

 

[Doc Savage Fan]

I was confused by that link also. I think you messed something up there. 🙂

Ooops. Edited post. Try this.

http://www.rasmussenreports.com/public_content/politics/elections/election_2010/election_2010_senate_elections/final_rasmussen_poll_results_2010_senate_elections

 

[Blackjack200]

I linked 2010 numbers that clearly showed the specific results for Ohio. Did you read my link? /confused

nm

 

[Charles Kozierok]

Ooops. Edited post. Try this.

http://www.rasmussenreports.com/public_content/politics/elections/election_2010/election_2010_senate_elections/final_rasmussen_poll_results_2010_senate_elections

Um. Even there he was off by 6 points. 🙂

And of course, towards the Republican.

Look through that list. In nearly every close race, he was too high towards the Republicans.

Some of his calls were truly awful, like Nevada.

 

[Blackjack200]

Ooops. Edited post. Try this.

http://www.rasmussenreports.com/public_content/politics/elections/election_2010/election_2010_senate_elections/final_rasmussen_poll_results_2010_senate_elections

Looks like he predicted a 24 point margin and the actual margin was 18?

 

[Farang]

Ooops. Edited post. Try this.

http://www.rasmussenreports.com/public_content/politics/elections/election_2010/election_2010_senate_elections/final_rasmussen_poll_results_2010_senate_elections

Someone correct me if I'm going too far out on a limb here, but I don't know if that qualifies as nailing the result. If a pollster tells me Portman 57%, Fisher 33%, and I'm assuming undecided 10%, is that pollster really accurate when Portman gets 57% on election day? He got 0% of the undecided vote?

 

[emperus]

Let me simplify, again.

You have two measurement methods. One give you an answer of 40 with and error of 4%. The other gives you an answer of 60 with an error of 5%. What some are trying to say is that the final results of these two methods is 50 with a 3.6% error. How is this possible?

That result doesn't even fit within the two measurements possibilities. You could still get a result of 59 or 42 or 56 or.......

Can you stop sidetracking the thread. I have linked you to a post with the math and I have linked to a notable economist explaining it. What else would a rational person need? You're wrong and there is nothing more to argue. Stop trolling.

 

[xBiffx]

Um. Even there he was off by 6 points. 🙂

And of course, towards the Republican.

Look through that list. In nearly every close race, he was too high towards the Republicans.

Some of his calls were truly awful, like Nevada.

For the Democrat candidate but he was dead on for the Republican candidate.

Ohio
Rob Portman (R)
57%
57%
10/26/2010

Lee Fisher (D)
39%
33%

So on average, wait nevermind not going there......

 

[Charles Kozierok]

Okay, I looked through that list. Considering only reasonably close races (within 20 points at most), I see 2 that he got exactly right and 15 that he didn't.

Of those 15, 14 were too high for the Republican, and 1 for the Democrat. That includes several cases of getting the winner wrong.

I don't see how anyone can seriously suggest that this is an argument for Rasmussen's credibility.

 

[xBiffx]

Can you stop sidetracking the thread. I have linked you to a post with the math and I have linked to a notable economist explaining it. What else would a rational person need? You're wrong and there is nothing more to argue. Stop trolling.

Try a second or third or maybe even fourth opinion. One opinion does not make for the truth or a mathematical law. Also, I was following up on a question. You stop sidetracking. Jackass.

 

[emperus]

Try a second or third or maybe even fourth opinion. One opinion does not make for the truth or a mathematical law. Also, I was following up on a question. You stop sidetracking. Jackass.

I'm sorry, you're an idiot. Where is your 1st opinion or theory let alone your second? I've already given you two, one by a PHD in statistics who teaches a statistics course and a 2nd by Nobel Laurette. You have not shown proof of 1 theory, or expert that agrees with you but continue arguing a point that all the evidence is against, not because you have some mathematical theory to base it upon because u just believe it to be so. Ask yourself are you arguing out of emotion or logic?

Sometimes you have to know when to shut up and listen. This is your time. Just shut up.

 

[Numeros]

Holy crap, hell of a first post! :thumbsup:

Thanks!

Nice work. I didn't know this until you posted it. It's right there in wiki:

http://en.wikipedia.org/wiki/Rasmussen_Reports#2008

It's amazing that Rasmussen's supposed top accuracy in the 2008 PV polling is still being publicized even this year. For example,

http://www.newsmax.com/Newsfront/pew-rasmussen-most-accurate/2012/09/11/id/451377
http://www.breitbart.com/Big-Govern...back-rasmussen-most-accurate-pollster-in-2008

Not only was Rasmussen off by 1.2%, but of course it was in McCain's direction. As was its state by state polling that year. As was its Congressional polling a heavy R lean in 2010.

This would be why people like Silver are treating Rasmussen as having a 2 point R lean house effect.

- wolf

Yes, that outdated preliminary Fordham report is still all over the internet... and on Fox News!

A couple of weeks ago, Scott Rasmussen appeared on Fox with Megyn Kelly. She pulled out the old Fordham list and said that Rasmussen was #1 in accuracy. Instead of correcting her, Scott Rasmussen just nodded his head and smiled.

 

[First]

Charles, each of these examples is for identical datasets. That is like looking at data from the same poll where they increased the size. As I said, each pollster uses different criteria and different questions to conduct their polls. Each comes up with a calculated polling error which can be different even if the sample size is the same. When you combine all these mixed polls you can't get more confidence given they are all completely separate events.

That's like writing two different computer programs to answer a question. You know the first one is wrong 5% of the time. You know the second one is wrong 5% of the time. When you ask the question to both, the error isn't going to be 5%, its going to be more. Each answer has a 5% chance of being wrong. So when you combine the answer there is no way you only have a 5% chance that answer is wrong, given that each could have been incorrect due to error within each program program.

I agree with regards to this type of polling that errors get smoothed out but in no way is the error of the accumulation of the polls the average or less than the highest error of an individual poll.

The polls generally all conform to the same polling standards and methods, most don't deviate substantially, so your point is entirely moot on that front, and is ridiculous anyway considering the much larger sample sizes you get during aggregation always reduces error, tard. Your stats courses failed you if you think otherwise.

Overall you are extraordinarily, boundlessly wrong, and you're being laughed at.

 

[First]

http://www.publicpolicypolling.com/main/2012/10/obama-leads-50-45-in-ohio.html

Another polls has him up 5 in Ohio, conducted on the 29th and 30th.

 

[Farang]

Thanks!


Yes, that outdated preliminary Fordham report is still all over the internet... and on Fox News!

A couple of weeks ago, Scott Rasmussen appeared on Fox with Megyn Kelly. She pulled out the old Fordham list and said that Rasmussen was #1 in accuracy. Instead of correcting her, Scott Rasmussen just nodded his head and smiled.

imagining this seemed hilarious so I had to find video:

http://www.mediaite.com/tv/chuck-todd-calls-rasmussen-polls-slop-scott-rasmussen-responds/