The number of dots has to do with the number of polls.
The non-Rasmussen pols greatly outnumber the Rasmussen polls.
So, the number of Rasmussen polls can't be throwing off the results of combining the polls. The only way for the Rasmussen polls to throw off the combined other polls is for their volume to be greater.Obviously, reading comprehension is my friend. I believe that is a relationship you need to work on though Sebastian.
Wow.
Seriously, wow.
This is not hard to understand.
Number of polls =! number of people polled.
I never said that it did.
|
Stuff
Well, it may seem that way to you; but I DON'T expect other people to do what I am unwilling to do myself. I DON'T want other people to do stuff while I sit back on the sidelines.
I don't think that I can be any clearer than that. However, I refuse to play your game of wordplay. No matter what I tell you I have done (true or not); you will just use that information as another tool to try and discredit me. Because no matter how much I have done, more can always be done; I have no illusions about this. But that does not invalidate my point. I'm not asking people to give all their extra cash to charity. I'm asking them to stop b*$&*ing that their tax money might be used to help other Americans out. Because face it; arguing against that boils down to selfishness pure and simple.
So keep asking me what I have done. And I will give you the same answer; none. of. your. business. Deal with it.
Bella Sara Charter Superscriber
|
But, polling is more accurate when larger samples are polled. So, this poster is arguing that Rasmussen's polling is bad because he has different results than others and is supporting that this shows a weakness by pointing out a strength of the methodology that Rasmussen uses.
Wow, you can't even read your own posts...
Just want to say to Sam. I agree with pretty much every thing that you have said in this thread. And kudos to you for sticking to your guns and speaking what you believe no matter what the fallout. Keep preaching the word! I'd write more, but I think you have the debate well in hand on our side of the fence. Keep it up my friend!
You like me! You really like me! :) Thanks.
I admit I find internet discussion fairly enervating, especially when I'm trying to write about a big, complicated issue with a long history. Plus I love my history stuff so I try to pull out all the stops for it. Then most of the time it falls on deaf ears or the point is lost in a stream of decontextualized back and forth over minutia.
Or someone gripes about how rude I am, which is probably true but doesn't entail my being wrong. ;) I mean I try to be nice when I can but if I have to compromise decency or honesty to do so, I don't think it's a fair trade.
|
Bella Sara Charter Superscriber
|
The number of dots has to do with the number of polls.
The non-Rasmussen pols greatly outnumber the Rasmussen polls.
So, the number of Rasmussen polls can't be throwing off the results of combining the polls. The only way for the Rasmussen polls to throw off the combined other polls is for their volume to be greater.
Ah...I finally see what you're saying. It's not supported by logic, math, or what the link says, but at least I understand what you're saying. Part of the problem is you keep using the concepts of "volume of polls" and "number of people polled" as if they are the same, which they aren't. That's why it's so hard to understand what your argument is.
Let's do some basic math. Here are some non-R polls:
4 4 4 4 4 4 4
Average: 4
Here are some R polls:
6 6 6
Average: 6
Now, you seem to be asserting that because there are more of the former than the later, the later can't affect the average of the combined group. That's mathematically false.
4 4 4 4 4 4 4 6 6 6
Averge: 4.6
Also, nothing, and I mean nothing, in the link that supports the idea that the polls are being combined by volume of people polled. You made that up to discredit the analysis and then attempted to paper it over with poor use of math.
Wow, you can't even read your own posts...But, polling is more accurate when larger samples are polled. So, this poster is arguing that Rasmussen's polling is bad because he has different results than others and is supporting that this shows a weakness by pointing out a strength of the methodology that Rasmussen uses.
The KOS posted this:
A key point here is not just that Rasmussen's numbers show lower approval ratings for President Obama than other pollsters, it's that the sheer volume of Rasmussen's polls allow them to steer the narrative away from the numbers being shown by everybody else.
I specifically stated that this "volume" referred to was the size of the polls:
[quote"The Thing from Beyond the Edge"]
But, polling is more accurate when larger samples are polled. So, this poster is arguing that Rasmussen's polling is bad because he has different results than others and is supporting that this shows a weakness by pointing out a strength of the methodology that Rasmussen uses.
To which you argued:
There is no claim that I saw that Rasmussen polls more people, only that they conduct a large number of polls. A large number of polls != more people polled or a larger sample size (they could be polling the same group of people multiple times). The statement that Rasmussen polls a larger sample is something you pulled out of thin air or got from another source - it is not supported by the link.
But, you are pulling things out of thin air as pointed out below:
[quote"The Thing from Beyond the Edge"]
Red dots are Rasmussen polls the other color are non-Rasmussen polls.
If you look at the dots, there are more non-red dots than red dots. There are a whole lot more. It isn't even close. The volume of Rasmussen's polls is the size of the polls or the poster isn't counting the number of polls.
Your argument is that when the poster stated that the "volume" of Rasmussen's polls was throwing off the combined results that he was referring to the number of polls and not the size. My argument is that the poster was referring to the size of the polls and not the number of polls. I back that up by counting the dots which refer to the number of polls and not the size of the polls. The Rasmussen number clearly is greatly outweighed by the combined other polls. So, the only way for the "volume" of the Rasmussen polls to throw off the combined totals is if the size of the polls were greatly larger.
Either the size of the polls is larger or the claim that the number of polls throws off the combined total is incorrect as stated above.
Again, no where did I state that number of dots referred to size of polling but rather that "volume" referred to size of polling.
The number of dots has to do with the number of polls.
The non-Rasmussen pols greatly outnumber the Rasmussen polls.
So, the number of Rasmussen polls can't be throwing off the results of combining the polls. The only way for the Rasmussen polls to throw off the combined other polls is for their volume to be greater.Ah...I finally see what you're saying. It's not supported by logic, math, or what the link says, but at least I understand what you're saying. Part of the problem is you keep using the concepts of "volume of polls" and "number of people polled" as if they are the same, which they aren't. That's why it's so hard to understand what your argument is.
Let's do some basic math. Here are some non-R polls:
4 4 4 4 4 4 4
Average: 4
Here are some R polls:
6 6 6
Average: 6
Now, you seem to be asserting that because there are more of the former than the later, the later can't affect the average of the combined group. That's mathematically false.
4 4 4 4 4 4 4 6 6 6
Averge: 4.6
Also, nothing, and I mean nothing, in the link that supports the idea that the polls are being combined by volume of people polled. You made that up to discredit the analysis and then attempted to paper it over with poor use of math.
Now you are speaking in hyperbole.
I stated that the results would not throw off the average and did not state that it wouldn't change it.
So, I think you need to work on your math.
Again, nothin in the link stated that the "volume" spoken of referred to the number of polls. You made thatt up.
Further, you continue to make unsubstantiated claims that they are using poor methodology.
Again, I refer to the past presidential election. 17 of the 23 polls overestimated Obama while Rasmussen and Pew were perfect. So, 17 of the 23 is a huge number that apparently skew to the left. there is far more evidence to suggest malpractice by other pollsters than by Rasmusen.
You fail again, Sebastian.
Wow. That's some very bad analytical reasoning.
Yes, it is bad analytical reasoning that you are displaying.
|
Bella Sara Charter Superscriber
|
Now you are speaking in hyperbole.
Nope, providing an example because 8th grade math seems to throw you for a loop.
Either the size of the polls is larger or the claim that the number of polls throws off the combined total is incorrect as stated above.
Incorrect. The link is talking about the average result of polls. If Rasmussen does a large number of polls, they will have a greater effect on the average than they would if they did fewer. That is how volume relates to the conclusion in the link.
Again, nothin in the link stated that the "volume" spoken of referred to the number of polls. You made thatt up.
To illustrate this point, I generated a series of scatter plot charts using pollster.com's index of polls. Every poll in pollster.com's index is represented on each chart by a dot, plotted horizontally by the date of the poll, and vertically by the results of the poll.
Where exactly is the number of people polled referenced?
Further, you continue to make unsubstantiated claims that they are using poor methodology.
I make no claims about their methodology, I provided examples of how they could do a large number of polls while polling a small number of people.
The link doesn't support your original argument and doesn't say what you claim it says (i.e., Rasmussen has a larger sample size). That's my only claim. It's still true. The end.
|
|
Wow, apparently the board ate my last response to The Thing.
So let me sum it up. You may protest all you want, but it's still none of your business. I don't expect anyone to do anything that I am unwilling to do myself. I don't want folks to go bankrupt giving to charity; I want folks to stop b~$*~ing when their tax money (which they have to pay regardless) is used to help better the fortunes of their fellow countrymen.
What I give is none of your business; so stop asking.
Did you not notice that they were perfect in the last election and best in the previous one?I'm not sure how that relates to your inability to read.
It doesn't. It relates to your inability to read.
You seem to forget that you posted this immediately prior:
Rasmussen conducts a lot of polls. They might just poll their good friends Billy Bob and Jimmy John Jones every day.
If they just polled their good friends they would not have been perfect in the 2008 presidential election or best in the 2004 one.
There is no evidence to support their polling being bad other than that it is "right" most of the other pollsters, 17 (out of 23) of which were left of it when it was perfect in 2008.
Wow, apparently the board ate my last response to The Thing.
So let me sum it up. You may protest all you want, but it's still none of your business. I don't expect anyone to do anything that I am unwilling to do myself. I don't want folks to go bankrupt giving to charity; I want folks to stop b%~&~ing when their tax money (which they have to pay regardless) is used to help better the fortunes of their fellow countrymen.
What I give is none of your business; so stop asking.
I will ask every time that you criticize others as not caring when they dislike how their tax money is spent. Money they could spend themselves to help others in a more efficient manner.
So, still waiting since you throew out another criticism...
|
|
Wow, apparently the board ate my last response to The Thing.
So let me sum it up. You may protest all you want, but it's still none of your business. I don't expect anyone to do anything that I am unwilling to do myself. I don't want folks to go bankrupt giving to charity; I want folks to stop b%~&~ing when their tax money (which they have to pay regardless) is used to help better the fortunes of their fellow countrymen.
What I give is none of your business; so stop asking.
I will ask every time that you criticize others as not caring when they dislike how their tax money is spent. Money they could spend themselves to help others in a more efficient manner.
So, still waiting since you throew out another criticism...
Well ask all you want; it's a free country. Just don't hold your breath.
|
Bella Sara Charter Superscriber
|
Did you not notice that they were perfect in the last election and best in the previous one?I'm not sure how that relates to your inability to read.It doesn't. It relates to your inability to read.
You seem to forget that you posted this immediately prior:
Rasmussen conducts a lot of polls. They might just poll their good friends Billy Bob and Jimmy John Jones every day.
If they just polled their good friends they would not have been perfect in the 2008 presidential election or best in the 2004 one.
There is no evidence to support their polling being bad other than that it is "right" most of the other pollsters, 17 (out of 23) of which were left of it when it was perfect in 2008.
You're seriously too vested in this.
Again, I direct your attention to:
I make no claims about their methodology, I provided examples of how they could do a large number of polls while polling a small number of people.
Which is what the quote above is.
I don't care which pollster is more accurate. For all I know, Rasmussen has a larger sample size. All I know is that the link Erik provided didn't state that and you incorrectly claimed it did. That's the sum total of my claim. You didn't read the link correctly and you continue to pretend like it talks about the volume of people polled when it talks about the volume of polls conducted.
Oooh, excellent use of "you're rubber and I'm glue." Clever stuff.Wow. That's some very bad analytical reasoning.Yes, it is bad analytical reasoning that you are displaying.
OK, the volume refers to number of size of polls because the number of dots would not throw off the averages.
Nothing in the post states that "volume" refers to number of dots.
Rasmussen has been incredibly accurate in the past even when its numbers were to the right of basically everyone else. There is no evidence to substantiate that they are improperly polling and past accuracy lends credence to the methodology used.
|
Bella Sara Charter Superscriber
|
OK, the volume refers to number of size of polls because the number of dots would not throw off the averages.
Yes it would. If they did a large number of polls and those polls were averaged with other polls, they would throw off the averages.
Nothing in the post states that "volume" refers to number of dots.
You mean other than this:
To illustrate this point, I generated a series of scatter plot charts using pollster.com's index of polls. Every poll in pollster.com's index is represented on each chart by a dot, plotted horizontally by the date of the poll, and vertically by the results of the poll.
Rasmussen has been incredibly accurate in the past even when its numbers were to the right of basically everyone else. There is no evidence to substantiate that they are improperly polling and past accuracy lends credence to the methodology used.
And I care because?
Again, all I'm saying is that you aren't reading the linked article correctly. It says nothing about the number of people polled, it only talks about the number of polls conducted. You are incorrect in your assertion that the only way for the Rasmussen polls to throw off the average of all polls is for them to poll more people. If they conducted more polls, that would throw the average off as well.
The article makes no claims as to the number of people Rasmussen polled. You brought that into the conversation and claimed that's what it said.
Did you not notice that they were perfect in the last election and best in the previous one?I'm not sure how that relates to your inability to read.It doesn't. It relates to your inability to read.
You seem to forget that you posted this immediately prior:
Rasmussen conducts a lot of polls. They might just poll their good friends Billy Bob and Jimmy John Jones every day.
If they just polled their good friends they would not have been perfect in the 2008 presidential election or best in the 2004 one.
There is no evidence to support their polling being bad other than that it is "right" most of the other pollsters, 17 (out of 23) of which were left of it when it was perfect in 2008.
You're seriously too vested in this.
Again, I direct your attention to:
I make no claims about their methodology, I provided examples of how they could do a large number of polls while polling a small number of people.
Which is what the quote above is.
I don't care which pollster is more accurate. For all I know, Rasmussen has a larger sample size. All I know is that the link Erik provided didn't state that and you incorrectly claimed it did.
Providing examples of how they could have large poll size with a small number of people is moot and means nothing to the debate unless it is questioning the methodology used.
I stated it referred to the size because it did. You are saying that I stated it incorrectly because I read "volume" as meaning poll size. There is nothing in the post that says "volume" referrs to the number of polls. The sentence reads as if it refers to the size of the polls. It really is that simple.
Stating so is not an incorrect claim. You may wish to interpret differently but there is nothing there for you to make your statements that I misread anything or do not have good reading comprehension.
Again, your poor logic fails, Sebastian.
|
Bella Sara Charter Superscriber
|
Providing examples of how they could have large poll size with a small number of people is moot and means nothing to the debate unless it is questioning the methodology used.
Uh...no. It shows how you can conduct a large number of polls without polling a lot of people. You keep arguing that the only way the number of polls could affect the average of all polls is if those polls were of a larger number of people. That's not true.
I stated it referred to the size because it did.
Okay, prove it then. I've already posted a quote (twice) that it is the number of polls (not the number of people polled) being averaged. I don't see a single reference to the number of people polled. Show me where that fact is stated. You assertion that the only way for Rasmussen polls to affect the average is if they poll more people is not correct. If their polls are different from the average and they conduct a large number of them, it will pull the average in that direction.
|
But, polling is more accurate when larger samples are polled. So, this poster is arguing that Rasmussen's polling is bad because he has different results than others and is supporting that this shows a weakness by pointing out a strength of the methodology that Rasmussen uses.
Just thought I'd chime in here. Polling more people does not necessarily make your poll more accurate.
If there is a problem with your survey methodology increasing your sample size won't fix it.
Classic example was a postal poll that used magazine subscription, telephone ownership and car ownership as a means of selecting the sample. 2 or 3 million people polled, showed the republican incumbent would win by a landslide. Gallup polled 50k and predicted Roosevelt would win. Sometime in the 30s I think.
Knew that sociology degree would be useful for something..
OK, the volume refers to number of size of polls because the number of dots would not throw off the averages.
Yes it would. If they did a large number of polls and those polls were averaged with other polls, they would throw off the averages.
Nothing in the post states that "volume" refers to number of dots.You mean other than this:
To illustrate this point, I generated a series of scatter plot charts using pollster.com's index of polls. Every poll in pollster.com's index is represented on each chart by a dot, plotted horizontally by the date of the poll, and vertically by the results of the poll.Rasmussen has been incredibly accurate in the past even when its numbers were to the right of basically everyone else. There is no evidence to substantiate that they are improperly polling and past accuracy lends credence to the methodology used.Can I care because?
Again, all I'm saying is that you aren't reading the linked article correctly. It says nothing about the number of people polled, it only talks about the number of polls conducted. You are incorrect in your assertion that the only way for the Rasmussen polls to throw off the average of all polls is for them to poll more people. If they conducted more polls, that would throw the average off as well.
The article makes no claims as to the number of people Rasmussen polled. You brought that into the conversation and claimed that's what it said.
It specifically refers to the "volume of Rasmussen's polls allow them to steer the narrative away from the numbers being shown by everybody else". The "volunme" of their polls reads as the size of their polls and not the number of polls performed. I have only seen it used that way in the past and that is how it reads.
As a note, I checked the FAQ to the composite index mentioned in the quote in hopes of finding out how the numbers were obtained but the page was under construction. No luck.
But, polling is more accurate when larger samples are polled. So, this poster is arguing that Rasmussen's polling is bad because he has different results than others and is supporting that this shows a weakness by pointing out a strength of the methodology that Rasmussen uses.
Just thought I'd chime in here. Polling more people does not necessarily make your poll more accurate.
If there is a problem with your survey methodology increasing your sample size won't fix it.
Classic example was a postal poll that used magazine subscription, telephone ownership and car ownership as a means of selecting the sample. 2 or 3 million people polled, showed the republican incumbent would win by a landslide. Gallup polled 50k and predicted Roosevelt would win. Sometime in the 30s I think.
Knew that sociology degree would be useful for something..
But, when you use good methodology, the larger poll is more accurate. Rasmussen uses good methodology. Give me a minute to look it up in the link that was used to support them not being good...
I would read "volume of rasmussen's polls" as the number of polls conducted, not the number of individuals polled.
I would not.
Also,
Wiki linked to discredit Rasmussen earlier...
Rasmussen's polls are notable for their use of automated public opinion polling, involving pre-recorded telephone inquiries. These types of polls have been shown to produce accurate results at low cost. But some have doubted their reliability. In 2004 Slate magazine said they “publicly doubted and privately derided Rasmussen” polls because of the methodology. However, after the election, they concluded that Rasmussen’s polls were the most accurate.At the end of the 2008 Presidential Election, there were eight national tracking polls and many other polls conducted on a regular basis. Polling guru Nate Silver reviewed the tracking polls and said that while none were perfect, "Rasmussen -- with its large sample size and high pollster rating -- would probably be the one I'd want with me on a desert island." After the election, Rasmussen's poll was rated as the most accurate, when compared to various other final pre-election polls.
I see little reason to doubt their methodology.
|
|
I would read "volume of rasmussen's polls" as the number of polls conducted, not the number of individuals polled.I would not.
I really think the guy is making a composite of poll results not a comparison of number of people polled. If that's what he's doing, the number of polls conducted and their result that matters and not the number of people polled.
That's my understanding of what he's doing.
How about this:
There is not a methodology shown for how the composite index used determined its numbers. So, we do not know if it is the average of each poll taken as a sum of polled people or an average taken with each poll having equal weight. Also, we do not know if the poster is correctly or incorrectly making conclusions about someone else's data because we do not know how the data is obtained.
|
Bella Sara Charter Superscriber
|
I would read "volume of rasmussen's polls" as the number of polls conducted, not the number of individuals polled.I would not.
I really think the guy is making a composite of poll results not a comparison of number of people polled. If that's what he's doing, the number of polls conducted and their result that matters and not the number of people polled.
That's my understanding of what he's doing.
Plus, if he were looking at the number of people polled, he'd use terms like "weighted" polls and, typically, the dots would be different sizes.
Sorry Thing, it just doesn't say what you want it to say. It's talking about the number of polls, not the number of people polled. There's absolutely nothing to imply the later.
|
Bella Sara Charter Superscriber
|
How about this:
There is not a methodology shown for how the composite index used determined its numbers. So, we do not know if it is the average of each poll taken as a sum of polled people or an average taken with each poll having equal weight. Also, we do not know if the poster is correctly or incorrectly making conclusions about someone else's data because we do not know how the data is obtained.
Uh no. It's not ambiguous at all. He's talking about number of polls. That may not be the best way to compare polls, but that's what's happening in that particular article.
I would read "volume of rasmussen's polls" as the number of polls conducted, not the number of individuals polled.I would not.
I really think the guy is making a composite of poll results not a comparison of number of people polled. If that's what he's doing, the number of polls conducted and their result that matters and not the number of people polled.
That's my understanding of what he's doing.
The poster is referring to someone else's composite index.
|
RPG Superstar 2009 Top 32, 2010 Top 8
|
Interesting. Matthew, would you say you are one of the people I referred to earlier? Those who would prefer a more powerful Medicaid as opposed to Medicare(State "insurance" over Federal "insurance", that is).
Well even Medicaid is a misnomer to call local as it's still macromanaged from the federal level. That aside, yes, I'd have less of an issue if the states established their own safety net, or lack of.
Funny story about medicaid/disability coverage:
"I don't understand. Some of my caseload in Ohio, who are on disability, I don't understand why they aren't working."
"What do you mean?"
"Well, if they were in Florida, they'd not be on disability, they'd be considered able to work."
"And?"
"Well, they're not working but they could."
"And..." (come on mom, you've almost made the leap)
"I just don't get it."
Uh no. It's not ambiguous at all. He's talking about number of polls. That may not be the best way to compare polls, but that's what's happening in that particular article.How about this:
There is not a methodology shown for how the composite index used determined its numbers. So, we do not know if it is the average of each poll taken as a sum of polled people or an average taken with each poll having equal weight. Also, we do not know if the poster is correctly or incorrectly making conclusions about someone else's data because we do not know how the data is obtained.
The numbers used are not those he generated. He generated the plots. Someone else generated the numbers. He is comparing the numbers generated by someone else. He is using the plots to show the difference and saying this difference is accounted for by the volume of the polls. The volume of a poll is its size. the number of polls is how many polls are performed.
|
RPG Superstar 2015 Top 16, RPG Superstar 2013 Top 16
|
I admit I find internet discussion fairly enervating, especially when I'm trying to write about a big, complicated issue with a long history. Plus I love my history stuff so I try to pull out all the stops for it. Then most of the time it falls on deaf ears or the point is lost in a stream of decontextualized back and forth over minutia.
I disagree with you frequently but on the above I think you and I are on the same wavelength, brother. I refrain from posting on a lot of these threads just because doing so in a meaningful way would require a significant commitment of mental resources, and the net result would be physical fatigue.
Or someone gripes about how rude I am, which is probably true but doesn't entail my being wrong. ;) I mean I try to be nice when I can but if I have to compromise decency or honesty to do so, I don't think it's a fair trade.
Here's an example of where we'd differ. Even wars can be fought politely.
|
Bella Sara Charter Superscriber
|
The volume of a poll is its size. the number of polls is how many polls are performed.
Nope. Asserting it doesn't make it true.
Look, you didn't read it carefully. That's fine. But you aren't going to convince anyone that it says something it doesn't. There are many reasons and ways to dispute the conclusions of the article, but mis-stating what was said and trying to say there is only one correct way to reference the number of polls v. the number of people polled isn't doing any good.
You're entitled to your own opinion, but not your own facts. The fact is, the article is discussing the number of polls conducted, not the number of people polled.
|
|
The numbers used are not those he generated. He generated the plots. Someone else generated the numbers. He is comparing the numbers generated by someone else. He is using the plots to show the difference and saying this difference is accounted for by the volume of the polls. The volume of a poll is its size. the number of polls is how many polls are performed.Uh no. It's not ambiguous at all. He's talking about number of polls. That may not be the best way to compare polls, but that's what's happening in that particular article.How about this:
There is not a methodology shown for how the composite index used determined its numbers. So, we do not know if it is the average of each poll taken as a sum of polled people or an average taken with each poll having equal weight. Also, we do not know if the poster is correctly or incorrectly making conclusions about someone else's data because we do not know how the data is obtained.
He's using a scatter plot with two variables. Neither variable is poll size. Poll size won't have any effect on his graph or any conclusion he could draw from it. It would be very odd for the poster to use volume to refer to poll size instead of the number of red dots (polls conducted).
I'll try to dig up my survey methodology text and see if volume has a definition, if the guy is a statistician he might be using the word differently.
Going back to the original argument.
Rasmussen was criticized as being wrong because its results were to the right of most other pollsters.
But, it has a history of being accurate while simultaneously being to the right of most other pollsters.
Example: 2008 election.
Rasmussen and Pew were dead on with their numbers. 17 of the 23 over-estimated Obama's margin of victory. $ underestimated it. So, more than four times as many were to the "left" as were to the "right" of their poll.
Also, they were the most accurate in the 2004 presidential election.
I see no reason to dispute their polling other than because it disagrees with what soemone wants it to be.
|
RPG Superstar 2015 Top 16, RPG Superstar 2013 Top 16
|
stuff
I gotta say I really respect that you Paizo staffers don't get your hands too dirty in the political threads. While it might be possible to draw some inferences about your political leanings from your posts, for all we know you guys could be far-left Marxist guerrillas or far-right neo-McCarthy-ites or anything in between. And I for one like it that way. It's smart business policy to avoid alienating your customers with politics. I don't know if you guys have an internal policy about staff participation on these forums, but it sure seems like you do. So kudos for your continued professionalism in that regard, and thanks for spending your time developing product instead of getting mired eyeball-deep in threads like this one.
|
|
Going back to the original argument.
Rasmussen was criticized as being wrong because its results were to the right of most other pollsters.
But, it has a history of being accurate while simultaneously being to the right of most other pollsters.
Example: 2008 election.
Rasmussen and Pew were dead on with their numbers. 17 of the 23 over-estimated Obama's margin of victory. $ underestimated it. So, more than four times as many were to the "left" as were to the "right" of their poll.
Also, they were the most accurate in the 2004 presidential election.
I see no reason to dispute their polling other than because it disagrees with what soemone wants it to be.
I believe some of the criticism stems from the early polling data not the final data. And if you read the later nate silver post, he has concerns about question formulation in polls since the election being biased. So non sampling errors, not sample size or composition.
|
Bella Sara Charter Superscriber
|
I see no reason to dispute their polling other than because it disagrees with what soemone wants it to be.
Does that apply to other polls that have reliable track records and don't support what you want them to be, or does it just apply to other people?
I mean, you did just spend a page arguing that an analysis of polls posted on the most left leaning website on the planet somehow supported your views. That's some pretty strong bias...
He's using a scatter plot with two variables. Neither variable is poll size. Poll size won't have any effect on his graph or any conclusion he could draw from it. It would be very odd for the poster to use volume to refer to poll size instead of the number of red dots (polls conducted).The numbers used are not those he generated. He generated the plots. Someone else generated the numbers. He is comparing the numbers generated by someone else. He is using the plots to show the difference and saying this difference is accounted for by the volume of the polls. The volume of a poll is its size. the number of polls is how many polls are performed.Uh no. It's not ambiguous at all. He's talking about number of polls. That may not be the best way to compare polls, but that's what's happening in that particular article.How about this:
There is not a methodology shown for how the composite index used determined its numbers. So, we do not know if it is the average of each poll taken as a sum of polled people or an average taken with each poll having equal weight. Also, we do not know if the poster is correctly or incorrectly making conclusions about someone else's data because we do not know how the data is obtained.
But, he is not making the calculations. Those come from the composite index link. He states that the change is due to poll volume but does not state that it is number of polls. Again, he is not making the calculations that I can see. He is using someone else's index.
Rasmussen does not say, "hey we did five polls all at the same time" so that people will weight the averages by having more Rasmussen polls. If you take a snapshot at at given time, you will have one Rasmussen poll.
The only way that the numbers come into play is if you create a trend line.
•If we have at least 8 public polls, we fit a trend line to the dots represented by each poll using a "Loess" iterative locally weighted least squares regression.
But, at each point in time looked at, Rasmussen is one of many polls just not necessarily the same one of many. So, there are 23 polls being performed. For example, in any three day period there are six polls performed and one is Rasmussen and the other five come from the other 22. At any given time it does not push the average one way or another due to its number of polls. The point is that it can adjust a "trend line" but cannot adjust a poll average with the number of polls performed. The only way it can adjust a poll average by its volume is for volume to be the size of the polls.
But, the KOS was referring to the narrative, and thus the trend line:
Obviously, tracking polls don't actually count for anything when it comes to tabulating ballots, but they do have a big impact on narrative, and the impact of Rasmussen's numbers on the overall composite was shocking to me, to put it mildly.
So, he is talking about the narrative and thus the trend line and thus the number because it is affected by making a trend based upon least squares at individual points, connecting the dots, and Rasmussen has more frequent polling.
/facepalm
I cede that point.
Sorry about that.
:)
The problem is that the poster is using the fact that it moves a composite trend line to decry it as being inaccurate. The case is simply not analogous to a poll conducted at any given time.
There is nothing in this to support poor methodology from Rasmussen. Tt amounts to Rasmussen is poor because it gives different results than the others every time they are matched together, but we just can't find any evidence to prove or even support it is poor.
I believe some of the criticism stems from the early polling data not the final data. And if you read the later nate silver post, he has concerns about question formulation in polls since the election being biased. So non sampling errors, not sample size or composition.Going back to the original argument.
Rasmussen was criticized as being wrong because its results were to the right of most other pollsters.
But, it has a history of being accurate while simultaneously being to the right of most other pollsters.
Example: 2008 election.
Rasmussen and Pew were dead on with their numbers. 17 of the 23 over-estimated Obama's margin of victory. $ underestimated it. So, more than four times as many were to the "left" as were to the "right" of their poll.
Also, they were the most accurate in the 2004 presidential election.
I see no reason to dispute their polling other than because it disagrees with what soemone wants it to be.
But, the criticism is severely flawed.
It is criticizing the ealy polls because they did not match the late result. There is no evidfence to support that the early polls were incorrect or that they should in any way match the final result.
Chief Creative Officer, Publisher
|
stuffI gotta say I really respect that you Paizo staffers don't get your hands too dirty in the political threads. While it might be possible to draw some inferences about your political leanings from your posts, for all we know you guys could be far-left Marxist guerrillas or far-right neo-McCarthy-ites or anything in between. And I for one like it that way. It's smart business policy to avoid alienating your customers with politics. I don't know if you guys have an internal policy about staff participation on these forums, but it sure seems like you do. So kudos for your continued professionalism in that regard, and thanks for spending your time developing product instead of getting mired eyeball-deep in threads like this one.
Believe you me, it's difficult sometimes. :)
Over the years I've come to appreciate that its WAAAAAY better (and more interesting) to focus on the things we share (GAMING) than the things we probably disagree about (politics and religion).
I am SO TEMPTED to push for an all-out ban on political and religious discussions, but so long as they remain civil and in the off-topic areas, I think they're ok. A lot of people seem to enjoy them, so I'd prefer to let the boards largely police themselves.
I've seen some very respectful discussions here where people on different sides of an issue explain their point of view in a rational, non-emotional way. I don't expect people to change their minds about politics or religion based on posts on gaming sites, but I think there is some value in being able to see an opposing viewpoint articulated by someone you _know_ you share interests with.
So, yeah, we don't have a formal policy on getting involved in political threads, but all of us know it's generally a pretty bad idea.
I want people to decide to buy or not buy our gaming products based on whether they think those products would be fun to play, not based on some political screed one of us posts to the boards in a fit of pique.
Glad to know our efforts (or lack thereof) are largely appreciated! :)
I see no reason to dispute their polling other than because it disagrees with what soemone wants it to be.Does that apply to other polls that have reliable track records and don't support what you want them to be, or does it just apply to other people?
I mean, you did just spend a page arguing that an analysis of polls posted on the most left leaning website on the planet somehow supported your views. That's some pretty strong bias...
I haven't stated that left leaning polls are wrong.
I haven't said that other polls in general are wrong.I have simply stated that the fact that Rasmussen's polls differ from the majority is not reason to dismiss them.
I have stated that their record of being correct when to the "right" of most others is good enough reason to not dismiss them for being to the "right" of most others.
Again, they (the KOS) were misleading. They were trying to say that because Rasmussen was more prevalent in its appearances in the trend line that should be dismissed as inaccurate. Or perhaps it should be dismissed because it does not help them tell the narrative they want it to tell. However, no where did they show it to be inaccurate and history has shown it to be repeatedly accurate.
They want to paint a narrative using a trend line but because the polls of Rasmussen basically show up each time they calculate their points of the trend line, it distorts what they WISH to portray. Each point in time is correctly weighted. Each point in time, Rasmussen is one of several who made a poll at that point.
The KOS is trying to dismiss its numbers becauyse it does not paint a trend line that it wants to.
|
|
stuffI gotta say I really respect that you Paizo staffers don't get your hands too dirty in the political threads. While it might be possible to draw some inferences about your political leanings from your posts, for all we know you guys could be far-left Marxist guerrillas or far-right neo-McCarthy-ites or anything in between. And I for one like it that way. It's smart business policy to avoid alienating your customers with politics. I don't know if you guys have an internal policy about staff participation on these forums, but it sure seems like you do. So kudos for your continued professionalism in that regard, and thanks for spending your time developing product instead of getting mired eyeball-deep in threads like this one.Believe you me, it's difficult sometimes. :)
Over the years I've come to appreciate that its WAAAAAY better (and more interesting) to focus on the things we share (GAMING) than the things we probably disagree about (politics and religion).
I am SO TEMPTED to push for an all-out ban on political and religious discussions, but so long as they remain civil and in the off-topic areas, I think they're ok. A lot of people seem to enjoy them, so I'd prefer to let the boards largely police themselves.
I've seen some very respectful discussions here where people on different sides of an issue explain their point of view in a rational, non-emotional way. I don't expect people to change their minds about politics or religion based on posts on gaming sites, but I think there is some value in being able to see an opposing viewpoint articulated by someone you _know_ you share interests with.
So, yeah, we don't have a formal policy on getting involved in political threads, but all of us know it's generally a pretty bad idea.
I want people to decide to buy or not buy our gaming products based on whether they think those products would be fun to play, not based on some political screed one of us posts to the boards in a...
Letting the boards police themselves is appreciated. Hopefully they stay civil enough for that to continue.