- May 22, 2005
- 122,906
threads merged.
looks like the anti-trolls are taking turns in the comments section. It's was elcid's turn and now I think mojosfun is going for the "king dumbass" crown
I wish I could be more clear here, so let me say it how it is: from a statistics perspective, it is the equivalent of denying the value of 2+2=4 because you don't know what "+" or "=" means, and you're really sure in your gut it should be 2+2=22. It is really, really embarrassing to see our community showing this kind of ignorance and then ignoring the people trying to help them understand because they don't like the results they're hearing.My issue here is that the sample size is no where near the 3.5% percentage indicated by Tyeraxus post nor does 0.00013793% of a population (800 of 5.8mil) guarantee of confidence interval of 95%as implied.
I wish I could be more clear here, so let me say it how it is: from a statistics perspective, it is the equivalent of denying the value of 2+2=4 because you don't know what "+" or "=" means, and you're really sure in your gut it should be 2+2=22. It is really, really embarrassing to see our community showing this kind of ignorance and then ignoring the people trying to help them understand because they don't like the results they're hearing.
The sample size is not supposed to be "3.5%" of the population. Polling 800 truly random voters in Maryland resulted in these results +-3.5%, and there is a ~95% chance the results are true within those limits. There's no bias in how these numbers were calculated; you plug them into a basic statistics equation and get results back.
Valid criticisms and questions are:
1. What were the questions? Were they loaded?
2. What was the polling method? Was it truly random?
That is what I meant. But, honestly, I'd be a little shocked if they were systematically screwing up like that... the protocol for doing random sampling of a population of people is well-known and -established.Would another valid concern be "Where do the respondents live?" or was that what you intended from the "random" question?
I wish I could be more clear here, so let me say it how it is: from a statistics perspective, it is the equivalent of denying the value of 2+2=4 because you don't know what "+" or "=" means, and you're really sure in your gut it should be 2+2=22. It is really, really embarrassing to see our community showing this kind of ignorance and then ignoring the people trying to help them understand because they don't like the results they're hearing.
The sample size is not supposed to be "3.5%" of the population. Polling 800 truly random voters in Maryland resulted in these results +-3.5%, and there is a ~95% chance the results are true within those limits. There's no bias in how these numbers were calculated; you plug them into a basic statistics equation and get results back.
Valid criticisms and questions are:
1. What were the questions? Were they loaded?
2. What was the polling method? Was it truly random?
I named elcid89 and IsabellsF "Macho Libre and Cinday Lauper" mojosfun is Capt. Lou Albano. It's the tag team match of the century against Capt. America and his sidekick 2A
One thing to take note of is this ...
OpinionWorks has totally changed their website overnight, removing their list of 'progressive clients' and making their business presentation a softer, gentler one. However, you can still find residule content on the internet Wayback Machine. GreenBook, an internet listing of market research firms, still has a partial client list.
They can run, but they can't hide.
Their corporate moto used to be "Finding Insights That Bring Results", now it's "Understand the attitudes and perceptions of the public. Know how to engage your audience." More like, manipulate the sheeple with slanted polling results and unsubstantiated data.
OpinionWorks and The Baltimore Sunpapers jumped the shark on this one.
Anyone else commenting should call this out loud and clear. Subscribers to the Sunpapers are being duped.
http://www.baltimoresun.com/news/maryland/bs-md-gun-poll-20130113,0,835835.story?track=rss
So... if the state gov plays out like this poll indicates, then we are in for some rough times.
My issue here is that the sample size is no where near the 3.5% percentage indicated by Tyeraxus post nor does 0.00013793% of a population (800 of 5.8mil) guarantee of confidence interval of 95%as implied.
it is not 3.5% sampling - it is a (if truly random and proper sampling) a rsult within a 3.5% error of the result from polling 100% of the population.
so if done correctly the poll result of 46% with a 3.5% error is really 42.5%-49.5%
this poll could be simply scewed by the fact that conservatives tend to not like answering polls, etc...
Try rewriting this into English that other people can understand, please. The method described should produce perfectly valid statistical results in the state of Maryland, as it has umpteen million times before.Your proposition is based on a perfect distributed sampling,(random as you refered to it), however not with standing other factors such bears an objective relationship to the sampleing percentage of the population.
What "assumption"? The article said it was 800 randomly-selected voters. I didn't contradict myself on anything. I said, here are a couple potentially valid criticisms of polls. Not, "this is a valid criticism of this particular poll".If you are going to continue to defend a contradictory study based on assumptions the the polled population was ideal, and then contradict yourself by calling into question if the sample group was ideal, then maybe you could kindly refrain from calling others ignorant.
Try rewriting this into English that other people can understand, please. The method described should produce perfectly valid statistical results in the state of Maryland, as it has umpteen million times before.
What "assumption"? The article said it was 800 randomly-selected voters. I didn't contradict myself on anything. I said, here are a couple potentially valid criticisms of polls. Not, "this is a valid criticism of this particular poll".
Do you have some sort of specific criticism about their polling method? Put up or shut up. Right now, your criticism of the poll appears to be wholly based on "I didn't like the outcome, so I am going to invent problems out of whole cloth."
I've had many conversations with elcid. I believe he works in some capacity for Gansler. He's avoided a direct answer,but I believe he's part of the problem. Go back and re-read his responses. They sometimes state "we're working on this or we're looking into that".Be very careful with him I believe he is dangerous to our cause!!!