|
|
Thread Tools | Display Modes |
#101
|
|||
|
|||
|
Ads |
#102
|
|||
|
|||
"Edward Dike, III" writes:
Actually, I wrote nothing in this post. Please me a bit more careful with your attributions as you cut, paste to suit your needs. ED3 "Tim McNamara" wrote in message ... | "Edward Dike, III" writes: | | | I also noticed that the document you refer to shows the percentage | | of "registered" voters who actually voted. According to | | | | http://www.census.gov/prod/2002pubs/p20-542.pdf | | | | the percentage of "registered" voters in the U.S. who actually | | voted in the 2000 election is about 85%. I couldn't find a figure | | for this election. Oops, sorry, Ed. |
#103
|
|||
|
|||
"Edward Dike, III" writes:
Actually, I wrote nothing in this post. Please me a bit more careful with your attributions as you cut, paste to suit your needs. ED3 "Tim McNamara" wrote in message ... | "Edward Dike, III" writes: | | | I also noticed that the document you refer to shows the percentage | | of "registered" voters who actually voted. According to | | | | http://www.census.gov/prod/2002pubs/p20-542.pdf | | | | the percentage of "registered" voters in the U.S. who actually | | voted in the 2000 election is about 85%. I couldn't find a figure | | for this election. Oops, sorry, Ed. |
#104
|
|||
|
|||
Mark Hickey wrote in message . ..
David Kerber wrote: In article , says... Here's a question for someone with a statistics background: Given a sample that predicts the behavior of a population within 3% of the true value with a 95% degree of confidence, what is the probability that the true value will in fact turn out to be 8% from the value predicted by the sample? I believe it's about 1%. That's right. I don't think so. There were individual polls, taken the same day, same state, both with 3% margins of error that were 8% apart - proving nothing but the fallibility of polls. You're confusing exit polls and pre-election polls. The pre-election polls are subject to errors in assumptions (whether someone will actually vote, what percentages of people represented by the sample are actually present in the sampled population, etc.); the exit polls are polls of people leaving the polling place, so they are not subject to nearly the uncertainty. Two pre-election polls can correctly measure their populations within a 3% MOE but give very different predictions because in reality they are measuring very different populations. The problem is not with the polls, it is with the pollsters assuming that they are measuring something that they are not. Exit polls are not subject to pollsters' errors to such a great degree. "Fallibility of polls" has got to be one of the stupedest things I've ever heard, but for people who don't have any desire for reality-based analysis, I'm sure it works. Me, I'd like to see someone explain how Bush conveniently beat worse than 1000-one odds in both Ohio and Florida to win "re-election", even as Karen Hughes was telling him he had lost. JP |
#105
|
|||
|
|||
Mark Hickey wrote in message . ..
David Kerber wrote: In article , says... Here's a question for someone with a statistics background: Given a sample that predicts the behavior of a population within 3% of the true value with a 95% degree of confidence, what is the probability that the true value will in fact turn out to be 8% from the value predicted by the sample? I believe it's about 1%. That's right. I don't think so. There were individual polls, taken the same day, same state, both with 3% margins of error that were 8% apart - proving nothing but the fallibility of polls. You're confusing exit polls and pre-election polls. The pre-election polls are subject to errors in assumptions (whether someone will actually vote, what percentages of people represented by the sample are actually present in the sampled population, etc.); the exit polls are polls of people leaving the polling place, so they are not subject to nearly the uncertainty. Two pre-election polls can correctly measure their populations within a 3% MOE but give very different predictions because in reality they are measuring very different populations. The problem is not with the polls, it is with the pollsters assuming that they are measuring something that they are not. Exit polls are not subject to pollsters' errors to such a great degree. "Fallibility of polls" has got to be one of the stupedest things I've ever heard, but for people who don't have any desire for reality-based analysis, I'm sure it works. Me, I'd like to see someone explain how Bush conveniently beat worse than 1000-one odds in both Ohio and Florida to win "re-election", even as Karen Hughes was telling him he had lost. JP |
#106
|
|||
|
|||
It is disingenuous to call this the "bigget turnout" because although the
most voters turned out many previous elections had a higher percentage turnout (still only a bit over 59%). It is also technically accurate to say that this election saw the largest vote ever AGAINST anyone running for president - including all incumbents. (One could also say that more people voted for Bush than the entire population of the planet 5,000 years ago...) -- Steve Juniper Berkeley, CA "Fascism should more properly be called corporatism, since it is the merger of state and corporate power." -- Benito Mussolini "Tom Kunich" wrote in message link.net... "Tom Sherman" wrote in message ... Or maybe both the candidates with a chance of winning did nothing to inspire people to vote. I guess that's why this was the biggest turnout in election history. |
#107
|
|||
|
|||
It is disingenuous to call this the "bigget turnout" because although the
most voters turned out many previous elections had a higher percentage turnout (still only a bit over 59%). It is also technically accurate to say that this election saw the largest vote ever AGAINST anyone running for president - including all incumbents. (One could also say that more people voted for Bush than the entire population of the planet 5,000 years ago...) -- Steve Juniper Berkeley, CA "Fascism should more properly be called corporatism, since it is the merger of state and corporate power." -- Benito Mussolini "Tom Kunich" wrote in message link.net... "Tom Sherman" wrote in message ... Or maybe both the candidates with a chance of winning did nothing to inspire people to vote. I guess that's why this was the biggest turnout in election history. |
#108
|
|||
|
|||
JP wrote:
Mark Hickey wrote in message . .. David Kerber wrote: In article , says... Here's a question for someone with a statistics background: Given a sample that predicts the behavior of a population within 3% of the true value with a 95% degree of confidence, what is the probability that the true value will in fact turn out to be 8% from the value predicted by the sample? I believe it's about 1%. That's right. I don't think so. There were individual polls, taken the same day, same state, both with 3% margins of error that were 8% apart - proving nothing but the fallibility of polls. You're confusing exit polls and pre-election polls. The pre-election polls are subject to errors in assumptions (whether someone will actually vote, what percentages of people represented by the sample are actually present in the sampled population, etc.); the exit polls are polls of people leaving the polling place, so they are not subject to nearly the uncertainty. Two pre-election polls can correctly measure their populations within a 3% MOE but give very different predictions because in reality they are measuring very different populations. The problem is not with the polls, it is with the pollsters assuming that they are measuring something that they are not. Exit polls are not subject to pollsters' errors to such a great degree. "Fallibility of polls" has got to be one of the stupedest things I've ever heard, but for people who don't have any desire for reality-based analysis, I'm sure it works. Me, I'd like to see someone explain how Bush conveniently beat worse than 1000-one odds in both Ohio and Florida to win "re-election", even as Karen Hughes was telling him he had lost. I believe Diebold CEO and Republican Walden O'Dell might have some answers to that. -- Tom Sherman |
#109
|
|||
|
|||
JP wrote:
Mark Hickey wrote in message . .. David Kerber wrote: In article , says... Here's a question for someone with a statistics background: Given a sample that predicts the behavior of a population within 3% of the true value with a 95% degree of confidence, what is the probability that the true value will in fact turn out to be 8% from the value predicted by the sample? I believe it's about 1%. That's right. I don't think so. There were individual polls, taken the same day, same state, both with 3% margins of error that were 8% apart - proving nothing but the fallibility of polls. You're confusing exit polls and pre-election polls. The pre-election polls are subject to errors in assumptions (whether someone will actually vote, what percentages of people represented by the sample are actually present in the sampled population, etc.); the exit polls are polls of people leaving the polling place, so they are not subject to nearly the uncertainty. Two pre-election polls can correctly measure their populations within a 3% MOE but give very different predictions because in reality they are measuring very different populations. The problem is not with the polls, it is with the pollsters assuming that they are measuring something that they are not. Exit polls are not subject to pollsters' errors to such a great degree. "Fallibility of polls" has got to be one of the stupedest things I've ever heard, but for people who don't have any desire for reality-based analysis, I'm sure it works. Me, I'd like to see someone explain how Bush conveniently beat worse than 1000-one odds in both Ohio and Florida to win "re-election", even as Karen Hughes was telling him he had lost. I believe Diebold CEO and Republican Walden O'Dell might have some answers to that. -- Tom Sherman |
#110
|
|||
|
|||
(JP) wrote:
Mark Hickey wrote in message . .. David Kerber wrote: In article , says... Here's a question for someone with a statistics background: Given a sample that predicts the behavior of a population within 3% of the true value with a 95% degree of confidence, what is the probability that the true value will in fact turn out to be 8% from the value predicted by the sample? I believe it's about 1%. That's right. I don't think so. I do... and so do these guys... http://www.realclearpolitics.com/Pre...chart3way.html There were individual polls, taken the same day, same state, both with 3% margins of error that were 8% apart - proving nothing but the fallibility of polls. You're confusing exit polls and pre-election polls. The pre-election polls are subject to errors in assumptions (whether someone will actually vote, what percentages of people represented by the sample are actually present in the sampled population, etc.); the exit polls are polls of people leaving the polling place, so they are not subject to nearly the uncertainty. They're subject to a very significant amount of uncertainty. As pointed out previously, the tendency for Republicans to participate in exit polls is considerably less than Democrats. Also, the firm doing the poll insists that the early hour polls showing Kerry ahead weren't representative of the final polls, and should NOT have been released as they were (by themselves) meaningless. Two pre-election polls can correctly measure their populations within a 3% MOE but give very different predictions because in reality they are measuring very different populations. They measure the same populations, actually. Either ~1,000 likely voters or ~1,000 registered voters. The problem is not with the polls, it is with the pollsters assuming that they are measuring something that they are not. Exit polls are not subject to pollsters' errors to such a great degree. It does eliminate the portion of the population that doesn't vote - but is hardly fool-proof. "Fallibility of polls" has got to be one of the stupedest things I've ever heard, but for people who don't have any desire for reality-based analysis, I'm sure it works. Duh. Check this assembly of poll statistics for Arizona done at about the same time: http://www.realclearpolitics.com/Pre.../az_polls.html Do you really think the voters were wavering by +- 10%, or do you suppose that it's just possible that the polls are indeed fallible. I think history has proven it. Me, I'd like to see someone explain how Bush conveniently beat worse than 1000-one odds in both Ohio and Florida to win "re-election", even as Karen Hughes was telling him he had lost. He didn't beat 1000 to 1 odds - check out the pre-election polls in the states and try (as much as Michael Moore will suggest you don't) to ignore the fact that you're operating under a partial, and flawed assumption. http://www.realclearpolitics.com/Pre.../fl_polls.html FL, Bush +0.6% http://www.realclearpolitics.com/Pre.../oh_polls.html OH, Bush +2.1% Both represent the average of the last polling done in both states before the election. Funny that the actual results look pretty much the same, huh? Must have been a Karl Rove conspiracy. Heh. Mark Hickey Habanero Cycles http://www.habcycles.com Home of the $695 ti frame |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Stupid Americans! -- Stupid... Stupid... STUPID!!! ___________ ylojceq | Tom Kunich | Rides | 4 | November 10th 04 04:26 AM |