RobertN
Legend
Posts: 35,485
Joined: Jan 2003
Reputation: 95
I Root For: THE NIU Huskies
Location: Wayne's World
|
RE: White House: "Silent Majorty" support President on Health Care
(09-15-2009 01:14 PM)Rebel Wrote: (09-15-2009 01:08 PM)RobertN Wrote:
1. Without knowing why it is not allowed now(there could be a very good reason), I think it should be.
Why don't you tell us?
Quote:2. He need INSURANCE because if he ***** up, he will be sued. If some doctor ****** up on one of your relatives(hope it doesn't happen), do you want someone to tell you how much your relative is worth. THat is what you are asking for-"rationing". I thought you didn't like rationing?
No way in HELL should they need as much as they carry. There's a reason doctors are leaving in droves and heading to states like Florida with Tort reform laws. You'll have them move to other countries.
Quote:3. THere are so many because it is for your safety. There will be more to come.
Hey buddy, NO F'N THANK YOU. I don't NEED the FEDERAL government acting as my f'n nanny. I'm not a lib.
Quote:4. I am for competition but what you want will be like the wild west.
No, I want what has made this country the greatest, most prosperous, and most envied country on the face of the Earth, Free-Market Capitalism. Doctors are swimming to f'n Cuba to treat patients. They're coming here.
1. If there was a reason that I heard of, I would tell you. Most things have a reason. I just don't know what it is. 2. Could be the tort reforms or it could be that their is a large population down there. As I said, you don't want anyone getting in between you and your doctor and complain that with a government plan you can't sue the insurance comppany. Your opinions just seem all over the place and not consistent. 3. So we should take away government food safety, drug safety, product safety too based on your opinion. You are starting to sound more and more like Fo all the time. 4. I am not sure what you say you want ansd was there intent was actually the founders intent(wild west free for all).
|
|
09-15-2009 01:34 PM |
|
I45owl
Hall of Famer
Posts: 18,374
Joined: Jun 2005
Reputation: 184
I Root For: Rice Owls
Location: Dallas, TX
|
RE: White House: "Silent Majorty" support President on Health Care
(09-15-2009 12:31 PM)Rebel Wrote: (09-15-2009 11:31 AM)I45owl Wrote: When polls sample substantially more than that, they only do so because of political objections about the sample size - they don't do it to increase accuracy. I doubt that you would find the results to be more than 3% different if you sampled 100,000. What could sway the results substantially are changing how the questions are phrased or asking the same question +/- one day. I take Mach's side with regard to the statistics.
So you're saying all dynamics, throughout the country, are essentially exactly the same? That the South polls just as the North and West? That people in North Dakota poll the same way people poll in West Virginia?
I'm saying that whatever the dynamics are in a group (in this case, the 320 million Americans), taking a random sample of sufficient size should fairly represent those dynamics (as long as there is no systemic bias of the sample, such as polling 500 North Dakotans and 500 West Virginians). And, I believe that 1000 is a sufficient sample size.
(09-15-2009 01:31 PM)Claw Wrote: When your poll results start nearing a 50/50 split, the 3% margin of error becomes significant. At that point an increase of sample size is recommended.
The wording of the questions is more potent than the mathematics. I think it is Peter Yen's "The Art of Asking Questions" that used to be the recommended study. I'm sure that's changed over the years.
You have to address the question: Why does the proportion matter more at 50/50 than it does at 10/90? If you are not talking about opinion polls, there's no statistical basis for that rule, although it does decrease the statistical margin of error. If you are talking about opinion polls, then the reason for the increase in sample size is primarily CYA - the statistical margin of error is almost certainly far less than other variables like the phrasing of the question or what the lead item on the news is.
|
|
09-15-2009 02:21 PM |
|
WoodlandsOwl
Up in the Woods
Posts: 11,813
Joined: Jun 2005
Reputation: 115
I Root For: Rice Owls
Location:
|
RE: White House: "Silent Majorty" support President on Health Care
(09-15-2009 01:00 PM)Rebel Wrote: (09-15-2009 12:54 PM)RobertN Wrote: Is "You liar" the next man on the right that will beat the Democrats in 2012?
He said, "You Lie" not "You Liar".
I bought a couple of T-Shirts.
http://www.madeinmultiples.com/imwithjoe.aspx
|
|
09-15-2009 02:49 PM |
|
Tripster
Most Dangerous Man on a Keyboard
Posts: 3,140
Joined: Nov 2008
Reputation: 16
I Root For: The Best Only
Location: Where the Action is
|
RE: White House: "Silent Majorty" support President on Health Care
(09-15-2009 12:54 PM)RobertN Wrote: Is "You liar" the next man on the right that will beat the Democrats in 2012?
Why not, he is the only man in the entire Chamber that TOLD THE TRUTH THAT NIGHT !!!!!!!!!!
Those ARE the Facts, AND they are Undisputed !!!!!!!!!
.
|
|
09-15-2009 06:57 PM |
|
Claw
Hall of Famer
Posts: 24,898
Joined: Feb 2004
Reputation: 1225
I Root For: Memphis
Location: Orangeville HELP!
|
RE: White House: "Silent Majorty" support President on Health Care
(09-15-2009 02:21 PM)I45owl Wrote: (09-15-2009 12:31 PM)Rebel Wrote: (09-15-2009 11:31 AM)I45owl Wrote: When polls sample substantially more than that, they only do so because of political objections about the sample size - they don't do it to increase accuracy. I doubt that you would find the results to be more than 3% different if you sampled 100,000. What could sway the results substantially are changing how the questions are phrased or asking the same question +/- one day. I take Mach's side with regard to the statistics.
So you're saying all dynamics, throughout the country, are essentially exactly the same? That the South polls just as the North and West? That people in North Dakota poll the same way people poll in West Virginia?
I'm saying that whatever the dynamics are in a group (in this case, the 320 million Americans), taking a random sample of sufficient size should fairly represent those dynamics (as long as there is no systemic bias of the sample, such as polling 500 North Dakotans and 500 West Virginians). And, I believe that 1000 is a sufficient sample size.
(09-15-2009 01:31 PM)Claw Wrote: When your poll results start nearing a 50/50 split, the 3% margin of error becomes significant. At that point an increase of sample size is recommended.
The wording of the questions is more potent than the mathematics. I think it is Peter Yen's "The Art of Asking Questions" that used to be the recommended study. I'm sure that's changed over the years.
You have to address the question: Why does the proportion matter more at 50/50 than it does at 10/90? If you are not talking about opinion polls, there's no statistical basis for that rule, although it does decrease the statistical margin of error. If you are talking about opinion polls, then the reason for the increase in sample size is primarily CYA - the statistical margin of error is almost certainly far less than other variables like the phrasing of the question or what the lead item on the news is.
If you have a 5% spread, then a plus/minus 3% error - which is actually 6%- is larger than your finding. The only way I know to reduce that margin of error is to increase the sample size significantly - but it's been 25 years since I studied this stuff.
|
|
09-15-2009 08:08 PM |
|
I45owl
Hall of Famer
Posts: 18,374
Joined: Jun 2005
Reputation: 184
I Root For: Rice Owls
Location: Dallas, TX
|
RE: White House: "Silent Majorty" support President on Health Care
(09-15-2009 08:08 PM)Claw Wrote: If you have a 5% spread, then a plus/minus 3% error - which is actually 6%- is larger than your finding. The only way I know to reduce that margin of error is to increase the sample size significantly - but it's been 25 years since I studied this stuff.
I guess in the sciences you're taught the difference between accuracy and precision (I assume your background in statistics is from journalism or politics). Accuracy being how close to the correct answer you are, and Precision being how close your results correlate to one another. You can treat the margin of error in a poll as precision and the accuracy as how well it represents reality when all is said and done.
You'll get better precision by increasing the sample size, but the accuracy won't be improved. The problem with polls is that the accuracy is affected to a large degree by how the questions are phrased. If you're question is "are you going to vote for McCain or Obama", then there's not much way to sway that (except randomly reversing the order of the ballot), so increasing precision makes sense. If your question is "do you like this (theoretical) health plan?" your accuracy can't be assumed to be better than +/- 15-20%, so increasing precision doesn't matter.
|
|
09-16-2009 08:57 AM |
|