Note from Hot Air management: This section is for comments from Hot Air's community of registered readers. Registration is currently closed. That means if you're not already registered, you can't comment. We will let you know if and when registration re-opens. Please don't assume that Hot Air management agrees with or otherwise endorses any particular comment just because we let it stand. A reminder: Anyone who fails to comply with our terms of use may lose their posting privilege.
Murtha is opening his mouth and contributing all he can.
meci on October 23, 2008 at 8:16 AM
How is a poll of 400 voters in PA’s 12th district a 4.9% margin of error, and national polls of a few thousand are 3%. Any statisticians around who can help explain?
Anyway, Murtha could be in trouble, and Obama is already in it deep in the 12th district and the rest of western PA. He’s gonna get mauled in Greene county (12th) and a bunch of surrounding areas where Hillary hammered BO in the primaries.
I think Murtha’s “racist” and “redneck” comments were a botched effort to lower some expectations and maybe motivate his own supporters to not assume it’s in the bag like the major media have been trying to say.
forest on October 23, 2008 at 8:21 AM
Is it worth it to have Obama win if Murtha loses? My mind says no, but he heart says yes.
BadgerHawk on October 23, 2008 at 8:24 AM
BadgerHawk on October 23, 2008 at 8:24 AM
I’m kinda hoping we’ll get a two-for. Let’s hope Murtha’s weakness will hurt Obama’s performance in the 12th.
forest on October 23, 2008 at 8:26 AM
It is beyond me how anyone, anyone can vote for Murtha. I find him so repulsive and disgusting that it would be like going into the booth and pulling the lever for Idi Amin.
We need to send two veterans to Washington: McCain and Russell.
We need to defeat the leftists that are ruining this country for the past two years.
jencab on October 23, 2008 at 8:34 AM
How is a poll of 400 voters in PA’s 12th district a 4.9% margin of error, and national polls of a few thousand are 3%. Any statisticians around who can help explain?
forest on October 23, 2008 at 8:21 AM
I don’t know. You would think that since PA’s 12th district is a lot smaller of a sample than the U.S. that it would be a smaller margin of error. However as someone who has taken statistics, there are a lot of factors that go into margin of error. Maybe they had to do something really drastic to actually give Murtha a decent looking lead, but it also gave the poll a greater margin of error.
Lets get this guy in and Murtha out. I live in Ill-noise and it doesn’t matter. I want Murtha out.
BobK on October 23, 2008 at 8:45 AM
They better get it right or lots of those “racist rednecks” are going to find themselves in an Obama re-education camp.
anniekc on October 23, 2008 at 8:49 AM
How is a poll of 400 voters in PA’s 12th district a 4.9% margin of error, and national polls of a few thousand are 3%. Any statisticians around who can help explain?
The margin of error of a poll is approximated by the square root of (.25/(samplesize)) x 2.
Emphasize approximated. A random sample is presumed to come from an infinite population, but as you say, a congressional district is a lot less infinite than a national population.
There are mathematical corrections to reflect the relative proportion of sample to population, but they tend not to change the outcome too much.
Now that you know sampling errors were computed fair and square here, remember that there are 2 sources of fluctuation in polling. Sampling error is one. The other is selection bias. Knowing that a truly random sample is nearly impossible, some polling firms use special “weighting” coefficients which give certain demographics more emphasis. Or the polling firm can decide that the mix of democrat to republican in the sample is not representative of expected voting patters, so they’ll weight one or the other more heavily.
The upshot is, polling rarely achieves a random sample, for all kinds of reasons. So the whole idea of the margin of error is pretty much moot. The best criterion for judging a poll is to know who paid for it.
jeff_from_mpls on October 23, 2008 at 8:56 AM
The best criterion for judging a poll is to know who paid for it.
jeff_from_mpls on October 23, 2008 at 8:56 AM
Haha, true. In this case, a conservative-leaning paper appears to have paid for it, but I suspect it’s more accurate than the liberal polls that are saying Murtha has a 20 point lead.
forest on October 23, 2008 at 9:08 AM
Russell needs donations fast!
I have concerns about Russell’s connections to the Religious Right, but he won’t be my Congressman and I absolutely despise Murtha. I just donated to Russell’s and encourage everyone who cares about the military to do likewise. Let’s drive that SOB Murtha out of DC and send him back to “racist” PA!!!
Defeating the traitor will be a nice little consolation prize for us, should Alinsky-Ayers win.
jeff_from_mpls on October 23, 2008 at 8:14 AM
That would be nice, wouldn’t it? The guy insults and prejudges our Marines, and his own constituents. A Repub with the same arrogant mouth would find himself already conceding.
Blowback
Note from Hot Air management: This section is for comments from Hot Air's community of registered readers. Registration is currently closed. That means if you're not already registered, you can't comment. We will let you know if and when registration re-opens. Please don't assume that Hot Air management agrees with or otherwise endorses any particular comment just because we let it stand. A reminder: Anyone who fails to comply with our terms of use may lose their posting privilege.
Trackbacks/Pings
Trackback URL
Comments
Defeating the traitor will be a nice little consolation prize for us, should Alinsky-Ayers win.
jeff_from_mpls on October 23, 2008 at 8:14 AM
Russell needs donations fast!
JellyToast on October 23, 2008 at 8:14 AM
Murtha is opening his mouth and contributing all he can.
meci on October 23, 2008 at 8:16 AM
How is a poll of 400 voters in PA’s 12th district a 4.9% margin of error, and national polls of a few thousand are 3%. Any statisticians around who can help explain?
Anyway, Murtha could be in trouble, and Obama is already in it deep in the 12th district and the rest of western PA. He’s gonna get mauled in Greene county (12th) and a bunch of surrounding areas where Hillary hammered BO in the primaries.
I think Murtha’s “racist” and “redneck” comments were a botched effort to lower some expectations and maybe motivate his own supporters to not assume it’s in the bag like the major media have been trying to say.
forest on October 23, 2008 at 8:21 AM
Is it worth it to have Obama win if Murtha loses? My mind says no, but he heart says yes.
BadgerHawk on October 23, 2008 at 8:24 AM
I’m kinda hoping we’ll get a two-for. Let’s hope Murtha’s weakness will hurt Obama’s performance in the 12th.
forest on October 23, 2008 at 8:26 AM
It is beyond me how anyone, anyone can vote for Murtha. I find him so repulsive and disgusting that it would be like going into the booth and pulling the lever for Idi Amin.
Rod on October 23, 2008 at 8:29 AM
We need to send two veterans to Washington: McCain and Russell.
We need to defeat the leftists that are ruining this country for the past two years.
jencab on October 23, 2008 at 8:34 AM
I don’t know. You would think that since PA’s 12th district is a lot smaller of a sample than the U.S. that it would be a smaller margin of error. However as someone who has taken statistics, there are a lot of factors that go into margin of error. Maybe they had to do something really drastic to actually give Murtha a decent looking lead, but it also gave the poll a greater margin of error.
MobileVideoEngineer on October 23, 2008 at 8:41 AM
You can find his donate button here.
http://russellbrigade.com/
Lets get this guy in and Murtha out. I live in Ill-noise and it doesn’t matter. I want Murtha out.
BobK on October 23, 2008 at 8:45 AM
They better get it right or lots of those “racist rednecks” are going to find themselves in an Obama re-education camp.
anniekc on October 23, 2008 at 8:49 AM
How is a poll of 400 voters in PA’s 12th district a 4.9% margin of error, and national polls of a few thousand are 3%. Any statisticians around who can help explain?
The margin of error of a poll is approximated by the square root of (.25/(samplesize)) x 2.
Emphasize approximated. A random sample is presumed to come from an infinite population, but as you say, a congressional district is a lot less infinite than a national population.
There are mathematical corrections to reflect the relative proportion of sample to population, but they tend not to change the outcome too much.
Now that you know sampling errors were computed fair and square here, remember that there are 2 sources of fluctuation in polling. Sampling error is one. The other is selection bias. Knowing that a truly random sample is nearly impossible, some polling firms use special “weighting” coefficients which give certain demographics more emphasis. Or the polling firm can decide that the mix of democrat to republican in the sample is not representative of expected voting patters, so they’ll weight one or the other more heavily.
The upshot is, polling rarely achieves a random sample, for all kinds of reasons. So the whole idea of the margin of error is pretty much moot. The best criterion for judging a poll is to know who paid for it.
jeff_from_mpls on October 23, 2008 at 8:56 AM
Haha, true. In this case, a conservative-leaning paper appears to have paid for it, but I suspect it’s more accurate than the liberal polls that are saying Murtha has a 20 point lead.
forest on October 23, 2008 at 9:08 AM
I have concerns about Russell’s connections to the Religious Right, but he won’t be my Congressman and I absolutely despise Murtha. I just donated to Russell’s and encourage everyone who cares about the military to do likewise. Let’s drive that SOB Murtha out of DC and send him back to “racist” PA!!!
JohnAGJ on October 23, 2008 at 9:11 AM
That would be nice, wouldn’t it? The guy insults and prejudges our Marines, and his own constituents. A Repub with the same arrogant mouth would find himself already conceding.
Bigfoot on October 23, 2008 at 9:59 AM
Can we possibly get so lucky?
promachus on October 23, 2008 at 10:19 AM