View Full Version : Reverend Wright May have Ended Obama
Lysander
03-20-2008, 05:42 PM
Well, the Blacks almost had it...sooo close but no cigar.
Clinton takes lead over Obama in Gallup poll Thu Mar 20, 10:11 AM ET
WASHINGTON (Reuters) - Democratic presidential candidate Hillary Clinton has moved into a significant lead over Barack Obama among Democratic voters, according to a new Gallup poll.
ADVERTISEMENT
The March 14-18 national survey of 1,209 Democratic and Democratic-leaning voters gave Clinton, a New York senator, a 49 percent to 42 percent edge over Obama, an Illinois senator. The poll has an error margin of 3 percentage points.
The poll was a snapshot of current popular feeling, but Clinton trails Obama in the state-by-state contest which began in January to select a nominee to face presumptive Republican nominee John McCain in the November election to succeed President George W. Bush.
The nominees are formally chosen by delegates at the parties' conventions in the summer.
Gallup said the poll lead was the first statistically significant one for Clinton since a tracking poll conducted February 7-9, just after the Super Tuesday primaries. The two candidates had largely been locked in a statistical tie since then, with Obama last holding a lead over Clinton in a March 11-13 poll.
Gallup said polling data also showed McCain leading Obama 47 percent to 43 percent in 4,367 registered voters' preferences for the general election. The general election survey has an error margin of 2 percentage points.
The Arizona senator also edged Clinton 48 percent to 45 percent but Gallup said the lead was not statistically significant.
(Reporting by David Morgan, editing by Vicki Allen)
To be fair though, it's still not over and Obama still leads in the delegate count. But if this is a start of a downward spiral, a crazy reverend is to blame.
Stretch
03-20-2008, 05:43 PM
Sample size of 1,209?
Really?
thefarmer
03-20-2008, 05:44 PM
Well, the Blacks almost had it...sooo close but no cigar.
That's a great turn of phrase.
Lysander
03-20-2008, 05:44 PM
Sample size of 1,209?
Really?
Yeah, that is kinda fishy. But my point is, if the Blacks will all just STFU it will be BEST. Brother Obama's got this made, just be cool and let him work his magic.
And this matters how? They are going to go back and re ballot all the states that have had primaries already.
Parkbandit
03-20-2008, 05:55 PM
Polls change like the weather.
Only small minds assume that all black people are voting for Obama.
Crazy Bard
03-20-2008, 06:27 PM
In the great words of P Diddy; We aint, go-in nowhere, we aint, goin nowhere
We cant be stopped now, cause its obama for 08!
Only small minds assume that all black people are voting for Obama.
:)
Music to my ears.
875000
03-20-2008, 06:46 PM
In the great words of P Diddy; We aint, go-in nowhere, we aint, goin nowhere
P. Diddy was a punk-assed bitch who tried to have Tupac killed.
http://www.latimes.com/news/nationworld/nation/la-naw-quad17mar17,0,4451053.story
Suppa Hobbit Mage
03-20-2008, 06:52 PM
Only small minds assume that all black people are voting for Obama.
Yeah I agree. I actually like Obama over Hillary though.
Valthissa
03-20-2008, 07:46 PM
Sample size of 1,209?
Really?
Why do people always question the sample size?
Remember the central limit theorem.
A sample size of 1000 gives a margin of error of 3%, which is why when a news source says the poll has a margin of error of 3% all they are really telling you is that the sample is greater than a 1000 and less than 5000 (at least I think 5000 is the next increment).
The difficulties associated with polling (and they are so many as to make this type of poll worthless in my opinion) lie in asking neutral questions, obtaining a random sample, etc., but not in the basic calculations of confidence intervals.
C/Valth
Lysander
03-20-2008, 08:41 PM
Yes my dear but here is a mind bender for you. Say the total population this survey is trying to ascertain is 1001. Then compare to how the survey is trying to ascertain 300,000,000. Do you think the survey will be more accurate for the 1001 or the 300,000,000? Think carefully now, all but 1 person was asked their opinion. What do you think?
875000
03-20-2008, 08:59 PM
Yes my dear but here is a mind bender for you. Say the total population this survey is trying to ascertain is 1001. Then compare to how the survey is trying to ascertain 300,000,000. Do you think the survey will be more accurate for the 1001 or the 300,000,000? Think carefully now, all but 1 person was asked their opinion. What do you think?
The closer the sample size is the a population, the greater the probability that it is correct assuming that the same correct statistical methodologies were used.
HOWEVER, that does not mean a smaller sample size is statistically invalid. You can run a statistically valid test with 1001 people and still produce results that have an extremely high probability of being accurate. The important thing is that the given sample size passes a certain threshhold.
In Layman's terms: Valthissa is right.
If you look at the sample size of almost all polls they are around 1000. Like every politics poll ever.
Xaerve
03-20-2008, 09:24 PM
I love seeing these threads on the various message boards.
"FATALISTIC MESSAGE BECAUSE OF NEW POLL"
Sample Size: 1,200.
Stretch
03-20-2008, 09:39 PM
Why do people always question the sample size?
Remember the central limit theorem.
A sample size of 1000 gives a margin of error of 3%, which is why when a news source says the poll has a margin of error of 3% all they are really telling you is that the sample is greater than a 1000 and less than 5000 (at least I think 5000 is the next increment).
The difficulties associated with polling (and they are so many as to make this type of poll worthless in my opinion) lie in asking neutral questions, obtaining a random sample, etc., but not in the basic calculations of confidence intervals.
C/Valth
Forget statistics for a minute. I still don't think 1,200 is a large enough sample size to control for all of the different factors going into the population (i.e. gender, race, income level). This is also the biggest problem I have with consumer surveys and shit that vendors try to hawk to the company I work at.
On the surface, you have some 200 MM (pulling a number out of my ass here) people of voting age. You ask 1,200 randomly selected people who they prefer. Then you ask another 1,200 randomly selected people who they prefer. You don't think there's a good chance that you're going to get different results with so few data points?
Warriorbird
03-20-2008, 09:40 PM
Obama... like Clinton... does his best when attacked. If Clinton had run a flat campaign against him I think she'd have sewed it up much sooner.
Parkbandit
03-20-2008, 09:47 PM
Obama... like Clinton... does his best when attacked. If Clinton had run a flat campaign against him I think she'd have sewed it up much sooner.
I disagree. Obama is one of the best speakers I've ever listened to.. he's engaging, he exudes personality and you really are almost forced into liking him. THIS is his greatest asset... not his ability to throw mud back at his opponent.
Valthissa
03-20-2008, 10:23 PM
Forget statistics for a minute. I still don't think 1,200 is a large enough sample size to control for all of the different factors going into the population (i.e. gender, race, income level). This is also the biggest problem I have with consumer surveys and shit that vendors try to hawk to the company I work at.
On the surface, you have some 200 MM (pulling a number out of my ass here) people of voting age. You ask 1,200 randomly selected people who they prefer. Then you ask another 1,200 randomly selected people who they prefer. You don't think there's a good chance that you're going to get different results with so few data points?
You can't forget statistics when you ask a question about random sampling. 1,200 data points yields a margin of error of 3%, period. It's the other variables - the wording of the questions, whether the sample is truly random, and others that lead to inaccuracies in polls. You might think that 1,200 points of data is not enough to accurately predict (accurate defined as the confidence interval selected for the poll) the result of 200 million people's vote. You would be wrong. It's not a matter of belief or opinion. You might as well say you disagree with the theory of quantum mechanics. At some point Kranar will rescue me - I've posted before that I'm not 5% the mathematician I was at 20.
C/Valth
Warriorbird
03-20-2008, 10:34 PM
Attacks give Obama the chance to give speeches. If the attacks weren't there I'm not sure the speeches would mean as much.
Stanley Burrell
03-20-2008, 10:37 PM
You can't forget statistics when you ask a question about random sampling. 1,200 data points yields a margin of error of 3%, period. It's the other variables - the wording of the questions, whether the sample is truly random, and others that lead to inaccuracies in polls. You might think that 1,200 points of data is not enough to accurately predict (accurate defined as the confidence interval selected for the poll) the result of 200 million people's vote. You would be wrong. It's not a matter of belief or opinion. You might as well say you disagree with the theory of quantum mechanics. At some point Kranar will rescue me - I've posted before that I'm not 5% the mathematician I was at 20.
C/Valth
3% margin of error is called negative data. Which is data, nonetheless.
http://www.fortunecity.com/tatooine/silverberg/484/data4.jpg
See?!
By the way, I've taken and easily sat-on-my-ass-passed Quantum Theory... And wholly withdrawn from actual Quantum Mechanics. What = Theory of Quantum Mechanics? If I have to postulate on what color atoms are, again, I shall cannon ball into the nearest atomic smasher for great justice.
Obama... like Clinton... does his best when attacked. If Clinton had run a flat campaign against him I think she'd have sewed it up much sooner.
I disagree. Obama is one of the best speakers I've ever listened to.. he's engaging, he exudes personality and you really are almost forced into liking him. THIS is his greatest asset... not his ability to throw mud back at his opponent.
Actually, below is why you should agree. Playing the victim has served Obama well thus far. If Obama's smart, he'll respond to every serious allegation with a speech. It would be Hillary's worst nightmare. He'll I'm impressed with Obama when he talks and I'm a white conservative (quasi republican) Texan.
Attacks give Obama the chance to give speeches. If the attacks weren't there I'm not sure the speeches would mean as much.
Methais
03-20-2008, 11:08 PM
I disagree. Obama is one of the best speakers I've ever listened to.. he's engaging, he exudes personality and you really are almost forced into liking him. THIS is his greatest asset... not his ability to throw mud back at his opponent.
He is THE MOST ELECTRIFYING MAN IN POLITICS!!!!!!!!!
Barack says KNOW YOUR ROLE AND SHUT YOUR MOUTH!!!!!!!!!!!!!11
Barack says he's gonna go to the Clintons' house, take the cigar Bill used on Monica, shine it up real nice, turn that sumbitch sideways AND STICK IT STRAIGHT UP HILLARY'S CANDY ASS!!!
IF YA SMEEEELLLLLLLLLLLLLLLLLALALALALALAEEEELLLLL!!!!!!! WHAT BARACK.....
....is cookin'.
http://totallyclueless.iwarp.com/images/therock1.jpghttp://08election.stickit2.us/wp-content/uploads/2007/09/barack-obama-1.jpg
EDIT: The Rock would be the perfect running mate for this guy imo.
Parkbandit
03-20-2008, 11:27 PM
Actually, below is why you should agree. Playing the victim has served Obama well thus far. If Obama's smart, he'll respond to every serious allegation with a speech. It would be Hillary's worst nightmare. He'll I'm impressed with Obama when he talks and I'm a white conservative (quasi republican) Texan.
One of the best speeches I've heard from him was after he won the Iowa caucus (I think it was anyway.. ) It was one of the most awe inspiring and motivational speeches by Obama to date.. and it had zero to do with responding to any attacks.
Whimsi
03-21-2008, 12:29 AM
If the attacks weren't there I'm not sure the speeches would mean as much.
You must be joshing. If the speeches weren't in response to attacks they'd have been even MORE impressive. Then noone could accuse him of only speaking out because he was on the defensive.
I really believe some of his speeches, whether or not he wins the nomination, will be quoted for generations. He is that good at speechifying.
Clove
03-21-2008, 07:09 AM
Why do people always question the sample size?
Remember the central limit theorem.
A sample size of 1000 gives a margin of error of 3%, which is why when a news source says the poll has a margin of error of 3% all they are really telling you is that the sample is greater than a 1000 and less than 5000 (at least I think 5000 is the next increment).
The difficulties associated with polling (and they are so many as to make this type of poll worthless in my opinion) lie in asking neutral questions, obtaining a random sample, etc., but not in the basic calculations of confidence intervals.
C/Valth
I agree with you on the possibility of accurate national statistics based on samples of less than 5,000 however this assumes an excellent method and execution.
Since we don't usually have much information on how the statistics were arrived at (beyond the sample size) and since if all other factors that contribute to accuracy are equal, greater sample sizes mean more accurate results. For this reason alone it's natural for people to be uncomfortable with small sample sizes.
Valthissa
03-21-2008, 08:26 AM
I agree with you on the possibility of accurate national statistics based on samples of less than 5,000 however this assumes an excellent method and execution.
Since we don't usually have much information on how the statistics were arrived at (beyond the sample size) and since if all other factors that contribute to accuracy are equal, greater sample sizes mean more accurate results. For this reason alone it's natural for people to be uncomfortable with small sample sizes.
Checkout the Pew Research Center sometime. They post extensive details on their methods. I think most of the polling companies have excellent methods (probability science just isn't that hard) - execution on the other hand seems open to debate.
Do you think people would really be happier if the margin of error was 2% instead of 3?
C/Valth (I think 3 posts in the same thread exceeds my limit)
Clove
03-21-2008, 08:31 AM
...Do you think people would really be happier if the margin of error was 2% instead of 3?
C/Valth (I think 3 posts in the same thread exceeds my limit)
Yup, and for the same reason people will buy anything that's 1 per customer, or buy one get one free. And on that note, 1% when talking about a group of 300 million is no small improvement.
Trouble
03-21-2008, 09:01 AM
I'm too lazy to read the thread but I highly doubt Wright will have any impact on Obama's campaign. Especially after that speech Obama made recently.
Kefka
03-21-2008, 09:12 AM
Well, the Blacks almost had it...sooo close but no cigar.
Hmmm. "The March 14-18 national survey"
Obama gave his speech on the 18th. Doubt you'd see a bounce on the day of the speech.
I'm too lazy to read the thread but I highly doubt Wright will have any impact on Obama's campaign. Especially after that speech Obama made recently.
I'm too lazy to respond to this post.
Suppa Hobbit Mage
03-21-2008, 09:31 AM
I'm too lazy to respond to this post.
I'm too lazy to...
Kefka
03-21-2008, 09:33 AM
http://images.huffingtonpost.com/gen/15702/original.jpg
Keller
03-21-2008, 09:34 AM
I'm too lazy to respond to this post.
I'm too lazy to point out the irony of reponding that you're too lazy to respond to a post that someone is too lazy to read a thread.
:love:
I'm too lazy to point out the irony of reponding that you're too lazy to respond to a post that someone is too lazy to read a thread.
:love:
:lol:
Checkout the Pew Research Center sometime. They post extensive details on their methods. I have a ton of respect for the Pew Research Center and their methodologies.
I also highly recommend How to Lie with Statistics by Darrell Huff.
CrystalTears
03-21-2008, 09:52 AM
However considering that this poll was done by Gallup, it gives some even more reason to criticize the small sampling.
Latrinsorm
03-21-2008, 10:26 AM
The question of representative sample size is something anyone here can check, really. Set up a thousand die rolls in Excel (or roll a d6 a thousand times if you're bored), graph it, throw on your error bars, and how see well the linear fit works.
The question of representative sample size is something anyone here can check, really. Set up a thousand die rolls in Excel (or roll a d6 a thousand times if you're bored), graph it, throw on your error bars, and how well the linear fit works.Taking issue with such a small sampling size doesn't necessarily indicate that there's a problem with the manner in which the results were arrived at. It is absolutely true that the more people interviewed, the smaller the sample error rate. But, other factors are just as important in judging the quality of a survey.
Latrinsorm
03-21-2008, 12:33 PM
I can't believe I had to edit that twice to get the right English. Anyway.
My point was that the best way for a person to convince himself or herself that a "small" sample size is statistically representative is to have it repeatedly demonstrated to them by a real-world example (such as rolling dice).
Yeah, we got that. The "other factors" I speak about can only be obtained from the person and/or organization conducting the poll. Basically, doing what you've suggested is completely unnecessary.
Clove
03-21-2008, 01:30 PM
I can't believe I had to edit that twice to get the right English. Anyway.
My point was that the best way for a person to convince himself or herself that a "small" sample size is statistically representative is to have it repeatedly demonstrated to them by a real-world example (such as rolling dice).
Sure, if people's opinions were analogous to random dice rolls (3.5 being the average of a 1d6 roll as any good gamer knows).
Valthissa
03-21-2008, 01:48 PM
Sure, if people's opinions were analogous to random dice rolls (3.5 being the average of a 1d6 roll as any good gamer knows).
but of course people's opinions are exactly analogous to random dice rolls - they both conform to a normal distribution (or do they have to use a T-distribution for opinion polls?)
C/Valth
Clove
03-21-2008, 01:52 PM
but of course people's opinions are exactly analogous to random dice rolls - they both conform to a normal distribution (or do they have to use a T-distribution for opinion polls?)
C/Valth
Nope, you're right there, they do follow a normal distribution.
Latrinsorm
03-21-2008, 05:15 PM
Yeah, we got that. The "other factors" I speak about can only be obtained from the person and/or organization conducting the poll. Basically, doing what you've suggested is completely unnecessary.I'm not sure Edaarin or Lysander (the people putting forth the question about sample size specifically) do, which would make it a worthwhile exercise in some cases. :)
Powered by vBulletin® Version 4.2.5 Copyright © 2024 vBulletin Solutions Inc. All rights reserved.