Obama Up By 9% In WP Poll??????

[red states rule]

On the surface it looks like Obama is breaking out. However if you look at the sample you will see the Washingon Compost oversamples Dems to get the results they wanted

and the Washington Post puts this poll, and gloats about the results on the front page



Washington Post-ABC News Poll [Details about Obama +9 "Poll"]
washington compost ^ | 9/24/08

Posted on Wednesday, September 24, 2008 4:45:09 AM by mathprof

This Washington Post-ABC News poll was conducted by telephone September 19-22, 2008, among a random national sample of 1,082 adults, 916 registered voters and 780 likely voters. The survey includes additional interviews with randomly selected African Americans, for a total of 163 black respondents. The added interviews (commonly referred to as an "oversample") were completed to ensure there were enough African American respondents for separate analysis; the group was not over-represented in the reported results from the full sample. The results from the full survey have a margin of sampling error of plus or minus three percentage points. Error margins are higher for subgroups. Sampling, data collection and tabulation by TNS of Horsham, Pa.


http://www.freerepublic.com/focus/f-news/2089189/posts

[mundame]

On the surface it looks like Obama is breaking out. However if you look at the sample you will see the Washingon Compost oversamples Dems to get the results they wanted


This Washington Post-ABC News poll was conducted by telephone September 19-22, 2008, among a random national sample of 1,082 adults, 916 registered voters and 780 likely voters. The survey includes additional interviews with randomly selected African Americans, for a total of 163 black respondents. The added interviews (commonly referred to as an "oversample") were completed to ensure there were enough African American respondents for separate analysis; the group was not over-represented in the reported results from the full sample. The results from the full survey have a margin of sampling error of plus or minus three percentage points. Error margins are higher for subgroups. Sampling, data collection and tabulation by TNS of Horsham, Pa.




I'm not seeing the extra Dems you claim in the sample they say is "random"?

They say above they added interview to get enough blacks for a SEPARATE analysis but did not include those in the original "random" sample.

I saw one poll last spring in which the pollster admitted deliberately adding Dems ----- and he was very defensive about it, which he had better have been, because he sure didn't make a case on TV for that which persuaded me.

However, if the sample is in a defendable way "random," I don't see how there's a problem.

That is, it HAS to be random because we do not know beforehand what the proportions are, Dem-leaning versus GOP-leaning. You can't put in 50% of each, we KNOW it's not that. That would direly skew the results.

Random sampling of likely and probable voters is the only way that makes sense, I think. If you think differently, please explain. I read that Freeper site, but none of that was anything but assertions that since they didn't like the result, it must be crooked. Which is silly.

[red states rule]

I'm not seeing the extra Dems you claim in the sample they say is "random"?

They say above they added interview to get enough blacks for a SEPARATE analysis but did not include those in the original "random" sample.

I saw one poll last spring in which the pollster admitted deliberately adding Dems ----- and he was very defensive about it, which he had better have been, because he sure didn't make a case on TV for that which persuaded me.

However, if the sample is in a defendable way "random," I don't see how there's a problem.

That is, it HAS to be random because we do not know beforehand what the proportions are, Dem-leaning versus GOP-leaning. You can't put in 50% of each, we KNOW it's not that. That would direly skew the results.

Random sampling of likely and probable voters is the only way that makes sense, I think. If you think differently, please explain. I read that Freeper site, but none of that was anything but assertions that since they didn't like the result, it must be crooked. Which is silly.

There is a link to the WP Poll

Item # 901

10% more were Dem then Republican

[mundame]

There is a link to the WP Poll

Item # 901

10% more were Dem then Republican


If it's a random sample, then at least 10% more probably ARE Dem ------ that there are more Dems than Gopers is well known by now, it's been polled repeatedly for the past couple years.

If you are saying they should artificially choose 50% GOP and 50% Dem, that seriously would be inaccurate!! It would gravely overcount the GOPers and the poll would be useless. Are you saying that?

The WaPo/ABC poll hired pollsters that claim they made a random sample. Are you denying this, RSR? Are you saying they fiddled the sample while claiming it was random? If so, you'd have to back that up, surely.

I think there is a lot of fooling around with polls, but the big ones done by the big newspapers and TV networks in combination, as they have been doing all this season, are usually good enough unless it's Zogby. If someone does something......adventurous, like that poll where they oversampled Dems last spring, other pollsters get on TV and attack him.

So let's wait and see the next big one before we call it a "breakout." I prefer the RCP averages of polls, myself.

I think the best polls are probably the private, secret ones the candidates do; these tend to leak out from time to time. Because the candidates really need to know.

[red states rule]

If it's a random sample, then at least 10% more probably ARE Dem ------ that there are more Dems than Gopers is well known by now, it's been polled repeatedly for the past couple years.

If you are saying they should artificially choose 50% GOP and 50% Dem, that seriously would be inaccurate!! It would gravely overcount the GOPers and the poll would be useless. Are you saying that?

The WaPo/ABC poll hired pollsters that claim they made a random sample. Are you denying this, RSR? Are you saying they fiddled the sample while claiming it was random? If so, you'd have to back that up, surely.

I think there is a lot of fooling around with polls, but the big ones done by the big newspapers and TV networks in combination, as they have been doing all this season, are usually good enough unless it's Zogby. If someone does something......adventurous, like that poll where they oversampled Dems last spring, other pollsters get on TV and attack him.

So let's wait and see the next big one before we call it a "breakout." I prefer the RCP averages of polls, myself.

I think the best polls are probably the private, secret ones the candidates do; these tend to leak out from time to time. Because the candidates really need to know.


Mundame, libs conduct their polls like they do their recounts

They keep doing them until they get the results they desire

They WP did the same thing in 2004 and they Kerry way ahead of Pres Bush. Then they are stunned when their guy lost

[mundame]

Mundame, libs conduct their polls like they do their recounts

They keep doing them until they get the results they desire

They WP did the same thing in 2004 and they Kerry way ahead of Pres Bush. Then they are stunned when their guy lost


Was that the famous Zogby poll? That was pretty funny, but he WAS the only one who got that result. I don't trust Zogby at all: his polls always, 100% slant Dem.

However, I don't believe the polls from the big media are slanted. I think they do want to know, and tell the people. After all, they've had several pro-McCain results since his convention bounce, and you didn't say those were inaccurate. Now one is showing something different.

I do like to see several to see a trend, but I don't think these big polls are crooked. As long as you keep Zogby out.

[red states rule]

Was that the famous Zogby poll? That was pretty funny, but he WAS the only one who got that result. I don't trust Zogby at all: his polls always, 100% slant Dem.

However, I don't believe the polls from the big media are slanted. I think they do want to know, and tell the people. After all, they've had several pro-McCain results since his convention bounce, and you didn't say those were inaccurate. Now one is showing something different.

I do like to see several to see a trend, but I don't think these big polls are crooked. As long as you keep Zogby out.

Maybe you are right the liberal media is not slanting the polls to Obama. Maybe they are following the rules of Afirmative Action

[Abbey Marie]

So now we are using Affirmative Action in polling too? :rolleyes:

[red states rule]

So now we are using Affirmative Action in polling too? :rolleyes:

Seems logical to me Abbey

But let's not forget racism

How many articles have already "predicted" and "explained" that will be the reason the messiah loses

What some MSM outlet needs to do is to practice CYA and report how "Obama will lose because he is an empty suit with questionable associations and Americans don't trust him or his judgment"

Not that is wishful thinking

[mundame]

Maybe you are right the liberal media is not slanting the polls to Obama. Maybe they are following the rules of Afirmative Action


Pretty cute. http://wade.hu/smiley/kategoriak/szemtelen/cheeky-smiley-023.gif

I guess they DID do affirmative action, in the sense that they sampled extra blacks for "separate analysis," they said.

As long as they didn't include the extras in the main poll results, and they SAID they didn't, that doesn't matter.

[Immanuel]

Maybe you are right the liberal media is not slanting the polls to Obama. Maybe they are following the rules of Afirmative Action


So now we are using Affirmative Action in polling too? :rolleyes:


Seems logical to me Abbey



Seems to me that this is a stretch RSR.

They said it was a random sample across America. Now, if they did a "random sample" by dialing only zip codes in New York, Illinois and California that would not be such a random sample but there is no proof that this was done.

Immie

[red states rule]

Seems to me that this is a stretch RSR.

They said it was a random sample across America. Now, if they did a "random sample" by dialing only zip codes in New York, Illinois and California that would not be such a random sample but there is no proof that this was done.

Immie

It is not hard Immie. All the WP had to do was poll4% fewer Dems and 5% more Republicans

They oversample Dems, get the tresults they want, and put the results on the front page as a news story

[mundame]

It is not hard Immie. All the WP had to do was poll4% fewer Dems and 5% more Republicans

They oversample Dems, get the tresults they want, and put the results on the front page as a news story


Sure, RSR, but DID they do that?

And do you have any proof they did? They SAID they took a random sample.

Believing things simply because you want them to be true is perhaps not the best reason for belief.

[red states rule]

Sure, RSR, but DID they do that?

And do you have any proof they did? They SAID they took a random sample.

Believing things simply because you want them to be true is perhaps not the best reason for belief.

I already answered that question

Here you go

There is a link to the WP Poll

Item # 901

10% more were Dem then Republican

[Immanuel]

It is not hard Immie. All the WP had to do was poll4% fewer Dems and 5% more Republicans

They oversample Dems, get the tresults they want, and put the results on the front page as a news story

I am sure that polling companies know all the tricks of the trade to get the results they want. But, if they do it too often and their results are always off base, then they lose credibility and eventually they are out of business. Of course, if they are an arm of a much larger business such as The Alan Guttmacher Institute, they don't have to be accurate because their funding comes from NOW and Feminist Organizations rather than people who want serious and accurate information.

Immie

[red states rule]

I am sure that polling companies know all the tricks of the trade to get the results they want. But, if they do it too often and their results are always off base, then they lose credibility and eventually they are out of business. Of course, if they are an arm of a much larger business such as The Alan Guttmacher Institute, they don't have to be accurate because their funding comes from NOW and Feminist Organizations rather than people who want serious and accurate information.

Immie

I could go along with that Immie, if the WP did not put this on their front page and call it a news story

They created the story, and hoped nobody would look at the breakdown which they placed deep in the poll

[mundame]

I already answered that question

Here you go

There is a link to the WP Poll

Item # 901

10% more were Dem then Republican


RSR. It doesn't matter if 10% more were Dem if 10% more of the population are Dem!!!!

In fact, to correctly reflect the population, if 10% more of the population is Dem (and it's at LEAST that, as I'm sure you know), they HAVE to have such a result, or the poll would be inaccurate.

The point is, is it successfully randomized so that it reflects the actual voting population?

You sound as though you want it to be a poll of exactly 50% Republicans and 50% Democrats ----------- even if a lot more voters are Democrats!! That wouldn't tell us anything about how the election is likely to go, would it?


Because at the voting booth, they don't let exactly 50% Republicans in and 50% Democrats in and then turn everyone else away, you know.

Suppose the population turns 75% Republican in 2010 ------- do you want the polls to rigidly question 50% Republicans and 50% Democrats? Would that make any sense?

If they randomized this poll to capture the actual voting ratio of the population, then it's a good poll. If they lied, surprise, surprise. If they made a mistake and it's a bad ratio they polled, we'll soon see that in the next poll and the next.

But I hope you aren't wanting some sort of 50--50 ratio that doesn't pertain in the actual voting population.

[red states rule]

Mundame, I know you hate McCain, but try to think clearly. The WP is the only poll that has Obama with a big lead

Here is the daily tracking poll

The Rasmussen Reports daily Presidential Tracking Poll for Wednesday shows Barack Obama attracting 49% of the vote while John McCain earns 47%. It’s the first time in more than two weeks that Obama has enjoyed a lead larger than a single percentage point (see trends). Both men are now viewed favorably by 55% and Rasmussen Markets data currently gives Obama a 53.8% chance of victory

http://www.rasmussenreports.com/public_content/politics/election_20082/2008_presidential_election/daily_presidential_tracking_poll


I guess you do not have a prblem with the WP stakcing the poll, getting the results they wanted, and running it as a news story

[mundame]

Mundame, I know you hate McCain

Not me, I'm an equal opportunity hater.



I guess you do not have a prblem with the WP stakcing the poll, getting the results they wanted, and running it as a news story

No, not unless they actually DID that and you can prove it. I don't have a problem with you just saying that because you wish it were true, but you can't show it happened.

I prefer data that points to objective reality, not a made-up world of wishes.

[red states rule]

Not me, I'm an equal opportunity hater.




No, not unless they actually DID that and you can prove it. I don't have a problem with you just saying that because you wish it were true, but you can't show it happened.

I prefer data that points to objective reality, not a made-up world of wishes.

Maybe you are just ignoring the WP own stats. I can provide the proof, but it is up to you to read it and accept it

[mundame]

Here is the daily tracking poll

The Rasmussen Reports daily Presidential Tracking Poll for Wednesday shows Barack Obama attracting 49% of the vote while John McCain earns 47%. It’s the first time in more than two weeks that Obama has enjoyed a lead larger than a single percentage point (see trends). Both men are now viewed favorably by 55% and Rasmussen Markets data currently gives Obama a 53.8% chance of victory



How you are calling this a GOOD result for McCain passeth my understanding: looks pretty dire to me, given that Rasmussen the Rong always slants heavily GOP. Darn, Rasmussen gives Obama a 53.8% chance of victory and you call that GOOD for McCain?????

Also, you say this is "THE" daily tracking report, but my same link (http://www.realclearpolitics.com/epolls/2008/president/us/general_election_mccain_vs_obama-225.html) shows the Hotline/FD tracking poll for the exact same time frame as the Rasmussen, and it puts Obama at +6!!


Let's report ALLLLLL the data, RSR, because that's the honest and right way to do.

[red states rule]

How you are calling this a GOOD result for McCain passeth my understanding: looks pretty dire to me, given that Rasmussen the Rong always slants heavily GOP. Darn, Rasmussen gives Obama a 53.8% chance of victory and you call that GOOD for McCain?????

Also, you say this is "THE" daily tracking report, but my same link (http://www.realclearpolitics.com/epolls/2008/president/us/general_election_mccain_vs_obama-225.html) shows the Hotline/FD tracking poll for the exact same time frame as the Rasmussen, and it puts Obama at +6!!


Let's report ALLLLLL the data, RSR, because that's the honest and right way to do.


So now Rasmussian is a biased poll :laugh2:

You are desperate

Oh well, here is the poll that counts, and Obama lost NH this morning

http://www.rasmussenreports.com/public_content/politics/election_20082/2008_presidential_election/election_2008_electoral_college_update

[mundame]

I can provide the proof, but it is up to you to read it and accept it


Pray do so, then; you haven't so far.

You just keep asserting that the poll is lying even though THEY assert that they randomized the sample.

[red states rule]

Pray do so, then; you haven't so far.

You just keep asserting that the poll is lying even though THEY assert that they randomized the sample.

Are you to dense to go to the WP poll and find #901?

http://www.washingtonpost.com/wp-srv/politics/polls/postpoll_092308.html?sid=ST2008092303897&s_pos=list

[mundame]

So now Rasmussian is a biased poll

They always have been; everyone knows that. Look at the whole line of polls over time: Rasmussen consistently scores rightward.

That's why it is remarkable that they are now showing Obama ahead --- that means to me he's WAY ahead, if even Rasmussen shows him ahead at all.



You are desperate

Not me. I figure we're doomed: we all lose whichever of these clowns wins.

So now you're a McCain partisan? I thought you said you weren't going to vote........

This is not the biggest surprise I've had all year, however. http://forums.offtopic.com/images/smilies/happysad.gif

[red states rule]

They always have been; everyone knows that. Look at the whole line of polls over time: Rasmussen consistently scores rightward.

That's why it is remarkable that they are now showing Obama ahead --- that means to me he's WAY ahead, if even Rasmussen shows him ahead at all.




Not me. I figure we're doomed: we all lose whichever of these clowns wins.

So now you're a McCain partisan? I thought you said you weren't going to vote........

This is not the biggest surprise I've had all year, however. http://forums.offtopic.com/images/smilies/happysad.gif

Are you related to MFM? If you are, it would explain alot

[mundame]

Are you to dense to go to the WP poll and find #901?

http://www.washingtonpost.com/wp-srv/politics/polls/postpoll_092308.html?sid=ST2008092303897&s_pos=list


I did; it shows nothing more than we've already talked about.

If you have some analysis you are making that you would like to share with us that shows this is a biased poll, I'm sure people would be interested to read that.

I fear you haven't got anything, and that's why you keep not saying it.

[red states rule]

I did; it shows nothing more than we've already talked about.

If you have some analysis you are making that you would like to share with us that shows this is a biased poll, I'm sure people would be interested to read that.

I fear you haven't got anything, and that's why you keep not saying it.

OK Ms MFM. Ignore the WP admits they oversampled Dems by 10 points and cheer how Obama has a 9 point lead :rolleyes:

[mundame]

Ignore the WP admits they oversampled Dems by 10 points and cheer how Obama has a 9 point lead :rolleyes:


I don't see how you can say that; it looks like an honest poll to me. And you aren't able to specify why you think they oversampled. And nobody else thinks so; it's widely published all over the news and is in the RCP average, so you are the only person who seems to think it's a bad poll.

I think you don't like the result they got, so you are asserting they must have cheated.


But real data always trumps wishful thinking.

You are free to believe in a world you simply make up, but such a world tends to drift way off from reality and I question the value of doing that.

[red states rule]

I don't see how you can say that; it looks like an honest poll to me. And you aren't able to specify why you think they oversampled. And nobody else thinks so; it's widely published all over the news and is in the RCP average, so you are the only person who seems to think it's a bad poll.

I think you don't like the result they got, so you are asserting they must have cheated.


But real data always trumps wishful thinking.

You are free to believe in a world you simply make up, but such a world tends to drift way off from reality and I question the value of doing that.

It is one thing to "assert" cheating - and it is another to pioint out the oversampling of Dems

Gallup and Rasmussen Tracking have Obama up by 2 and 3 points

We still have 41 days to go

[mundame]

Maybe you'll like this poll better -------- or maybe not.

The Fox News Poll just this minute came up. 1:40, 9/24/2008

Poll: Obama Regains Lead (http://www.foxnews.com/story/0,2933,427241,00.html)

Latest FOX News poll shows Democrat recapturing lead — 45 percent to 39 percent — over McCain.
**********************************

So has this poll been cooked too, RSR? http://pages.prodigy.net/indianahawkeye/newpage08/4.gif

[red states rule]

Maybe you'll like this poll better -------- or maybe not.

The Fox News Poll just this minute came up. 1:40, 9/24/2008

Poll: Obama Regains Lead (http://www.foxnews.com/story/0,2933,427241,00.html)

Latest FOX News poll shows Democrat recapturing lead — 45 percent to 39 percent — over McCain.
**********************************

So has this poll been cooked too, RSR? http://pages.prodigy.net/indianahawkeye/newpage08/4.gif


Polling was conducted by telephone September 22-23, 2008, in the evenings. The
total sample is 900 registered voters nationwide with a margin of error of ±3
percentage points. Results are of registered voters, unless otherwise noted.
LV = likely voters
Democrats n=369, ±5 percentage points; Republicans n=306, ±6 percentage points;
independents n=188, ±7 percentage points
Obama supporters n=409 ±5 percentage points; McCain supporters n=350, ±5
Whites n=751, ±4 percentage points; Blacks n=91, ±10 percentage points
Investors n=412, ±5 percentage points
http://www.foxnews.com/projects/pdf/092408_poll1.pdf

What I see here is registered and not likely voters. I am not saying Obama is not in the lead, but the lead is only about 2 to 4 points

[actsnoblemartin]

the only poll that matters, nov 4th, 2008

[mundame]

Whoa!!! I have to throw a bone to you, RSR.

The Wall Street Journal poll came out this morning. They show Obama with a two-point lead, which they call a statistical tie, because it's inside the margin of error.

But the interesting thing is their final paragraph, in which their pollsters disapprovingly analyze the same WaPo/ABC poll that RSR so disliked yesterday. The Journal found a higher error than RSR was claiming! (My emphasis:)

********************************************
The Journal/NBC News poll (http://online.wsj.com/article/SB122228902289272455.html?mod=todays_us_page_one) comes a day after a national poll conducted for the Washington Post and ABC News found Sen. Obama with a nine-point lead. Mr. Hart, the Journal's Democratic pollster, argued that these results were skewed because their sample included a disproportionate number of Democrats. Democrats had a 16-point advantage in the Post/ABC poll, which is considerably higher than most polls have found historically and even this year.

The Journal/NBC poll found an eight-point advantage for Democrats.

********************************************

You didn't detail what you were talking about yesterday, RSR, presumably because your Freeper source didn't either, but this is exactly the same issue that showed up when another poll was challenged earlier this year on the same basis, that it oversampled Dems, and that one was (I think!) charged with actually adding Dems, rather than just hitting too many by chance.

No one is saying here that WaPo/ABC is cooking the books on purpose. They are saying that their sampling vacuumed up too many Democrats, more than the proportion there actually are in the population, and thus gave a result that the whole voting population would not give, if polled (the whole population will never be polled, not even on November 4, of course, but only the voting population matters).

The Journal pollsters think the Dems have about an 8% numerical superiority in the voting population at this point and that therefore their poll reflects the whole voting population at this time. WaPo/ABC pollsters would presumably defend at least their randomization tactics, I don't know whether they are saying there actually are 16% more Dems, but that does seem high, and their poll is an outlier in terms of the big gap for Obama.


I'm guessing the Journal poll is on the low side re Obama's lead and the WaPo/ABC poll is high.

So you were onto something, RSR, though next time you might want to try to explain what you are trying to say.