ALERT!! Disregard all Time polls!! New evidence of flawed methodology!!!

[JohnnyCougar]

I was examining the specifics of the latest Time polls, done by Schulman, Ronca, & Bucuvalas, inc. In their last THREE polls, the sample they used voted 41% for Gore and 53-54% for Bush in 2000!!!!

Now, you may be asking yourselves, "Why the FUCK would any reputable polling organization or shill magazine report these results as the truth?!?!" That doesn't sound like a representative sample if the people you were polling voted Bush over Gore by 12 POINTS!!

Unless a ton of people who voted Gore died (6% of total electorate)and were replaced by a younger segment that overwhelmingly supports Bush, something is wrong here!

What does this tell us?? When corrected and weighted for the 2000 vote totals (a virtual tie b/t Gore and Bush) this poll would show KERRY AHEAD!!!!!!! This poll is a TOTAL LIE!!!!

Tell everyone, tell the world....KERRY IS AHEAD!!! And stop pouting for god's sakes.

On edit: Here's the poll....see for yourselves!!! http://www.srbi.com/condensed-data-2004-5.pdf

[Carla in Ca]

1. It was that poll that made me promise myself

I wouldn't read any more of them...just too over the top. I knew they were wrong but I worry about the momentum it creates, especially if it is not quickly disavowed.
I'm going to stay focused, not look at polls and wait for the debates.

[Seabiscuit]

3. I don't read anything in print anymore...

or listen to anything on TV anymore about politics or polls. The only reliable source for anything anymore is the internet.

[BushOutNow]

24. let's hope they're wrong

I too worry about the effect they have. There are lots of sheeple that flock to bandwagons.

[featherman]

2. that is very strange

Someone needs to explain that to me. What are they playing at?
Agree, Johnny, that the internals seem to show that a sample of voters who voted for Bush by 12% to 13% now favor him by only 11%. I suppose, realistically, this sample has to be weighted in some way for this poll to have any validity but I don't see it. Need an expert here.

[JohnnyCougar]

5. What we need is to e-mail blast this to all news organizations

so the stop reporting LIES!! The sample greatly underrepresented self-proclaimed Democrats as well. I'm pretty good at reading polls, I'm getting my masters in I/O Psychology. I have had so many research methods classes, i dream about statistics in my sleep. :) There is no doubt that this poll has a flawed methodology. We just have to make sure the media know about this. After all, practically 2 out of 3 people think Bush is going to be reelected! This perception needs to change!

[DarthDem]

4. Great Job

Hey, thanks for digging this up. Like a lot of others, I knew this poll was total refuse but didn't do the research to back up my beliefs like you did. Thanks for doing us all a solid service.

What an absolutely criminally negligent poll. The degeneration of the fascist media is really, really disturbing.

[JohnnyCougar]

6. Hehe...I'm giving a presentation about it to my class tomorrow!

I can't wait!

[WiseMen]

7. PLEASE! Why were they OK when Kerry was 300+ in the Electoral College??

[aquart]

25. They weren't lying then?

But they panicked and changed their methodology?

It isn't six of one and half a dozen of the other. It isn't both sides lie.

It's one side is a criminal fascist enterprise and the other side is people who still believe we have and need our constitution.

I don't see a resemblance between those two sides. They are not the same coin. They are not equal in any way.

[DanGough]

8. Polling data

The polls are used to "push" opinion one way or another. Notice how the media told the public before hand that Kerry would not get a bounce out of the DNC and it happened. The media also said Bush would get a bounce and of course he did. Many pollsters are pushing the polls in the tradition of Lee Atwater.

[Seabiscuit]

12. BINGO! BINGO! BINGO!

Your post says it all. End of story. Congrats!!!

[Sugarbleus]

29. Wasn't there something written recently about how Bushco fired

or asked that someone be fired from a certain Polling outfit? Whorge didn't like the way the polls were reading.....against him..hmmm

Thanks again, Bush Crime Family, for tampering with elections apparatus!! *insert the finger here*

[newyawker99]

30. Hi DanGough!!

Welcome to DU!! :toast:

[Gore1FL]

9. I don't think you can "re-weigh" a poll

The problem is the sample. -- While the methodolgy works in a generic sense -- The numbers probably favor Kerry more than the poll indicates (and Bush less so) -- there is no way to really measure the results of what it would be without the flaw.

Still good news -- but not really quantifiable in a meaningful way.

[JohnnyCougar]

10. I don't see why not

That's what Zogby does. You can re-weigh the poll unless there is some sort of massive change in the population in that Gore voters simply vanished in the last 4 years. You can add weights to different segments of the poll, it's the same as stratafied random sampling.

[LittleClarkie]

11. Remembering, of course that Zogby

was the only accurate poll in 2000.

I picked up an Investor's Daily News today. A conservative newspaper, it interestingly showed Bush and Kerry in a 47 to 47 dead heat, and said that the bounce didn't really exist.

Hmmm. If even a self-proclaimed conservative newspapaer is reporting a dead heat, then perhaps the rest of the country could catch on. Eh?

[Gore1FL]

14. It can give you an idea -- but it blows the MOE all to hell

The issue is that the only thing to base the skew on is the bad population sample that was the problem to begin with.

Again -- for a macro adjustment you can prbably get away with it, but I wouldn't "go to the bank" on the specifics, only the trend.

[JohnnyCougar]

16. fair enough...

blows the MoE to hell, but we all see what direction this is going....NO BUSH BOUNCE.

[Gore1FL]

18. Roger that!

except of course for his bounce back to Crawford!

[grasswire]

13. bring me up to speed, would ya?

Are you saying they used the same people for these polls as they did for Bush Gore? Huh? Or the same demographic distribution?

[JohnnyCougar]

17. I'm not sure who they polled in 2000

but now they are polling much more Bush voters than Gore voters. If you poll 54% Bush voters and 41% Gore voters, of course you are gonna get poll results that show Bush in 2004 with an 11 point lead.

[JohnnyCougar]

15. another interesing fact...

45% in the 9/3/04 poll considered themselves "evangelicals" or "born again." So that means Bush has the support of 7% of people who are NOT evangelicals...that's terrible. Someone should find out the actual percentage of evangelicals in the population.

[grasswire]

21. screwy

According to the City University of New York's American Religious Identification Survey, the number of Americans who identify themselves as "evangelicals" more than quadrupled from 1990 to 2001, from 242,000 to 1,032,000.

[featherman]

19. letter to countdown

Dear Keith,
Your staff might want to look into the internals of the TIME poll you featured on Monday.
It is readily available but I'll include a link for your people in case they don't know how to find it.
Here's the poll internals: http://www.srbi.com/condensed-data-2004-5.pdf
(Key internal on pg 19)

TIME Poll results as prominently featured on Countdown (9/13/04):
Bush leads by 11%

Question asked of the sample: (voted 2000)
"Regardless of how you feel today, who did you vote for in 2000"
Bush - 53%
Gore - 41%
Nader/other 2%
Don't know/refused 4%

Now here are the poll results asked of likely voters (voted 2000)
Bush/Cheney - 52%
Kerry/Edwards - 41%
Nader/other - 0%
Don't know/refused 7%

I hope this speaks for itself to save me the trouble of making snide remarks about the obvious.

Best wishes,

[JohnnyCougar]

20. Thanks...

I'll mail him too...maybe if enough people send this to him, he'll break the story, and the media will have to take a RESPONSIBLE look at the polls they yabber about. I'm gonna send this to Matthews, too. Maybe he'll think twice before asking if the campaign is over.

[TexasBushwhacker]

26. This is so bogus!

Are people's memories THAT short?

TIME Poll results as prominently featured on Countdown (9/13/04):
Bush leads by 11%

Question asked of the sample: (voted 2000)
"Regardless of how you feel today, who did you vote for in 2000"
Bush - 53%
Gore - 41%
Nader/other 2%
Don't know/refused 4%

Now here are the poll results asked of likely voters (voted 2000)
Bush/Cheney - 52%
Kerry/Edwards - 41%
Nader/other - 0%
Don't know/refused 7%

Now, we KNOW that in fact, the results of the 2000 election were 48% for Gore, 48% for Bush, with Gore leading by over 500K of the popular vote. Some Time's poll, either through inadequate sample size or whatever, shows that their sample said they voted for Bush by a margin of 12%, when in fact the margin was ZERO. So why would their poll showing a Bush lead of 11% for the upcoming election have an validity at all?

Then take into the account the good old "likely voter" polls. Likely voters are voters that are not only registered, but actually voted in the last election. Doesn't that exclude any newly registered voters, ie. 18 to 21 years olds that were too young last time? What about the thousands of black voters that weren't allowed to vote in Florida? They don't make it to the "likely voter" group either. Then there is the 25% of America that uses ONLY a cell phone, who pollsters cannot call. It all adds up to biased polls that give bias results.

[NJCher]

22. here's the deal

I already brought the Time poll to the attention of the host of our local NPR (WNYC), Brian Lehrer, and he did a segment on it. Time and Newsweek use a similar methodology. I am providing the link to the show and my original e-mail to him.

The reason this poll seems to be weighted toward republicans is that after the convention, more people identified themselves as republicans. Using this method, the same phenomena would have happened after the Democratic convention.

Here is the description of the show and a link if you would like to listen:

Patrick Murray, Acting Director of the Star Ledger Eagleton Poll, and professor of Public Policy at Rutgers University discusses the four point difference between Kerry and Bush in the polls in the state of NJ and Larry Hugick, Chairman of Princeton Survey Research Associates on criticism over a recent Newsweek poll skewing Republican Princeton Survey Research Associates.

http://www.wnyc.org/shows/bl/episodes/09102004

scroll down to "Swinging in New Jersey."

Here are e-mails I sent the host of the show. I have edited the first one for length:

You had a segment on your program the other day where you had either the Newsweek or Time pollster who reported bush's big bounce after the Republican convention. The polls talked to more Republicans than Democrats--by a substantial margin!

This fact was discussed in depth on various sites on the Internet the previous evening. Several reputable pollsters--among them Rasmussen and Zogby--have issued analyses of the skewed sample in these polls. Both concur that there is no way that bush received an 11 per cent bounce.

In correcting for this skewing of the sample, pollsters are now saying the lead was about 2 per cent.

I had the information from Rasmussen and called the show to tell your listeners. Unfortunately, I could not get through, despite trying throughout the hour.

I think you owe it to your listeners to correct the record.

---------------------------

I wanted to thank you for airing that segment where you explored the polling issue with Larry Hugick and Patrick Murray. It was very informative and taught me a number of things.

I've referred a number of people to the show, all of whom have found it similarly informative.

-----------------------------

The bottom line is that the bounce fades because after the influence of the convention wears off, people are more likely to return to their original position, whatever that was. Here are a few links that illustrate the fact that the "bush bounce" is disappearing:

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=132x790743

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=102x819865

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=132x794271

------------------

I think it's a good topic for a show. I hope Keith Olberman picks up on it.


Cher

[Dookus]

23. One thing to keep in mind

is that people who are voting for Bush this year may SAY they voted for him last time, but aren't telling the truth.

[aquart]

27. So you think lots of people have converted TO Bush FROM Gore?

And your proof of that laughable statement would be...........?

Please tell me how many people you personally know who did NOT vote for Bush last time who ARE voting for him this time?

Get back to me.

[Dookus]

32. I said no such thing

sheesh...

[BootinUp]

28. My guess

Is not that people lied, just that people that lean republican but don't always vote, were fired up by the convention and said immediately after the convention, they will vote for sure. At least I can see that accounting for some of the swing. Likewise, Dem leaners may have been turned off and saying they won't vote as some sort of protest.

[JohnnyCougar]

31. well if people were lying in that poll

it's not valid anyway. So why take it's content to mean anything meaningful? The whole thing is nothing but horseshit.

[Dookus]

33. it's well known

that asking people how they voted in the past does NOT reflect the actual vote in the election being asked about.

In '74, you couldn't find 20 people who said they voted for Nixon.

[JohnnyCougar]

34. alone, this may be questionnable

but coupled with the low ammount of democrats and liberals in the poll, it just doesn't add up. And if you lie about who you voted for in 2000, than what is to say you're not lying when you say who you're voting for in 2004?

Look...I have discredited these polls. The more the media finds out these polls are horseshit, the better it is for us. No one but the newscasters and their sheep believe these polls to be accurate.

So go contact every news organization you know and tell them the Time polls CAN'T be valid!!

[lancdem]

35. That is true

I also remember reading that in the early 1960s, a poll showed 70 percent of respondents claimed to have voted for Kennedy when he got less than 50 percent of the vote in 1960.

[JohnnyCougar]

36. So the people polled are lying...

then why should we believe anything they say?