NAYSAYERS, YOUR TURN: The Washington Post/National Election Pool Exit Poll

[TruthIsAll]

THIS KNOCKS OUT ALL THE STRAWMEN CREATED OVER THE LAST 2 MONTHS:

a) "Clustered sample": Not a word about it.
The MOE is 1.0% (See the note at the bottom).
This is close to the formula: MOE = 1/sqrt(13047) = 0.87%

b) "Not a Random Sample":it IS a RANDOMLY-SELECTED sample(see below).
c) "Small sample-size": 13,047 were polled. BIG national sample.
d) "Exit Polls Not Accurate": "Final percentages may shift SLIGHTLY".
e) "Data Not Weighted": 54% women/46% men, etc....

Naysayers are invited to comment on this caption from the Washington Post site.

Applying the weights for men (46%) and women (54%), Kerry is the winner by a 2.56% margin:

Kerry 50.78%
Bush 48.22%
Nader 1.00%

Gender Pct Kerry Bush Nader
Men 46.00% 47.00% 52.00% 1.00%
Women 54.00% 54.00% 45.00% 1.00%

Total 100% 50.78% 48.22% 1.00%

Here are some more weightings.
Kerry wins them all:

....................... Bush Kerry Nader
EDUCATION... 48.05% 50.21% 1.17%
INCOME........ 48.12% 51.42% 0.95%
RACE........... 47.86% 50.94% 1.00%
AGE............. 48.17% 50.53% 1.00%
PARTY.......... 48.24% 50.24% 0.90%
IDEOLOGY.... 48.15% 49.85% 1.00%
RELIGION.... 48.19% 51.55% 1.19%
MILITARY.... 47.62% 51.20% 1.00%
DATEDECIDE 47.95% 51.23% 0.54%
ISSUES...... 47.92% 50.80% 1.28%

AVERAGE..... 48.00% 50.80% 1.00%


GRAPHICALLY:

[GetTheRightVote]

1. Where is this info track from but it does seem you have a point

:kick:

exit poll info new or old ??

[TruthIsAll]

8. KERRY WINS REGIONALLY: 50.0%- 47.5%- 1.0%

KERRY WINS REGIONALLY: 50.0%- 47.5%- 1.0%

..........................................WEIGHTED
REGION       PCT       Bush Kerry Nader    Bush Kerry Nader
East	       22%	  41%	58%	1%	9.0%	12.8%	0.2%
Midw         25%	  49%	50%	1%	12.3%	12.5%	0.3%
South	       31%	  54%	44%	1%	16.7%	13.6%	0.3%
West	       21%	  45%	53%	1%	9.5%	11.1%	0.2%

TOTAL		 100%			      47.5%	50.0%	1.0%

[zann725]

28. Okay, JK won...let's inaugurate him!

Enuf said!

[chorti]

35. east region is really 'fixed'

They really goofed up when they tried to fix the East region results. For the gender breakdown, they have
...............Bush .. Kerry .... Nader
Male (48%).....45%....53%.....1%
Female (52%)...42%....57% ....0%

For an overall 55-43 Kerry win.

But then for first time voters check this out. It looks like they missed "correcting" this one:
Have You Ever Voted Before?
.............Bush .. Kerry .... Nader
No (13%)....37%....62%.....*
Yes (87%)...42%....57% ....1%

This puts Kerry winning at a 58-41 gap, the same margin which the unfixed exit poll has for the East region.

The other really strange result from the "fixed" East region - "When did you decide who to vote for?"
Today (6%) - Bush 50%, Kerry 47%, Nader 1%.

This is the only region where Bush won this group of voters, according to the fixed polls. Yet Kerry won this group of voters in the most populous East state - New York - 64 to 34. This would mean that excluding New York, Bush would have won the East-region election-day-deciding-voters by a margin of about 58 to 39, assuming New York has one-third of the voters in the East region. Even being conservative and assuming NY has only one-fourth of the East voters, means Bush wins this group of voters, 56 to 42. In either case I say this is impossible. In every other region Kerry was winning the last-minute undecideds by a margin of 50-45 to 56-40 (even in the fixed poll - the reality is probably a much bigger margin). Why would it be so different on the East Coast and New England? And then, why would New York follow the rest of the country but Massachusetts and Pennsylvania did not? And if Bush won this group of voters by a 14 to 19 percent margin in the East region minus New York, why is it that he did not win by this margin among this group of voters in any of the individual states in the region (even in the 'fixed' poll results)?

[TruthIsAll]

45. Chorti, we need more of your kind - ANALYZERS , NOT PROSELYTIZERS

You're digging, digging. Good. You are digging much deeper than I. You are analyzing, not just rebutting without relevant arguments.

One poster questioned the need for the same total vote percentage calculations based on the various exit poll categories:

"Why break down the characteristics, just to build them back up? It's the SAME sample".

Well, you just showed why.

[Pam-Moby]

2. Very interesting.

Thanks for posting this...do you have a link?

[blue neen]

3. The evidence looks compelling to me.

NO Naysayer here, TIA!

[dzika]

4. Interesting... It's coming right off of the WaPo site.

The Headers show the last modified date:

(Response Status) HTTP/1.1 304 Not Modified
Accept-ranges bytes
Age 1573
Connection Keep-Alive
Date Mon, 03 Jan 2005 05:51:02 GMT
Last-Modified Wed, 03 Nov 2004 05:50:27 GMT
Server Sun-ONE-Web-Server/6.1


Would that be 12:50 AM ET?

[Clark2008]

5. Yes... I can look at the post here on DU

and see that it's set for GMT 5:57 and it is 12:57 a.m. EST.

[Sick_of_Rethuggery]

31. I can't find the exit poll on wp.com..

could you post a link to it?

Thanks...

[dzika]

6. "preliminary exit poll results"

Can someone explain what that means?

[bleever]

7. What a kick.

[Mistwell]

9. There is a reason it is a preliminary analysis

Since it does not have ACTUAL demographic tracking, only PREDICTED demographic tracking...and the prediction can be off. Once the final analysis was issued they reported that, for example, a lot more men ACTUALLY voted than were predicted to be voting prior to the election. Turnout for Republicans was also reported in the final analysis as higher in a material way than initially predicted.

Now I am not saying the final analysis is correct. However, I am saying it COULD be correct, and in order to prove that it is false we MUST use the raw data for the final exit poll.

There is simply no other way to do this. No amount of analysis and statistical breakdown will result in a CONCLUSION without the final raw data. No matter how often or how loudly anyone says that the existing exit poll data we have "proves" the election was stolen, the fact of the matter is that we can not "prove" it until we have the actual raw data, and not preliminary analysis of preliminary data.

[dzika]

11. That's mostly what I thought.

Thanks for your response. I'm sure this question has been answered many times so I appreciate your patience.

[TruthIsAll]

14. Circles. Circles. Circles. Catch 22. Strawmen. A Foggy Day in DU Town.

You cannot prove fraud unless you have the data.
But you can't have the data.
So how are you going to prove the fraud?

OK:
We will just use the data we have.
Until THEY can prove the data is NOT valid.

[understandinglife]

17. I just wish I could get them to stop running in those circles just ....

....long enough to explain these historical facts to me (and everyone else here):

How do they explain the fact that the US paid our tax $$ to do the exit polls in the Ukraine; and how they might explain the consequences of those exit polls.

I would like for all of them to just read these few paragraphs:

"The U.S. state department last week said it had spent $65m over the past two years financing groups in support of democracy in Ukraine, part of the $1bn spent for the same purpose globally each year.

"Our money doesn't go to candidates. It goes to the institutions that it takes to run a free and fair election," said a spokesman, Richard Boucher.

The U.S. embassy said it - together with seven other western embassies, including Britain's - had funded an exit poll which showed Viktor Yushchenko was ahead in the first run-off by 11 points, and helped to spark the mass protests." (more at links)

As quoted from:

Inquiry Sought into Claims of U.S. Funding

By Nick Paton Walsh
The Guardian U.K.

Monday 13 December 2004

Ukrainian MPs are seeking a parliamentary investigation into allegations that money from the U.S. government was used to help fund the opposition during the recent electoral campaign.

Links:

http://www.guardian.co.uk/ukraine/story/0,15569,1372536 ...

http://www.truthout.org/docs_04/121404C.shtml

oh, and 'they' can google and find other references, including the fact that no less than 3 exit polls were performed during the REVOTE on 26 Dec 2004 (and one of the pollsters is a Republican org).

And, the darnedest thing is that I have yet to read about a Ukrainian citizen who actually gives a shit about their franchise of democracy demanding 'raw' exit poll data before they forced a REVOTE -- not a one. And, I do not recall any member of the shrubco organization demanding access to 'raw' exit poll data as they argued with Putin and did everything possible to ensure a REVOTE.

Dang; what am I missing here TIA: what am I missing? I'm sure I'm going to have another sleepless night trying to figure it out. NOT.

Peace.

"Who bought the green shoes: daddy or Karl?"

[Mistwell]

23. Fair enough

Hey, fair enough. I just caution against hyperbolic verbage such as "proves" and "definite" and "obvious" and "clearly" and "conclusively".

I do think we can get the raw data in a request for production discovery demand without conclusive proof. We just need to get past the criteria for summary judgement, which is a LOT lower than needed to show a conclusively proven case.

[TruthIsAll]

25. When did I use those words? I just crunch the numbers.

The results speak silently for themselves.

OK, I'LL ADMIT, I RAISE MY VOICE NOW AND THEN TO GET ATTENTION.

[Mistwell]

52. In this thread alone

You said "THIS KNOCKS OUT ALL THE STRAWMEN CREATED OVER THE LAST 2 MONTHS"

Yet, it doesn't. It might be a good response to much of the naysayer claims over the last 2 months, but it does not "knock them all out". It does not "prove" that Kerry "won", though it does form one element in a case to be made that Kerry won.

You gotta know that you use this kind of language here, in most threads that you start.

[Eye_on_prize]

27. Yes yes yes. we work with what we've got to leverage what we need

[bunny planet]

33. Exactly, you go to war with the data you have

not the data you would choose to have.

didn't little Donny Rumsfeld say that or something similar?

[unpossibles]

16. not if you look at it this way

if the published results from 11/03 were weighted with the actual votes, as was stated then the breakdown of votes by party/gender/etc will also be weighted to the same result.

So when they say they ajusted the exit polls to fit the actual results, why don't the demographics reflect that?

[merh]

10. I think I love you TruthIsAll!

You do wonderful work and boy howdy, do I love your sources!!!! Now to post a link to this thread and that lovely graphic to a dem that thinks pursuing election fraud is a waste of time -- fixing the problems of the party is what is important to him :puke:

I love having documentation to support my position - KERRY WON, the party is not in shambles and he (the poster) can go chew his toenails for all I care! :argh:

Thanks again! :loveya: :hug: You are the best!

:yourock:

[New Earth]

12. am i the only one who ever wondered

what TruthIsAll looks like? I'm so used to seeing that JFK avatar that it's grinded in my head that he's as cute as him.....LOL

[merh]

15. LOL, I have wondered the same thing!

But then again, I am curious about so many here. The important thing is the sense of what type of people they are from their posts. Some are very beautiful people and others are nasty, mean people.

[fooj]

21. Another fine job!! Keep the positive energy flowin'

:bounce: Makes me feel like dancing!!! Thanks for your advice in previous thread! Nice to have someone watching my back! :loveya:


WHAT ARE THEY HIDING???:think: :think: :think:

[Bill Bored]

13. I refuse to answer on the grounds that

I've already started ANOTHER THREAD about this! :)
<http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x225847>

[merh]

20. Which thread?

I want to go play.

[meatsack]

18. Why break it down

I don't even get what you are trying to prove with all these. It's all from the same sample of people. They break it down into various groups, but it's still the same sample.

You are just rebuilding the data they broke down. This isn't proving again and again that Kerry won. It's showing that in one poll he won. Of course the Kerry % is similar each time.

............... Bush Kerry Nader
EDUCATION.... 48.05% 50.21% 1.17%
INCOME........48.12% 51.42% 0.95%
RACE..........47.86% 50.94% 1.00%
AGE...........48.17% 50.53% 1.00%
PARTY.........48.24% 50.24% 0.90%
IDEOLOGY.... 48.15% 49.85% 1.00%
RELIGION.... 48.19% 51.55% 1.19%
MILITARY.... 47.62% 51.20% 1.00%
DATEDECIDE 47.95% 51.23% 0.54%
ISSUES...... 47.92% 50.80% 1.28%

If you had access to the raw data it would be EXACTLY the same number in each column. Cause it's all from the SAME SAMPLE.


I'm not saying this exit poll is wrong. I'm saying this analysis is doing nothing. I believe in the exit polls, but all this analysis is telling us nothing, merely rebuilding the poll after they broke it down.
http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=224309&mesg_id=225567

[unpossibles]

19. again, what this shows is that they are not consistent

if the published results from 11/03 were adjusted with the actual votes, as was stated as why the polls were wrong, then the breakdown of votes by party/gender/etc will also be weighted to the same result in the final published results, right?

So when they adjusted the exit polls to fit the actual results, why don't the demographics reflect that?

[TruthIsAll]

22. Because they DID NOT KNOW how to do it QUICKLY.

Too many relationships, too many factors, they would have to break the internal consistency in all the characteristics.

These guys feel they can get away with EVERYTHING.

They got careless. But the Internet is more powerful than anything they have been up against. This is not Shock and Awe. The data was downloaded by sharp-eyed bloggers, so they could not backtrack and complete the full demographic transformation.

Things were moving too quickly. The dynamic was changing. Kerry's margins were increasing. Big turnout skewed the samples.

There was no time for subtlety. They used a sledgehammer on a finely tuned electorate. They broke it quickly, but they could not put it back together again as fast.

One analogy is cheating on the final exam. You can easily copy the answer but not the full derivation or proof.

[meatsack]

24. isn't this the unadjusted numbers

These are the unadjusted numbers. It's blindingly clear that these show a Kerry victory.

If you did the same analysis with these numbers:
http://us.cnn.com/ELECTION/2004/pages/results/states/US/P/00/epolls.0.html
and found a category that indicated Kerry won, then you've got a story. Then that would show cheating. At the moment all you've done is pull apart the unadjusted polls and put them back together. Your math isn't wrong, but the idea is pointless.

I'd just like to re-emphasize I've like your previous works, this one just isn't really showing us anything.

[TruthIsAll]

30. It's not pointless. It shows polling consistency and confirmation, n/t

.

[meatsack]

55. what?

I'm starting to doubt you really understand polling.

If there are 1000 people in the room, and 600 voted for candidate 'y' and 400 for candidate 'x'. I can ask them a series of questions breaking them up into sub groups & demographics. No matter how the groups are sorted, there are still 600 'x' supporters and 400 'y' supporters. That's all you've done. The reason the numbers aren't exactly the same is that they've dropped the decimal places from the percentages.

[Pobeka]

26. The 2000 presidential vote breakdown also shows it clearly.

Kerry got 57% of those who didn't vote in 2000 -- Bush only got 41%

More voters stayed with the dems (91%), than stayed with the reps (90%).

Given that Gore won the popular vote in 2000, Kerry had to get even more of a margin of the popular vote based on these numbers.

[Carolab]

29. Looking at Kerry's gray areas and Bush's gray areas makes me realize

that Kerry was campaigning as he was in order to capture "Bush's base". It's pretty clear to me that the "powers that be" in either party are concerned with wealthy, college-trained, protestant, married white males and females. That's the base their "special interests" (i.e., corporations) care most about. Kerry could have served them just as well as Bush but could also have managed to put the "dominionists" out of power. I doubt that he and the DNC/DLC would have done much to help the true Democratic base, however. They would have lost their "money backing" if they did, after all. And I am sure they didn't want to have to deal with the "progressive push" from the base, either...probably figure it's easier to push back the right wing extremists (and let the party moderates help that along) and then shoot for 2008. I just don't see how they think they can play for any more time, with things as bad as they are.

I read an article earlier that both parties have engaged in exit poll and actual vote tally manipulations. I guess that the polls/tallies could be managed in favor of the Democrats next time if the machines are removed from the Republicans' control. But it doesn't mean that winning will make them any more responsive to the true base.

[brindis_desala]

32. excellent analysis! You hit the nail Smack on the head!

That's why this issue is so important. The Democratic leadership has to decide- do they want to preserve white privilege (real and imagined) or defend democracy. It's really that simple... but it's not going to be easy...

[Walt Starr]

34. The exit poll data was WRONG

Until you have concrete evidence to the contrary, it's all just bullshit.

[Carolab]

36. The exit poll data we are discussing was "unretouched"

they massaged it later to conform to the "actual" votes, remember?

The polls were right until they skewed them.

[Walt Starr]

37. Where is this data

You don't have access to it!

[TruthIsAll]

38. The WP/NEP data is right there in the orginal post. n/t

If I can read it in the WP site, and its from Edison/Mitofsky, I have enough access to use it, no?

What are you not seeing?

[Salomonity]

53. How can you know?

The NEP people asked a few thousand people who they voted for, and various other questions.

If they asked a true random sample, we'd have proof of a stolen election.

They didn't. First, because of clustering--the various voters sampled tended to come from the same precincts, so that radically enlarges the margin of error. If one of the polled precincts happens not to be representative, the poll is no good.

Suppose you've polled 15 precints in a given state. That just tells you what happens in each of them--the problem is, there's no particular reason to think that those precints are a representative sample of precincts in the state. I'm unaware of them having been chosen randomly; my understanding is that Mitofsky picks them for his own reasons.

They don't have to be representative, because the raw results aren't actually used by anyone. As election returns come in, they weight some precincts more than others, to control for increased or decreased turnout in some regions.

The reason exit polls are accurate is that after the fact, you can adjust the poll to make the presidential vote numbers coincide with the real data, and then hopefully the other columns also line up.

[m.standridge]

39. the most convincing scenario to me

is that Kerry won in the Electoral College.
And only in the Electoral College.
But there is enough, among the various exit polls, to suggest this.
Ohio was pivotal. But I think Kerry probably also carried New Mexico, though possibly even closer than Gore got it before.
And I think he also carried Nevada this time, though it was so hair-close that it might understandably be awhile before that can be confirmed in post-election analyses.

It's in Ohio, though, that the real answer lies. There are so many things about Ohio, entirely separate from the statistical analyses, to make it look crooked:
1. In late October "someone" broke into Democratic offices in Toledo and stole data out of computers. NO MONEY WAS TAKEN.
2. During the original election counting, the Warren County elections office was closed to non-GOP observers for at least several minutes in the name of "Homeland Security". The office claimed the FBI had warned them of a possible pending terror attack. THE FBI SAYS NO SUCH WARNING WAS ISSUED. Meantime, for several minutes, ID was required for all observers, while GOP counters alone were in the building. The only Democrat in the building, was there at the behest of Sec'y of State Blackwell.
3. TRIAD computer techs entered at least two county election buildings, took data out of computers and encouraged the staff to engage in what would be election fraud by using original tabulations instead of any recounted or hand counted tabulations, all going into the recount.
4. Voting machines were withheld from heavily-Democratic precincts in Cuyahoga county, for no clear or justifiable reason, and against the requests for the machines by local officials.

Stop there. What does that look like to you? Doesn't it look like someone was trying to cheat in Ohio?
Don't be naive. We don't need statistics to prove this thing. It was in the headlines of the news media. Bush was cheating to carry Ohio, otherwise he wasn't going to win in the Electoral College.

[Walt Starr]

41. If Kerry won in the electoral college

We'd be calling him president-elect Kerry.

He lost in the electoral college. Fraud or no fraud there was n o proof in time to alter the outcome, ergo, Kerry lost in the electoral college.

[Name removed]

43. Deleted message

Message removed by moderator. Click here to review the message board rules.

[Walt Starr]

44. TIA, Kerry DID NOT WIN!

Face the facts. HE lost. Bush won. If you had something concrete, December 13 was the LAST day you had to use that evidence to alter that fact.

Personal attacks about smoking something notwithstanding!

[davidgmills]

48. Depends on what you mean by winning

Bush won the mechanical tabulation

Kerry won the intent of the people

[Snivi Yllom]

42. LOLOLOL

OMG, you are on a roll today!

Cutting through the cyberfog with a 100 million candlepower spotlight.

[TruthIsAll]

51. Mollie Ivins, one of my favorites.Great to have you here. n/t

tia

[me b zola]

46. The election was fixed

Until you have evidence to the contrary, it's all bullshit.:hi:

[deacon2]

54. Walt, exit polling is historically "accurate" within small MOEs

that's why election overseers use it to detect fraud. Sadly, it takes more faith to maintain your viewpoint than it does to acknowledge the obvious: the polling numbers point toward massive "irregularities" and they should be investigated and explained. That's democracy. Then you can say Kerry lost - as you seem to enjoy saying so much - and we can all kiss your ass. Okay?

[chorti]

40. Gore voters much less likely to vote in 2004

According to the final (fixed) exit poll data, Bush won the election not because he kept his voters from switching over, or grabbed Gore voters away from Kerry, or gained new voters, but rather he was able to better mobilize his own base to return to the polls. I don't find much anecdotal evidence to support this but this is what the 'fixed' exit polls suggest - New voters went overwhelming to Kerry (54 to 45); about an equal number of Gore voters switched to Bush as Bush (2000) voters switched to Kerry; Nader, etal (2000) voters went overwhelming to Kerry (71 to 21 to 8 other).

So how did Bush win? His 2000 voters were more likely to return to the polls in 2004 (regardless of who they ended up voting for). Let us assume that for every 100 Bush voters in 2000, 95 returned to the polls in 2004. (Some passed away, others did not vote for whatever reason.) Then we can say, if we are to believe the final exit polls, that for every 100 Gore voters in 2000, only 81 returned to the polls in 2004. (Ranging from 80 in the West region to a little over 82 in the South.)

Why this happened should be the great debate in the Democratic Party. I can think of only 3 reasons.

1) Among the year 2000 voters, the Democrats had many more elderly voters who passed away from 2000 to 2004. There is a little evidence for this ... Bush this time won the 65+ crowd 54 to 46 compared to him getting only 47% of the 65+ vote in 2000. But it is hard to imagine how this would create such a huge gap in return voters.

2) Many Gore voters were disallusioned and stayed home in 2004. I don't see this. I just don't see the anecdotal evidence that this happened. Perhaps someone can inform me otherwise. I can understand why it might have happened, given the candidate and the party, etc. But what I was amazed at was how many people were totally disallusioned with Kerry/the Dems, etc. and yet decided they absolutely had to get out the vote for Kerry.

3) Fraud.

Any other reasons out there? Why were Bush 2000 voters 13 to 15 percent more likely to return to the polls in 2004 than Gore 2000 voters? Anyone...

[rodeodance]

47. so-more women voted for Kerry than Bush. also

[davidgmills]

49. TIA this is funny

Your nemisis on Stones Cry Out just emailed me at 2:55 Central time saying:

"david, it is difficult to continue a discussion with you when you don't read what others have written and you simply do not understand statistics."

I got to email him back and tell him that I can read this and it says the MOE is 1% and it was the day after the election and before the spin began.

[TruthIsAll]

50. This WP summary utterly destroys them. Their heads are exploding.

David, they have spun themselves into oblivion.