KERRY'S WINNING WEIGHTS: PARTY ID, IDEOLOGY, RELIGION, MILITARY, ISSUES

[TruthIsAll]

Once again, the internal consistency of the first 80% of the
exit polls (before the final contaminated 20%) is confirmed. 

Just weight the numbers. Kerry wins. 
Every which way to Sunday.

Average of the categories (party, ideology, religion, military
background, most important issue, when the voter decided):
Kerry: 50.81%
Bush: 48.01% 
Nader: 0.98%

In the prior thread (Race, Gender, Education, Age, Income),
the average worked out to: 
Kerry: 50.78%
Bush: 47.99%
Nader: 1.02%

I love the consistency, don't you?

Categ  Bush	Kerry	Nader
PARTY	48.2%	50.2%	0.9%
IDEOL	48.2%	49.9%	1.0%
RELIG	48.2%	51.6%	1.2%
MILIT	47.6%	51.2%	1.0%
DECID 48.0%	51.2%	0.5%				
ISSUE	47.9%	50.8%	1.3%				

AVG	48.01% 50.81% 0.98%				
							
							
							
		HORIZONTAL			WEIGHTED		
PARTYID	Vertical	Bush	Kerry	Nader	Bush	Kerry	Nader
Dem    	     38%	9%	90%	1%	3.4%	34.2%	0.4%
Repub 	     36%	92%	7%	0%	33.1%	2.5%	0.0%
Indep 	      26%	45%	52%	2%	11.7%	13.5%	0.5%

	       100%				48.2%	50.2%	0.9%
							
IDEOLOGY							
Liberal 	22%	12%	86%	1%	2.6%	18.9%	0.2%
Moderate 	45%	41%	57%	1%	18.5%	25.7%	0.5%
Conserv 	33%	82%	16%	1%	27.1%	5.3%	0.3%

	      100%				48.2%	49.9%	1.0%
							
RELIGION							
Protestant	53%	56%	43%	1%	29.7%	22.8%	0.5%
Catholic 	27%	49%	50%	1%	13.2%	13.5%	0.3%
Jewish	          3%	23%	77%	0%	0.7%	2.3%	0.0%
Other 	          7%	20%	75%	4%	1.4%	5.3%	0.3%
None 	         11%	29%	70%	1%	3.2%	7.7%	0.1%

	     100%				48.2%	51.6%	1.2%
							
							
MILITARY							
Yes 	      18%	55%	43%	1%	9.9%	7.7%	0.2%
No 	      82%	46%	53%	1%	37.7%	43.5%	0.8%

	      100%				47.6%	51.2%	1.0%
							
							
WHEN DID YOU DECIDE?							
Today 	          6%	40%	53%	5%	2.4%	3.2%	0.3%
Last3 days 	3%	41%	53%	4%	1.2%	1.6%	0.1%
Last Week 	2%	51%	48%	1%	1.0%	1.0%	0.0%
Last Month 	10%	38%	60%	1%	3.8%	6.0%	0.1%
Before 	         79%	50%	50%	0%	39.5%	39.5%	0.0%

	     100%				48.0%	51.2%	0.5%


MOST IMPORTANT ISSUE							
Taxes 	       5%	53%	47%	0%	2.7%	2.4%	0.0%
Education 	4%	24%	75%	0%	1.0%	3.0%	0.0%
Iraq 	      15%	23%	76%	1%	3.5%	11.4%	0.2%
Terrorism 	18%	85%	15%	1%	15.3%	2.7%	0.2%
Econ/Jobs 	20%	16%	82%	1%	3.2%	16.4%	0.2%
MoralVals 	21%	78%	19%	3%	16.4%	4.0%	0.6%
HealthCare 	8%	20%	79%	0%	1.6%	6.3%	0.0%

	     90.9%				43.5%	46.2%	1.2%
							

[Old and In the Way]

1. Have a question for you on this, TIA.

I just posted in a separate thread, but I'd be interested in your comment-

From what precincts/states would the "fixed" samples have come from? The reason I ask is that it would seem that the last data samples would be coming from the West Coast. We know Kerry won those states, so why should that sampling data be different? If anything, it should have been even more favorable to Kerry than Bush.

[TruthIsAll]

2. I don't have the answer, but I agree 100% with your logic

In fact, I have read that Kerry's margin was INCREASING as the evening progressed - probably due to the West Coast coming in.

[Quakerfriend]

20. Once again, Thx TIA for all your hard work!

And, BTW, for all of you who know nothing about stats: When exit polling are done on such a large # of people they are very, very accurate. :9 Repeated sampling increases validity!! And, I'm quite certain that these exit polls are designed to include the # of samples that will provide for 99% accuracy.

[TruthIsAll]

34. GRAPH: NATIONAL EXIT POLL WEIGHTED BY CHARACTERISTIC

[bones_7672]

3. AH GEEZ! ANOTHER EGGHEAD STAT SHEET THAT MEANS SHEEET!

The vote is over. The certification of electors is over. There is not a court in this country that will allow even the aroma of an exit poll into its courtroom. NOT NOW, NOT EVER!! After Thursday, GWB is Prez, and, again, NO COURT, NO FEDERAL COMMISSION, WILL ALLOW AN EXIT POLL AS ANY KIND OF EVIDENCE, NOT NOW, NOT EVER. Rant over, now it's back to Casablanca.

[Darkhawk32]

4. If you don't like TIA's threads, then don't freakin' read them! n/t

[TruthIsAll]

6. Are you in panic mode or what? Chill out. Just numbers.

What a joke.

[euler]

42. What a joke. Yes, I agree

[mod mom]

9. It's not just the statistics-statistics helps strengthen the argument.

[Imajika]

26. Actually...

if I was in any way convinced by TIA's research, I wouldn't care if Bush was certain to be inagurated on Thursday or not. I would be screaming at the top of my lungs to anyone who couldn't completely ignore me that the election was stolen. Further, I'd be at every anti-fraud rally I could attend, calling every law enforcement agency that was even remotely relevent, likely be arrested for civil disobedience, calling for mass uprisings, and otherwise doing everything I could to see Kerry rightfully take his place as President.

I also don't believe that no court or federal commission would not allow exit polls as any kind of evidence. I actually believe they might. But exit polls alone, and statistical analysis of them, just don't cut it. Real evidence would have to emerge as the primary proof of fraud. Not just any evidence either - clear, obvious, understandable proof of the fraud conspiracy would have to be shown. Exit polls would just be one part of any fraud case. The evidence would have to hold up in court, to intense scrutiny, and be convincing to some in the MSM and perhaps even some Republican partisans.

So I don't see anything wrong, in principle, with TIA examing data and posting some of his findings. The problem I have is that I don't really believe TIA knows what he is talking about. He may be great at mathematics and tireless at plugging numbers into Excel, but he doesn't understand polling and polling results. He doesn't seem to accept that his assumptions could be fundamentally in error. I'm not sure if he even really knows if the data he uses is complete or correct. I don't believe he understands elections or comprehends that there are unbelievably good election experts working for the Democratic Party and Kerry campaign who understand the ins and outs of elections thousands of times better than himself and simply do not buy into fraud theories. Lastly, he doesn't really have a grounded view of American politics, how rapidly political fortunes change, and how an electorate is dynamic rather than static. For TIA it is enough to believe something, and then take whatever data he has and try to use it to prove he is correct.

"The vote is over. The certification of electors is over."

True, but that doesn't mean the numbers shouldn't be analyzed as a way of determining what what wrong, and where if an investigaton were to take place would be the most effective areas to look for fraud.

Imajika

[Name removed]

45. Deleted message

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

[Machiavelli05]

40. LOOK AT HIS SCREEN NAME

Obviously a member of teh Skull and Bones.. quick hang him!

[48pan]

5. The Problem with Exit Polls

Exit polls are voluntary. The willingness of different types of voters to answer an exit poll skews the results. People who are pissed off by the government are much more vocal than those who are satisfied and thus are far more likely to agree to answer an exit poll.

Angry people seek out the exit pollers. Satisfied people just want to go home.

The exit polls were wrong because the sample was skewed. That is the most obvious and most likely reason the votes didn't match the polls.

[TruthIsAll]

7. Is that so? Bizarro World. n/t

.

[48pan]

8. It was at my polling place.

The activists (young Democrats) sought out the exit pollers. When people they knew approached the polls, they mentioned the exit polling and reminded them to try to be part of it. The Republicans were not doing the same thing.

At least in my ward, the exit polls were a joke.

[mod mom]

10. How do you know the repugs weren't?

[48pan]

13. In my ward we all stand in the same spot. I stood there for...

6 hours and our guys were making a point of talking to the exit pollers. The Republicans weren't. We were loud, excited and angry. They weren't. Our voters were young and energetic. Theirs were older and more sedintary.

It was a running joke all day that we were going to get Kerry up to 90 percent in the exit polls. I don't know how the exit polls came out in my ward, but I suspect we had a serious effect on the results.

[Darkhawk32]

11. Dude, that is why there is "weighting"

The numbers TIA has are weighted numbers.

[clydefrand]

12. Educate me please, explain weighted numbers. Thanks, Fran

[48pan]

14. Here's why that wasn't working.

We told our people to claim to be Republicans who had voted for Kerry. We were screwing with the exit poll numbers in every way possible. It was something to do other than stand and hold a sign.

[Darkhawk32]

16. They factor that behavior out too.... you were nullified n/t

[48pan]

25. How did they do that?

How can they factor out people who lie to them? How do they know who lied? Answer: They don't. They just make money doing exit polls so they don't want to admit that the exit polls aren't scientific, reliable or accurate.

[Darkhawk32]

29. They've factored into the equation a % of lying and such.... n/t

[48pan]

32. How did they do that?

They have no clue who is lying or how many. No stats course I've taken has addressed unknown error. The only way to factor it out is to choose a purely random sample large enough to not be skewed. Unless you know the source of error before polling, you can't factor it out. The turnout in this election was different than any since exit polling started. The exit pollers weren't ready for what happened. Their sample wasn't random. They blew it.

This is, of course, just my opinion and you are welcome to disagree. I think the fact that exit polls throughout the US were all wrong supports my view. Ohio and Florida were not remarkably different from other states. Here in New Hampshire, the totals were much closer than the exit polls indicated.

[kerstin]

43. Huh?

The exit polls were not off nationwide. Only in the states where there was no paper trail. Plus even though Kerry won New Hampshire, New Hampshire showed the largest differential between the exit poll figures and the vote tallies (9%, which is inconceivable).

Just who are you anyway?

[48pan]

46. So what you are saying is...

NH votes have no paper trail and there was massive vote fraud in NH?

Uh... I hate to tell you this, but there are no black box machines in the whole state. Large parts of the state are paper balloting only. The highest tech voting we have is optical scanning machines with paper ballots that are easily checked and recounted.

The fact that NH exit polls were lousy pretty much proves that the exit polls were unreliable and inaccurate.

I'm an employee of the University of New Hampshire and I live in a small city of about 30,000 residents. We had a voter turnout above 80% here. The close elections for President and Governor were the main draw. We elected a Democratic Governor by about the same margin that we went for Kerry. NH went for Bush in 2000.

The only real voter fraud I see in our area is non-resident college students (largely Democratic) voting in Durham and Dover (towns near the college). They should be voting by absentee ballot in their own towns and states. On balance, it favors the Democrats.

[48pan]

50. I found this at Slate - About Exit Poll Errors

http://slate.com/Default.aspx?id=2111460&MSID=E2159B3DB4E440728A0614F509B5724D

[New Earth]

17. are we still on planet Earth?

:shrug: :hi:

[TruthIsAll]

18. I see, 48pan. From your one experience, you extrapolate the RULE

Makes a lot of sense.

The 48pan Rule: Democrats holler, Repukes just slither away.

Love that robust analysis.

[burned]

23. sssssss

:)

[48pan]

24. I can only report what I see.

We scammed the exit polls. It was easy. Please don't shoot the messenger.

[merh]

28. And what fraternity are you with?

:shrug:

[kerstin]

48. Since you're evidently so proud of "scamming the exit polls"...

how do we know you're not scamming us now?

The fact remains that the exit poll figures were not outside the margin of error in every state as you imply, and certainly not to a consistent degree as you also seem to be suggesting.

[Old and In the Way]

36. Proof for this assertion?

Why would we possibly do that? I heard a lot of kooky theories, but this wasn't ever suggested on the board before the election....and if it had, most here would have counseled against it.

Sorry, the old people I know that screw with elections are Republicans.

[Mistwell]

44. Not so

That is not so. In fact, full weighting WAS the change made to the final exit poll. The reason this was a preliminary poll was because full weighting had NOT happened yet, in addition to 20% of that days vote along with all absentee votes were not included. In the final poll, full weighting was applied, and suddenly the poll matched the actual reported vote within the margin of error.

[dzika]

15. Did you take any pics or other details? I would like to archive it.

The more evidence you can provide the better. There may be something in there to help our case... can you tell us which ward and state? Were you there all day? How long did you observe this behavior?

Please, be specific. Maybe you should start a new thread about this.

(TIA, thanks for slicing and dicing. I love it.)

[clydefrand]

19. How can this help our case?

"The more evidence you can provide the better. There may be something in there to help our case"

Do you mean skewing the exit poll data by lying helps our case?

Please explain your statement. Just trying to understand.

[dzika]

22. I understand your concern.

If 48pan has valid details that would show that exit polls are not valid then let them be posted. So far, I haven't seen any evidence that can show this.

I just don't think that should be subject of this particular thread.

Sorry if my wording was wrong.

[1democracy]

21. Another topic...

I believe the exit polls. And I love number analysis.

What numbers did Rove have? Was he watching the exit polls on his computer on Nov 2? What kind of data was he getting? Did he do his own poll? On what basis would he call the TV stations to tell them they had it wrong? Was he getting actual ballot numbers? Can we get a copy of his data???(Ha)Inquiring minds want to know!

[48pan]

30. The most accurate numbers

are the turnout numbers in the individual wards. Wards tend to be one sided. For instance, in my town, wards 2,3, and 5 are heavily Democratic and wards 1, 4 and 6 are heavily Republican. If wards 2,3 and 5 turn out more heavily than 1,4 and 6, we know that we will win. We don't have to count a single vote to know the results.

From what I've read, Rove was using the turnout data by ward to predict the results. His people did a vast amount of canvassing in heavily Republican areas in the battleground states. Then, they could use the count of voters, which was readily availble and verifiable, to predict the results.

We do the same thing here every two years. At about noon, we take the counts at the polls. If the numbers in our wards aren't good, we get on the phone lines and get our people out. We also have our own people inside the polling places recording who has voted and who has not. At 2PM each voting day, we start calling and visiting Democrats who have not voted yet. Using this technique, we have won some elections by getting our voters out later in the day.

Exit pollers can't adjust for this factor. A good organization can make a big difference in the results in a few hours.

[48pan]

27. Why would I take pictures of our own people scamming...

some moron exit poller?

No. We didn't take pictures or document scamming of exit pollers. We loathe exit pollers. Look what happened in Florida in 2000. The press used exit polls to call the election early, before voting was even closed. Some of us think exit polling should be abolished.

I think the press should not be reporting results while voting is still taking place.

[Old and In the Way]

37. You loathe exit polls?

We vote on Republican machines. We have Republican SOS who moonlight as Bush-Cheney Re-Elect chairpersons. We have a corporate media that is biased towards the Republican Party....and you loathe the one indicator that might deliver us an honest picture of how the voters really voted?

And you accuse us of rigging the exit polls.

We need more Democrats like you in the Party. :eyes:

[Old and In the Way]

35. So Republican's wouldn't have the same motivation to let the pollsters

know who they voted for? They wouldn't be just as motivated to make sure that their votes weren't stolen by Democrats?

I kind of doubt it. But if I was on the stealing end, I'd sure be pushing this explantion.

[stirringstill]

31. Pollster dodging

Exit polls can be skewed for a number of reasons such as choice of precincts polled but the reason you noted is nothing but spin. Historically Republican pollster dodging has never been shown to be the case for skewing exit polls which have been by Mitofsky for decades. In addition, exit polls are designed to minimize and to correct for people not wanting to be interviewed. Non-responders are logged as attempts and usually the gender is also noted. Well, Mitofsky's raw data which the networks will not release would contain evidence of such a bias yet Mitofsky and the Republican spinners have spouted this as a possibility and yet offered no proof. Till they pony up, they are only spinning. So the "obvious and most likely reason" is bullshit until proven not be.

[48pan]

33. Whatever you say.

I think calling the exit polls accurate and the actual counts wrong is bullshit until proven otherwise. We saw no evidence of vote fraud here but the exit polls wern't close. They gave Kerry victory by a wide margin and the final vote count was quite close.

I guess it's just a matter of opinion. To each his own.

[me b zola]

38. You've got it back wards

The election can not be validated until proven. We have been screaming for fair and transparent elections. If they can't prove it, it is not valid, if not valid, then we need a revote. Simple as pie, if you believe in Democracy.

on edit: Due to my urge to puke at a couple of the responses, I almost forgot to thank-you,TIA. Once again you come through.:toast:

[48pan]

47. I see it as proven

"The election can not be validated until proven. We have been screaming for fair and transparent elections. If they can't prove it, it is not valid, if not valid, then we need a revote. Simple as pie, if you believe in Democracy."

Every recount has shown that Bush won Ohio. Since they don't use black box voting, the burden of proof has been met. The evidence is in. No analysis of invalid exit polls is going to change the actual count of paper ballots.

We did a lousy job in Ohio. The Republicans did a better job of getting their votes out and we lost. We have to move on and look to the next Congressional election. If we sit around and whine, we look like sore losers.

In my opinion, that hurts us in future votes.

I'm with you in a big way in Florida. Black box voting without paper ballots in unacceptable and we should make sure it is banned before the next election. Any good programmer worth his salt can set the program to count every third vote for one candidate to be counted for another. It would be easy to make subtle changes in the program that change close election results without being obvious. All Florida black box machines should be confiscated and checked.

[IndyOp]

39. Very nice! Thanks TIA for helping make the world make sense again! (n/t)

[meatsack]

41. Somethings wrong then.

These percentages should be exactly the same.
Categ Bush Kerry Nader
PARTY 48.2% 50.2% 0.9%
IDEOL 48.2% 49.9% 1.0%
RELIG 48.2% 51.6% 1.2%
MILIT 47.6% 51.2% 1.0%
DECID 48.0% 51.2% 0.5%
ISSUE 47.9% 50.8% 1.3%

Bush's column should remain the same percentage, and Kerry's column should be the same all the way down.

If all this data is from the same survey and every respondent answered every question, there shouldn't be any difference.

example. 1000 people are polled
510 for Kerry -51%
480 for Bush -48%
10 other -1%

now lets break it down by sexes. If you don't add or remove anyone from the sample, you have to end up with 510 people voting for Kerry regardless of sex, and 480 people voting for Bush. (these numbers are just examples)
Kerry breakdown
280 women, 230 men -510 people total
Bush breakdown
240 women, 240 men - 480 people total
Other breakdown
5 women, 5 men
Breakdown 47.5% male 52.5% female
Male breakdown
48.4% Kerry, 50.5% Bush 1.1% other
Female breakdown
53.3% Kerry, 45.7% Bush 0.9% other

Now if we were to just do a straight averaging of male/ female vote preference we would get
50.85% Kerry 48.1% Bush. But we know the hard numbers in this case should show exactly 51% Kerry 48% Bush.
Bush got a slight advantage as we gave the male sample the same weight as the female sample. But since there were more females voting we can
'weight' their numbers to show the increased turnout.
27.98% of voters, voted for Kerry and were women
22.99% of voters, voted for Kerry and were men
50.97% for Kerry, (the imprecision is me working with only 1 to 2 significant digits.) That's more like it, that matches what we know to be true.

Now if I did the same thing, with a 'closed' sample, no matter what the question was, (military, income, etc) as long as everyone answered my weighted results should always equal:
51% Kerry, 48% Bush, 1% other.

So why are TIA's numbers different? They shouldn't be 'consistently' close, they should be EXACTLY the same if they're from the same exit poll. While the poll, on the whole shows fraud compared to results, this analysis is pretty much worthless. (TIA I like your previous work, but I'm not sure what you are trying to show with this one, of course the weighted numbers will look the same, they're from the same poll!)

What's interesting is the inconsistency of the numbers, but that could probably be attributed to people not answering all the questions and therefore a different sample size for the categories.

[TruthIsAll]

49. The numbers should be exactly the same? Really? Prove it.

They are close, as you would expect.
That they are NOT exactly the same makes the analysis all the more real

For them to be exactly the same the numbers would have to be fudged.
Do you know how to fudge the numbers to make them the same?

Show us.

THEY SHOULD NOT BE EXACTLY THE SAME. THE REASON?
THESE ARE EFFECTIVELY DIFFERENT POLLS OF SEPARATE VOTER CHARACTERISTICS.

WHY WOULD YOU EXPECT DIFFERENT POLLS TO MATCH EXACTLY WHEN WEIGHTED?

[GregD]

51. So, what then are the odds of a unheard tree falling in a forest?

I think it is absolutely marvelous that TIA has gone to this effort. But I really would ask, what now? What is being done with all this wonderful work.

I've asked TIA privately whether all this was being gathered in a single web site, and to contact people who have high level contacts (who I can connect you with). To these PMs I have received either the briefest responses, or no response at all.

So, I ask again. What is being done with all this? Is there a central site that has posted all of it so that serious investigators who will not spent hours and hours digging through the archives and various sections of DU to find them all? Has any of it been forwarded to the Judiciary Committee? Has a printed presentation of all this been prepared?

Sorry to be so direct, but I see a lot of talking and squawking by DUers aplenty, but not much that convinces me that this clearly valuable information is being used as ammunition in this battle to win the hearts and minds of that evasive Senator that will stand for us on Thursday.

Is anyone carefully organizing all of TIA's research in a clean and thoughtful layout, ready for printing/binding/presentation?

[Needtodosomething]

52. BUT BUT BUT

But come on Bush won so many christian votes....we need to be more godly...not like the black and hispanics who voted against him...I mean none of them are christians or catholics...Right???

Please note the sarcasm.