KERRY WON: No matter how you WEIGHT (RACE,GENDER, EDUCATION, AGE,INCOME)

[TruthIsAll]

From the FIRST PRISTINE 11,000 polled by 7:30:


VOTE BY EDUCATION							
		2004					
							
TOTAL							
							
	            Vertical  Bush	Kerry	Nader	 Bush	Kerry Nader
No High School (4%)	4%	47%	52%	1%	1.9%	2.1%	0.0%
H.S. Graduate (22%)	22%	48%	51%	1%	10.6%	11.2%	0.2%
Some College (31%)	31%	51%	47%	1%	15.8%	14.6%	0.3%
College Grad (26%)	26%	50%	48%	1%	13.0%	12.5%	0.3%
Postgrad Study (17%)	17%	40%	58%	2%	6.8%	9.9%	0.3%
							
	   				             48.1%    50.2%	1.2%
							
							
VOTE BY GENDER							
							
Male (46%)	          46%	51%	47%	1%	23.5%	21.6%	0.5%
Female (54%)	         54%	45%	54%	1%	24.3%	29.2%	0.5%
							
					              47.8%   50.8%	1.0%
							
VOTE BY RACE AND GENDER							
White Men (36%)	  36%	58%	41%	1%	20.9%	14.8%	0.4%
White Women (41%)	  41%	52%	47%	1%	21.3%	19.3%	0.4%
Non-White Men (10%)	10%	28%	69%	1%	2.8%	6.9%	0.1%
Non-White Women (13%)	13%	22%	77%	1%	2.9%	10.0%	0.1%
							
					               47.9%	50.9%	1.0%

VOTE BY AGE							
18-29 (17%)       	17%	43%	56%	1%	7.3%	9.5%	0.2%
30-44 (27%)	        27% 	50%	49%	1%	13.5%	13.2%	0.3%
45-59 (30%)	        30%	47%	51%	1%	14.1%	15.3%	0.3%
60 and Older (26%)	26%	51%	48%	1%	13.3%	12.5%	0.3%
							
					              48.2%  50.5%	1.0%


VOTE BY INCOME							
Under $15,000 (9%)   9%	33%	66%	1%	3.0%	5.9%	0.1%
$15-30,000 (15%)	  15%	39%	59%	1%	5.9%	8.9%	0.2%
$30-50,000 (22%)	  22%	47%	52%	1%	10.3%	11.4%	0.2%
$50-75,000 (23%)	  23%	53%	45%	1%	12.2%	10.4%	0.2%
$75-100,000 (13%)	  13%	50%	49%	0%	6.5%	6.4%	0.0%
$100-150,000 (11%) 11%	53%	45%	2%	5.8%	5.0%	0.2%
$150-200,000 (4%)	   4%	53%	47%	0%	2.1%	1.9%	0.0%
$200,000 or More (4%)  4%	58%	41%	1%	2.3%	1.6%	0.0%
							
					                48.1%   51.4%	1.0%

[goclark]

1. Keep It Coming TIA nt

[TruthIsAll]

20. TO SUMMARIZE: NOTE THE CLOSENESS

Category	Bush Kerry	Nader

GENDER	      47.76%	50.78%	1.00%
EDUCATION	48.05%	50.21%	1.17%
INCOME	      48.12%	51.42%	0.95%
RACE/GENDER	47.86%	50.94%	1.00%
AGE	      48.17%	50.53%	1.00%
			
Average	      47.99%	50.78%	1.02%
Median	      48.05%	50.78%	1.00%

[molly]

24. The most surprising to me was Income

considering the tax cuts.

[Sick_of_Rethuggery]

2. Inspired by your tenacity,

I was just going to do this exact calculation today (occurred in sleep last night); you beat me to it, as usual. :-)

We need to take this to the Dems in Congress and reasonable Republicans...

[Name removed]

3. Deleted message

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

[Kansas Wyatt]

6. Shrub also captured all of the imaginary votes.....

That gave him the election.

[Carolab]

8. Whaaaaa?

Where do you get that?

[troubleinwinter]

4. TIA, how much of your calculations are you forwarding to Conyers?

LOTTTTTTS, I hope.

[TruthIsAll]

7. YOU can forward it. I'm too lazy. n/t

.

[troubleinwinter]

10. "Lazy"? That's a laugh!!!!

I hope SOMEONE has been forwarding them!! I don't understand most of them myself (just get the general drift), and wouldn't know what exact analyses to send.

[rwork]

5. Do you take

contributions?We really appreciate all you have done.

[Carolab]

9. So much for the polls being "skewed by women", huh?

Look at *'s totals for women.

If you say there were too many women, then *'s totals would go DOWN if you reweighted for that.

[Old and In the Way]

11. I think they used the 2000 election demographics for developing the

sampling plan for the 2004 election.

[corbett]

26. I Thought W Stands For Women!

Oh yeah, underfunding education, reestablishing the global gag rule and hoping to overturn Roe v. Wade are not women-centric positions.

My bad!

[molly]

27. Regression of women to the 1950's

[bleever]

12. Small edit:

data in the rows that start "White men" and "Under $15,000" are shifted one column to the left by accident.

Thanks as ever!

:thumbsup:

[Mistwell]

13. "From the first" skews numbers

You know, I know, and EVERYONE knows that taking only the FIRST numbers that come in does NOT reflect the complete electorate. Indeed, we all know that Democrats tend to vote earlier in the day than Republicans on average, so we know pretty well that the first numbers will not match final numbers.

So, why do you insist on using numbers you KNOW are wrong?

Look TIA, I appreciate the work you do, and I want you to continue. But why do you keep ignoring the hard issues with these things? The hard issue is GETTING THE FINAL EXIT POLL RAW DATA. Until we have that, all you are doing is playing with nearly random numbers that you know for sure from the start do not reflect the entirety of the people who cast votes.

[helderheid]

15. There weren't that many more votes

after these numbers that could skew it as far for * as did on Nov 2nd.

AUDIT!

[Mistwell]

18. Really?

How do you know that? What data do you have that shows there were not that many more votes? As far as I can tell, it's fully about 25% additional voting that happened after that point, and that doesn't include absentee ballots at all. So, from where I stand, it's a TON of people...WAY WAY more than needed to make those numbers meaningless from a margin of error perspective.

[pointsoflight]

21. The final exit poll numbers are based on 13,600.

The numbers posted here are based on 11,000 of those 13,600.

There is, by the way, a later report based on 13,000 responses in which Kerry still leads 51% to 48%. So with just 600 more responses gathered, it somehow switched to 51-48 Bush. That's an impossibility--you can't get that switch even if you give Bush all 600 of the final responses! In order for to get the 6% switch, you have to actually take votes away from Kerry.

[Mistwell]

25. 13600 minus 11000 /= 600

It's 2600 more. 11000 is only 80% of 13600, making it WELL within the margin of error. You still have 20% of the vote to go, along with ALL absentee ballots (which made up a decent chunk of the voting this year in Ohio), and then it doesn't account for ANY demographic stats also collected at the time (which are used to complete exit polls - in all countries). I'm sorry, but this whole thread starts with an inaccurate sample and then tries to make broad statistical generalizations about the entire vote in Ohio by taking less than 80% of the small sample used by the exit poll company, and excluding all demographic adjustments.

You CANNOT make a conclusion about this race based on only part of the exit poll data. It's hard enough to try and make conclusions with ALL the exit poll data...but IMPOSSIBLE without using at least an entire days worth of votes.

[rdmccur]

16. Sigh

[pointsoflight]

19. Those numbers reflect 11,000 of the 13,600 responses.

These are NOT the first numbers that came in. These are numbers from very late on election day.

[Mistwell]

30. It's missing A TON

It's missing all the following:

1. 20% of the remaining exit poll numbers, all from later in the day when Republicans historically vote much heavier than Democrats.

2. All absentee data, which was a higher number of votes than usual and a materially large segment of the total vote.

3. All demographic data used to tweak the numbers FOR ALL EXIT POLLS IN ALL COUNTRIES at the end of the day. In other words, they use two counters - one person counting votes as people exit, and another counting gender, race, etc... as people exit. At the end of the day, they adjust the vote count by the demographic count to come up with a concluding count. NONE of that is present in this data.

And none of that mentions that we don't even have a verified sample to begin with, as all of this seems to be third had stuff.

I'm sorry, but we are using a materially flawed sample.

I WANT the exit poll data to pan out to show CONCLUSIVELY that there was fraud. I'm just sick and tired of continuing to use clearly flawed data to do that. It's enough already. We need the raw data, PERIOD. We can get that with a lawsuit that requests the raw numbers in discovery. That should be the focus. Not this bullshit game people keep playing to make the numbers fit our desires and dismiss the blatantly obvious gaping holes in our analysis.

Are we the party of science, or not? When rightist dismiss global warming data based on the kind of flawed analysis I see in this thread, we call them on the mat for it. What the heck is going on when we start using that same sort of flawed data manipulation to achieve our own ends?

[molly]

31. I've been out of this loop a lot, but it seems to me that

I have seen a lot of numbers - possibly raw data - posted here. Can someone back me up?

[helderheid]

14. add to that

the approval rating of 48% before AND AFTER the election - yeah, sure * won.

My sister in law and her boyfriend visited us from the Netherlands over the holidays. By the time they left, they were convinced Kerry won too. It's amazing how little information is out there on this with the exception of the internet so only those paying attention are getting this!!

[keepthemhonest]

17. you are definately

inspirational, keep em coming.

[fooj]

23. Kudos to a class act patriotic warrior!!

Thank you for all that you do!

"If you lose hope, somehow you lose the vitality that keeps life moving, you lose that courage to be, that quality that helps you go on in spite of it all. And so today I still have a dream..."-MLK Jr.

Your tenacious efforts allow us all to hope and dream for a better tomorrow. :yourock:

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

[chapel hill dem]

22. naive question

We had early voting here in NC and I voted about a week before Nov 2.
The local TV news said about 1/3rd of all North Carolinians voted early.
Did the exit polls include the early voters? Were there separate exit polls for early voters?
Thanks!

[corbett]

28. On The Whole, Early Voters Were Not Polled

In Florida, KKKarl was worried because the early voting numbers went strongly for Kerry but it's my understanding that, given the difficulties in tracking the results, early voters generally were not included in exit polls.

[corbett]

29. Kicking For The Truth!

Keep up the good work!