FL Vote analysis: Using registration, machine type, NEP Party ID stats.

[TruthIsAll]

This analysis indicates that Kerry won Florida. I applied
national exit poll vote percentages based on Party ID to
analyze Kerry's equivalent votes using  Florida registration
percentages as Party ID weights.

This is the breakdown of Florida Votes by Machine Type:
Touchscreen, ESS Optiscanners, Diebold Optiscanners

Reg = registered voters (millions)
Vote/Reg - Total Votes as % of Registered
DemA/E, RepA/E: Actual to Expected vote (based on
Registration)
 

MACHINE	Reg	%DEM	%REP	Vote	%DEM	%REP	Vote/Reg DemA/E RepA/E
Touchsc	5.576	40.89%	36.77%	3.864	51.30%	47.77%	69.3%	79.1%	128.6%
ESSOpti	1.442	41.90%	39.65%	1.050	39.36%	59.95%	72.8%	106.5%	151.3%
DiebOpt	3.283	41.93%	38.69%	2.370	43.56%	55.73%	72.2%	96.1%	143.4%
									
Total	10.301	41.37%	37.79%	7.284	47.06%	52.12%	70.7%	87.5%	136.9%
												
................................................................															
Left: National Exit Poll -
Party ID Weighting 	

Right: Florida Vote data
Voter Registration Weighting

Based on the registration weights: 

Kerry won Florida by 207,000 votes assuming Exit Poll 11027
percentages.							

He won by 223,000 votes assuming Exit Poll 13047
percentages.							
He lost by 29,000 votes assuming Final Exit Poll 13660
percentages.							
			
						Florida	7.284	Total							
7:38pm	11027					7:38pm									
PARTY ID	
   Mix	Bush	Kerry	Nader			Mix	Bush	Kerry	Nader					
Dem	38%	9%	90%	1%		Dem	41.37%	9%	90%	1%					
Rep	36%	92%	7%	1%		Rep	37.79%	92%	7%	1%					
Ind	26%	45%	52%	3%		Ind	20.84%	45%	52%	3%					
															
Total	100%	48.24%	50.24%	1.52%		Total	100%	47.87%	50.72%	1.42%					
	122.26	58.98	61.42	1.86			7.284	3.487	3.694	0.103					
							Kerry	0.207							
															
12:22am	13047					12:22am									
PARTY ID
	Mix	Bush	Kerry	Nader			Mix	Bush	Kerry	Nader					
Dem	38%	9%	90%	1%		Dem	41.37%	9%	90%	1%					
Rep	35%	92%	7%	1%		Rep	37.79%	92%	7%	1%					
Ind	27%	44%	52%	4%		Ind	20.84%	44%	52%	4%					
															
	100%	47.50%	50.69%	1.81%			100%	47.66%	50.72%	1.63%					
	122.26	58.07	61.97	2.21			7.284	3.471	3.694	0.118					
							Kerry	0.223							
															
2:05pm	13660														
	Mix	Bush	Kerry	Nader			Mix	Bush	Kerry	Nader					
Dem	37%	11%	89%	0%		Dem	41.37%	11%	89%	0%					
Rep	37%	93%	6%	1%		Rep	37.79%	93%	6%	1%					
Ind	26%	48%	49%	3%		Ind	20.84%	48%	49%	3%					
															
	100%	50.96%	47.89%	1.15%			100%	49.70%	49.30%	1.00%					
	122.26	62.30	58.55	1.41			7.284	3.620	3.591	0.073					
							Kerry	-0.029							
...............................................................		

This is a breakdown by Machine type, again using percentages
from the 11027, 13047 and 13660 exit poll versions.
													
7:38pm		Touchscreen	3.864			ESSOptiscan	1.050			DieboldOptiscan		2.370	
PARTY ID
	Mix	Bush	Kerry	Nader		Mix	Bush	Kerry	Nader		Mix	Bush	Kerry	Nader	
Dem	41.37%	9%	90%	1%		41.37%	9%	90%	1%		41.37%	9%	90%	1%	
Rep	37.79%	92%	7%	1%		37.79%	92%	7%	1%		37.79%	92%	7%	1%	
Ind	20.84%	45%	52%	3%		20.84%	45%	52%	3%		20.84%	45%	52%	3%	
															
	100%	47.87%	50.72%	1.42%		100%	47.87%	50.72%	1.42%		100%	47.87%	50.72%	1.42%	
	3.864	1.850	1.960	0.055		1.050	0.502	0.532	0.015		2.370	1.135	1.202	0.034	
	Kerry	0.110				Kerry	0.030				Kerry	0.067			



12:22am															
PARTY ID
	Mix	Bush	Kerry	Nader		Mix	Bush	Kerry	Nader		Mix	Bush	Kerry	Nader	
Dem	41.37%	9%	90%	1%		41.37%	9%	90%	1%		41.37%	9%	90%	1%	
Rep	37.79%	92%	7%	1%		37.79%	92%	7%	1%		37.79%	92%	7%	1%	
Ind	20.84%	44%	52%	4%		20.84%	44%	52%	4%		20.84%	44%	52%	4%	
															
	100%	47.66%	50.72%	1.63%		100%	47.66%	50.72%	1.63%		100%	47.66%	50.72%	1.63%	
	3.864	1.841	1.960	0.063		1.050	0.500	0.532	0.017		2.370	1.130	1.202	0.039	
	Kerry	0.118				Kerry	0.032				Kerry	0.072			

										
2:05pm			
PARTY ID
	Mix	Bush	Kerry	Nader		Mix	Bush	Kerry	Nader		Mix	Bush	Kerry	Nader	
Dem	41.37%	11%	89%	0%		41.37%	11%	89%	0%		41.37%	11%	89%	0%	
Rep	37.79%	93%	6%	1%		37.79%	93%	6%	1%		37.79%	93%	6%	1%	
Ind	20.84%	48%	49%	3%		20.84%	48%	49%	3%		20.84%	48%	49%	3%	
															
	100%	49.70%	49.30%	1.00%		100%	49.70%	49.30%	1.00%		100%	49.70%	49.30%	1.00%	
	3.864	1.920	1.905	0.039		1.050	0.522	0.517	0.011		2.370	1.178	1.169	0.024	
	Kerry	-0.015				Kerry	-0.004				Kerry	-0.009			
															

[brettdale]

1. Send this to all media

Pleae please please send this to all media, its only when a large group of people contact the media, will make them listen.

[Carolab]

2. Link to media blaster

http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=104&topic_id=3479316

[TruthIsAll]

3. IS THIS A COINCIDENCE OR WHAT?

Wasn't Bush leading the final Florida Exit Poll by something very close to 49.7-49.3%?

Were the Final 13660 NEP Party ID vote percentages applied to the state exit polls using the state registration mix as weights?

Or vice versa?

[Carolab]

5. TIA, assuming you've seen this?

http://rigged-aggregators.blogspot.com/

[LatePeriduct]

15. Open Voting Consortium is asking for your help and more!

I was checking the blogsite again and he's updated with a whole ton of really important info, including Open Voting Consortium is trying to get donations for a brand new type of election system.

The election registration system is going to be 100% open and transparent, and could replace these funky default systems for good once its certified!

That and a whole lot more going on, I think we may soon get to the people behind this.

"They are putting it up for sale and for immediate dispersion and even solicitation. The department of Agriculture, worst of all, knows all about it and has always authorized this. Giving everyone using the technology, the full power and unprecedented control to purge voters and invoke the citizen's rights and values themselves.

"Gaining access to voter-registration data is getting easier. In the past, anyone who wanted statewide voter lists would often have to collect them individually from county election offices around the state. This made the services of data collectors valuable, since they would collect the data and aggregate it in a single file.
But in an effort to standardize data collection and reduce double votes by people who register in more than one county, the Help America Vote Act, passed by Congress last year, requires that states develop a centralized, statewide voter-registration database. This will make it even easier for political entities and marketers to collect huge voter files.

Jim Dempsey, executive director of the Center for Democracy and Technology, said candidates and political parties are entitled to use voter-registration data for campaign purposes.
But, he said, the information should not be given to anyone else for commercial use.

"There's a fundamental principle here that information collected for one purpose should not be used for another purpose," Dempsey said. "There's a risk to the integrity of voter rolls. Because if people know that their voter registration is being sold to marketers, they are less likely to give accurate voter-registration information."

According to the California Voter study, to be published early next year, states varied widely in the data they collected and disseminated. The study included data from the District of Columbia.
Although many registration forms warned voters about possible fines or jail time for lying on the form, only four out of 49 states that required registration told voters that their data became part of the public record and was therefore open to public inspection.

Among the data collected, two states required voters to provide their mother's maiden name upon registration. Thirty states either required or requested all or part of a voter's Social Security number. Eight states required the full Social Security number.

That last number is likely to increase. The Help America Vote Act requires states to collect a personal identification number such as a driver's license or Social Security number. States are supposed to redact the Social Security number before distributing voter lists to buyers.

Alexander said that if voters were told up front how their data might be used, they likely would be more guarded about the optional information they provided.
Therefore politicians are reluctant to mandate disclosure because it might limit the data that they and their parties could collect.

"If we were talking about banking data and third parties being able to profit from it, there would be a huge debate, as we recently had in California over the uses of financial data," Alexander said.
She said one reason this debate isn't happening is that "the people who ultimately decide how voter data should be allowed to be used are the politicians.... Politicians need to reign in the laws, yet they're the biggest consumers of this data."


http://wired-vig.wired.com/news/business/0,1367,61507-2,00.html?tw=wn_story_page_next1

You heard absolutely correct, the people data aggregators allow the immediate solicitation of voters forms. It has been abused in this way now for years, by the companies like Accenture and ChoicePoint and has been used to perform the real fraud. The central state lists they are talking about, is of course the BOE system. The BOE controls these registrants, and aggregates them using preferred aggregators. It is supplied directly into their system, by the data mining aggregators. Put it all together and what you have now, is a full unstoppable democracy eroder which moves faster than 20,000 different light beams per second.

We must learn the truth and stop the Department of Agriculture, this much is certain. I have been contacted by the Open Voting Consortium. They need everyone's help for donations, please contribute what you can. We need completely transparent registries and aggregators in the electoral process, or soon enough our civil rights will be washed away.

OPEN VOTING CONSORTIUM LETTER:

Dear Friends,In NOV of 2000, about 12% of the ballots in America were invisible - created with secret software.

In 2002, it was about 18% In 2004 over 30%. That's more than 30 million invisible ballots created with secret software in the most recent and controversial election!Even where we have paper ballots, those are mostly counted with black boxes that are hard to audit.

We have a registration system that systematically discourages participation.If this type of voting system is fine with you, no further action is required - this is what they want you to have.

If, on the other hand, you want a transparent, auditable, inclusive voting system, you need to do something. One of the things you can do is show your support for the Open Voting Consortium.You may be tired of hearing pleas for money, and we're tired of asking. But we must do it. We have to build up support for open voting until the job is done. Right now, we have matching funds pledged to Open Voting Consortium so your gift will be matched dollar-for-dollar - doubled.Open Voting Consortium is helping to expose the myth and show a better way. Please contribute now.

The Amazon link is here:http://s1.amazon.com/exec/varzea/pay/T43MQSRHDPEDQ/104-1192253-1034308 If you'd prefer to use PayPal you'll find a link on our home page atwww.openvotingconsortium.org.Thanks again to all!

Alan DechertPresident, Open Voting Consortium http://[e-mail]

9560 Windrose LaneGranite Bay, CA 95746

Everyone support them with what you can!

[LatePeriduct]

42. HUGE UPDATE: He just wrote it today!

-cut break-

"Then the state department didn't tell anyone what they were allowed to do with those lists because according to an outdated 1996 law which is also enforced by the Agriculture Department, they are allowed to do whatever they wish under certain guidelines.


"Computer access in another public official's office would circumvent the supervisor's "exclusive" control of the registration books."

"Therefore, it is this division's opinion that the exemption contained in Section 98.211, Florida Statutes, carries forward when this information is provided to others, and a property appraiser may neither permit the public to have computer access to voter registration records he has obtained from the supervisor of elections nor extract information from such records for purposes of public disclosure."

"However, it is this jurisdiction's legal opinion that the voter registration cancellations received from other jurisdictions and the lists of resident deaths received from the county and the Bureau of Vital Statistics are not part of the voter registration books and are, therefore, not exempt from disclosure under Chapter 119, Florida Statutes. Likewise, the lists of convicted felons and persons adjudicated mentally incompetent furnished by the clerk of the court to the supervisor pursuant to Section 98.301(2), Florida Statutes, are also not a part of the voter registration books."

"A magnetic tape is comparable to a written list and is distinguishable from "inquiry" or "on-line" computer access. Therefore, our opinion in DE 88-29 would not prohibit the distribution of a magnetic tape to the requesting public official if the Oath of Acquisition is completed by the public official."

In other words, the law of the entire state explicitly mentions that the public has no right to know who has died or if the statistics are off. Only the supervisor and their cohorts, can control how deceased voters or phantom votes show up in the elections. So it is perfectly legal, and not to mention perfectly corrupt. It happened now in multiple elections. Whoever wrote this outdated trash form of a law is no doubt a republican, following the new world order of religious whackos.

"Read the amazing software developed by Jeb's board of directors that took much of the code from Nick Chalko's work."


Yes that is what we have to work with now. Exactly as I mentioned in the beginning. By just the registration ID number, they now have the exact way to decide who's vote goes through and who's does not. Even with the exit polls, Accenture has retrieved enough profile information for automated exemption. Someone's so called voter ID may be "invalid" because he/she were of the wrong background now, and the system which runs automatically will never tell them. Oh positively, they show up on first come exit polls and early results but they will never be anything more than an error in the final results."

http://rigged-aggregators.blogspot.com

[LatePeriduct]

4. This is amazing....

Unbelievable!

[berniew1]

6. What are you assuming??? Are you assuming Florida same as national

vote pattern??
or do you have Florida exit poll data you are using??

If the latter, where does it come from and where can I find it?
I'm extremely interested in getting some exit poll data for Florida.
It appears you used the registration data that I supplied.
Is that correct?

note that the dem/repub mix for optical scan counties vs touch screen counties is not the same. The touch screens are the big urban counties that are more heavily Dem. The optical scan counties have a few big urban counties but include more of the small rural counties- some of which have more Dems but historically vote different(more Repub) according to historical official results.
The data file in my analysis of Florida vote has totals for touch screen counties and optical scan. http://www.flcv.com/fla04EA.html

[berniew1]

7. Here are totals for touch screen counties in Florida

year...Rep.V..... Dem V....... Total.......Rep V... Dem V
2000..1,480,816..1,731,521....3,285,940...45.1%.....52.7%

......Rep R.....Dem R...Oth R....Rep Reg......Dem Reg....Total R
2000...39.7%....41.9%....18.4%...1,853,194....1,956,095..4,665,937 TO = 70.4%....(2000-2004).new rep reg =197,425..new dem reg =324,279
year...Rep V.....Dem V........Tot V....Rep V....Dem V
2004..1,960,115...2,091,990...4,089,161 47.9%...51.2%

......Rep R.....Dem R...Oth R....Rep Reg......Dem Reg....Total R
2004..36.8%.....40.9%...22.3%...2,050,619.....2,280,374..5,576,264
TO = 73.3%,new rep votes=479,299, new dem votes= 360,469
2000 to 2004


that is; Dems got 62% more new voters, but Repubs got about the same percent more new votes than Dems

big swing to Dems on registrations, big swing to Repubs on votes

(and remember there was much registration fraud in florida- switching of Dem and other to Repub without their permission)

You can get the optical scan totals by difference from totals

[berniew1]

8. Dems had more visible get out the vote campaign and more

succesful at signing up new voters; and seemed to have the biggest get out the vote campaign. But what happened? Where did the votes go?

[TruthIsAll]

9. Yes

I'm assuming Florida had the same vote pattern as national. I believe it's a fair assumption, along with Party registration as weights.

I don't remember where I got the data, maybe it was from you. It was Nov-Dec.

[berniew1]

10. Does the more detailed data by machine type I supplied change anything?

[LatePeriduct]

11. I think it should show..

Somewhere in there exactly how the votes are arranged. That's why I think we need real experts looking at it by machine type, like those from UsCountVotes.

[TruthIsAll]

25. Your reg/vote changes from 2000 to 2004 further dramatize...

the extent of the fraud.

Increase in Dem registrations = corresponding % increase in Repub votes?

It don't compute.

[TruthIsAll]

12. What is the probability that....

Kerry would win 51.30%-47.77% in Florida counties(3.864 mm
votes) which used touch screen computers, but lose by
42.27%-57.03% in counties where Optical scanners (3.429mm
votes) were used?

In other words, what is the probability that Kerry's
TouchScreen vote would exceed his 47.06% total vote
(TouchScreen + Optical Scanner) by 4.24%?

Note: 
In TouchScreen counties, Dem registration is 40.89% vs. 36.77%
Repub.
In OptiScan counties, Dem registration is 41.92% vs. 38.98%
Repub.

Reg   = registered voters (in millions)
DemR  = registered Democrats
RepR  = registered Republicans
DiffR = Demr - RepR

Votes = Total votes
DemV  = Kerry votes
RepV  = Bush votes

Reg	DemR	RepR	DiffR	Votes	DemV	RepV
Touchscreen	
5.576	40.89%	36.77%	4.12%	3.864	51.30%	47.77%

Optiscan
4.725	41.92%	38.98%	2.94%	3.420	42.27%	57.03%

Total	
10.301	41.37%	37.79%	3.58%	7.284	47.06%	52.12%
							
Assume a 1.0% Margin of error (this is being extremely
conservative):

Prob	= NORMDIST(0.513,0.4706,0.01/1.96,FALSE)

Prob	 = 7.877E-14, less than 1 in 12.7 Trillion.

[Sancho]

13. The smoking gun...which I haven't seen...

is the inconsistency within the key counties. I live in Clearwater. It makes no sense that the numbers of reported Betty Castor voters, school tax referendum voters, etc. have numbers that differ wildly from the Presidential election.

In other words, just like voter registration is a prediction of voter outcome, why would someone vote for such clearly Democratic issues while certain key elections like President and Senate have swings and errors that only occur within certain machines or races.

I think voter registration in Florida is confounded by the primary system. Some democrats register as Republican (reported here commonly) because it's the only way to vote in primaries where there are no Democrats running - a common occurance in Florida in regional elections. If anything, the prediction should be more swing in favor of Kerry from new registrations.

[Sancho]

14. Example: in case I didn't make sense...

Pinellas County Florida 2004

Bush "wins" 225,686 to Kerry 225,460 (226 votes) BUT...

Castor beats Martinez 264,802 to 151,527

Questions: Besides the voter party registration, why are there almost 20,000 more votes for President than Senator in a county where Castor was a local college president (Univ. South Fl.) so name recognition isn't an issue?

The local TAX INCREASE passed 62% to 38% - almost the same as the Senate race - and there were 40,000 more votes for President than the tax referrendum....were almost ALL the people who didn't vote for either Senator or the tax increase Republicans who only wanted to vote for Bush and skip the other elections? Very weird!

[Iceburg]

16. Very interesting differential ... what technology is used in this

county? Do they rotate the ballot order on a) the presidential contest and b) other contests?

[Sancho]

19. My theory?....here goes

They hacked the tabulators, but didn't hit all the "local" elections, so IF we could get to enough raw data, it would be clearer what was manipulated...

[Sancho]

31. Touch screens...and we had an interesting experience...

My machine worked ok. More importantly, my wife was at another machine across the room and I heard her complain loudly, "All I want to do is vote for Kerry and Castor!" Apparently, after voting for 10 screens of candidates (judges, etc.)...the review screen reversed her democratic votes and she kept "revoting" and it kept on defaulting to the Republican straight ticket! She kept on complaining to the poll workers for 5 minutes after I went outside until they 1.) cut the machine off to take it out of circulation, 2.) rebooted it and declared it was working correctly, 3.) let her "revote" at a different machine 4.) offered no explanation for what to do about the "data cards" already submitted from that machine. We were very early in the morning at the first of the line. I'll bet that machine was back in use as soon as we left!

Hacking tabulators is one thing, but a random machine that reversed votes or a program that defaulted to Bush every 20th person, etc...would be all that was needed to swing Florida! Different "cheating" could be hypothesized if raw poll data were available.

[FogerRox]

34. I worked in Broward and folks were complaining about vote

switching from KERRY/Castor to BUSH/MArtinez thru out the eearly vote period

[berniew1]

36. Why didn't someone document which precincts and which machines

and how many voters saw it???????????

Do you have info on any of this?
Do you know someone who does?
I'd like further info on this.

[berniew1]

35. I'd like to talk to you; or interact directly about Pinellas

email me at the DU contact page or contact me somehow
or tell me how to contact you

you can't be emailed at contact page yet

[Sancho]

38. I must be new to the list...

Won't let me send private email...

[mgr]

33. Touch screens preset to record Kerry when Bush selected?

Seriously, the difference is between large urbanized counties with strong democratic support versus smaller counties with republican support. If you take the county results from 2000 and compare them, the distribution of the vote will be approximately the same.

TIA, I just cannot take you seriously anymore--compare the national exit poll results with the results from Florida. That is a completely inane comparison, and characteristic of the analytical skills of a first year freshman, not someone with two masters.

It only costs $79 to purchase a CD with the precinct level data for NEP, so why would you make this comparison? Too cheap, or just don't know how to build a state sample from the precinct data? You have had sufficient time, since its been available since March or earlier.

Berniew, you need to agglomerate the precinct data from both the national election precincts and the state precincts, it may provide a sufficient MOE to separate from the reported results given the appropriate F test and/or goodness of fit test. I'm working on getting that data since I have been cash strapped for the past five months, something about my daughter totaling her car, otherwise I would have your answer.

Mike

[berniew1]

37. Are you sure NEP precinct data is available?? I've been asking how to get

it for weeks??
How does one get it?

And if its available, why hasn't someone been working with it.
I would think if this were available Jonathan Simon and the U.S. Count votes people would be working with it; rather than just the national data.

[berniew1]

40. Mike, what you say about analysis isn't accurate- look more closely

but I'm interested in getting the NEP data if it is available. I had heard it wasn't available.

[passy]

53. Could a simple vote switch have occurred.

If you take *'s numbers to be Kerry's does it make the whole thing look right?

[LatePeriduct]

17. Jimmy Carter has shown his true colors.

Don't let them fool you.

The dominionist pawns are out to prove elections are crooked because of Voter fraud, they want to take over. It will allow Accenture and the rest to now have full control over voter manipulation if the voter fraud laws become passed.

Don't let them- start spreading the truth.

http://www.bradblog.com/archives/00001332.htm#comments

[Bill Bored]

18. TIA, 7:38pm 11027 had regional breakdowns.

Why not use party IDs from that instead of the national, or add that to your analysis?

I assume FL is in the Southern region, although it seems like something out of the Wild West lately!

[TruthIsAll]

20. FL is a microcosm of the U.S. , not the south...

Thats why Kerry won it by the same percentage as his total vote.

[Bill Bored]

21. True, but here are the numbers anyway:

No matter how you voted today, do you
usually think of yourself as a: (n=3,773)
Total KERY BUSH NADR Rep2000
Democrat 37 88 11 1 0
Republican 41 5 95 0 4
Independent or something else 22 42 54 1 7

Might change your assumptions somewhat, but then I think you're right about it being a microcosm, so never mind.

Now if NEP had asked the party affiliation question in the STATE POLLS, we might have something. Did they by any chance?

[berniew1]

22. Is this Southern Region?? Could you rearrange it so its understandable?

I can't figure out what the categories are the way its arranged???

for example the last line seems strange

which states are in the southern region?

[berniew1]

23. I don't understand your info at all?? totals don't make sense

some of the numbers don't make sense(to me)

no matter how I look at it, it just doesn't seem to make sense

independent category adds to about 125 or so

and what is the Rep2000 category, none of the numbers or totals make sense to me?


is the compilation horizontal or vertical, neither tallies it seems to me?

[Bill Bored]

24. Yes Southern Region

These are Horizontal NEP numbers from 2004. I'm not sure exactly how to read the Rep2000 col. though. Sorry.

-------Total KERY BUSH NADR Rep2000
Democrat 37   88   11   1   0
Republic 41    5   95   0   4
Independ 22   42   54   1   7

[berniew1]

26. they don't tally either verticle or horizontal??? what are they?

[berniew1]

27. so in southern region, 54% of independents voted for Bush??

very different from national
and likely very different from Florida I think
sure wish I had Florida info

[Bill Bored]

30. Here the Eastern Region

-------Total Kery Bush Nadr Rep2000
Democrat 42   90    9   1   -2
Republic 28   12   88   -   -3
Independ 30   57   39   3   -8

[berniew1]

39. Is this 2004 NEP data? which states are Eastern?

Florida to Maine? everything east of Miss. River?

are dems really 42% of total registrations in Eastern region?

[berniew1]

41. In Eastern region, more Rep voted for Kerry than Dem for Bush

This is an interesting result; along with only 39 % of Indep voting for Bush

[Bill Bored]

32. FL may be considered Eastern. I posted that below. nt

[TruthIsAll]

28. Rep 2000 col is probably Bush % change from 2000

Where did you get these numbers?
Can you get them for the other regions?

Are these from Final 13660? If so, they are bogus.

[Bill Bored]

29. They are from the 11,017, Time Created: 11/2/2004 7:33:46 PM

Yes, Bush change from 2000 makes sense.

I have the other regions and the National. The PDFs are on Scoop. I'm sure you must have them, TIA. There is no list of states for each region so it's possible FL is East instead of South. You might be able to tell by adding up the samples from the regions and plugging in states weighted by electoral votes or something. I don't know why they couldn't provide a list.

[Sancho]

43. Another pattern I found....

I ran some plots, but I don't know how to post them. Regardless, it seems obvious that an increase in turnout per precinct must be a positive correlation with number of votes...and when I correlate Bush and Martinez I get .99 and very little variability (Pinellas County).

When I do the same for Kerry/Castor, the correlation is less (.89) and the error is way more than expected around the line. In other words, why would an increase in voters in a district do anything except increase the number of votes for any candidate and especially for some candidates instead of others...(the only change expected is the slope or shape of the line). The increase in error appears to be due to manipulation of the vote in SOME precincts in Pinellas...without the raw data, it's the best I can thing of to find a pattern.

[LatePeriduct]

44. Castor still turns up ahead...

But this whole area denies ever doing straight-ticket voting right, or did they do it illegally?

It just seems to be the quickest explanation. Betty Castor is receiving more votes than everybody else, and more undervotes for president or "bush" votes.

[berniew1]

45. Don't you have the votes by precinct? they are available.

Knowing what precincts to look at indicates which precincts there needs to be an audit carried out for

compare voter logs to official votes; check signatures;
compare votes in 2004 to votes in 2000;
compare Kerry vs Castor and martinez vs Bush
other votes
check for undervotes; machine glitches
can you get info on absentees by precinct
do regression for early votes vs election day votes and see how the
questionable precincts compare to estimated

[Sancho]

46. I agree....

Our election supervisor is a jeb appointment who ran unopposed in 2004. With the TS's we had boxes of little white computer "disks" that each voter put into the machine...once dumped in the box, there's no record (that I know of) which voter was at what machine or even which precinct except for the "tabulator" stuff that appears to be the data on the web site.

Since I can see an error trend for the Dem. candidates that is different than Rep., I think there must have been a non-linear error (what I mean is that if 100 more people voted and the Rep. candidate gained by some "constant" amount with little variability - like 40 or 50 or 60 votes per 100 in that district) while the Dem. candidate in some precinct seemed to have more variance (meaning that sometimes there was a steady increase but other times an increase in 100 voters seemed to produce large differences in proportion per precinct). This is similar to TIA's error analyses...sort of an estimate of likely vs. unlikely consistency in the precinct. Over 300 or 400 precincts, the Rep. are what we'd expect (.99), while Dem. aren't.

The error is explainable with by speculations about swing, voter misuse of TS, etc., but VERY unlikely. If the Rep. and Dem. acted the same, I'd be less suspicious of manipulation and look to the machines or instructions for the error...as it is, I can put a TS manipulation together with the error variability and see it's consistent with a TS that defaulted or "undervoted" the Demo. candidates.

[berniew1]

50. Do the little white disk contain the recorded votes? Do they represent

a permanent "paper trail" that could be audited? recounted?

If not, why not?

[adolfo]

48. Absentees by Precinct

>can you get info on absentees by precinct


Very good question bernie. Interestingly, this information is not easily available but does exist. Contact me if you need Florida precinct level data or go to:

http://www.recountflorida.com/ivrs_index.php

Adolfo(at)recountflorida.com

[berniew1]

49. I have precinct data for some counties for election day, but not

early voting or absentees. So its hard to figure out whats going on with a precinct. There were a lot of early voters and absentees in some precincts, as much as 40%. And it would be very instructive to compare the numbers for the 3 types of voting. and to be able to figure out the undervote pattern. And check to see if there appear to be default patterns.

I'm extremely interested in getting precinct info for early voting and absentees if it exists.

Since early voting is not by the precinct, do they keep track of which precinct the voters were from? and thus turnout by precinct for total voting? I only have it for election day voting.

[berniew1]

51. I didn't find any precinct data on their site?

[LatePeriduct]

47. kick

[berniew1]

54. Does anyone know which states are in the 4 NEP regional EP data?

I have the 4 regional NEP samples for East, South, Midwest, West
but don't know which states are represented by each of these,
and which states were actually in the sample for each of these.

Can anyone shed any light on any of this??

But I don't think any of the regionals are very relevant to Florida

When is the raw data going to be released that has state and precinct data? Why isn't more being done to get it released?
Is there a precedent for not making the EP data public?

[LatePeriduct]

52. kicked

[berniew1]

55. The case for using U.S. total Exit Poll data for Florida

for the last 2 national elections, the Florida vote by party in the Presidential vote has consistently and closely followed the U.S. total. Gore carried Florida in 2000 according to the 2000 exit poll data and the 2001 Media recount by approx. the same percentage as for the national total.

Likewise the 2004 official Florida exit poll data adjusted to the official vote was about the same as the official national vote count adjusted to give the final official counts. Thus the unadjusted national exit poll data should be a relatively good estimate of the "unadjusted" actual Florida votes.


Anyone see any problem with this analysis??

[berniew1]

58. But only at the top level, no further down than sex & race

The analysis by machine types is not valid, imo. Florida has major differences in voting patterns by area, region; and the main difference in the official vote counts by the 3 machine types you analyze are caused more by regional differences than machine types. The 3 major machine types tend to be used consistently based on region. And this masks significant differences that I think exist due to machine type, that can't be dealt with unless the larger regional differences are factors out somehow. There are ways to do this statistically using such as factor analysis, etc. but its a major undertaking and requires lots of data, from past voting patterns,etc.

The countywide statistical analyses of Hout and myself identified apparent swings by county that are consistent with the vote irregularities reported by thousands of voters to the Election Protection and Common Cause hotlines, and SOE officies. These analyses found the largest differences in touch screen counties.
http://www.flcv.com/fla04EAS.html

But I don't think nationwide exit poll data can be applied to counties or regions of Florida(or any other state) as was attempted here. The major regional differences which are the biggest difference in vote patterns in Florida overwhelm and obscure any other difference, if some method isn't used to take them into account, such as factor analysis.

Using exit poll data, the best you could do if you are familiar with Florida voting patterns is to assume the north Florida rural counties vote like the South Region(Alabama and Georgia which border)
And that the big urban coastal counties vote like Eastern or National. But this would be very difficult to do without having a lot of historical vote pattern information on which counties vote like Southern, Eastern, National, etc. No one has the information to do such a study that has any usefulness/validity so far. Beyond the countywide statistical analyses based on Florida data that Hout and myself have done. The best that can be done with exit poll data is to apply the top level Florida exit poll data that we have or the top level national exit poll data, if an adequate case can be made from past voting information that they are statistically similar.

[berniew1]

56. Why were the time of poll chosen different? why the 12:22 am and 2:05 pm

poll data for some categories, but not for others?



why did you include the exit poll version that was adjusted to
match the vote totals??

[kster]

57. kick.n/t