The True MATH: Confirming Election Fraud 2000-2008 (TIA) - x

[tiptoe]

The True MATH: Confirming Election Fraud 2000-2008 — TruthIsAll

...
The calculations ... confirm that exit-polled "new voters" ('did not vote' in prior election) comprise the difference (RV – LV) between pre-election registered (RV) and likely voter (LV) sample-size.
...
2004 Final Pre-election Polls

The Kerry/Bush vote split of the difference (1335) in sample between the RV and LV subset (57.8–42.2%)
matched the 12:22am Preliminary National Exit Poll  'New Voter'  shares: 57–41–2%.

The 1769 difference — i.e. 5-poll RV samples (10310) exclusive their LV subsets (8541) — comprised 17.2% of the total (RV) sample.
In 2004, there were approximately 21.4 million newly registered and other new voters (i.e., 'DNV' 2000) — 17% of 125.7m votes cast.
Of the 21.4 million, approximately 13.8m (11% of 125.74) were first-time voters. Kerry won 55% of first-timers.
The average pre-election poll projected turnout of registered voters was 82.8% (117m of 142m registered).
The census reported an 88.5% voter turnout (125.7m of 142.1m registered).

 

 

 

Actual Poll (%)

 

0.75

0.25

Undecided Voters Allocated

   RV
1-Nov
31-Oct
31-Oct
31-Oct
30-Oct




   LV
1-Nov
31-Oct
31-Oct
31-Oct
30-Oct

Poll
CBS
Fox
Gallup
ABC
Pew

Total
Average

Poll
CBS
Fox
Gallup
ABC
Pew

Total
Average

Sample
1125
1400
1866
3511
2408

10310
2062

Sample
939
1200
1573
2904
1925

8541
1708

Kerry
46
48
48
48
46


47.2

Kerry
47
48
49
48
48


48

Bush
47
45
46
47
45


46

Bush
49
46
49
49
51


48.8

Spread
-1
3
2
1
1


1.2

Spread
-2
2
0
-1
-3


-0.8

 

Kerry
50.50
52.50
51.75
51.00
52.00


51.55

Kerry
49.25
51.75
49.75
49.50
48.00


49.65

Bush
48.50
46.50
47.25
48.00
47.00


47.45

Bush
49.75
47.25
49.25
49.50
51.00


49.35

Spread
2.0
6.0
4.5
3.0
5.0


4.1

Spread
-0.5
4.5
0.5
0.0
-3.0


0.3

Projected
Turnout
83.5%
85.7%
84.3%
82.7%
79.9%


82.8%

 

Pre-election Polls Final RV and LV Samples

 

Registered Voters (RV)

 

Likely Voters (LV subset)

 
CBS
Fox
Gallup
ABC
Pew

Total
 

Total
1047
1302
1754
3335
2192

9629
93.4%

Kerry
518
672
896
1685
1108

4878
47.3%

Bush
529
630
858
1650
1084

4751
46.1%

 

Total
901
1128
1542
2817
1906

8294
97.1%

Kerry
441
576
771
1394
924

4106
48.1%

Bush
460
552
771
1423
982

4188
49.0%

 

 

RV – LV [Less-Likely Voters]

 

"RV minus LV" Voters

CBS
Fox
Gallup
ABC
Pew

Total

Total
145
174
213
519
286

1335

Kerry
76
96
125
291
184

772

Bush
69
78
88
227
102

563

 

Kerry
52.6%
55.2%
58.8%
56.2%
64.3%

57.8%

Bush
47.4%
44.8%
41.2%
43.8%
35.7%

42.2%

Spread
 5.2%
10.3%
17.6%
12.4%
28.7%

15.6%

 

• If Final NEP weightings indicate a mathematically impossible number of
   returning voters, then simple logic dictates the weightings are impossible.
• Since impossible weightings were necessary to match to the official vote count,
   then the official vote count must also be impossible.
• Since the vote count is impossible, then all demographic category cross tabs
   must use incorrect weights and/or vote shares to match the count.

Polled Pre Vote: 10,310. 'RV-LV' sample: CBS + Gallup + ABC + FOX + Pew

 

Post-Election 12:22a Prelim NEP

‘DNV’ Voters

57%

41%

 

Polled Exit Vote: 13,046. Shares = 3.4m Dem margin  (17% DNV == 21.4m)

Post-Election   1:25p  Final NEP

‘DNV’ Voters

54%

45%

 

'Forced' to match the vote-count = Dem margin dis-count, 1.5m votes  1 2 

 

NATIONAL EXIT POLL
...

Full article: http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x512608

 

[Occam Bandage]

1. This stuff always reminds me of "mathematical proof" that "Darwinism" is wrong.

Start with a politically-driven assumption, then plug in numbers left and right until you get to that assumption.

[Nicholas D Wolfwood]

2. This obsession of the OP's is rather unhealthy.

Even if the numbers could be believed, you only have circumstantial evidence, which is based off of exit polling, the validity of which is not exactly beyond reproach. Without something far more concrete than mathematical masturbation, there's nothing that anyone can or should do with this information.

[tiptoe]

3. "Even if the numbers could be believed,..." Enjoy the mathematically impossible vote counts. nt

[BzaDem]

4. The RCP average of 15 pre-election polls take before the election matched within 3/10 of a percent

http://www.realclearpolitics.com/epolls/2008/president/us/general_election_mccain_vs_obama-225.html

RCP Average of 15 pre-election polls: Obama +7.6 percent
Actual result: Obama +7.3 percent

The result you want us to believe: Obama + 15 percent

This stuff gets old.

[tiptoe]

5. "This stuff gets old." What ‘gets old’ is persistent failure or refusal by some DUers and co-opted,

useless, "broken" mainstream media 1) to differentiate final pre-election RV and LV poll results, 2) to note that RealClearPolitics removed all Registered Voter (RV) poll results that appeared until a few weeks after both ‘08 and ‘04 elections, and 3) to appreciate the implication of 2008 and 2004 ‘forced’ Final National Exit Polls requiring the use of impossible demographic weightings in order to match respective cumulative state-recorded vote counts.

If Final NEP weightings indicate a mathematically impossible number of returning voters, then simple logic dictates the weightings are impossible.

Since impossible weightings were necessary to match to the official vote count, then the official vote count must also be impossible.

Since the vote count is impossible, then all demographic category cross tabs must use incorrect weights and/or vote shares to match the count.

Just by virtue of the impossible weightings within the 2008 Final NEP, the 2008 vote count itself is impossible, and the 7.27% official margin, therefore, baseless.

Did you ever compare the final RV polls with the LV subsets?
How come RCP does not show the final RV polls among the 15?

Were you aware of the fact that Obama has a 13-point margin, 52-39-1.5, in RV polls (see Gallup, ABC/WaPo final pre-election polls...
(assume recorded 1.5% for third parties)? There were 7.5% undecided.

The pollsters gave Obama an average 62-37% margin in undecided voters.
That indicates Obama 4.5%, McCain 3.0% of the undecided 7.5%
So Obama's total share (based strictly on published pollster data) is 56.5% (52+4.5).

How come you did not mention Registered voter (RV) polls? Were you aware that "Likely Voter" polls are subsets of RVs and that the data is from the SAME poll?

Can you see why RV polls (which include all registered voters) are better indicators, especially when there are a large number of new voters (as in 2004 and 2008)?

RCP avoids mentioning any RV poll results among the 15. They only display the LV subsets, a few of which (Rasmussen and Battleground) are VERY BIASED Republican polls which were outliers designed to bring down Obama's average.

Like others reading and blindly accepting RCP, you were misled into believing and accepting the results -WITHOUT DOING ANY CONFIRMING ANALYSIS.

Your limited view was formed by a media which has not EVER done the thorough analysis that TIA has done here.

 

[grantcart]

6. 'Some Duers'? I think that there is a very broad consensus that your methodology

doesn't hold water.

You speak of 'simple logic'.

Simple logic shows the following;

1) Sophisticated polling analysis nailed the election ahead of time.

Never has the actual results been predicted so closely as it has by fivethirtyfive.com.



Discussed here http://www.fivethirtyeight.com/2009/04/comparing-election-outcomes-to-poll.html

Even the few states that were off had reasonable explanations, like Hawaii was such a landslide state that many people didn't feel the urgency.

That alone basically kills your argument but lets go on;

2) You are suggesting a national conspiracy in which the conspiracy is worked out evenly across 50 states.

You are arguing that the discrepancy is achieved nationally even though that would mean that it would have had to coordinated 50 state conspiracies and that all 50 not only were equally effective but all shared the same rate of fraud.


3) Your argument has to include that even though this conspiracy would include thousands of polling officials in both Democratic and Republican states that none of the actors have been found and that we can only know that something was done because of a statistical discrepancy.

4) You also have to argue that the side that won, the Democrats and President Obama, are unaware of this because obviously if they were aware of it then they would take actions to fix it before the 2010 election.


If you had found a certain result that existed in a few states where Obama significantly underperformed then it would have been a more interesting case.


Your methodology is not unlike the Creationists and Intelligent Design, like you they start with the conclusion and try to find the best path to explain it. You start with a conclusion and then find the math that seems to fit it the best. The fact that the results were accurately predicted before the election and that your premise requires 50 coordinated conspiracies is simply not credible.

[BzaDem]

8. Right, so 15 pollsters doctor their likely voter samples to help Republicans.

Gotcha.

Conspiracy theorists will never be convinced that a conspiracy did not occur. They do damage to the movement for vote integrity (which I support), since even if there is straightforward evidence that a conspiracy did not occur, they will not be convinced. They reduce the chance that when actual vote/counting fraud occurs (which can definately happen), something will be done about it. For example, they will ignore 15 pollsters (some of which have been polling for decades) all because one consistently wrong exit poll (the NEP) gives a margin even higher than the highest margin reported in any Registered Voter poll that you cite.

I am not going to go through your analysis line by line and counter it, since regardless of what evidence I exhibit to the contrary, you will continue to believe a conspiracy has occured and you will continue to post misleading analysis based upon a flawed poll.

[PBS Poll-435]

7. Hater

nt

[Sancho]

9. The pollsters' behaviors are the more interesting to me.

We all know how most pre-election and exit polls were funded and the history of E-M etc., so we don't need to rehash that here. I haven't worked much with political polls, but I have a background of many years of social science research and creating surveys is always part of the process. When respondent errors (reluctant responders, etc.) or sampling errors (demographics misrepresentations as often mentioned by TIA, etc.), or generalizability errors (nonignorable nonresponse and WPE contrasts), are repeated in the method over and over (2000, 2004, 2008...); then it becomes clear that there is too little effort to fix the design issues.

No matter how much results are "adjusted" and regardless the adjustment methods, you can never be confident of the conclusions without a good design to start with...there are many articles on this. One fun one is:
Journal of Educational and Behavioral Statistics, Vol. 14, No. 2, 121-140 (1989)
Eelworms, Bullet Holes, and Geraldine Ferraro: Some Problems With Statistical Adjustment and Some Solutions
Howard Wainer

TIA's analyses, some WPE analyses, DRE undervotes, butterfly ballots, Clint Curtis/ Tom Feney stories, plane crashes, and anecdotal reports from voters are part of a pattern of complaints of potential manipulation of elections. As John Snow illustrated a century ago, sometimes experimental scientific studies don't reveal a cause but the pattern is revealing where to look! The poll adjustments may or may not be part of a conspiracy, but it seems obvious that the exit pollsters most in question from media called elections have made little or no reported efforts that serious, neutral researchers would likely do to investigate, identify, or improve the errors.

Most of the post hoc arguments and data mining of poll results are interesting, but the best bet would be polls with results that didn't require so much artificial adjustment. I realize the adjustments are the practice and habit of pollsters; and all disciplines have their own methods, but for such a serious issue as potential manipulation of the Presidential election, the lack of change in polling in response to critics is the best evidence that there is a desire to avoid controversy or hide the truth or protect the contract with the media.

Several of us have listed potential improvements on DU and in other places; some are simple while others are more complex statistical/survey development suggestions so we don't need to reargue exactly what could be done here. The point is simple. If there is an issue (see Wainer above), then the researcher is obligated to do what is necessary including "heroic efforts" to improve the design. On DU, in journals, and at professional conferences I have seen lots of explanations and excuses - I have not seen efforts to find and fix. That seems an answer to me that I've seen in other practices (like smoking causes cancer or whatever). Repeated research with the same problems over an over has its own underlying cause!

[tiptoe]

10. Forced '08 EP match w VoteCount = impossible # of return-voters & implies grossly-distorted '04 vote

 

2008

Final National Exit Poll  category: "Vote for President in 2004"
('forced' match to recorded vote-count shares == mathematically impossible numbers of returning Bush and Other voters, after ~4.8% mortality)

2008 Final National Exit Poll

Voted
'04
 

Adjustable only w/in MoE
Replicable Vote Shares
 

Obama

McCain

Other

 DNV
 Kerry
 Bush
 Other

71%
89%
17%
66%

27%
9%
82%
24%

2%
2%
1%
10%

 

2008 Recorded Vote Count

Obama
52.9%
 

McCain
45.6%
 

Other
1.5%
 

100%

131.37

 


←

2008 Final NEP

Returning '04-Voter Mix
Weightings Needed to Reconcile
NEP category Vote Shares with
Vote Count aggregate shares

13%
37%
46%
 4%

 


→

 

Dem
52.6%
69.13


12.13

43.26
10.27
3.47

Rep
45.5%
59.80


4.61

4.37
49.55
1.26

Oth
1.9%
2.44


0.34

0.97
0.60
0.53

 

 
 
 

48.61
60.43
5.25

 

 
 
 

42.5%
52.9%
4.6%

 






→

Impossible Return-Voter #

Kerry
48.61

Bush
60.43

Other
5.25

 
 

2004 Recorded Vote Count

59.03

62.04

1.22

 

Implied 2004 Vote Count

52.01

64.66

5.62

 

Calculated 2004 True Vote

66.81

57.23

1.70

 

[Every Man A King]

12. You just post numbers and pretend you have a point

What exactly are you saying happened in the 2008 elections?

[tiptoe]

15. "just post numbers"? ..."pretend"?? work on it. nt

[Every Man A King]

11. 88.5% voter turnout ? Huh?

[Dreamer Tatum]

13. Using estimates, guesses, assumptions, and generalizations to prove a point

can always be relied on to do just that.

Epic waste of time. Fail.

[tiptoe]

14. Try multiply, add, divide and compare official recorded vote & Final NEP figures. "Waste of time"...

...usually just for Freepers and other dolts.

Most DUers shouldn't have problems of epic proportions.