Yesterday exit poll figures for 4 pm on 11/2 were published on
http://scoop.co.nz/mason/features/?s=usacoup Below is the data from a pdf file of an official exit poll document which we are not “allowed” to see.As usual the overall % is not shown but is easy to calculate: Kerry 49.9% Bush 50.1%.The date is not shown or has been removed, but the number of respondents (2826) appears on the document. This is only 36 less than the total in the “final” changed data that appeared on CNN at 1.30am on 3/11.
So this is pretty much the final exit poll - before the figures were “fixed” to fit the results! These figures show pretty much a dead heat in Florida! As these figures are slightly less favorable to Kerry than those published yesterday they may already be weighted –but as nobody is telling, it is difficult to know!
Exit poll summary
Kerry Bush
49.9% 50.1%
The exit poll data meant a cliffhanger.
But the declared votes was a walkover for Bush
Kerry Bush Declared Bush lead
3 574 509 3 955 656 381 147 votes
47.47% 52.53% 5.05%
This is 5.02% more margin for Bush than the exit pole
Conclusions
1/ No normal weighting is ever going to explain the 5% discrepancy between the exit polls and the declared result.
For example if the men/ women ratio needed to be weighted because there was evidence that men were underrepresented in the sample, increasing the 46/54 ratio shown in the document to a 49/51 ratio, only gives Bush an extra 0.2% of the vote.
2/ Either the poll was rigged or the exit poll was appallingly wrong
3/ All the exit polls that were published on the evening 2/11 have disappeared. (This data form Florida is one of the few exceptions)
4/ CNN, NBC, CBS are not releasing any information – and consider the whole subject as hype.
5/ Nobody is saying how many times the exit poll data was fixed or why.
Warren Mitofsky one of those behind the NEP, $10 exit polling operation has stated that
http://mayflowerhill.blogspot.com/2004/11/mayflower-hill-exclusive-warren.html the polls were skewed 1.9% compared to the results. Looks like more to me!
If they really want to restore confidence then they need to face this issue and release the raw canvassing data on a precinct-by-precinct basis for all thestates that were canvassed. But don’t count on it – it might turn up some more surprises!
The following is page 1 of 8 that I have in in pdf
(I would have posted the whole pdf file but didn't know how)
Florida General Exit Poll Page 1 of 8 Total KERY BUSH NADR Rep2000
Are you: (n=2,826) Male 46 47 52 1 -2
Female 54 52 48 0 3
Sex by race (n=2,775) White male 33 42 57 0 -
White female 38 46 53 0 -
Non-white male 13 59 40 1 -
Non-white female 16 64 36 0 -
Are you: (n=2,791) White 70 44 55 0 -2
Black 12 87 12 1 5
Hispanic/Latino 15 46 54 0 5
Asian 1 - - - -
Other 2 34 66 - -
AGE (60 and over) (n=2,825) 18-29 17 60 39 0 -1
30-44 27 48 51 0 1
45-59 28 44 55 0 6
60 or over 27 49 50 0 -1
Age 65 & over (n=2,825) 18-64 81 49 50 0 2
65 or over 19 51 49 1 -3
2003 total family income: (n=2,524) Under $15,000 9 61 38 1 -
$15,000-$29,999 15 61 37 1 -
$30,000-$49,999 22 53 46 0 -
$50,000-$74,999 21 47 52 0 -
$75,000-$99,999 14 40 60 0 -
$100,000-$149,999 10 46 54 - -
pages 2 to 8 omittted but I have them in pdf