Sunday, January 09, 2005

exit, stage left

* "More fun with exit polls - New data gives ammunition to critics who believe the presidential election was stolen but still doesn't settle anything."
read it if this stuff interests you - but theres one issue that struck me "At midday Eastern time, NEP interviewers had spoken to about 8,000 voters" and "at 7:33 p.m. Eastern time, when it had interviewed 11,027 people." (the leaked data is here - it hasnt been verified that this is the actual data - but no-one has disputed them.)

anyways - so i had a closer look - surprised that 68% of interviews were done b4 midday eastcoast time - given that there was a whole afternoon, and the rest of the country to go... and its not obvious where the midday thing came from - so that may have been a mistake. it seems there are 3 reports - 4pm, 7.33pm, and then 1pm on the following day. (altho they werent stamped with a timezone - so that isnt quite clear)

(i havent really wanted to get into 'investigating' the data - knowing that it is prolly fraught with a whole bunch of problems - but i just wanted to check the 'midday' thingy and i kinda got caught up)

anyways, heres from the NEP website" "The National exit poll *was* conducted at a sample of 250 polling places among 11,719 Election
Day voters representative of the United States. In addition, 500 absentee and/or early voters in 13 states were interviewed in a pre-election telephone poll."
ok - so that seems pretty straightforward.

here is what the exitpolls say:
3.59pm 8349 interviews / jonk 51% / stupidmurderoussniggerer 48%
7.33pm 11027 interviews / jonk 51% /stupidmurderoussniggerer 48%
1.24pm(*nov3*) 13660 interviews / jonk 48% / stupidmurderer 51%

* so heres the thing, they officially tell us that they conducted 12,219 interviews, but at 1pm Nov3, their apparently massaged numbers include 13660 interviews. why is their a 12% discrepancy? is this a smoking gun?

* could it be that their 7.33pm numbers which showed jonk ahead by 3 points, with 90% of the interviews conducted, simply didnt allow for the possibility of a sufficient swing - and so they had to create another 1400 interviews the following day to give some semblance of possibility? seriously. is this a smoking gun?

* to make matters worse, their comment "The National exit poll was conducted at a sample of 250 polling places among 11,719 Election
Day voters" is a pdf from nov2 titled "MethodsStatementNationalFinal"
from the pdf - its very specifically says that they conducted 11,719 (+500) interviews - where the fuck did they find 1400 interviews the next day? statistics is sposed to be their expertise - do you reckon they wished that they hadnt sent out that memo on election night?

* if im onto something here, then it blows away any bollox about the wrong weightings and all that, and it blows away any discussion about whether exit polls are accurate or not - it specifically catches them with their hand in the cookie jar trying to diddle the results. exposte. which means that they believe in exit polls sufficiently to know that they couldnt get away with what they had - and therefore had to change them. no wonder they are so snarky and wont release any of the information.

having said all that, ive just taken a closer look at some of the regional numbers - east&west - which are as follows:
EAST
3.59pm 1746 interviews / jonk 58% / stupidmurderoussniggerer 41%
7.33pm 2077 interviews / jonk 58% /stupidmurderoussniggerer 40%
1.24pm(*nov3*) 2888 interviews / jonk 56% / stupidsniggerer 43%

WEST
3.59pm 1747 interviews / jonk 53% / stupidmurderoussniggerer 45%
7.33pm 2203 interviews / jonk 53% /stupidmurderoussniggerer 45%
1.24pm(*nov3*) 2640 interviews / jonk 50% / stupidsniggerer 49%

which is suspicious - cos at exactly the same time (to the second), virtually the same number of interviews are recorded on the east and west coasts - we'd expect those to be different given the time lags and all - which would point to these 'leaked' results being fraudulent. damn that karl rove.

something is amiss. either the 'leaked' data is fraudulent, or these people ought be wearing orange jumpsuits.

________________________________________________________________________________
war makes me mad, sad, bad. tsunamis makes me sad.
wotisitgood4.blogspot.com

No comments: