Polling data sources:
National Pre-election polls
http://www.pollingreport.com/wh04gen.htm http://www.economist.com/media/pdf/YouGovS.pdf Ohio exit poll
http://www.exitpollz.org/cnn2004epolls/Pres_epolls/OH_P... National Exit Poll (pdf)
11/2/04, 3:59pm 8349 respondents: Kerry 51-Bush 48
http://www.exitpollz.org/mitof4zone/US2004G_3737_PRES04... 11/2/04, 7:33pm 11027 respondents: Kerry 51-Bush 48
http://www.exitpollz.org/mitof4zone/US2004G_3798_PRES04... 11/3/04, 12:22am 13047 respondents: Kerry 51-Bush 48
‘
http://www.democraticunderground.com/discuss/duboard.ph... 11/3/04, 1:25pm 13660 respondents: Kerry 48-Bush 51
http://www.exitpollz.org/mitof4zone/US2004G_3970_PRES04... POLL SAMPLE-SIZE AND MARGIN OF ERROR
The Law of Large Numbers is the basis for statistical sampling.
All things being equal, polling accuracy is directly related to sample size. The larger the sample, the smaller the margin of error (MoE).
The MoE for a random sample of size N is .98/sqrt(N)
In an unbiased random sample at the 95% confidence level,
there is a 95% probability that the true population mean
falls within the MoE (or within 1.96 standard deviations) of the sample mean.
In pre-election state polls, about 600 were sampled in each state.
That gives a 4.0% MoE for any given state. That may seem high, but...
the combined national sample size is 30,000 (50 states * 600).
So the combined MoE is equal to .98/sqrt(30000) or 0.56%.
In 18 pre-election national polls, the sample-size ranged from 800 (3.5% MoE) to 3500 (1.7%).
The total 27,229 sample size reduces the combined MoE to 0.59%.
The post-election state exit polls sampled 73,607 nation-wide,
with respondents ranging from 600 (4% MoE) to 2800 (1.8%).
The total 73,607 sample-size gives an MoE of just 0.37%.
Assuming (for the naysayers) a very generous 40% cluster effect,
the combined MoE is 1.4*(0.37%) or 0.52%
In the National Exit Poll of 13047 respondents, the MoE is 0.88%.
Once again, assuming a 40% cluster effect, the MoE is 1.4*.88 or 1.2%.
Kerry won the exit poll (gender demographic) by 50.8% - 47.8%.
Therefore, there is a 95% probability that Kerry won BETWEEN 49.6% and 52.0%
of the true vote. And 97.5% that his vote EXCEEDED 49.6%.
According to the 1:25pm Final Exit poll (13660 respondents),
WHICH WAS MATCHED TO THE VOTE, Bush won by 51.1 - 47.9%.