It makes it all so easy to raise the eyebrows of those who refuse to believe.

Historically, exit polls have been much more accurate than standard polls. In prior calculations, I have conservatively used the standard polling MOE (margin of error) methodology to compute probabilities based on deviations from the exit polls. Using standard polling MOE's does not make much sense intuitively. We should consider historically proven EXIT POLL accuracy.

So let's be a little more realistic this time and assume that the Exit Poll margin of error in each state is 2.0%, even though all historical evidence indicates that it's less than half that (see Germany, Ukraine, France etc.) So we are still being very CONSERVATIVE in this assumption.

Now let's take another look at the election results:
For Bush, vote tallies in 23 states were outside the 2.0% MOE - all in his favor.
For Kerry, vote tallies in 2 states were outside the 2.0% MOE - all in his favor.

So here we go again. I'm almost sorry to do have to do this. Let's calculate the probability (for Bush) of this occurring under the 2.0% MOE criteria.

I ran the numbers in Excel as before, with a single input parameter changed to calculate the odds that at least 23 states would be beyond the the MOE.

We input 22 as the number of states inclusive. The probability is still .025 that the MOE would be exceeded. See the description of the BINOMDIST below.

So, once again, let's calculate the probability:
Probability = P = 1-BINOMDIST(22, 51, 0.025, TRUE)

.........................P = 0.000000000000000000000000000000E+00

It's a big fat ZERO. ZILCH. NADA! Thirty of them!
Excel won't print any more than thirty decimals!

To compute the odds, we must calculate 1/P. But wait! P = ZERO
We can't divide by ZERO. No can do. What does this mean?

It's BEYOND INFINITESMAL! It's IMPOSSIBLE! THE DEVIATIONS IN FAVOR OF BUSH COULD NOT HAVE OCCURRED BY CHANCE!

THE PROBABILITY THAT IT COULD HAVE BEEN DUE TO CHANCE IS
ZERO!

Thank you Bil Gates for giving Excel to America. Oh, and thank you, too, Lotus. I used Lotus 1-2-3 long before I used Excel.

Notes:

BINOMDIST(number_s,trials,probability_s,cumulative)

Number_s is the number of successes in trials = 22 (at most 22 would fall within the MOE, at least 23 above)

Trials is the number of independent trials = 51
Probability_s is the probability of success on each trial = .025.

If cumulative is TRUE (which it is), then BINOMDIST returns the cumulative distribution function, which is the probability that there are at most number_s successes; if FALSE, it returns the probability mass function, which is the probability that there are number_s successes.

Size refers to the exit poll sample size for the given state.
The percentages are Kerry's Exit Polls and reported Votes.

Click here for responses to this analysis at DemocraticUnderground.com