I wanted to share this information from my other post:
I am a wall street quantitative analyst, I decided to do a simple statistical significance test of Ron Paul receiving 0 votes in Sutton County.
First, the facts:
Sutton county: 378 votes cast
Ron Paul average across the state: 0.0774
Assumptions:
We will use a binomial distribution (as n is not very large).
We will assume that this county is representative of the averages of the rest of the state.
Probability that Ron Paul receives a vote: 0.0774
Probability that he does not: 1 - 0.0774
According to a simple binomial distribution probability calculator:
http://www.stat.tamu.edu/~west/apple...omialdemo.html
The probability that Ron Paul receives "0" votes: 0.000
The actual probability is greater than zero, but less than 1e-3. An exceedingly small number.
With an approximately 50% probability, Ron Paul should have received roughly 28-29 votes.
This is statistically extremely improbable. The binomial assumption is valid. The assumption that can't be validated, however, is that this county is representative of the entire state.
There is definitely something fishy going on. Statisticians, physicists, etc. please feel free to contribute.
