Re: Vote recount begins in Mexico

From: Kathy Dopp <kathy_dot_dopp_at_gmail_dot_com>
Date: Fri Aug 11 2006 - 19:05:32 CDT

> >
> > The real questions are:
> >
> > 1. If the 11839 polling booths are really
> > selected at random, and if there is no actual
> > correlation between fraudulent polling place vote
> > count and number of votes cast at the polling
> > place, what is the probability that a recount of
> > these polling places will uncover at least one instance of fraud?

OK. Frank found time to run the numbers for the Mexican 9% audit
today, plus Frank wrote another little Matlab script to calculate the
straight-forward probabilities (because N=130,000 total vote counts
seems to exceed the capacity of Excel hypergeometric function when I
tried it).

The good news is that the 9% audit that Mexico is currently
conducting, if it is scientifically conducted, is plenty adequate to
detect vote miscount down to 0.1%. Smaller audit sample sizes would
even probably be fine. The probabilities of detecting a level of
miscount that could be outcome altering was virtually 100% down to a
0.1% margin between candidates.

I am working on trying to obtain a copy of MatLab for myself so that I
can run numbers and make modifications myself.

If someone wanted to give us the number of total ballots cast at the
say 5,000 largest Mexican polling locations, we could calculate new
numbers to see if the probability would go down slightly, but it does
appear that the Mexican audit amount (about 9%) is more than adequate.

The one caveat I have is that audits should IMO be designed to audit
states or counties separately rather than an entire country because
errors are specific to local election administration.

OVC-discuss mailing list
= The content of this message, with the exception of any external
= quotations under fair use, are released to the Public Domain
Received on Thu Aug 31 23:17:06 2006

This archive was generated by hypermail 2.1.8 : Thu Aug 31 2006 - 23:17:10 CDT