Progress on New Audit Calculation

From: Kathy Dopp <kathy_dot_dopp_at_gmail_dot_com>
Date: Thu Aug 03 2006 - 02:50:31 CDT

OK. I've worked a lot of hours and have figured out how to make a
spreadsheet automatically and transparently using Frank Stenger's
solution to my election integrity equation (& wholly inelegantly)
calculate the correct audit percentages and audit amounts that are
necessary to meet the goal of detecting amounts of vote miscount that
could have wrongly altered any election result. This spreadsheet is
limited to counties with total number of vote counts under 2,000 (This
may leave counties like LA county out for now, although I could add in
an estimate for them). I kept it to 2,000 total vote counts, N, due
to the long time that it takes to re-calculate all the cells in the
spreadsheet using this rather inelegant method.

Frank Harris has written an extremely elegant program in mathlab which
accurately calculates the audit percentages that needs some revisions
to it still, but which Frank is too busy now for a while to make, and
I still must clearly write up the pseudo-code or description for Frank
to modify his program. Eventually we could put a calculation program
on NEDA's web site and OVC could put one on its web site too. I don't
have mathlab or any programs installed on NEDA's server or my computer
that have the math libraries that include the gammln function that we
use yet.

The election integrity audit calculation could be tweaked further I'm
sure, but this is a huge step in the right direction for determining
audit percentages to ensure the integrity of election outcomes. I did
a comparison between the exact calculation and the estimate for audit
amounts that I noticed while reading the Brennan Center appendix on
parallel Election Day machine tests. The Brennan Center formula for
estimating, over-estimates the amount of parallel machine or vote
count audits by much as 70% so election officials will very much
appreciate these more exact results, which may be more likely to be
implemented since they are correct.

I've just posted my new spreadsheet now if anyone wants to look at it.
 Don't try to download it without a high speed connection as it is
approx 14 MBs. It is really only a temporary solution until we get a
real program working and available on our web site as soon as we can,
along with all the other projects on the plate.

I could use help making it more effecient by those who are spreadsheet
whizzes, if anyone has the inclination.

http://electionarchive.org/ucvAnalysis/US/paper-audits/

It is called: KDgammaLN_AuditCalculator.xls

Directions: Simply input three inputs into the first worksheet (total
number of vote counts, the desired probability, and the max vote shift
rate per vote count assumption), press F9 to recalculate and watch the
tables and charts of properly calculated audit percentages regenerate.

I have really got to make myself work on more urgent things, since
this is a long term solution because no states are going to implement
this by November, 2006 - but this fascinates me, so I find myself
working on it and ignoring other things.

----
Kathy Dopp
http://electionarchive.org
National Election Data Archive
Dedicated to Accurately Counting Elections
Subscribe to announcements by emailing election-subscribe@uscountvotes.org
Please donate or volunteer.
"Enlighten the people generally, and tyranny and oppressions of body
and mind will vanish like evil spirits at the dawn of day," wrote
Thomas Jefferson in 1816
_______________________________________________
OVC-discuss mailing list
OVC-discuss@listman.sonic.net
http://lists.sonic.net/mailman/listinfo/ovc-discuss
==================================================================
= 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 23:17:03 2006

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