some thoughts on audit procedures

From: Kathy Dopp <kathy_dot_dopp_at_gmail_dot_com>
Date: Sun Jul 02 2006 - 12:44:50 CDT

Here is an outline of preferred methods for auditing vote counts

1. CENTRAL COUNT OP-SCAN (CCOP) (with help from Jonathan Simon)
statistical random sampling of optiscan ballots (overseas, military,
mail-in, absentee,..). This method looks for statistical anomalies rather
than real discrepancies. The sampling is expanded to a full hand-count if
statistical sampling shows miscounts are probable. Breakout by ballot-type
and by precinct as much as possible.

2. DRE w/ VVPT
hand-count 100% of paper records associated with randomly selected vote
counts - either precinct counts or machine counts can be randomly selected.
This method finds actual discrepancies. (statistical selection of individual
ballots would not be practical w/ DREs w/VVPT)

can be audited either via actual audits of vote counts (#2 above) and/or
statistical sampling of ballots (#1 above) for precinct-count optical scans
(I am as yet unclear as to the advantages and disadvantages of each for

NOTE: With DREs w/ VVPT some types of election tampering can only be
detected by voters who notice that the voter verifiable paper ballot record
does not match the electronic record. Cameras should be kept at each polling
location to record this if it happens. There is no remedy if this hack is
discovered, other than to re-conduct the election. This hack is described
in the Brennan Center Report on pp. 10-11. See

NOTICE that the statistical sampling of ballots described above compares
hand-counts with the official reported machine counts from the election. It
does not recount the ballots electronically after the election. This method
is similar to doing an exit poll but uses actual ballots rather than voter

I have not given thought to securing ballots, chain of custody issues that
perhaps should be mentioned in any audit proposal.

This Utah proposal and this short math paper describes primarily #1 above,
plus some general ideas applicable to ensuring that audits are independent
and scientific:
some help from Joycelynn Straight)
some help from Ron Baiman)

I just realized today, that a combination of statistical and actual audits
of vote counts, is best for most counties, depending on the voting systems
they use.


Kathy Dopp
National Election Data Archive
Dedicated to Accurately Counting Elections
Subscribe to announcements by emailing
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

= The content of this message, with the exception of any external
= quotations under fair use, are released to the Public Domain
Received on Mon Jul 31 23:17:02 2006

This archive was generated by hypermail 2.1.8 : Mon Jul 31 2006 - 23:17:09 CDT