Re: Fwd: logic and accuracy tests

From: Douglas W. Jones <jones_at_cs_dot_uiowa_dot_edu>
Date: Fri Aug 25 2006 - 13:30:24 CDT

>> From: Paul Malischke <malischke@yahoo.com>
>> Date: Aug 24, 2006 8:11 PM
>> Subject: logic and accuracy tests
>>
>> Hello,
>> One major arguement that the clerks make against post-election
>> audits of vote counts is that they do logic and accuracy
>> pre-tests. How can we counter this argument?

Pre-election testing focuses on different aspects of the problem
than post-election audits, at least, as normally conducted.

See my web page for a discussion of pre-election testing in the
context of other kinds of voting system testing:
    http://www.cs.uiowa.edu/~jones/voting/testing.shtml

Because pre-election testing cannot possibly involve exhaustive
tests of each voting machine, it is, in many ways, more of a
dress rehersal for the election than anything else. If properly
conducted, it's also an opportunity for a final proofreading of
the ballots (so precinct 13's ballot doesn't get mounted on the
machines for precinct 12).

Exhaustive testing of an old-fashioned lever voting machine made
by AVM corporation required 311 votes for each candidate on each
machine. That's many more than 311 test ballots. A former
Maryland election administrator explained this to me, with
a great story about serious machine failures in 1968 that
went undetected by AVM's recommended test scripts and led her
to design the exhaustive test, and also led her to initiate
Maryland's move away from hard-to-test mechanical voting machines.
Exhaustive tests of computerized machines are even harder.
Formally speaking, in fact, truly exhaustive testing of computer
systems verges on the impossible.

Given this, we have to supplement pre-election testing with other
kinds of tests. Post-election audits, of the sort that involve
reconciliation of all of the different sources of information, are
a powerful tool. In part, such an audit verifies that procedures
were followed, or at least, that all the paperwork was filled out.
Were the numbered security seals actually used? Were the seal
numbers properly recorded? Do the numbers recorded when the seals
were applied match the numbers recorded by the people who broke
the seals? Do the number of signatures in the pollbook really
match the number of ballots issued to voters? Do the independent
records of those votes match, and are there the right number of
such records? (Even DRE machines produce independent records at
the point where the results are printed on paper, sent off by
hand carried electronic cartridge, retained in the machine's
own memory, and perhaps transmitted by modem).

Such audits gain their highest value if the audit is conducted by
someone outside the election department. For example, post
election auditing has traditionally been done in Miami by the
county's Audit and Management Department, not the Election
Department.

Of course, VVPAT types of technologies strengthen post election
auditing, but I want to see serious pressure put on states to
require post election auditing no matter what kind of voting
technology is used. There's nothing like an auditor's report
saying "with this voting system, we were unable to reconcile the
data from the machine with that from the paper records" to get
people thinking. That's the kind of wording that came out of
a post-election audit in 1973 that led Miami to discover a
significant string of programming and procedural errors in the
iVotronic.

                Doug Jones
                jones@cs.uiowa.edu

_______________________________________________
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 23:17:09 2006

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