Electronic Voting Systems "Audited"

From: Charlie Strauss <cems_at_earthlink_dot_net>
Date: Sat Aug 20 2005 - 02:51:52 CDT

Below is an intriguing article. An organization tried the following
experiment which they mistakenly call parallel testing.
Voters voted on an Diebold Accuvote, and then they were asked to vote
again in a different manner.

the result showed a 4% difference.

The article wants you to believe this shows that the accuvote is

Such a conclusion is of course unjustified till we reject more
plausible explanations.

If we assume that in fact the accuvote were perfectly accurate then I
can think of three possible reasons for the discrepancy.

1) the experiment was misconducted and is a pile of steaming poo
2) The voters are mischevious and lie.
3) Voters intrisicly make huge errors

While 1 and 2 are highly likely contributing factors it's worth
considering 3) further.

we can divide 3) into three possibilites:
3a) Different presentation formats induce different voting behavior.
3b) voters are poorly skilled
3c) voters memories of past decisions are ephemeral

My own belief is that 3a) and 3b) have been previously observed.

regarding 3b) Undervote rates for major elections on paper suggest
that voters probably mismark their ballots perhaps 1 out of every
hundred contests.

regarding 3a) my own analysis of the florida elections showed that
voters behaved differently in precints using optical scan and

Now when I say behaved differently I DONT mean their mean voting
preference were different. You can argue about that till you are
blue in the face and you'll never be able to convincingly eliminate
the demographic differences between precints even with so-called
matched pairing.

What I do mean is that if you look at the variance around the mean in
the presidential races in each of the districts in florida you see
that e-voting shows a much lower variance than optical scan.

To me this says that if you ask people their vote preferences in a
serial order, you tend to get more uniform voting behaviour. Whereas
if you ask it in a full-face format you get more variation.
Presumably when voters can see the party membership of the clowns in
the congressional races, it biases their presidential vote. Or
perhaps people trying to serial vote become exasperated and hit the
party-line vote to end their misery. Who knows. But the variances
are remarkably different even if you only compare districts of the
same sizes (thus arguing against the e-voting could be associated
with more homogenous communities)

So anyhow, I think the real explanation is all of these; from the
steaming poo theory to serial-vote order problem. Or you can believe
the accuvote is not so accurate. Obviously they need a real parallel
testing. if it's this bad it would show up.


"Enron by the Sea" shows strange electoral anomalies - a 4 percent

Diebold's Un-Accu-Vote

Now, a nonpartisan citizens' group that conducted a parallel election
has requested a recount of 11 precincts. This time, the issue isn't
unmarked bubbles, but the accuracy of Diebold Accu-Vote optical scan
voting machines and the Diebold GEMS central tabulator used to count

ES&S and Sequoia are the other side of the coin to Diebold. Closed
source software and control of the equipment before, during, and after
the election.

The Citizens Audit Parallel Election (CAPE) asked voters exiting polls
to vote again and sign a log book attesting to the accuracy of their
second vote. Sealed parallel election ballots were counted at KGTV's
studio with a TV camera crew filming the counting process.

Now that is a great way to audit an election. Lets do this in New

Nearly 50 percent of all voters participated in the parallel election,
which included five polling places representing 11 precincts. The
sample included more conservative than liberal precincts, with
participation as high among Republicans as among Democrats. The tandem
election results showed what most feel to be startling results.

"There is a shift of four percent of the vote, consistently," Joe
Prizzi, (engineer and physicist,) reported at a press conference held
by CAPE in front of City Hall. Frye received 50.2 percent of the votes
cast in the parallel election - enough for an outright victory if
those results reflect the outcome citywide. CAPE also found that the
official count added approximately 2 percent to each of Frye's two
Republican opponents, Jerry Sanders and Steve Francis.

In addition, CAPE examined the only other ballot measure, a
proposition over a war memorial cross on public land. The
proposition's vote total also appeared to have been padded by 4
percent in the official election tally, which was certified Friday
August 19 by San Diego County's newly appointed Registrar of Voters,
Republican Mikel Haas.

Math is non-partisan

A team of statisticians from California State University- Northridge -
have analyzed the data from CAPE, concluding that the probability of
luck or chance as the cause of the observed four percent deviation is
less than one in 1,300 - or .000678.

Activists suspect fraud. "I am troubled by the prospect that we are
losing our democracy very quickly. We've been voting on machines that
were never intended to be tools of democracy," said Brina-Rae
Schuchman, media spokesperson for CAPE, noting that Diebold machines
utilize "secret software."


Secret SOFTWARE is closed source software and ALL of New Mexico's
equipment uses "secret software." We can't check it or tell what the
software is doing. CAPE has the best alternate method to audit
elections. Leave the old corrupt system and use a completely separate
system to vote and count ballots.


CAPE isn't the only group to accuse Haas of withholding public
information. Jim March of Black Box Voting and a Republican maintains
that the Registrar refused his request during the election to obtain
audit logs, which would show whether records were kept of each user
who accessed the Diebold GEMS central tabulator.

In an interview with the East County Californian before the election,
Haas stated that he would allow citizens to observe the central
tabulator counting votes. But on election night, March found the
tabulator screen had been placed eight feet away, behind glass and
readable only through binoculars, literally. According to March, an
activist who was with him brought binoculars and was able to clearly
make out the screen. March's request to have the screen moved closer
was refused, so he entered the secured tabulating room.

March was arrested and charged with a felony count of obstructing an
election official. The charge was later dropped. "This was a violation
of my civil rights," said March, who plans to sue County election
officials for violating his right under California law to observe an
election and his right to access public records.

Computer experts hired by Black Box Voting to penetrate voting systems
in Leon County, Florida (with permission of an election official)
demonstrated the ease of reprogramming Diebold optical scan voting
machines and changing votes through the Diebold central tabulator -
the same voting systems used in San Diego during the recent election.

Informed of these facts, Haas nonetheless allowed hundreds of San
Diego poll workers to keep voting machines at home overnight -
including programmable memory cards protected only by seals that could
easily be removed with pliers and resealed.

March and other observers contend that San Diego's central tabulator
was hooked up to the Internet on election night. An Internet
connection would violate Diebold's own procedures manual, which
states: "The GEMS server should not be connected to any network that
has an external Internet connection." State certification required
that manual procedures be followed.

"If that manual isn't followed, it's an illegal installation," says
March. "They ran a completely illegal election."


Many reports of central tabulator's totals were sent over the internet
  from Soccoro county to Santa Fe's SoS's web site.

Is it illegal to have central tabulators connected to any kind of
network in New Mexico. Shouldn't all voting equipment have any modems
or I/O port capability removed, except memory cards?

If you can access the tabulator over a network or internet or wireless
network/internet the tabulator's election totals can easily be changed.


Caught with tabulater plugged in

Asked by this reporter if the central tabulator was hooked up to the
Internet, Haas replied, "Yes. That's so we can get our results out to
the Internet, so people can see. It's firewall protected."

But after being informed that hooking the tabulator up to the Internet
would potentially render the election illegal, Haas backpedaled and
said he may have been mistaken about the tabulator's Internet
connection. "I'm not that technical," he noted, then suggested that
perhaps the machine was transmitting results to a secondary unit.

Activists plan to monitor the recount, but the potential for problems
remain. "We are very worried about tampering," Alter admitted. "That's
why we want the count videotaped."

Those fears evoke comparisons to Clermont County, Ohio, where Raw
Story reported that a recount of the 2004 presidential election
revealed that stickers were placed over the Kerry/Edwards oval on
opti-scan ballots. Those ballots were then fed into machines after the
hand recount. Witnesses have stated that beneath the stickers, the
Kerry/Edwards oval was selected.

Subtler forms of tampering might include substituting entire batches
of ballots, described Alter, who plans to monitor the recount.

Soon, San Diego's Registrar hopes to eliminate the opti-scan system
entirely and retrofit warehoused TSx touchscreen machines with paper
trails--if the new Republican Secretary of State, Bruce McPherson,
opts to recertify the TSx system previously decertified by Democratic
Secretary of State Kevin Shelley.


Delivering the election - not just in Ohio

The nation's first parallel election was conceived by Ellen Brodsky,
an election official in Coconut Creek, Florida. Held at a single
precinct during a May 2005 special election on a gambling initiative,
the Florida parallel election drew a 67 percent participation rate and
revealed significant discrepancies, leading to revelations of
programming issues with touch-screen voting machines.

San Diego's far broader parallel election was the brainchild of Judy
Alter, an emeritus professor in the department of world arts and
culture at UCLA who participated in the New Mexico recount after the
2004 presidential election. In Santa Fe, Alter detected a shift of
third-party candidate votes into the Bush/Cheney column.

"That pattern has now been identified in eight states," Alter told Raw
Story in an exclusive interview, adding that numerous other
indications of electronic fraud have been found. "This is why I'm
leading Study California Ballots, because we have to actually count,"
Added Alter.


Count ballots not trust equipment to do it.


------------------------ Yahoo! Groups Sponsor --------------------~-->
<font face=arial size=-1><a href="http://us.ard.yahoo.com/
">Fair play? Video games influencing politics. Click and talk back!</

More info on Verfied Voting New Mexico
Yahoo! Groups Links

<*> To visit your group on the web, go to:

<*> To unsubscribe from this group, send an email to:

<*> Your use of Yahoo! Groups is subject to:

OVC discuss mailing lists
Send requests to subscribe or unsubscribe to arthur@openvotingconsortium.org
= The content of this message, with the exception of any external
= quotations under fair use, are released to the Public Domain
Received on Wed Aug 31 23:17:29 2005

This archive was generated by hypermail 2.1.8 : Thu Sep 15 2005 - 11:44:12 CDT