more on UK Independent on Florida E-Voting

[David Farber <dave () farber net>]

Begin forwarded message:

From: "Steven M. Bellovin" <smb () research att com>
Date: September 29, 2004 7:25:46 PM EDT
To: dave () farber net
Subject: Re: more on UK Independent on Florida E-Voting

In message <5C9850EC-1264-11D9-B62A-000393D166C6 () farber net>, David 
Farber writ
es:
> More phantom non-voters showed up in an election in Palm Beach
> County in January. Again, those supposedly present but not voting
> (137 people) greatly exceeded the margin of victory (12 votes).
> That persuaded a local Democrat Congressman, Robert Wexler, to
> sue LePore and the state of Florida to force them to adopt a
> paper trail. The case is pending.

This is not the first report of large-scale undervoting with
touchscreen electronic voting machines.  I suspect that a paper
trail will prove nothing, because (in this case) I suspect that
the machines are accurately recording what happened:  no votes were
cast by those people.  But the reason for that is a bit more complex.

With a paper ballot, hand-counted or mark sense, you know when you've
voted for someone.  WIth a lever machine, you have to pull a large
handle that both records your vote *and* opens the curtain -- and the
curtain has metal rods in it, so it's hard to just push it aside.  In
other words, the voting machine itself helps protect against inadvertent
undervotes.

WIth the new elecgtronic machines, however, the curtains are just that,
curtains, and they're meant to be pushed aside.  To record your vote,
you have to press a special button, but there's no mechanism to make
sure you do so.  That makes it easy to forget.  I chatted with some
of the poll workers in my precinct about this during this year's
presidential primary (New Jersey's was held in late June, long after
the race was decided, so turnout was *very* light -- hours after the
polls opened, my wife and I were voters #5 and 6.  They had plenty
of time to talk....)  They confirmed that the was a significant problem.
Now, properly-trained poll workers will notice that something is wrong,
because the machine does have an interlock preventing the next voter
from being authorized while a previous vote is pending.  If turnout
is light, or if the poll workers aren't watching for the problem 
(visible
via external lights), undervotes are very easy.  But I can easily
envision problems during a busy election, when things are very hectic
and it's hard to call back someone who, after waiting on a long line,
is rushing for the door.  What should the poll workers do?  Press the
"cast vote" button?  Reset the machine?  What are the *procedures*,
both in Florida and around the country?

There are two morals here.  The first, of course, is that procedures
matter; the second is that subtle human factors issues can be
extremely important.  Most of the fuss has been about the software;
very little attention has been paid to how these machines are actually
used.

-------------------------------------
You are subscribed as interesting-people () lists elistx com
To manage your subscription, go to
 http://v2.listbox.com/member/?listname=ip

Archives at: http://www.interesting-people.org/archives/interesting-people/