Re: Alternatives to a single bar code

From: Steve Chessin <steve_dot_chessin_at_sun_dot_com>
Date: Wed May 05 2004 - 18:46:32 CDT

>From alan@openvotingconsortium.org Wed May 5 11:31:03 2004
>From: "Alan Dechert" <alan@openvotingconsortium.org>
>To: <voting-project@lists.sonic.net>
>Subject: Re: [voting-project] Alternatives to a single bar code
>Date: Wed, 5 May 2004 11:30:56 -0700
>
>Arthur
>
>> Yes/no ballots do not need rankings. Multiple choices do.
>> Generally, when there are lots of races and questions in a ballot,
>> most of them are yes/no (plus no choice made) questions, not multiple
>> choice races.
>>
>No. I am asking to see a specific ballot. Doug Jones has a bunch on his
>web site--none with very many contests.

Well, are you designing for the past, the present, or the future? For
"typical"/average ballots, or for "worst case" ones? For this year,
next year, or five or ten years from now?

When they design a bridge, they don't do it for average traffic flow.
They do it for peak traffic flow (all fully-loaded tractor-trailers),
and then add a hefty safety margin.

The June 1978 ballot had at least 13 statewide measures on it, because
that was the year of Prop 13. (And I'm not sure that it was the last
one on the ballot, either.) So lets assume at least 13 of those
measures.

Also, let's assume a future where IRV is used for every race, because
that's where the momentum is (San Francisco in 2004, Santa Clara County
and Berkeley in 2006, more of Alameda County in 2008, entire state by
2014), and because Kevin Shelly likes it and his HAVA plan favors
systems that accomodate ranked ballots. (See
<http://www.ss.ca.gov/elections/HAVA_finalplan_12-03.pdf> PDF pages 16
and 17. Note the references to "alternative voting systems".)

I picked the November 1998 ballot as an example that would generate a
long ballot. See http://vote98.ss.ca.gov/Final/Statement_of_Vote.htm
for details.

Every November 1998 ballot had, as partisan races, Governor, Lt Gov,
SoS, Controller, Treasurer, AG, Insurance Commissioner, State Board of
Equalization, US Senator, US Representative, and State Assembly. Half
had a State Senator as well. In 1998 California had 8 ballot-qualified
parties. So let's assume 8 names on each contest, plus a write-in. So
that's 12x9 or 108 ranked candidates that need to be captured.

[I'm treating "write-in" as one candidate, but of course different
voters can write in different names, and you have to capture the actual
name written in. Do you encode the written-in name in the bar code?
Or just that there was a write-in? How do you give audio feedback on
write-in candidates? I haven't tested the Sequoia Edge audio feedback,
but I'm guessing it just spells it back to you, the same as you typed
it in.]

For the non-partisan races, we had four Supreme Court justices (yes/no)
on each ballot, and as many as eleven appeals judges (yes/no) on a
ballot. So that's another 15 decisions.

Oh, and then there's the non-partisan Superintendent for Public
Instruction. Then and today, that's a two-round runoff election, but
no reason that shouldn't be converted to IRV as well. Add another five
candidates plus a write-in.

Then, in someplace like Redwood City, I might be voting on 3 Supervisor
seats, each with, say, 5 contestents, and 4 city council people, with,
say, a total of 12 running. (Again, this is ranked-choice voting.)
Plus County Clerk and District Attorney, each with three contestants.
So that's another 33 rankings to capture. Oops, I forgot write-ins, so
make it 42.

Oh, local measures. Perhaps another 10 local ballot measures (5
county, 5 city). Maybe 5 contestents for 3 at-large school board
seats. Then there's Board of Education (3 contestants for 1 seat),
Community College District, Water Board, and two special districts
(open space, and mosquito abatement), each with one seat and three
contestants. Oh, and write-ins. So that's another 20 rankings.

So we have 176 rankings to capture, and 38 yes/no races. And that's
without the "hefty safety margin". Probably want to double or triple
that to get the safety margin.

BTW, the decertified Poll Star system had 432 (12x36) punch positions.
That could conceivably have been 216 two-person races. So you need
to at least be able to capture that much.

--Steve
==================================================================
= The content of this message, with the exception of any external
= quotations under fair use, are released to the Public Domain
==================================================================
Received on Mon May 31 23:17:18 2004

This archive was generated by hypermail 2.1.8 : Mon May 31 2004 - 23:18:16 CDT