-Aspen's historic May 5, 2009 IRV election audited as single ballots- 5/5/09 Aspen CO held an instant runoff election (IRV) for mayor and 2 council members. Interpreted contents of each ballot, scanned by True Ballot, were publicly released. Open records requests for a CD of image scans were denied. Aspen has been sued to protect records from destruction and to allow inspection of the scanned ballot files. A Court of Appeals ruling holds that unidentifiable ballots are public records.

Search this and related blogs

Tuesday, April 13, 2010

Second Meeting of Aspen 2010 Election Commission- comments

At 3PM on Thursday April 8, 2010 the Aspen Election Commission met with 5 people in attendance- two of them the Aspen City Attorneys and other others being Jack Johnson (ex-City Council and currently a candidate for County Commissioner), Elizabeth Milias (ex-Aspen Election Commission) and Harvie Branscomb (Eagle County Canvass Board and editor of this blog).

The meeting was very calm and reasonable and productive.  The Commission decided to take in written complaints/compliments/suggestions by May 8 and looks like it will have a public brainstorming session sometime after that. The Commission primarily discussed how it would interface with the public, how it might help avoid any retribution for coming forward with complaints, and whether or not it would weigh in on the method of IRV voting or the choice of voting methods. Apparently one conclusion was to wait for a determination of what voting system would be used in the next election before further discussing IRV related details, but in the mean time to ask citizens to come forward with comments.

[Here are comments on the meeting by Elizabeth Milias, followed by an email sent by Harvie Branscomb to the Commission following the meeting.]

I found the following to be interesting points made at yesterday's meeting of Aspen's Election Commission:

• The "confidential memo" that outlines the EC's roles and responsibilities (as prepared by City Attorney John Worcester) remains confidential and privileged. The EC has received a copy, but its contents are not yet available to the public. It is up to Council to release this document to the public.

• The EC has determined that its individual members can have private meetings with citizens who wish to discuss election issues.

• The EC recognizes that there exists the potential for reprisals against citizens who bring forth election complaints and challenges, but the commission will not be part of this behavior.

• Citizens are welcome to write, call, present to, and lobby EC members regarding election issues, and despite Marilyn Marks being deemed "corrupt" for doing just this following the May 2009 election, this is now ok.

The following points were discussed as potential topic ideas for the 4/20/10 EC Worksession with Council:

• Civilized discourse
• No use of names when discussing how voters voted
• What is the EC's role in election reform?
• What are Council's expectations of the EC? What are the roles' parameters?
• Confirm with Council that this EC's term is legal until July 2011
• Get a commitment from Council that there will be a binding ballot issue on the November ballot on IRV
• Discuss/determine the auditing and certification process
• Debate using a hand-count for final/official tally of votes

I was personally very concerned with the EC's aversion to dealing with issues pertaining to IRV, instead wanting to wait until after the November election when IRV might be repealed. It seems that they realize what a mess the recent IRV election was and simply don't want to get involved if we change the law (as if Aspen's election problems will disappear if IRV does). Sadly, this means that they are not terribly inclined to dig into the past election, rather just look forward. The Clerk was particularly forceful in her efforts to "table" IRV issues until after November. She made many attempts to bring most of the election issues (Zimet complaint, Marilyn's issues, etc) under the IRV umbrella, however, the EC realized that several of these issues are not solely tied to IRV. There was also a strange aversion by the EC to having/taking any role in election reform efforts.

Another concern is the apparent willingness of the EC to look to Council for direction, rather than the other way around. This will play out at the work-session -- I hope I am wrong!


From: Harvie Branscomb
Sent: Friday, April 09, 2010 1:04 AM
To: 'John Worcester'; 'Jim True'; 'Kathryn Koch'; 'ward.hauenstein'; 'Bob Leatherman'
Subject: law that requires report of vote patterns

Election Commission:
I think I heard a small but significant  error in describing what is in Colorado statute today during the meeting.

The yellow marked sections of statute below show how the Colorado  Revised Statutes  requires a report containing individual vote patterns to be issued when using a ranked voting method. Note that this  report is confusingly named a “ballot image report”.  The first citation  I show below is the municipal code, that then refers directly to the quoted section of Title 1.

Until today when I think I heard Mr. Worcester state that there was no such requirement for a report of “cast vote records” or “strings”, I had assumed that Aspen chose TrueBallot, Inc. in order to be able to produce the report required by C.R.S. 1-7-1003. TrueBallot has an almost unique service to execute what they refer to as  a machine enabled hand count that can produce a “cast vote record” or in CRS parlance a “ballot image report” with a data record for every ballot showing the “order in which the elector ranked the candidates”.  TrueBallot did produce that report  (buried inside a Microsoft Access database within a zip file) for the May 5 election.  The hand count I described today in the meeting  i.e.  “sort and stack”  would not produce such a report, but it would serve as a very good independent count (hand interpretation followed by hand accumulation  of numbers of sorted ballots) that would satisfy anyone who had a question about the accuracy of the TrueBallot approach or the IRV methodology  (i.e. an effective audit of the hybrid electronic/hand operated approach utilized by TBI and an effective means of performing a recount.)    The TrueBallot approach to tabulation  itself would be more effective if it were run more slowly, so that the public present or watching on TV could really in practice verify each interpretation of each ballot- and of course this tabulation process could also be replicated by any member of the public in their own time at their own location  if both the “cast vote records” or vote patterns were made available (as they were) as well as the captured scans of the ballots (as they were not in Aspen’s case).    For a complete verification it is necessary for election officials to establish trust that the ballot images themselves accurately portray the contents of the physical ballots.

To be clear, there are three levels of tabulation/verification I am talking about- the first being 1) the initial tabulation via for example TBI’s visual but also electronic process that creates an interpretation of each ballot and then proceeds to separately accumulate these interpretations into an overall set of results  (supplemented by a truly random verification of physical ballot to ballot image), 2) a chance for home replication of the interpretations and tabulation  using two sets of digital records  created by TBI, and 3) a sort and stack hand count of the physical ballots (must be conducted under election judge supervision) that would implement the actual IRV rules in placing the ballots in piles.  Having the three of these in place would make the ideal verifiable election.

Note that one big advantage of having the “cast vote records” in public hands is the possibility for the public to report precisely which ballot images and corresponding “cast vote records” might have been interpreted in  error.  This is far superior to receiving a report from the public of a different result or outcome without having any way to reconcile the differences with the official interim counts.  All three of these methods can be also used to great benefit for a conventional election method such as vote for one plurality.  The fact that many existing voting systems hide the individual ballot interpretations (e.g. Diebold Accuvote)  is not to their credit, and make elections much more difficult to assess and to verify.  I think that performance of all three of these levels of verifiability   is well within the range of existing Colorado law and also practical for Aspen, except that access to ballot images has recently been disputed in  Aspen.

 Colorado’s Title 1 deliberately allows use of improved audit practices on a county by county basis- but this law doesn’t directly affect municipalities. However, it appears that municipalities, particularly home rule,  have their own freedom of choice in regard to many election procedures such as means of public verifiability.

The Millard Zimet complaint should not be taken as a reason to step away from capturing or releasing a report of  “cast vote records” or vote patterns.  The sensible reaction to the Zimet claim is to deliberately  shuffle ballots as they are removed from precinct ballot boxes and to shuffle the ballots again before any central scanning takes place maintaining a separation by their original origin such as precinct and mail-in.  This makes any association of ballots to voter by use of the voter sign in order ineffective.    The Millard Zimet claim must not be used as  a reason to depart from either IRV or from single –ballot – audit ability.

Thank you very  much for a productive meeting.

FYI on one other point, I strongly recommend not using the “override function” that requires use of the Accuvote key.  The reason is that Accuvote will correct more than one “error” without revealing what  it is doing  to the election judge or the voter. It is far superior to have the voter, or a duplication board rewrite the ballot such that it passes the criteria set by the machine.  The key is still needed to respond to certain kinds of jams and mis-feeds.     This is a longer conversation than there is space for here.

Harvie Branscomb

31-10-617. Ranked voting methods.

(1) Notwithstanding any provision of this article to the contrary, a municipality may use a ranked voting method, as defined in section 1-1-104 (34.4), C.R.S., to conduct a regular election to elect the mayor or members of the governing body of the municipality in accordance with section 1-7-1003, C.R.S., and the rules adopted by the secretary of state pursuant to section 1-7-1004 (1), C.R.S.

(2) A municipality conducting an election using a ranked voting method may adapt the requirements of this article, including requirements concerning the form of the ballot, the method of marking the ballot, the procedure for counting ballots, and the form of the election judges' certificate, as necessary for compatibility with the ranked voting method.

Source: L. 2008: Entire section added, p. 1252, § 4, effective August 5.

1-7-1003. Conduct of elections using ranked voting methods - instant runoff voting - choice voting or proportional voting - reports.

(1) A ranked voting ballot shall allow an elector to rank as many choices as there are candidates. However, if the voting system cannot accommodate a number of rankings equal to the number of candidates, the designated election official may limit the number of choices an elector may rank to the maximum number allowed by the voting system; except that the number of choices shall not be less than three.

(2) A ranked voting ballot shall allow an elector to rank up to two write-in candidates. A vote for an unqualified write-in candidate shall not be considered a mark for a candidate.

(3) (a) In an election in which one candidate is to be elected to an office, the ranked voting method shall be known as instant runoff voting. The ballots shall be counted in rounds simulating a series of runoffs until two candidates remain or until one candidate has more votes than the combined vote total of all other candidates. The candidate having the greatest number of votes shall be declared the winner.

(b) In each round of counting ballots in an election using instant runoff voting, each ballot shall be counted as a vote for the remaining candidate ranked highest by the elector, and the candidate with the smallest number of votes shall be eliminated.

(c) If two or more candidates tie for the smallest number of votes, the candidate to eliminate shall be chosen by lot.

(4) In an election in which more than one candidate is to be elected to an office in a multiple-seat district or on a governing body that includes multiple at-large seats, a local government may conduct a ranked voting election using the single transferable vote method, in which a winning threshold is calculated based on the number of seats to be filled and the number of votes cast so that no more than the correct number of candidates can win. The ballots shall be counted in rounds, with surplus votes transferred from winning candidates and candidates with the fewest votes eliminated according to the methodology established by the secretary of state by rule, until the number of candidates remaining equals the number of seats to be filled. A local government may also conduct an election pursuant to this subsection (4) using the principles of instant runoff voting specified in subsection (3) of this section to ensure that each elector has equal voting power and that an elector's lower ranking of a candidate does not count against the candidate to whom the elector gave the highest rank.

(5) (a) In an election conducted using a ranked voting method, an explanation of ranked voting and instructions for electors in the form approved by the secretary of state by rule shall be posted at each polling place and included with each mail-in ballot.

(b) A local government that conducts an election using a ranked voting method shall conduct a voter education and outreach campaign to familiarize electors with ranked voting in English and in every language in which a ballot is required to be made available pursuant to this code and the federal "Voting Rights Act of 1965", 42 U.S.C. sec. 1973aa-1a.

(6) In an election using a ranked voting method, the election judges shall not count votes at the polling place but shall deliver all ballots cast in the election to the canvass board, which shall count the votes in accordance with this section and the rules adopted by the secretary of state pursuant to section 1-7-1004 (1).

(7) (a) For an election conducted using a ranked voting method, the designated election official shall issue the following reports:

(I) A summary report listing the total number of votes for each candidate in each round;

(II) A ballot image report listing for each ballot the order in which the elector ranked the candidates, the precinct of the ballot, and whether the ballot is a mail-in ballot; and

(III) A comprehensive report listing the results in the summary report by precinct.

(b) The secretary of state may by rule establish additional requirements for the reports issued pursuant to this subsection (7).

(c) Preliminary versions of the summary report and ballot image report shall be made available to the public as soon as possible after the commencement of the official canvass of the vote pursuant to subsection (6) of this section.

No comments: