You are here: CopsWiki>Common Web>CommonProjects>ElectionIntegrity>SanDiegoCountyCanvassProcedure (2010-02-11, Raymond Lutz)EditAttach

San Diego County Canvass Procedure

Latest News

COPs has published a report on the activities of the San Diego County Registrar Of Voters and has published an extensive report, which can be found here --> COPs Rov Report

Our Concerns

  • Sloppy procedures
    • Instead of checking the number of ballots as they are processed and correcting problems early in the game, the registrar prefers to avoid incremental checks and instead relies upon a last-minute reconciliation step, which has no procedure or documentation. This is contrary to standard quality assurance recommendations that suggest that any procedure test early and test often to maintain quality.
    • The number of ballots received in mail and provisional ballots are not checked with the number actually scanned, and no scanner tapes are created to maintain an audit trail. The 1PMT does not limit the possibility that the election could be "fixed."
    • There are no operator notes or other documentation that we are aware of, that describes the decisions made throughout the canvass process, including why specific precincts are rerun through the system, what issues were resolved by trouble teams, and why it should be appropriate to claim that the election is complete and error free.
    • The number of mail ballots delivered to each precinct is not tracked. There is no way to know how many mail ballots were delivered to the polling place nor whether any were lost.
  • The number of ballots scanned does not equal the number of voters who voted
    • 65% of the precinct sign-in rosters do not agree with the count of voted ballots. In our 5% sample, 98.8% of precincts in our sample were in error. Only one precinct had no errors such that the number of ballots counted by precinct workers agreed with the number of voters who signed-in and also agreed with the number of ballots scanned.
    • After the number of ballots is determined by poll workers by hand counting them, reconciling the number of voted ballots with the number of blank ballots received, the number unused, the number spoiled, and the number used for provisional voters, the count of ballots actually scanned by the scanners differs in most precincts by a significant number of over 5 ballots unscanned per precinct.
    • The total number of ballots scanned by RoV workers on election night is not compared with the number of ballots provided, as counted by poll workers, and there is no effort to validate the count of ballots as provided by the polling place.
  • Poor security procedures
    • Tamper-proof seals are broken or incorrect in 6% of the cases, and no follow up investigation was apparently performed, and inadequate record keeping is performed to track seals by number to insure that tampering does not occur.
  • The One Percent Manual Tally proves there is a problem.
    • In the One Percent Manual Tally (1PMT), 29% of the precincts were defective in one way or another, and in several cases, rerunning the ballots through the machines confirmed that the original scans were at fault, and when rescanned, additional ballots were then included. However, the Registrar claims there was no error. In fact, the conclusion should have been that the scanning and counting process was seriously flawed.
    • The report of the 1PMT does not conform to the suggested format as published by the Secretary of State, but includes much less detail. If the format suggested by the Secretary of State were used, at least 400 pages for election night ballots and 400 pages for mail-in ballots would be expected. Instead of 800 pages, only five pages were provided, and the Registrar declined to improve their report to meet the format recommended by the Secretary of State. In comparison, Los Angeles county produced a 1300 page report.
    • The reports for the 1PMT provided by Technical Services (Charlie Wallis) could be the "good reports" while sufficient other adjustments are made election wide to change the results while still maintaining correspondence with the 1PMT.
  • The central Tabulator Audit Log:
    • did not contain initial commands to zero all the races.
    • Does not contain enough information to reconstruct the data sent to the tabulator to check that the correct data was transferred to it correctly, either through digital media on election night or by serial communication cable at other times.
    • Has several instances of crashes documented, but there was no explanation by Technical Services as to why the tabulator crashed. We were not provided with a contact at Premier Election Systems to follow up on these crashes.
    • Documents a large number of re-runs of precincts, but there is no explanation as to why these precincts are rerun.

Recent Articles and Media: (Add | All Media)

Project Form edit

Project Name? San Diego County Canvass Procedure
Project Description? Detailed written procedures for the canvass in San Diego County
Project Founder? Raymond Lutz
Project Curator? Raymond Lutz
Project Parents? Election Integrity
Project Status? Active
Project Type Issue Oversight
Thumbnail Link?

Edit | Attach | Print version | History: r40 < r39 < r38 < r37 < r36 | Backlinks | Raw View | Wysiwyg edit | More topic actions
Topic revision: r40 - 2010-02-11 - 19:26:58 - Raymond Lutz
 
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding CopsWiki? Send feedback