| Articles: Home List of articles on fixing elections, vote fraud, election manipulation |
Minnesota state statute 206.83 Testing of voting systems says, "The official in charge of elections shall have the voting system tested to ascertain that the system will correctly count the votes cast..."
The Diebold ES Accuvote 2000 optical scan system testing used by Ramsey County was by my judgement a reasonable test for the public of the optical scanning of ballots and the vote count at the scanner level.
However, there are omissions in the public test that seem to to be violating state statutes for testing voting systems. The central server which totals all the votes was not tested in the public test. The central server is the point all scanner counts go to, totals from each precinct vote reader are transmitted to the central server by modem using the public telephone network. The total from each reader is counted together by the central server. According to Ramsey County staff, there would be a contract Diebold technician present during the election, mostly for central server problems. Diebold GEMS central servers have been found to be vulnerable to many problems, including unauthorized software installed by Diebold and unaudited manipulation of the totals using MS Access.
If the central server is used to count votes and is part of the voting system it must be tested or it cannot be used as part of the voting system. If the central server software is not used, I still saw no test of the final vote count, the method of counting the final vote is unknown to me and after a several requests to the county staff and my Board member Raphael Ortega I was told by Ken Iosso, Ortega's aide "I talked to Joe Mansky and he told me that the official vote totals come from the "ticker tape" from the ballot machines at each polling place. The totals that come from the central server are the "unofficial" totals and until they are checked against the paper tapes they are not official."
Well, now I have some idea what is going on, but still... The central server and checking process to count the total votes used in the election is not tested in the public test, the details of the checking are not transparent. So, it still looks to me that the central server is used in the voting system, it is still untested and uses a checking process that is not transparent to me as a part of the voting public that has followed the due process of checking the voting system through the public test and inquiry of my county board member. I did not see the procedure to get the final count described in the voting judge guide or any other materials. The state statute seems clear to me, the voting system must be tested in the public test. I request that the process used to get the vote totals be documented and tested in the public test. Please, board member Raphael Ortega, bring this to the county board for action.
Here is the full state statute for reference: Minnesota Statute 206.83 Testing of voting systems.
"The official in charge of elections shall have the voting system tested to ascertain that the system will correctly count the votes cast for all candidates and on all questions within 14 days prior to election day. Public notice of the time and place of the test must be given at least two days in advance by publication once in official newspapers. The test must be observed by at least two election judges, who are not of the same major political party, and must be open to representatives of the political parties, candidates, the press, and the public. The test must be conducted by processing a preaudited group of ballots punched or marked to record a predetermined number of valid votes for each candidate and on each question, and must include for each office one or more ballot cards which have votes in excess of the number allowed by law in order to test the ability of the voting system to reject those votes. If any error is detected, the cause must be ascertained and corrected and an errorless count must be made before the voting system may be used in the election. After the completion of the test, the programs used and ballot cards must be sealed, retained, and disposed of as provided for paper ballots."
My county is a large county in Minnesota, over 500,000 people. It has a good sized, competent, well trained staff that can handle complex procedures. Even with this skilled and experienced staff they must contract a Diebold person to hold their hand with the central vote server the day of the election. A central point to change vote totals seems to be a technical weakness in the staff of a large sophisticated county.
I am not sure that a county with fewer resources would be able to reasonably use vote scanning systems at the scanner level without a lot more outside help. This would seem to make a smaller county more vulnerable to election problems, even at the precinct level.
If the counting scanning machines were to be totaled by hand with a documented procedure and that procedure was to be tested in the public test, then Ramsey County would meet the criteria of "Minnesota Statute 206.83 Testing of voting systems." Ramsey County could also save money by not hiring a Diebold contractor and remove a vulnerability in the voting system by not using of untested centralized reporting and totaling software that is controlled by an outside interest.
BlackBoxVoting.org has found many vulnerabilities in the Diebold scanners, not just the central counting system but even the precinct level ballot readers that can be programmed to "fix" an election. Ballot readers can be programmed to pre-load negative and positive number of votes so the "Zero Tape" test can total "zero" but the ballot reader is set to put the thumb on the scale. For example, candidate A is starting out with -37 votes and candidate B has +37 votes allowing a lovely "Zero" start to the ballot counting. If 200 votes are cast split 100 for A and 100 for B, then the total is 63 for A and 137 for B. 200 votes are totaled and B wins. In the Public Test I saw the "Zero Tape", but I could not verify the ballot readers were not pre-loaded for an election.
I could not validate any software or systems at the "Public Test." I saw no source code and did not check for the vulnerabilities mentioned by BlackBoxVoting.org. The targeting of specific precincts by programming ballot readers at the precinct level is a significant risk in the current system that was not covered in the "Public Test."
From the public test I can only conclude scanners are a significant risk and do not have the transparency needed to assure fair elections as currently done in Ramsey County. Are other "computerized" systems better? I would say no. Many other systems do not even have a paper ballot trail that the scanners do have and have costs that are very high to service the number of voters a single card reader can process. At the actual polls I saw people feeding a vote into the scanner between 5-10 seconds each. That is 6-10 people per minute because the ballots are marked before scanning and do not occupy machine time. Touch screen voting machines include the ballot marking time, several minutes, even 5-10 minutes per ballot which is exponentially slower or exponentially more expensive to get the people through the system.
Public tests and voting procedures are mandated by statute and governed by rules written by the Minnesota Secretary of State. The public test was of the Diebold ES 2000 Accuvote scan ballot readers used in the voting precincts of Ramsey County. One test reader was used for each combination of the candidates and questions that would appear in the Ramsey County ballots, more than the minimum described in the Secretary of State Public Accuracy Test rules. The people conducting the tests were election judges, the county staffer who programmed the reader was present for questions. Dorothy Mcclung and Joseph Mansky the staff people in charge of elections were present for questions. The public test was basically: - run the first report from the ballot readers ( a zeroing report), - run the deck of 60 ballots through the ballot readers, - run the result report from the ballot readers, - compare the result report to a spread sheet of expected results. At the test on October 29, 2004 at 1PM, County election judges and staff were present to answer any questions about election procedures, an election judge manual was available for distribution, questions about election monitors and challenge procedures were carefully covered and the challenge procedure described at the test was followed in the election, I heard nothing during or after the election to indicate otherwise on the subject of voter challenges. I was the first person to ever show up for a public test in the memory of the County staff. The second showed a few minutes later, a lawyer from the Kerry-Edwards campaign who had a list of questions about voting procedures, including the systems. The Kerry-Edwards representative was especially interested in the vote challenging procedure. The procedure described by county staff was: - all challenges must be presented to the head election judge in writing, - challenges must be of personal information of the voter only, - no "automatic" challenges by looks, voice accent, clothing, etc. would be used by judges, - no lists of voters could be brought in to the voting area to use for challenges, - no direct questioning of voters by challengers, only election judges would ask a set of questions, a challenger can talk to a voter in the presence of the election judge. - the judgement of the head election judge was final, - any monitor that made a pattern of false challenges or became an obstruction to voting would be removed from the precinct voting area. Apparently this challenge procedure was in used in the operation of the Ramsey County November 2, 2004 election with no trouble during the election. A County staffer does the ballot layout and programming for the vote counters and the server. No Diebold people are needed. There is a strong user group in the Metro counties that uses the same voting system. Ballots are sealed after the elections. Absentee ballots are the same ballot as everyone else and are run at the precinct on the same counter as everyone else. Absentee ballots are delivered to the precinct to be hand run during a slower time during midday. A recount will count the ballots that are not read correctly by the counters. Otherwise ballots that are an undervote by the counter will not be counted. Spoiled (rejected by the counter) votes go into a spoiled vote envelope. Felt tip pen could smear on ballots, the county now uses a ballpoint. Voting counters are cleaned before each election season in June-July. Ink and paper dust are the typical cleaning problem. Fewer people than the "dirt limit" are in each precinct so no problems with dirty scanners are usually encountered during the primary - general election season. Ballots are not individually numbered but are sealed in shrinkwrapped 50 count numbered sets. Diebold ES 2000 reader has a battery backup that lasts a couple hours. Voting can occur, the paper votes are held, then run through the machine with battery power. I estimate about 1,500-2,000 votes per hour can be run through one counter, about one every 3-5 seconds. A rough audit of the number of votes per precinct is to compare the ballot receipts to the number of ballots cast. The number will rarely match perfectly, because of spoiled ballots or if people walk out without putting the ballot in the counter. Another check is to compare the precinct voter books (signed by the voter before voting) with the number of ballots cast. The numbers should be very close, however. Precinct election judges have an incident log for stuff like a power outage. Separate bin in under the counter to hold uncounted votes if the counter cannot be used. Like a faulty memory chip or power outage. There are duplicate memory cards for each ballot counter so each can be replaced. The memory cards are put in with a numbered "tamper seal". Modem ports are sealed on the counters until the machine is ready to send information to the central server at the county office. After voting is finished an audit tape (report) is run, then the information is sent by modem to the central server, which must be offline until after polls close, then two more tapes run. Diagnostic tests are done on each counter once per election season. The usual problems are: power supply, modem, printer, ballot reader (sensor), all can be easily removed and replaced. The scanner uses timing marks on the side of the ballot to match the marked circles to the candidates and scans a narrow line close to the center of the vote count marked circles. Public test details: -60 ballots were used with each reader, write-ins were included in the test ballots. -4 ballots in each set of 60 were to be rejected by the reader and tested the following: - a single overvote, - all overvotes, - all timing marks scribbled over, - another timing mark scribble I did not record in my notes. State Voter Registration System Notes: The new State of Minnesota Voter Registration System did "crash" during the public test at 1:45 PM Friday October 29, 2004. County staff came into the public test area to inform the other county staff. The staffer was told to shut down all the machines, wait a few minutes and try again. This was said by County staff to be a daily occurrence when there were many registrations. Verification of voters is done in a separate batch job at the state. 10-15% of registrations are sent back to the county to be rechecked. Less than 1% of the rechecked registrations are not added to the database as voters. 125-150 rechecks per hour are done by the county to verify voters. Typical problems encountered at the state is one of the drivers license or state ID off by one character, Date of birth differences with the same name as someone else, any difference of name, DOB, or id between the voter registration and the Public Safety database. Peak Ramsey County registration rate before the election is 2000-2500 registrations per day. 2/3 of the registrations are done on voting day, 50,000-60,000, so 5000-7500 registrations will be rejected by the state. Minnesota has voting day registration.
List of articles on fixing elections, vote fraud, election manipulation