|Articles: Home All 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 ES & S Model 100 optical scan system testing used by Hennepin 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 if one assumed that the code worked perfectly. I would not make that assumption as I could not inspect the code used by the QNX based ES&S devices, and the code was not offered for inspection.
The test was basically the same as Ramsey County, all precinct variations in the political district had a scanner and a test deck of ballots. All were run to get a "Zero tape" then the test deck of ballots was read one at a time and the final count run. There were no problems found with the test decks.
However, there are omissions in the public test that seem to to be violating state statutes for testing voting systems. As was missing from the Ramsey County public test in 2004, in the Hennepin County public test 2006 the central server which totals all the votes was not tested. The central server is the point all precinct scanner counts go to, totals from each precinct vote reader are transmitted to the central server by a wireless system. The total from each reader is counted together by the central server for "preliminary totals" and reported to the media. Apparently "official totals" are counted differently, I never saw the test of the final procedure to get the "official totals" or the preliminary totals.
The main difference in testing in Hennepin County is that the public voting test can be at the city level, I witnessed the Minneapolis public test. The machines were essentially the same as Ramsey County, optical scan machines, just a different vendor, again, the central county counting system is not tested, the same problem as I found in Ramsey County. The ES & S scanners are now used in most of the counties in Minnesota, Diebold has 4 of the metropolitan Minneapolis-St. Paul area counties.
The ES & S scanners used in most Minnesota Counties have the same vulnerabilities as all optical scanners, the non-government outside consultants used in the central counting and as trouble shooters for the election and the proprietary code used to run the scanners and central count systems. Chances of a central count being overturned would be very small, and the checking process is probably also involving consultants and can get very expensive and complicated and could take a lot of time and the results could be very complex and unclear. I would guess that getting to the "consultant" is the best way to manipulate an election at the county or regional level, other actors would be the people with control over the consultant group and the proprietary software, that would be ES & S in this case.
Is this a probable risk? One only has to check the state "lobbying" expense reports to find out what it takes to get influence in the Minnesota voting process, ES&S and Diebold are the only vendors in the state, ES&S reported lobbying expenses of about $40,000+ during the HAVA equipment buying gold rush in the last few years ($0 in 2003, $7000 in 2004, $40000 in 2005,) to change most of Minnesota non-metropolitan counties to electronic optical scan voting, Diebold reported nothing in lobbying expenses. In 2005 83 counties now have ES&S equipment for $40 million in HAVA and state funds, Diebold has the same four counties as before.
Was the lobbying effective? Well, close to a 1000 times payback seems to show that it was. So how much risk to move an election one way or the other using consultants or a change in code in the scanners or the central counting system or just a mistake? The risk of overturning a declared election is very low, getting caught seems to be very low, getting caught and actually punished is lower still. To find if it is deliberate or not may be very difficult expensive and slow, and the process easily influenced by other actors in the media, judiciary and political bodies. I see the headlines already: "Computer Glitch Found But Election Stands." Just an error in the process or a deliberate election fix, it does not matter, the process to audit and determine a problem would take too long for an effective countermeasure to work, most actors would escape unscathed. I cannot ever remember any state wide election overturned for problems found, once a winner is declared it is too hard to change unless there is a revolution.
I could not validate any software or systems at the "Public Test." in Hennepin or Ramsey County. 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." In the larger counties of Hennepin and Ramsey the city and county staff did much of the programming. Smaller counties will rely on the ES&S technicians for setting up the precinct scanners. The targeting of the central vote counting systems is an even larger risk as it is untested and involves the outside non-public actors, the consultants for all counties.
Absentee ballots differ from "regular" ballots because they are folded to be put in an envelope. Folded ballots were not tested on the ES & S Model 100. This may be an issue as the absentee ballot is generally folded several times and then unfolded flat to be run through the ballot counter. Wrinkly ballots are more difficult to run through the scanner, I observed one ballot with an almost imperceptible wrinkle on the edge, it had a bit of a problem going through. I would suggest adding a test deck of ballots folded in the manner of a typical absentee ballot to check the accuracy of the scanner to handle and properly count folded ballots. I would add some more atypical folds too. I do not remember a folded ballot test in Ramsey County either so I think this is a common oversight in testing.
Pen scribbles in the "timing mark" area seemed too uniform for a good test. (Timing marks are black rectangles like a big bar code on the side of the ballot that tell the scanner where the votes should be located.) The scribbles were large and dark and did not seem to be what a voter may do to a ballot by mistake, such as a test scribble to get the ink flowing in a pen on the side of the ballot or a line or smaller dot on the side of the ballot that mimics a timing mark or even a "write in" word or several other types of errors. Or even a possible fraud technique of placing a fake mark in the timing marks to misread the votes and change the results.
From the public tests I have observed I can only conclude optical scan voting equipment, by ES & S or Diebold, is a significant risk and does not have the transparency needed to assure that fair elections are currently done in Hennepin County. Outside companies that "own" the means of voting with closed proprietary systems can influence results. In small counties state wide the risks are greater as the opportunity to manipulate the results are easier in smaller counties with fewer expensive technical staff that can properly oversee the ES & S consultants.
Are other "computerized" systems better? I would say no. Many other systems do not even have a readable paper ballot trail that the optical scan systems do have and have costs that are very high to service the number of voters a single card reader can process. Because the large counties have a larger technical staff the possibility of a consultant influencing results at the precinct level is less, for smaller counties with a smaller less technical staff that are unfamiliar with programming the optical scanners it is a much larger risk that the consultants can influence an election and be undetected.
List of articles on fixing elections, vote fraud, election manipulation