Chair, Bruce Danielson has been doing data research on the voting lists AND votes cast, here are some of his early findings he presented in the meeting today;
Since December we have been researching the voting process in Sioux Falls, Minnehaha County and the state of South Dakota. Each member of this committee has special skills and knowledge they have used as part of these discussions.
With this committee’s knowledge, my assignment was to research the data side of our voting process. I could go into a great deal of detail and bore each of you with too much detail at this time but won’t.
I have managed IT shops, done data recovery, plus studied and mined data for over 30 years. This data and process examination brought to light many weaknesses in the way our data has been compiled, processed, updated and saved. I have been disappointed with every government level involved. Every office involved with this data played its part in the process failures.
My Political Science and History degree gave me an insight into the processes we examine. The many years being involved in the political side of the voting process and my natural skepticism of modern voting systems brought me to the Minnehaha County resolution board for the first time in 2012. I needed to see it up close and personal.
The City of Sioux Falls petition drives in 2012 and 2013 gave me the need to buy voter lists to verify our signatures. The data issues I found in these files caused me to take a deeper look at the data side of the election process. The first files were created in late August then September of 2013 and February 2014. I was appalled by the sloppy nature of this data. In February 2014 I attempted to contact the Secretary of State’s office to inform them of the problems but no one would return my telephone calls. We learned to work with the hundreds of bad addresses we found.
As part of this report I yesterday loaded the December 2012 pre-conversion AS400 data used for setting up South Dakota’s TotalVote database. It is a database of old design but the data was fairly clean and the street directions appear to be correct (N, S, E, W).
Using data supplied by the Secretary of State’s office this week, my research has found at least 1087 repairs performed to correct some of the address issues. I have found more and I will be supplying these reports to the Auditor’s office for continuing updates.
The 2014 City of Sioux Falls and June primary elections were held using the same terrible data I was supplied with. I am not picking on this election but it is the most detailed information I have to analyze.
With this data research still in progress I have found a few highlights and questions needing attention:
- With all the bad data found, how could anyone do a voter purge in 2013 and feel secure it was correct? Did anyone controlling data the care?
- Who audited the purge to verify the process was correct before the action?
- How could data be modified and uploaded into a new database system without any quality control?
- Who wrote the procedures for the Driver’s License office and Department of Social Services to use to register voters? Where are the signer’s cards?
- There were approximately 121,000 Minnehaha County (including 97,341 Sioux Falls) records uploaded into TotalVote beginning in December 2012. By September 2013 there were 110048 records supplied to me of which 94,213 were for Sioux Falls. These records were uploaded with 385 different dates for the first 9 month period of operation. To use this file, I corrected this field to show 181 dates. A difference of almost 200. This encompassed thousands of voter information records with field errors.
- We had reports of people not being registered to vote any longer, how do people verify?
- What were the voting histories of those purged voters? (i.e., did they tend to be voters with long-standing histories of regular voting?) We don’t know.
- We have been asked what the demographics were and could they be independently verified?
- Do we know how many of those purged voters re-registered before the general election?
- Are there any non-malicious, purely data-based characteristics that could explain the update errors or purges as a “fat-fingered” error (e.g., could the system have purged voters based on the table issues, length of name, alphabetical sequence of name [some systems freak out and it deleted all names between J and M], birth month, day or year, precinct number, etc.)?
- Will the voting center data collected be included in the voter history table for the historical ability to audit? If the Auditor’s office is to be an Absentee polling place shouldn’t they be using the epoll book like the other voting centers?
- I asked for complete files with deceased, purged and moved data so I could balance the auditable information. The City Election comparison data is showing the date of 04/08/2014 with 31,676 votes cast vs. 31,957 reported. It would have been easier to audit with the full database.
- I have studied the City of Sioux Falls cityward field in both Lincoln and Minnehaha County data each uses different identification. Why does this matter? How about voting centers having different ballots styles to choose from and the epoll book not recognizing the correct cityward to give out correct ballot.
- Sioux Falls precinct data (prename) is identified by both counties differently. (SIOUX FALLS 1-11 vs. PRECINCT-0110)
- We heard many stories of people getting the wrong ballots, unusual voting patterns and being excluded from voting, the press stressed a few of them. I believe some of the causes start in the SOS TotalVote system.
- The street master system in Sioux Falls does not make not systemic sense. It requires the user to make assumptions not always based on principles of logic.
- In an attempt to understand the county / city GIS system, a file was supplied for me to work with. In a matter of minutes almost 600 errors were found. These were turned in for repair.
- The city, counties and state have been or are investing in major technology upgrades but seem to be forgetting to have audit paths as part of this process.
As part of our final report, we will be breaking this down further with more detailed recommendations. At this point every office I have dealt with in this process has a share of the blame for the issues we have had here. Thank you for your time. The floor is open to questions.