Election Updates

At yesterday's VTP research summit, there was a lot of exciting research presented. It's refreshing to see all of the research that is now being done just within the VTP group on such a wide variety of problems; not to mention that there is a lot of other really interesting and important research being done by other research groups and scholars now across the world!

We got a taste of the international dimension of research on voting technologies in our morning session, where we met electronically (via a three-way videoconference, Cambridge-Geneva-Florence). Our group traveled to the Swiss House in Cambridge, and we talked for nearly two hours with three top researchers and policymakers in Europe about the various pilot projects and technology-development efforts going on in Switzerland (mainly) but also in other European countries. The three Europeans participating were Michel Chevallier (Chancellery of Geneva), Alexandre Trechsel (University of Geneva) and Michel Warynski (Geneva Canton); on our end, participants were Ted Selker, Ron Rivest and myself.

Mr. Chevallier started by talking at length about the e-voting initiatives in Switzerland, their background and their progress since initiation. Professor Trechsel discussed his research on the social and political implications of e-voting (especially in Switzerland), in particular the results of their survey-based evaluations of the e-voting initiatives. Mr. Warynski concluded the European presentations by talking about the security features of the e-voting initiatives, and his argument was that the security features were well-developed and well-functional.

On our end, Ron, Ted and I focused on our current efforts to improve research and the electoral process. Interestingly, even though we made no prior effort to coordinate our presentations, in the end we ended up talking from three different perspectives about improving the research and adminstrative processes so as to achieve more openness, transparency and confidence in the American electoral process.

My presentation (pdfs of my slides are available) focused on better collection of election outcome and election adminstration data. Ron's presentation focused on verification, and Ted's (pdfs of his slides are available) was on monitoring. Again, it was interesting to see the convergence in our presentations, especially given that the three of us come at these problems from three quite different research backgrounds (Ron from crypto and computer security, Ted from usability, and I from social science).

A quite lively question and answer session also resulted from the discussion, also involving members of the audience who were participating from the Cambridge end. At some point we hope to have a recording of the proceedings, and if we do get that I'll make that available here (it should also be available from the VTP website.)

Next week there will be information posted to the VTP website about the rest of research summit. We had (by my count) fourteen research presentations after we returned to the MIT campus from Swiss House, and unfortunately not nearly enough time to adequately understand nor discuss the amazing research being done on voting system reliability, election administration, ballot design, voting system security, and usability. But we knew that we would have little time going into the event, and our goal was to try to have a discussion of the broad array of work that we and our various research groups are doing on these issues. As additional information is posted to the VTP website, I'll follow up and talk about some of the research projects, but all were really remarkable. We're making steps toward building a real research literature in some of these areas!

But until that information is up, I'll have to be content just leaving readers with a photo of the summit that Ted took early in the day of Jonathan Katz (Caltech) giving a presentation of work that he, Sarah Hill, and I have done regarding ballot recounting.

Michigan has been taking steps to verify the information contained within their Qualified Voter File (QVF), a statewide voter registry that receieved much attention in the wake of the 2000 presidential election when folks were looking for examples of statewide voter files. There is a brief story out about what Michigan is doing, and according to this story "approximately 169,000 obsolete voter records have been marked for deletion. Records are also being checked against the U.S. Social Security Administration's Death Index for removal."

Here is the process that Michigan is using to check the accuracy of the QVF:

Since last July, using funds from the national Help America Vote Act, the Department of State has reimbursed local clerks more than $220,000 for the costs of mailing new voter ID cards to registered voters-the main method for verifying those records in the QVF which are up-to-date.

If a voter ID card is returned as "undeliverable," the clerk sends a confirmation notice to the voter's last address on file. The notice informs the voter that his or her registration will be canceled if he or she fails to respond to the notice or vote over the course of the next two federal election cycles.

As a result of the mailings, the Bureau of Elections, working with local officials, has tagged nearly 169,000 obsolete records for possible removal. Under federal law, two national election cycles must pass before an obsolete record may be removed from the state's database.

County clerks forward a list of deceased adults to the clerk of each city or township within the county so the voter registrations may be cancelled. Recognizing that this is a burdensome task for election officials, Land instructed the Bureau of Elections to use data from other sources, such as the Social Security Administration's Death Index, in order to identify records of deceased voters in Michigan's QVF. Approximately 50,000 obsolete records were removed based on this data. To ensure accuracy, mailings based on the information will be sent to the address listed in the file for individuals in the QVF with voting history. The mailing informs the reader that the voter registration of the named deceased individual has been canceled and provides contact information if it is inaccurate. Roughly 13,000 notifications will be sent through this process.

It'll be interesting to see what other states do to check their lists, as other states get further down the path towards implementation of their statewide voter registries.

In the last several days there have been several efforts to improve the voting opportunities for military voters. First, the Overseas Vote Foundation issued a press release noting that they have created an online system for voters to find their local election official. This is critical because contacting these people is the first step to applying for and getting an absentee ballot. As they note:

Overseas Vote Foundation (OVF) today announced a new online voter service, the OVF Election Official Directory: http://www.ovf-eod.org. The Election Official Directory is a complete, online database of the 7,838 US Local Election Officials (LEOs) which is also offered as a licensed service to other organizations. It joins RAVAsm in the OVF suite of secure, online voter services available at http://www.overseasvotefoundation.org.

"With the Election Official Directory, OVF goes further than any other service to make all the contact information overseas and military voters need accessible - and in greater detail," said Janet Day-Strehlow, OVF's Election Official Directory Program Manager.

Overseas and military voters' questions commonly concern three main areas: 1) if a particular voting residence address will be accepted; 2) whether voter registration forms have been received; 3) and lastly concerning ballots, whether they were sent out, received and counted. "Questions such as these can typically only be answered by a local election official and the Election Official Directory facilitates that direct voter contact to those who usually can best provide the answers," stated Susan Dzieduszycka-Suinat, Executive Director, OVF.

The Election Official Directory will save voters and voting officials time and trouble in finding out who to contact with voting questions while giving them options to decide how best to make contact. OVF maintains the data and communicates regularly with local and state election officials to ensure that the Election Official Directory is up-to-date. "LEOs do not always have email, but we hope this service encourages more of them to take up the option. Email is best for communications with overseas and military voters as it avoids cost and time zone issues," added Ms. Day-Strehlow.

On a related point, the Washington Post reported that:

Twelve senators are asking Defense Secretary Donald H. Rumsfeld to make it easier for troops abroad to vote, saying the current system is outdated. Seven Republicans and five Democrats wrote to Rumsfeld this month, asking the department to create a new voting system that would allow members of the military to easily "request, receive, download and print" absentee ballots regardless of their location.

"Military absentee voting is still conducted in the same way it was conducted during World War II and the Korean War," the senators wrote. Samuel Wright of the National Defense Committee, a group that advocates for electing more veterans to Congress, said he has counted 7,838 different state and local election offices that administer federal elections and communicate with troops overseas. Wright said it would be ideal if military abroad could receive ballots by e-mail, fill them out and send them to their local election offices by regular mail.

Mike and I wrote about the problems associated with military voting in our book Point, Click, and Vote: The Future of Internet Voting. We have also examined the issues associated with non-internet, electronic voting platforms, such as emails, the web, and fax. The problem remains, as it has for 80 years, the difficulty of transporting ballots into and out of faraway countries. Wihtout some electronic transport, it is difficult to get ballots to voters and back from them in the short window that typically exists for voting. This is clearly an issue that the federal government needs to address, especially since it looks as though a very large contingent of forces will still be in Iraq in 2008.

I'll be heading to MIT tomorrow, for a two-day workshop with our MIT colleagues on the Voting Technology Project. We'll be spending Friday in two separate sessions. The first will be a workshop at the Swiss House in Boston, where we will be electronically conferencing with Swiss researchers and policy makers discussing their voting initiatives. The second session will be short presentations from many of the researchers involved in the project on both coasts.

There is a brief program on the VTP website. We are going to put materials from the VTP research presentations on the VTP website after the workshop, and are planning on making audio recordings of the presentations which should be downloadable from the VTP website as well. Watch this space for updates!

To continue our collection of stories relating to the Illinois primary yesterday, this morning there are a number of new stories that we'll keep following.

One of the more interesting stories concerns questions about the chain of custody of ballots in Cook County, questions being raised by candidates in a close race in Cook County, reported by Chicago's NBC5.

Not surprisingly, given that Cook County was implementing new voting technologies and new procedures to use those technologies, it seems that the election night counting process was slower than usual, despite what seems to be a relatively low voter turnout rate.

The issue seems to revolve around three problems, according to the NBC5 story:

... there were three separate problems, referred to as "election administration processes" by Neal. They are 1) paper ballots that were kicked out by scanners and, hence, not counted, 2) computer drives that held counted ballots but were not able to transmit the votes electronically, 3) flash drives that were not able to transmit from touchscreen voting machines.

Furthermore, the reporter

reported that moving trucks were arriving at election headquarters with uncounted ballots and scanners after midnight. He said there was no indication that the contents of those trucks were under any kind of guard, although the trucks were being met by election officials.

Election officials asserted that the chain of custody was protected, and that there were observers who were following the ballots at each step of the process.

In the end, this is an interesting example of how election officials do need to develop end-to-end plans to protect the security of ballots and all election materials, and to insure that those plans are followed to the letter. A key part of planning needs to be contingency plans, plans for when things go wrong, even minor issues like the three process problems that may have arisen in some Cook County precincts. Dealing with those procedural problems, and insuring that there is a clear chain of custody that can document the problems and steps to resolve them, is critical for confidence in the process and for insuring that elections can be accurately audited.

UPDATE: There is also a story in the Chicago Tribune that discusses the chain of custody issue. Here's more from that story:

Cook County Clerk David Orr said remote transmission of results had failed in "dozens" of polling places, slowing results. He said judges were sending some data cards downtown by cab for counting, after they were unable to merge and transmit results from the new touch-screen and optical scanning systems.

And in another precinct mentioned in the story, "The precinct's 61 Democratic paper ballots were dropped in a lock box, Masak said, and judges were told to drive them downtown."

It's now getting late enough in the day for initial reports to start to filter into the media of problems in the Illinois primary. But other than the election judge with her thumb stuck in the ballot scanner (who is oddly being helped by a candidate running for office), most of the reports indicate minor problems.

For example, WBBM radio in Chicago reports:

Earlier today, there were more glitches than usual in the early hours of voting.

Election officials said the problems were mainly logistical - election judges failing to show up, the wrong materials delivered to the precincts and other early bugs in the system.

If you do go to the WBBM site, make sure to listen to the short reprt there on "Changes in Voting Machines."

The Chicago Tribune is collecting reports of problems, again which are troubling but not widespread.

I'll keep updating this as the day progresses and additional reports come in.

UPDATE: 3:30pm (Pacific): A Chicago Tribute story from this afternoon reports "a few glitches", and notes that the Illinois attorney general's office reports "no widespread reports of election irregularities."

UPDATE: 4pm (Pacific): A report from ABC 7 in Chicago had some interesting details about the Illinois primary:

• They do report some problems with the new voting equipment, but note that the problems have been isolated: "ABC7 News has received reports from about half a dozen precincts of optical scanners malfunctioning."
  
• "On Chicago's Northwest Side, an election judge somehow got her finger stuck in the machine as Democratic gubernatorial hopeful Edwin Eisendrath tried to vote. She eventually was okay."
  
• Turnout is extremely low, especially in central Illinois were heavy snow kept voters at home.
  
• But here was an interesting report of attempts to confuse voters:
  

  
  Back on Chicago's South Side, hooded men were handing out what they claimed were the official palm cards for the 11th Ward. But Cook County Commissioner John Daley warned voters that they were fraudulent.
  
  The blue card is the real deal. The red card is a phony.
  
  "These candidates on the ballot we are not supporting. You can see none of these candidates are on the sample ballot that we've given out," said John Daley, Cook County commissioner.
  

  
  

This afternoon's AP story also focuses on the low turnout, and reports of few technical problems.

Doug Chapin forward to me this link of images from the Illinois primary in Chicago today. Pay particular attention to:

1. Picture 2, which shows a Democratic gubernatorial candidate pulling the thumb of an election judge out of a scanning device in a precinct. How she got her thumb in there is a mystery in the first place, and why they are allowing a candidate to monkey around with a vote scanner is another question.
   
2. Picture 4, where we see exposed wiring which the judge could easily disconnect (see it runs right behind his elbow, then behind the chair).
   
3. Picture 5, where we see an exposed wire running through what looks like an obvious part of the foot traffic pattern in a precinct.
   
4. Picture 6, where a ballot scanner didn't work.
   
5. Picture 9, where judges are looking, with a voter, at his voted ballot?
   
6. Picture 10, jammed scanner.
   

I'm sure we'll get more updates later.

UPDATE (9:25am, March 22, 2006): The Chicago Tribune interviewed Michelle Shafer of Sequoia Voting Systems about the judge's stuck thumb:

Shafer disputed a Tribune photo published on the Internet Tuesday that was described as showing a woman with her thumb stuck inside an optical scanner at a polling place. The photo also showed Democratic gubernatorial challenger Edwin Eisendrath helping pull her finger out.

"There is not a slot to get your thumb stuck," Shafer said. "There is a cover with a button back there."

Whether or not it is possible for her thumb to get stuck, there still remains the question as to why a candidate running for office is in such close proximity to a balloting counting device on Election Day.

Also note (unfortunately) that the Tribune is dynamically updating their photo database linked above, so that the numbering is now not the same as it was yesterday. The judge's thumb photo is now number 32 of 46.

Voters showing up in polling places today in Illinois are finding new voting technologies in their polling places, in counties large and small.

For example, in Cook County, voters there will be using voting technology provided by Sequoia Voting Systems. An article in today's Chicago Tribune summarized well how things will work for voters, and how they may face a problem with ballot design:

In what will be the first major hardware change in more than two decades, most Cook County voters will mark circles on paper ballots and feed them through an optical scanner, while those who are blind, disabled or just curious can use the touch screens, which are needed to meet a federal requirement for the disabled. Blind voters can use an optional audio ballot and keypad on the machine.

The new optical-scan machines will spit out ballots that are "overvoted," meaning more than one candidate has been incorrectly marked. But they will allow "undervotes," where people fail to mark a selection.

Election officials are especially concerned that voters will fail to realize the paper ballots have two sides. They plan to play instructional videos on a continuous loop at polling places.

As has been their strategy in the past, Sequoia is providing much assistance to Cook County as they roll out the new voting technologies, with many of the company's executives and other folks on hand to help with problems; according to the Tribune's story, the company will have perhaps 70 people on hand to help. Providing additional assistance will be a slew of technology-savvy workers from city and county government offices, approximately 100.

Of course, counties throughout Illinois are making a similar transition to Cook County, even though much of the focus has been on Cook County as it is one of the larger election jurisdictions in the nation. Of particular interest to me is following the transition in Stephenson County --- Freeport, Illinois --- where my parents once lived and I had the pleasure of spending a few nice summers and college breaks.

According to the Freeport Journal-Standard Stephenson County is also gearing up to use new voting technology, but they are using machines from ESS (63 M-100 machines, which cost the county $375,000 to purchase). No word on whether Stephenson County has vendor support for this roll-out, nor what types of contigency planning they have in place as problems crop up today.

One interesting side note here is that, while Stephenson County's chief election official is managing this transition, she is also facing reelection on the ballot today. She is listed on the Republican ballot (where she faces a challenger from her party), and there are no Democratic candidates who are on that party's ballot. So here's an example of an partisan elected administrator of elections, which has been the subject of recent research by Thad, Morgan Llewellyn and myself, in our paper "Who Should Run Our Elections? Public Opinion About Election Governance in the United States".

In any case, we'll keep an eye on what happens in Cook and Stephenson Counties today, and on events in Illinois more generally. While neither of us are on the ground in Illinois today, I'm certain we'll get email and media coverage of the primary election and any major snafus that arise.

The National Association of County Recorders, Election Officials and Clerks (NACRC) will host the following meetings in 2006:

• Spring Educational Conference, May 10-11, in Sacramento, CA.
• 2006 Annual Conference, August 3-6, Hilton Downtown Chicago, Chicago, IL

I just received my copy of the National Research Council's "Asking the Right Questions About Electronic Voting", which was recently published by the National Academies Press. This represents the final publication of the NRC's Committee on a Framework for Understanding Electronic Voting, a committee chaired by former governors Dick Thornburgh and Richard Celeste, and directed by Herbert Lin.

While this publication concludes the initial stage of the work of this committee (which I was on), there is now some planning going on for future work by the committee, including a planning meeting of some sort to be held in May. More details on that meeting when I have them.

There was an election today in Belarus and even before voting started there were claims of fraud. Interestingly, the fraud was not in the voting or the vote counting; instead, the state completely controlled the media in the election, ensuring that opposition party members either received no publicity or negative publicity and the ruling party received glowing reports. For example, this is what the NY Times reported today:

State television, meanwhile, offered hours of pro-government fare, including a series of interviews with analysts and observers, including a few Westerners, who praised the conduct of the campaign and the voting today. One observer, who said he was British, said, "There are several things that Great Britain could borrow from Belarus and use for its elections."

Later, as votes were being counted, state television showed unflattering pictures and video of the opposition and its supporters, including photographs of the several youth organizers from a Russian opposition party who came to Belarus to support the opposition here. The men were shirtless, drinking and smoking what appeared to be marijuana. This was followed by video of two transvestites engaged in a passionate kiss, and a suggestion from the news presenter that the opposition has the support of homosexuals.

NPR also noted that arresting and beating up opposition campaigners and candidates tends to promote a win by the incumbents.

ABC News reported that: The elections commission said 81 percent of the 7 million eligible voters had cast ballots by noon, clearing the 50 percent mark needed to make the election valid. Yermoshina said about 30 percent voted last week in early balloting, which is seen by the opposition as especially vulnerable to fraud.

Those of you who know Mike and I also know we love cool pictures of voting. The BBC has a set of election photos that illustrate the voting process, complete with food and vodka!