This page contains links to papers, presentations, information, and other websites concerning voting technology and systems.

Raw Documents

• The vast majority of computer science professionals support a non-electronic (i.e., paper) and voter-verified record of each voter's ballot. To this end, the Association for Computing Machinery -- the largest and most prestigious organization for computer science professionals -- adopted a public policy stance in August of 2004, supporting this position. This stance was supported by over 95% of ACM members polled, a staggering rejection of paperless systems. The statement is available in text (complete with a link to the original poll) or in Adobe Acrobat PDF format.
  
  Disclosure: I am a member of the ACM.



• I have a bunch of documents from the North Carolina State Board of Elections (SBOE) that I plan on posting online. These include documents from their training sessions, presentations by and to state election officials, and voting equipment test procedures. I will post the information as I have time to process it

  1. A copy of the North Carolina voting systems request for purchase (RFP) [PDF, 403.9 KB], along with the three addendum: one, two, three.
  
  
  2. The agenda of the August 18th, 2005 SBOE meeting and training seminar. A presentation on S223 was given at this meeting by Gary Bartlett (next item).
  
  
  3. See what the SBOE really thinks of our Public Confidence in Elections bill.

     Highlights

     • Notice that the estimated cost of developing the RFP was $400,000 to $600,000; how much of that went to Keith Long, formerly of Diebold in Georgia, Maryland, and Ohio?
     
     
     • "Decertification could be a four year process."
     
     
     • Also note that they expect the cost of source code review to be between $1.2 and $1.8 million; while this is reasonable, are they really going to follow through with that kind of audit?

  
  
  4. A report on precinct sizes put together by an "Edward Packard" at the SBOE. For the most part it just presents information about the number and size of precincts. However, there is a decent amount of focus on the limitations of "mega-precincts", or precincts with over 2,000 voters.
     
     Update: It seems that there is an "Ed Packard" who is an elections administrator for the Secretary of State of Alabama.

     Highligths

     • "The general consensus among election officials is that 1,500 voters is a desirable target for the base size of a voting precinct."
     
     
     • "While large voting precincts, ''mega-precincts'' (precincts with over 2,000 voters), may be useful under certain circumstances, they have their own limitations which may adversely affect activities on election day[.]"
     
     
     • "Even when centralized, individuals may have to travel farther to reach the polls, especially if the precinct is not compact. Additionally, highly urbanized areas experience problems with traffic congestion that coincides with the peak voting hours before and after traditional work hours."
       "When they reach the polls, voters in large precincts face the prospect of long lines, particularly during the peak voting hours. In past elections, the State Board of Elections office has received many calls from individuals reporting congested polling places and excessive waiting to cast a ballot. In some cases, these potential voters indicated they did not have the chance to vote due the inconvenience, and even burden, arising from these factors."
       "''Mega-precincts,'' especially those that encompass large geographic areas, are more likely to be split among various districts for the same offices. For example, two or more state senate or state house seats may represent different portions of the same precinct. As a result, polling officials in this ''split precinct'' must properly administer multiple ballot styles and ensure that individuals cast ballots only for the races in which they are qualified to vote. Historically, the use of multiple ballot styles in a split precinct has the created confusion on the part of some election officials and voters.

  
  
  5. A presentation about new and pending election legislation at the federal level. This was given by Don Wright, the General Counsel of the SBOE. Nothing too exciting here.
  
  
  6. The agenda of the August 19th, 2005 SBOE training seminar.
  
  
  7. A presentation on how to conduct Board of Elections meetings, including rules governing closed-door meetings.

     Highlights

     • "Open voting, the public has the right to know how each member voted."
       "May use ballots to vote, but they must be signed by the member voting."
       "Freedom of any person to photograph, film, tape record, or otherwise reproduce any open meeting."




• Wonder what the North Carolina SBOE has been doing with our Help America Vote Act Money? Check out an update from the NC SBOE [360 KB, PPT] from April 11, 2005. While it's somewhat vague, it's better than nothing. Note that the SBOE is prepared to push for Internet voting as part of their FY05 research.



• Look at the CalTech/MIT voting study [PDF, 44 KB] from 2001, which concluded that hand-counted paper ballots, optically scanned ballots, and lever machines were the most reliable voting technologies. DRE machines did about as well as punch cards.


• Have a look at the IRS Form 990s for The Election Center, a group that claims to be a non-profit organization, educating election officials about the latest laws, regulations, and technologies. Although apparently since 1997 [PDF, 1,252 KB] they've quadrupled the amount of their income they get from private investments and large donations; up to one-sixth by 2003 [PDF, 1,080 KB].


• Examine the documentation about massive flaws in Diebold Touchscreen Voting Computers; Diebold's own internal e-mail archives that acknowledge their product is hopelessly riddled with flaws.


• Have a copy of the 1990 [PDF, 452 KB] or 2002 [PDF, 4,332 KB] Federal Election Commission (FEC) standards for voting systems. Neither are too impressive.

Statistical Analysis of North Carolina Voting Technologies

• A cost comparison (592 KB) purchasing optical scan machines versus touchscreen machines. This is based on the ES&S price sheet available on the North Carolina Association of County Commissioners web site, and the 2005 per-precinct registration data available from the North Carolina State Board of Elections web site.


• A link for formatted election results from the 2000 and 2004 North Carolina general elections.
  
  The above link also includes data for the 2000, 2001, 2004, and 2005 general elections in Virginia.
  


• An analysis of undervote rates broken down by technologies.
  
  An analysis of undervote rates in Virginia broken down by technologies.
  


• An analysis of how undervote rates in North Carolina changed between 2000 and 2004. Charts are avaiable both with and without Carteret County included. An undervote (or residual vote) is one in which a voter goes to the poll but does not cast a vote because they can't bring themselves to vote for either -- or the only -- candidate, they don't know enough about the candidate, or they do cast a vote but technology in use does not record that they cast a vote.


• Check out per-precinct undervote rates for state-wide races. This data is a combination of the raw results [ZIP, 1580 KB] and the raw voter history/turnout statistics [ZIP, 4148 KB] from the North Carolina State Board of Elections website.

Presentations

• See what the Carteret Coalition for Verified Voting [MOV, 7367 KiB; MPG, 10898 KiB] has to say about paperless voting; several of its members were among the 4,500 who lost their vote in the 2004 General Election.


• I am a member of the National Committee for Voting Integrity. In May of 2004 I helped edit and contribute to the written testimony we submitted to the U.S. Election Assistance Committee.


• In December of 2004, I gave presentation about software engineering and security [PPT, 452 KB] to a Joint Special Committee of the North Carolina State Assembly. I continuted to serve as a technical advisor to the Committee throughout its work, resulting in bills H238 and SB223 in the North Carolina General Assembly.


• Starting in March of 2005, I have given presentations (2005/03/10 [PPT, 100 KB], 2005/03/28 [PPT, 216 KB]) about software engineering and preliminary statistical results from the 2004 General Election to various citizens' groups around North Carolina. The bottom line is that paperless machines are more expensive to own and operate, while performing worse than other alternatives (including punchcards).


• In April of 2005, I was privileged to give a presentation [PPT, 1092 KB] to the IBM North Carolina Technical Experts Council, a collection of high-ranking IBM executives, researchers, and product developers.

Other Information

• Just how effective and "worth it" is a voter-verified paper ballot, especially if it's only used by a small number of voters? Incredibly effective. Also available as a stand-alone document [PDF, 72K].


• An interview with an election official from a locality that uses optical scan devices. They discuss the process of preparing for the election, from tech support to machine testing to selecting and working with the ballot printing company. Very useful for any locality that is thinking of purchasing an optical-scan-based election system.


• Correspondence [PDF, 458 KB] between Don Wright, General Counsel of the North Carolina State Board of Elections, and Karen G. Gantt, Counsel for Diebold Election Systems. It seems that a Diebold technician ran the nuts-and-bolts of the 2004 General Election in Gaston County, North Carolina.


• Want to see how easy it is to create and edit electronic ballot images without knowing how the computer creates the original image?


• Want to know just one way someone might alter a Presidential election, thanks to DRE machines? This write-up was referenced in Bruce Schneier's monthly Crypto-Gram newsletter in June 2004.