The Voting Transparency Project : free software for fair elections
 
Polling Place Simulator v0.12 demo
 
What's new in the simulator
v0.12: GIF images for better compatibility with older browsers and Internet Explorer, plus other code and presentation cleanups.
Animate applet: nicer clock, improved UI, auto-start, internal changes to aid future versions.
The e-mail address for bug reports/comments appears at the bottom of this page.

About the project
The Voting Transparency Project creates tools for the study and evaluation of local public elections. Our hope is that these tools will help elections officials, citizens and equipment makers better understand and manage their own elections processes, supporting informed debate, and revealing potential problems with both technology choices and the operating practices around the technology.

or (in more clever language) We create tools and processes that support the design, operation, analysis and debugging of the tools and processes that run public elections.

Elections officials get two or three opportunities per year to test their equipment and processes - during live elections when so much is at stake and experimentation is not appropriate. This prompted our first release, the polling place simulator: a tool to estimate polling place capacity needs, and to illustrate the effects of changes in conditions on polling wait times and queue sizes. With the simulator and future projects, officials and interested citizens can experiment with election-day resource planning.

The project had a Spring, 2005 home with the Computing Culture Group at the MIT Media Lab.
We thank the group and Director Chris Csikszentmihalyi for their support.

Goals

  • Build tools to help citizens and officials debate, study and understand the local voting proces
  • Constrain the vocabulary used in discussions of the design and execution of voting processes, to encourage focus on the voting process
  • Support comparisons of voting processes, technologies and policies
  • Encourage collaboration between citizens, officials, and vendors

Overview of Deliverables

  • Polling Place Simulator - predicting wait times and polling place capacity through simulation of the polling place. Can long lines be avoided or must voters sometimes stand in line all day to cast their votes? Can elections officials head off long voter waiting times by practicing with a simulation beforehand?  
     
  • A public web page and database for storing voting process models, to collect voting processes as observed or reported
  • Analysis of reported processes, and assistance in completing process reports, also through the web page
  • Pre-emptive materials to prevent problems before they occur, including a guide for voters going into a polling station to help them audit their actual experience and report anything not done correctly, with a feedback mechanism so that the process at that place can be analyzed for execution failures and errors or omissions in the model.
  • Process visualization tools, primarily process diagrams at the start
  • Analysis and reporting of possible problems/fraud opportunities/missing audits with the reported process
  • Polling place calculator - a shorthand version of the simulator, applying queueing theory to run many scenarios rapidly
  • Fully interactive simulations - the animated simulator, with interactive controls

Publications

  • One-page quick summary of the Voting Transparency Project and Polling Place Simulator (PDF, 63K): vtp-1page-20050506.pdf
    Draft documents were written for Prof. Chris Cskiszentmihalyi's "Technology, Information and Social Control" course at the Media Lab, Fall 2004
  • DRAFT: Toward a Systematic Approach to Voting Process Design and Analysis (PDF, 2.1mb): voting-20041210.pdf
  • DRAFT: Extended abstract and table of contents only (revised 12/10/2004, PDF 63K): voting-abstract.pdf

    Audiences
    Our audiences include concerned members of the public, board of elections staff, vendors, and anyone with an interest in designing, operating, or evaluating the casting and tallying of ballots in public elections.

    Discussion / Research
    Often much of the process at a board of elections is evolved and is carried on by habit, rather than by intentional top-down design. Boards of elections have many responsibilities beyond running elections, and are not usually expected to be expert in process design. Nonetheless, running an election is an exercise in security, logistics and planning, all of which are part of process.

    Elections occur infrequently. Thus the general public has little experience knowing what to expect when voting and cannot always recognize errors by poll workers. Elections officials have very few opportunities to study their own process and react to process errors. Because the only trials of the process are live trials, the stakes are too high to allow for much experimentation in process improvement.

    In practice, votes seem to be counted more or less accurately, and elections appear to work acceptably much of the time. This generally acceptable result seems to have come despite the lack of rigor in the design and management of the process, not as a result of it. However, when a process isn't studied, we can only say that things appear to work well - often nothing more definite can be said.

    New electronic voting machines reassign big chunks of "process" that were formerly the province of boards of elections and voters, to voting machine builders. This transfer of control is not often acknowledged and has happened more or less unnoticed, ironically because the machines hide so much more information about the process than previous mechanical/manual practices.

    We will provide a vocabulary and visual language for voting process analysis, attempting to capture and analyze a model of the voting process at a local polling place. Then we will provide a feedback channel so the performance of the process in operation (compliance with the theoretical model) can also be revealed.

    By working with boards of elections, we will improve our own understanding of the elections process and refine our models, particularly with regard to simulations which can be run against real data provided by the elections officials.

    This work is not about "voting machines" but about the whole act of voting, from the time a voter appears at a polling place to the time that voter's intentions are recorded to the satisfaction of all stakeholders. "Voting machines" are involved, of course, but are only a part of the process. Every voting machine presents a different set of properties, concerns, behaviors and vulnerabilities. Some machines encapsulate the entire process of voting and vote-counting. Others merely augment a process that involves other machines, people, and external processes.

    Our study tools are intended to work with any kind of voting process, whether the process involves marks on paper and hand-counting, traditional mechanical systems, or the newer all-in-one touch screen devices.

    CONTACT
    Jim Youll: jyoull [at] alum.mit.edu

    COPYRIGHT
    Contents copyright (c) 2005-2008 voting transparency project, cambridge, ma

  • tml>