Convergent Usability Evaluation, p. 3
In June 2004, two non-profit organizations, Computer Professionals for Social Responsibility (CPSR), and the Verified Voting Foundation, received a grant from the Quixote Foundation to develop a web-based application to help monitor U.S. elections and capture data concerning voting problems. The Election Incident Reporting System (EIRS)  was to be used in the November 2004 U.S. election.
Intended users of EIRS fall into three classes:
- call-center volunteers who enter data about voting incidents into EIRS
- call-center managers who manage user accounts
- analysts who use the election incident data  to detect and remedy voting problems.
Most users entering election incident reports were volunteers who either: a) took incident reports by phon-e from poll observers or members of the voting public, or b) entered data from paper forms filled out by poll observers or call-center volunteers. Also, members of the general public could file voting incident reports directly at an auxiliary website (VoteProblem.org) although such reports were stored and reviewed separately.
Components of EIRS UI
User-visible aspects of EIRS fall into four categories:
Data-entry forms: EIRS provides many forms: election incident reporting, equipment test reporting, election official questionnaires, volunteer registration, and others.
Web pages: EIRS includes dozens of web pages, containing everything from navigation controls, instructions, and background information to data-input forms and data-output displays.
Map displays: The primary way EIRS currently organizes election-incident data is geographically – by state, county, and voting precinct. Accordingly, maps are used heavily as a display medium.
Incident detail displays: EIRS users can “drill down” through states, counties, and precincts to the details of particular election incidents.
Distributed, Fluctuating Development
VerifiedVoting and CPSR quickly built a team consisting mainly of volunteers to design, implement, test, and document the application. Team members were located all over the U.S. – there were even members in Europe and Australia – so they worked together mainly by Internet and phone. During the development period, designers and developers continued to join the project, and some left it.
Despite the fact that the team was physically distributed and fluctuating, it designed and developed EIRS on “Internet time” with a hard deadline: election day, Nov. 2.
To date, EIRS has been a success. A preliminary version of it was ready in time to help document voting problems in the Florida Primary election on August 31, only two months after development had begun. In the general election, volunteers organized by the non-partisan Election Protection Coalition, working at call-centers all over the U.S., handled over 175,000 calls, about half of which were on Nov. 2. Call-center volunteers used EIRS to log over 40,000 election incident reports, most of which were logged on election day.
In addition to collecting data that will be useful for research and
policymaking, EIRS helped in real time during the election. It
allowed election monitors to detect and remedy many serious problems at
polling places nationwide. Call-center operators also helped tens
of thousands of voters find polling places and resolve other minor
voting obstacles .
Last modified December 22, 2005 05:38 PM