Secondary menu

Electoral Insight - Technology in the electoral process

Electoral Insight – June 2000

Voting by Internet in the United States

Voting by Internet in the United States

Marc Chénier
Senior Policy and Research Officer, Elections Canada

and

Wayne Brown
Co-Editor,
Electoral Insight, Elections Canada

Arizona's on-line voting started on March 7, 2000
Arizona's on-line voting started on
March 7, 2000.

Arizona Democrats have become the first Americans to use the Internet to cast legally binding ballots. The traditional polling sites were open on March 11 for the state's Democratic primary, but registered party members also had the option of voting on-line from their homes, offices, schools or libraries between March 7 and 10. The major goal of using the Internet was to try to increase the traditionally low voter turnout for the primary and the participation of young Arizonans.

The Arizona Democratic Party worked with a New York-based company, Votation.com (now called Election.com), to carry out the historic election. In February, all of the state's registered Democrats (more than 800 000) were mailed a PIN number, similar to the numerical codes used for bank ATM machines. Those who wanted to vote electronically could then log on to the party's Web page or the Votation.com site and enter the PIN. They were also asked to verify their identities by entering a social security number or date of birth.

There is some opposition to this use of the Internet. The Voting Integrity Project (VIP), a civic group based in Virginia (http://www.voting-integrity.org), applied for an injunction to prevent on-line voting in the Arizona primary. The group argued that the initiative would discriminate against and dilute the voting strength of minority groups, who are less likely to have access to the Internet.

VIP also claimed that the Democratic Party had failed to obtain clearance from the federal Department of Justice before instituting new voting procedures, contrary to the Voting Rights Act. The injunction request was denied. The Court found that the evidence about the existence of a "digital divide" was not specific enough to determine the extent of such a divide in Arizona or its possible impact on the March primary. Although the VIP did not appeal the injunction ruling and the vote went ahead, the case is continuing. In its ruling, the Court stated the election could be set aside later if found to be contrary to the Voting Rights Act.

Internet voting by the Navajo Nation at Window Rock
Internet voting by the Navajo Nation at Window Rock.

In 1996, only 12 800 Arizonans participated in the Democratic Party's presidential preference primary. In the 2000 primary, 40 000 electronic ballots were cast, more than half of the total number of votes received. However, the total number of ballots (Internet and paper) still represented only about 10 percent of registered Democrats in the state and less than one percent of the electronic ballots were cast by voters in the 25-and-under age group. While the number of votes cast this year rose dramatically, the proportion of the increase that can be credited to the new option of Internet voting cannot be fully assessed. Other factors included additional days of voting this year and increased public interest in the campaigns of the presidential candidates.

The Arizona Democrats' new electronic voting process encountered a number of problems. There were reports that some Democrats claimed they did not receive a PIN, software conflicts that caused some ballots to come up blank, and erroneous messages that informed voters they had already voted.

Internet voting at the University of Arizona
Internet voting at the University of Arizona.

The newsletter Election Administration Reports described the Arizona experiment as a "rocky start". Not many Macintosh users were able to vote early, because the browser in most of those computers was blocked by Votation.com's election security system. When Democrats who lost or misplaced the PIN numbers they received in the mail called to obtain the number (without which they could not vote by Internet), many were unable to get through because the Macintosh users were also calling. The newsletter also says the party did not initially know how many polls it would have or where some of them would be located. The number of polling places in the state had to be increased significantly in the month before the election, to try to satisfy the complaints about Internet voting diluting minority votes. In turn, those additions and changes in the locations of polling stations may have led to errors in the list of polling places mailed to voters. Further, it took two-and-a-half hours after the polls closed to report voting results and, even a week later, there was no information about how many Internet votes had been cast.

On the other hand, State Democratic Chairman Mark Fleisher was quoted as saying the voter turnout was by far the largest ever for the party's presidential primary and that the Internet option "really created excitement about being part of history, and it made voting easier and more accessible than ever." In the primary vote, Vice-President Al Gore easily defeated Senator Bill Bradley, who, in fact, dropped out of the race midway through the fourday period of Internet voting.

While Arizona made history with its binding vote that selected thirty-one delegates to the national Democratic convention, other states have also taken the first steps toward Internet voting.

In January, Republicans in Alaska used the Internet for a non-binding vote. Registered voters received an information package that included log-in codes, with which they could get into a Web site to register for the election. At the registration Web site, the voter entered his or her full name, address and voter registration number. The information provided by the voter was then verified and the company conducting the vote (VoteHere.net) gave the voter three 8-digit codes. On voting day, electors logged on with their three codes and cast their votes on-line. Thirty-five voters used the on-line option.

Also in January, the California Internet Voting Task Force, commissioned by the State to undertake a ten-month study of Internet voting, released its report (http://www.ss.ca.gov/executive/ivote/). The task force was comprised of 34 technology experts, political scientists and civic leaders. It concluded that the implementation of Internet voting would increase access for millions of potential voters in the state who have not regularly participated in elections. However, it also found that there are technological threats to the security, integrity and secrecy of Internet ballots.

The task force stated that the possibility of virus and Trojan Horse software attacks on home and office computers used for voting is very real and could result in problems ranging from a denial of service to the submission of electronically altered ballots. To minimize those threats, it recommended that election officials should provide unique operating system and Web browser software to voters.

The task force also found that the use of digital signature and encryption technology could protect the integrity and secrecy of ballots transmitted over the Internet. All identifying information used to electronically verify the identity of a voter would be stripped from the ballot before tabulating the votes. The task force also warned that the ballots of voters who use the Internet through a local area network could have their privacy breached by a network administrator who might gain access to the voter's computer before the ballot is encrypted.

According to the task force's report, one of the most difficult tasks for an Internet voting system is the authentication of voters. To ensure that every voter has the opportunity to cast a ballot and none are able to vote more than once, the task force recommended that election officials should initially test Internet voting technology through the use of voting machines that are under the direct control of election personnel in traditional polling places.

The task force stated that it is technologically possible to utilize the Internet as an additional method of voting. But it added that currently it is not legally, practically or fiscally feasible to develop a comprehensive remote Internet voting system that would completely replace the paper process now used for voter registration and voting. The task force stated that the election process would be best served "by a strategy of evolutionary rather than revolutionary change."

Meanwhile, in New York, another task force is reviewing the possibility of Internet voting in elections at all levels and how to block potential computer hackers. Governor George Pataki has also requested an assessment of whether Internet voting would create an unfair disadvantage for people in low-income or minority neighbourhoods.

The U.S. military is working with the states of Florida, South Carolina, Texas and Utah in a pilot program to allow some 300 absentee military voters to cast on-line ballots in the presidential election this fall. In the wake of California's report, the Defense Department has decided that votes will be cast only from "virus-free" machines at military bases. This may be the first use of Internet voting in a binding public election.

U.S. President Bill Clinton has asked the National Science Foundation (http://www.nsf.gov) to conduct a one-year study of the feasibility of on-line voting in future elections. The Foundation will assemble legal scholars, technical specialists and experts on the democratic process to review whether cybervoting can increase voter turnout, while at the same time ensuring there is no fraud or loss of privacy. Clinton has also urged the technology industry to help reduce the potential "digital divide" problem.

The Americans are leading the world in the number of e-voting companies, pilot projects and technical studies related to Internet voting. These efforts merit close scrutiny by electoral agencies in Canada and other countries.

Sources: The Arizona Republic, Election Administration Reports, Election.com, Voting Integrity Project


Note: 

The opinions expressed are those of the authors; they do not necessarily reflect those of the Chief Electoral Officer of Canada.