California: Popularity of vote-by-mail adds extra complication to counting votes accurately | California Forward

Are absentee ballots the new hanging chads? More than 4 million presidential votes were lost in the 2000 election which was notoriously plagued by the hanging chads fiasco. Although voting technology has since vastly improved, the steady rise in absentee voting may undermine any gains in accuracy. Why is this important in California? Because last year’s presidential election was the first statewide general election in which a majority of Californians, 51 percent or 6.8 million to be exact, voted absentee. By comparison, less than 3 percent of California ballots cast in the 1962 general election were submitted by mail.

Voting Blogs: Pew’s Election Performance Index | Heather Gerken/Election Law Blog

A few years ago, I proposed creating a “Democracy Index” that would rank states and localities based on how well they run elections.  Since then, the Pew Charitable Trusts, a nonpartisan organization well known for promoting data-driven governance, has tried to put these ideas into action.  It created the nation’s first Elections Performance Index, which was released this week.  The EPI measures state performance based on seventeen indicators, which include the length of lines, the accuracy of voting technology, and the percentage of voters who experienced problems registering or casting an absentee ballot. The process for creating the Index was remarkable – as serious and professional an undertaking as I’ve witnessed.  Pew itself devoted significant funding and top-notch staffers to the project.  It also assembled an extraordinary group of advisors, which included some  of the top state and local election administrators in the country.  The legendary Charles Stewart, the former chair of MIT’s political science department, served as the data expert (though that seems a bit like calling a Ferrari a “car”).  The Pew staff and advisors — along with numerous outside experts Pew called in to poke and prod and test and challenge the validity of the indicators – narrowed down a list of almost fifty potential performance indicators to the seventeen you see on the website.  A huge amount of effort was put in to be sure the indicators were measuring something meaningful, and that the data gave us genuine signals rather than noise.  I am frankly amazed that Pew came up with so many good measures – it’s a testament to the creativity of the team, especially the political scientists who were involved.

National: Despite e-voting improvements, audits still needed for ballot integrity | Computerworld

Technology and process upgrades implemented since the controversial 2000 presidential election have made electronic voting machines more secure and reliable to use, the Caltech-MIT Voting Technology Project said in a report last week. Even so, the only way to ensure the integrity of votes cast with the systems is to have mandatory auditing of the results and of all voting technologies used in an election, the 85-page report cautioned. Rather than setting security standards for election equipment, the better approach for safeguarding ballot integrity is to hand-count a sufficiently large and random sample of the paper records of votes cast electronically, it said. “The 2000 United States presidential election put a spotlight on the fragility and vulnerability of voting technology,” the report said. “It became clear that providing robust, accurate, and secure voting systems remained an important open technical problem” for the United States. The Voting Technology Project is a joint initiative between MIT and Caltech and was launched originally to investigate the causes of the voting problems in Florida in 2000 and to make recommendations based on the findings.

Editorials: A symposium at MIT looks at the integrity of computer voting | SentinelSource.com

Twenty-five years ago, as election officials around the country were discovering wondrous new ways to tabulate votes, a group of computer scientists got together in Boston for an impressively titled “First National Symposium on Security and Reliability of Computers in the Electoral Process.”

The session aired concerns about the integrity of computer-based voting methods and machines. In addition to computer scientists, the participants included election administrators from around the country, academics and equipment vendors. The subject remained fairly esoteric for several years until the 2000 presidential election, when voting machine irregularities and related incidents in Florida cast a bright light on the security of votes.