OLR Research Report

December 27, 2006




By: Joseph Holstead, Associate Analyst

You asked for information on the optical scan voting machines that 25 Connecticut towns used in this past November's election. You also asked if there were any problems in other states that used the same optical scanning machines.


Voters in the 25 Connecticut towns used the AccuVote Optical Scan voting terminal (AV-OS), made by LHS Associates of Massachusetts, a subsidiary of Diebold Election Systems. To vote, voters in the 25 towns filled in ovals on a paper ballot next to the names of the candidates of their choice and submitted the ballots into the AV-OS device, which automatically read and recorded the votes.

Prior to the election, there were wide-spread concerns about the AV-OS' susceptibility to tampering. However, an October 2006 University of Connecticut study (which also discussed the machines' vulnerabilities) found that Secretary of the State Susan Bysiewicz's office had taken steps to avoid any problems. On December 7, 2006, the secretary of the state's office reported that an audit of the machines found them to be secure and reliable. The entire state will switch to the machines by the end of 2007.

LHS Associates is a New England dealer and its machines were used in New Hampshire, Vermont, Massachusetts, and Maine, according to its elections service manager, Sue Reynolds. Like Connecticut, there were some concerns about the potential for tampering in Vermont, but we found no post-election reports of problems with LHS Associates-produced machines in these other states.

Generally, optical scan machines are the most common form of voting technology in America today. In fact, according to an October 2006 pre-election report by (a non-partisan election reform information project), only nine states were not using some type of optical scan voting machines as part of their overall voting systems usage.

Though states have apparently experienced few problems with optical scan voting, concerns with it and other electronic voting in general have been raised across the country, though human error seems to have played some part with most issues, according to For example, many voters in one Maryland county had problems voting early in the state primary because an election worker forgot to send voter access cards with Diebold direct recording-electronic (DRE) systems, commonly known as “touch screens.” After the election, concluded, “[g]enerally, the midterm election's voting system troubles can be described in three categories: human error, machine error and unknown.”

Overall, touch screen technology, rather than the optical scan machine, appears to have caused the majority of problems.


Use in the November 2006 Election

On August 4, 2006, the secretary of the state announced that her office had entered into a contract with LHS Associates to provide optical scan voting machines to replace lever voting machines statewide. In September, she identified 25 municipalities that would use the optical scan voting machines for the November 7, 2006 election. The towns were: Ashford, Bethlehem, Bolton, Durham, East Hartford, East Haven, East Lyme, Hartford (in four precincts), Litchfield, Mansfield, Middletown, Monroe, Montville, Newington, Newtown, Old Lyme, Salisbury, Southington, South Windsor, Tolland, Vernon, Westbrook, Wethersfield, Wilton, and Wolcott. Copies of both of her press releases are attached.

On October 30, 2006, the University of Connecticut Voting Technology Research Center released a report stating the AV-OS machines “can be compromised in a matter of minutes by tactics such as neutralizing one candidate so his or her votes aren't counted or swapping votes of two candidates.” But, according to a Hartford Courant article, “the authors of the report credit the secretary of the state's office for implementing new security procedures to protect the machines.” A copy of the UConn report is attached, or visit:

After the election, on December 7, 2006, the secretary of the state's office and UConn's Voting Technology Research Center released election audit results, which found that the AV-OS devices “performed extremely well and were proven to be a safe and secure form of voting technology.”

The audit reviewed election results in 17 randomly selected precincts in Newington, Middletown, East Hartford, Hartford, Wethersfield, Wilton, Monroe, East Haven and South Windsor, according to the Hartford Courant. Bysiewicz said that the few discrepancies the audit found were due to voters incorrectly filling out the ballot (e.g., writing a check mark rather than fully darkening the ovals next to their candidates' names).

Secretary Bysiewicz has said that all Connecticut towns and cities with May municipal elections will receive their new optical scan voting machines starting in early 2007. The rest of the municipalities with November elections will get their machines after that, with training expected to start statewide in Spring.

Background on the Move toward New Voting Technology

After the voting irregularities in Florida during the 2000 presidential election, Congress passed federal election reform legislation entitled the Help America Vote Act of 2002 (HAVA). President Bush signed bill into law (H.R. 3295, P.L. 107-252) on October 29, 2002. HAVA included a number of voting requirements that states had to meet, as well as funding programs for which they could apply.

Under the law, states could apply for federal funding for activities to improve elections, which Connecticut did. The secretary of the state announced on September 13, 2006 that, “[t]he state will use $15.7 million in federal funds to purchase optical scan machines and privacy booths for every municipality by 2007, and also to cover the cost of all necessary training.”

The secretary's early research and study of voting machine technology and earlier purchasing process activity by her office are described in OLR Report 2006-R-0047 (a copy of which is attached or visit:


While the issues discussed below did not involve LHS Associates machines, they illustrate some of the difficulties experienced as the nation moves to meet HAVA-mandated changes.


As the October 2006 report states, an election worker in Maryland's Montgomery County failed to send voter access cards to polling places with the Diebold DRE machines, making the machines temporarily unusable.

“The cards were delivered to most polling locations by 8:30 a.m. and polling places were up and running by 10:00 a.m., but many voters were turned away at the polls, made to vote on paper ballots, and even on scrap paper when the paper ballots ran out,” according to the report. A judge ordered polls to stay open an hour past closing time to accommodate voters and Governor Robert Ehrlich was reported to support a switch from the Diebold DRE machines to a paper-based system.


Arkansas had several problems implementing its electronic voting options. As the report states, “a phase-in of new voting machines took place prior to Arkansas' May [2006] primary, but the $15 million purchase and installation of ES&S [Election Systems & Software, Inc.] voting systems, including touch-screen machines with attached voter-verified paper audit trails, was chaotic at best.”

The report goes on to say, “implementation bordered on disastrous in the early going as the state has had persistent troubles with its voting machine vendor and its products, from missing absentee ballots and misprogrammed optical-scan readers to 'defective software' in touch-screen voting machines.”

ES&S did not dispute the findings of an Arkansas panel that examined voting problems during the May primary, which concluded that the company was at fault. The company in turn did not charge the state for nearly $400,000 in services it provided; and in the end, the panel did not recommend “scrapping” the ES&S system entirely.

North Dakota

According to the report, North Dakota uses only optical scan vote counters and a hybrid ballot-marking device accessible to voters with disabilities, both made by ES&S. The report goes on to say that “[n]ews reports reported a generally smooth primary at the polls in June [2006] on the new machines, [but] [s]ome problems were reported with the counting process….”


In 2005, several counties used electronic voting machines with voter-verified paper audit trails [VVPATs] for the first time, and problems with the system and its implementation lingered. In the 2006 primary in Cuyahoga County, for example, it was reported that the Election Science Institute found a number of problems with the Diebold machines as well as trouble with poorly trained poll workers and their handling of the VVPATs. The study found problems might not be fixed before the November 2008 presidential election, never mind by November 2006. Diebold disputed the report's findings, according to

Post-Election Briefing

While human error played a roll in some of the problems with electronic voting in the November elections (and primaries before that), there were several examples of problems with the touch screen machines themselves during the November 2006 election. An survey found “[a]ccording to VotersUnite, a citizen watchdog group that tracked media reports of machine problems during the midterm, hundreds of reports of election-day problems with electronic voting machines should dispel the notion that things went smoothly.”

The November briefing goes on to say, “[t]he most frequent problem was vote-flipping, in which voters pushed the screen for one candidate and another lit up, or when a vote for one candidate turned into a different one when a voter reviewed the ballot. The VotersUnite Web site linked to news reports of vote-flipping in a dozen states, including Colorado, Florida, Illinois, Texas, New Jersey, and Ohio.”

Creating a “paper trail” seems to be a possible solution for the touch screen voting issues. The November briefing concludes that a “number of state legislatures could take up the issue of VVPATs again, armed with stories of vote-flipping and other anomalies and/or malfunctions. While 22 states already require DREs to have paper trails, Florida, Georgia, Texas, Pennsylvania, Maryland and other election 'trouble spots' in 2006 do not have requirements in place.”, “produced by the Election Reform Information Project, is the nation's only non-partisan, non-advocacy website providing up-to-the-minute news and analysis on election reform,” according to its website: A copy of its November briefing is attached, or visit: