The Machinery of Democracy: Voting System Security, Accessibility, Usability, and Cost



About This Report

From the Introduction:

In an effort to address the most serious concerns about new voting technology, the Brennan Center for Justice at NYU School of Law assembled four Task Forces of the nation’s leading experts in the areas of security, accessibility, usability and cost, to perform the first ever comprehensive and empirical analysis of electronic voting systems. The analysis focused on the three principal types of voting systems being purchased today: Direct Recording Electronic (“DRE”) systems, DREs with voter-verified auditable paper trails (“DREs w/VVPT”), and Precinct Count Optical Scan (“PCOS”) systems. To support Task Force analyses, the Center researched state and local election laws, reviewed voting system contracts, and conducted interviews with hundreds of election officials. The result of this work is a four-chapter report that offers policy makers, election administrators,and members of the public a more nuanced and complete understanding of new voting systems than ever before.

Two themes emerge from our four-part analysis. First, there has been surprisingly little empirical study of voting systems in the areas of security, accessibility, usability, and cost. The result is that jurisdictions are making purchasing decisions and adopting laws and procedures that bear little correlation to the goals they seek to accomplish. Advocates urge security measures that provide questionable security value, while ignoring steps that provide the best chance of catching the simplest attacks on the integrity of an election. Jurisdictions purchase accessible voting machines that do not yet fully address the needs of their disabled communities and without obtaining contractual guarantees that new accessibility features will be added at little or no extra cost as they become available. Counties make decisions about ballot design and instruction language without performing usability testing to avoid voter confusion and mistake. And state and local election officials often purchase voting machines by looking almost exclusively at initial costs, with little regard to long-term costs, which will almost always make up the vast majority of the voting system’s total cost.

Second, there is not yet any perfect voting system or set of procedures. One system might be more affordable than others, but less accessible to the disabled; some election procedures might make systems easier to use, but less secure. Communities across the country will have to decide what is most important to them: how much are they willing to pay for secure, usable, and accessible systems? Will they sacrifice usability for security? Accessibility for cost? In some cases, the decisions will be mandated by law. In others, there will be difficult choices to make. Election officials and the public should be aware of the trade-offs they are making when choosing one voting system or set of procedures over another, and they should know how to improve achievement of all four values, irrespective of which system they choose.


I was a member of the task force that worked on the security section of this report.

This is a local copy. The definitive version is available at