Voting Matters--Securing our elections Jan 07

With current voting systems a single person in the wrong place could change votes and rig an election. We can’t be certain that past elections haven’t been rigged. Ever wonder why the audit procedures for slot machines are better than voting machines? Let’s look at the Brennan Center Task Force on Voting System Security’s study on voting system vulnerabilities. This study is available at http://www.brennancenter.org

The Task Force divided voting machines into 3 basic types: Direct Recording Electronic (DRE), DRE with Voter Verified Paper Trail (VVPT), and Precinct Count Optical Scan (PCOS). All three types have significant security and reliability vulnerabilities. Fortunately, the most serious vulnerabilities of each system can be substantially reduced if proper countermeasures are implemented.

Unfortunately, few jurisdictions have implemented any of these countermeasures, and few states have codified any of these countermeasures into law. The Task Force analyzed voting system vulnerabilities and concluded that some apply to all electronic voting systems, and others only to specific technologies. Vulnerabilities common to all systems include insertion of malicious code to change close races and the use of wireless technologies in voting systems. DREs without a paper trail are vulnerable to any kind of electronic attack. For VVPT and PCOS, a paper trail alone is not sufficient; it only has value if an audit procedure is in place. Even with the paper trail and routine audits some attacks cannot be detected.

The Task Force made several recommendations. The first is conduct routine audits of the paper and electronic records after every election using a transparent and random selection process. Next is performing random “parallel” testing of voting machines on Election Day to detect software attacks. On the technology side, they recommend banning wireless components for voting machines. Voting system programming and administration should be decentralized and closely audited. Clear and effective procedures must be in place for reporting and investigating evidence of fraud or error.

On another note, Representative Holt’s H. R. 550, an amendment to the Help America Vote Act (HAVA) to require a voter verified paper trail, is a partial response to the Brennan Center’s recommendations. The Holt Amendment requires a voter verified paper trail for any voting machine, which addresses some of the major concerns of the Brennan Center study, but not all of them. Unfortunately, there is no requirement for independent audit or verification. For a more detailed discussion look on http://www.blackboxvoting.org/

In the next issue I will discuss in greater detail the vulnerability analysis section of this report. This will help in understanding the strengths and weaknesses of proposed legislation both at the state and national level.

Voting Matters--Securing our elections Feb 07

Last month I discussed some of the key points of the Brennan Center Task Force on Voting System Security’s study on voting system vulnerabilities. This study is available at http://www.brennancenter.org This month we’ll look a bit more at which threats the Brennan Center felt were most important and how we can ensure free and fair elections on electronic voting systems.

Remember the common vulnerabilities among all electronic voting machines? If the goal is to change a close election it is easiest to install attack software or corrupt programs. Furthermore, if a voting machine has a wireless access it is far more vulnerable to attack.

For voting machines without voter verified paper trails, there is no way to independently audit the results. If the voting machine does have a paper trail, or is an optical scan machine, it is useless without an audit process.

The threat is real: installing software or corrupting programs requires very few participants and can easily swing a close election. It requires installing software, perhaps during the setup or calibration of a machine, in order to selectively miscount the results. It could even be loaded or triggered by the wireless component of a machine. For your average election judge it would be an invisible attack.

Voter verified paper trails are another problem. Several studies have shown that few voters will verify their own ballots. Unfortunately, without a mandatory audit program it is unlikely that fraud or error will be caught by the voter. Of course a smart programmer can also modify the paper and electronic tallies at the same time. If the voter even notices, they may just assume that they made the mistake.

Precinct count optical scan machines show the highest accuracy and reliability, but central count optical scanners are subject to fraud and error at the central counting station. The former have under and over vote protection; the latter do not.

The best protections are a combination of testing and audit strategies. On Election Day, testing randomly selected voting machines should be routine. In addition, audit the paper trail and match it to the electronic tally. With multi-day early voting in Texas, this would have to be a daily procedure. Both the audit and testing process should be performed by independent auditors with observers present. The audit and testing should be video recorded and retained.

Some facts to think over prior to lobby day in Austin: only 2 states ban wireless access to voting machines. 26 states require voter verified paper trails, but only 12 of them require audits. Texas doesn’t require any of them. All of Dallas County’s early voting and ADA voting machines have wireless access; none have voter verified paper trails.

Next month I will review the National Institute of Standards and Technology white paper on voting machine vulnerability and recommendations about accrediting voting machines. http://vote.nist.gov/