Last updated: Wednesday, September 13, 2006

For more information and the full text of the study, see http://citp.princeton.edu/voting.

Why did you study these Diebold machines? Why not study some other voting technology?

We studied these machines because they were available to us. If we had gotten access to another kind of machine, we probably would have studied it instead.

Are other voting machines more secure than the ones you studied?

We don't know. We certainly hope they are more secure — elections are relying on them — but there isn't enough evidence yet to answer this question.

A voter-verified paper trail is the most important safeguard that can make e-voting machines more secure.

Why did you publish all of this information about security attacks? Aren't you just teaching the bad guys how to cheat?

We could not in good conscience withhold this information from citizens and public officials. People need to know that there is a serious problem. Details help people understand the problem and evaluate the credibility of our conclusions. They also help our colleagues in the research community understand the implications of voting machine design and evaluate new ideas.

The vendors have been claiming their systems are secure, and elections are now depending on these claims. Our report points out that the vendors have significant problems and helps create momentum toward solving those problems. Citizens paid for these systems and deserve to know what they got for their money.

That said, we are holding back some details that a real attacker would need to carry out an attack. We have thought carefully about how to balance the risk of abuse against the public's need to know.

You're computer scientists. Why are you opposed to this new computerized technology?

As experts, we understand the limitations of technology. We're not opposed to all use of computers in elections, but we do insist on having adequate safeguards in place. Computers can be part of a secure, user-friendly election system; we hope to vote on such systems in the future.

Have the vote-stealing methods you discuss ever been used in real elections?

Probably not, but we don't know for sure. We haven't seen evidence that these attacks have been used, but one lesson of our report is that the design of these voting technologies makes attacks relatively easy to cover up.

Do you think any recent major U.S. elections were stolen?

No. We know some people are claiming this happened, but we don't find their evidence convincing.

Isn't this all just partisan politics? Aren't you just unhappy with how recent elections have gone?

Our goal is to make elections more accurate. That shouldn't be a partisan issue, and when others try to make it partisan we do our best to ignore them.

The purpose of an election is to accurately measure the intent of the voters. The challenge is to convince the losing candidate and his supporters that he truly lost the election. Sufficient evidence can only come from a combination of properly-engineered technology and robust procedures for handling it. We can all benefit from a system that can supply that evidence.