More things to Do With the “Last 4”

voting-station.jpg
Apparently, in Ohio, you’ll be able to vote if you know the last 4 digits of an SSN. As the Cleveland Plain Dealer reports:

Voters who don’t have identification will be able to vote at next week’s election by presenting the last four digits of their Social Security number and casting a provisional ballot.

Will they be distributing lists to the polling places? If so, when if the lists are stolen, and people can access credit cards, phone records, and lord knows what else, will the loss of control be reported?

Via Jonathan Adler at Volokh, “Ohio Voter ID Case Settled.” Photo by Mike Benedetti.

The Hugo Chavez Test for Voting Machines

malcomx.jpgAt first I thought that the stories around Sequoia Voting Systems and Smartmatic having connections to Hugo Chavez were silly. I still do think that, but I also think that they’re coming out for an important reason: we have lost trust in the machinery of voting, and that is a criminal shame.

The right to vote, and to have one’s vote counted is fundamental to how and why we accept our government, even when it makes colossal mistakes. This is an ideal which people around the world recognize and aspire to. The imprint of legitimacy which an election confers on a leader is important enough that even the Soviets faked elections so they could claim that mantle.

If we had voting systems that were trustworthy, transparent and understood by those operating them, then we could buy our voting machines from Hugo Chavez or Mahmoud Ahmadinejad and not have to worry a lot about it. We do not, and cannot. We have transitioned from paper ballots and their understood problems into a brave new world of computerized and untrustworthy voting systems, and we are poorer for it.


I propose we call this the Hugo Chavez test, and see how all new voting technology fares under the test. We could realistically consider buying paper ballots, punch cards, or other verifiable voting technologies from the Chavez government, and be reasonably confident in our ability to test them and be sure we were getting what we specified. (I’m confident someone will point out an exceptionally clever trick, so read the comments.) I’m also confident that we can’t say the same of any computerized system on the market today. Our ability to audit them is simply too lacking, and the skills to do so too rare.

The photo is Malcom X, because we sometimes forget that within living memory, not all Americans had a right to vote. We forget that that right was important enough for Malcom X to declare 1964 might be be the “year of the ballot or the bullet.” That the ballot is so powerful that men ready to commit acts of violence could be placated by giving them the right to vote. It’s an important right, and the value of trust that our votes are counted accurately and securely is nearly incalculable.

Diebold goes open source

Well, not intentionally.
Seems that multiple versions of source code (including the one used to run the 2004 primaries in Maryland) were delivered anonymously to a former legislator who has been critical of Diebold.
Note that this is not the same source examined by Avi Rubin, et. al., and found wanting from a security perspective.
The Baltimore Sun has more.

Detecting Election Fraud

odd-frame.jpgThanks to my lovely spouse, I came across a series of fascinating papers by Walter R. Mebane, Jr. a professor of Government at Cornell. These papers use statistics, specifically Benford’s Law, to detect election fraud. Now I know statisticians, and I am no statistician (and boy howdy is my higher level math rusty), but the papers were still relatively easy to read and follow. Dr Mebane’s most recent paper, Election Forensics: The Second-digit Benford’s Law Test and Recent American Presidential Elections, was written for the Election Fraud Conference, which was held last week out in Salt Lake City, Utah. I haven’t had a chance to read through any of the other papers yet, but I’m sure they are equally interesting.