John Tanner at the Voting Rights Section of the DOJ has resigned. I have linked to Talking Points Memo, which provides some background about his tenure. I will add a link to the Washington Post once they do a story about it.

Ohio released a voting system report today. You can read the report at the link above. I haven’t read the full report, just the summary but it seems interesting. I will post a longer piece about this in the next week.

Given that, in 2004, part of the election controversy related to polling numbers that did not turn out to be correct, I thought this link might be of interest. It is to a Slate.com article, but there is a link off the page to http://www.pollster.com, which is run in part by Charles Franklin, professor of political science at Wisconsin and a really smart guy.

The California Secretary of State’s Office released the RFP today for the development and implementation of the VoteCal System.

The DoJ has filed an amicus brief in the Crawford case. Our own Michael Alvarez’s work is quoted on pg. 22.

I don’t have a horse in this race, and I’d never presume to speak for Michael, but I do want to make sure that scholarship is fairly represented.

It’s my reading that ABK are misquoted, or perhaps better, incompletely quoted. ABK found no differential racial impact but did find a differential impact by income and educational levels.

Here is the DoJ statement, followed by Alvarez, Bailey, and Katz’s full quote.

DOJ pg 22.

There is no basis to consider such extrarecord materials, but the most recent studies nevertheless undermine petitioner’s assertion of disparate impact. The Cal. Tech./M.I.T. study petitioners cite as evidence of the statute’s burden (Dem. Br. 34), actually states that “there does not seem to be a discriminatory impact of the requirements for some subgroups, such as nonwhite registered voters.”

ABK, pg. 21:

We find no evidence that voter identification requirements reduce participation at the aggregate level. At the individual level, voter identification requirements of the strictest forms — combination requirements of presenting identification and matching signatures, as well as photo identification requirements — have a negative impact on voter participation relative to the weakest requirement of stating one’s name. In general, there does not seem to be a discriminatory impact of the requirements for some subgroups, such as nonwhite registered voters. However, we do find that for registered voters with lower levels of educational attainment or lower income, stricter voter identification requirements do lead to lower turnout.

My colleague, Peter Ordeshook, forwarded me this morning to some election results, posted by what I believe is “The election Commission of the Karacaevo- Circassian Republic.” I say “I believe” — which I will say throughout this post, because I’m translating material from the website into English using AltaVista’s “Babel Fish Translation” website.

The website that Peter directed me to is here, though I do have a pdf screenshot of the page from about 1pm Pacific just in case the original page changes.

What is interesting to note here is, based on my ability to translate this page:

• Line 1 gives the number of voters in this district, by what I believe are voting or polling stations (the columns).
• Line 2 gives the number of ballots received (and line 4 the number given out to voters).
• Line 10 gives the number of votes going to the United Russia” party.

It does not take an advanced degree in math to see the odd result here: the number of voters in this jurisdiction (and in each polling place or station) is equal to the number of ballots received, and all of the ballots received were cast for United Russia. That’s 100% turnout, with 100% of ballots cast for United Russia.

Of course, if readers have better or different translations of the entries in this set of election results, or if readers have other examples of these oddities, please let me know.

• Update 1 (December 10, 2007), 1:20pm: Here is the pdf screenshot of the results page.

At the recent Pew Forum in Atlanta, there was quite a bit of discussion about Florida’s voter registration systems. The Tallahassee paper reported:

An outspoken Florida elections official Saturday predicted “mass confusion and chaos” at the polls because of errors in voter registration databases, but the state’s top elections official said Florida “is not going down the crapper” next November.

The Miami paper noted that:

[Secretary of State Browning] acknowledged, however, that at least 14,000 Floridians across the state who registered to vote in the last year may be blocked from voting in the presidential primary because they didn’t supply the correct information.

Those voters have until Dec. 31 to get their information corrected in order to be eligible to vote in the Jan. 29 presidential primary. But, Browning told reporters, that number is small — less than 1 percent — compared to the 1.5 million people who attempted to register to vote.

Ned Foley has often asked the question about whether perfection is possible in elections and how do we devise systems that address the lack of perfection. 2008 may be the year we have to figure out what our tolerance for errors in voter registration actually are.

This is one way to keep your unruly relatives from voting.

Alan Skyles was dumbfounded when the Snohomish County Auditor’s Office told him that he couldn’t vote because he was listed on voter registration rolls as a convicted felon. He was finally allowed to vote in the November 2006 election, but only after the records proved that he wasn’t a felon….Prosecutors last month charged his sister, Deborah Deanne Thorpe, 54, of Snoho­mish, with two counts of official misconduct. They allege she made several unauthorized changes on the computer in the voter registration rolls, including the one that vexed her brother. Thorpe was hired as a temporary worker in November 2005 and had access to county computers, according to an Everett police report filed with the charges.

So, what happens when a lever machine screws up? Read on!

The outcome of Tuesday’s $70 million school construction proposition will be lost forever in the turn of the voting machine counters. As the counters on both machines hit 999, they each rolled back to 000 like an overused car odometer. The question that will never be answered is, for which side were those 1,998 votes cast?

A number of South North Dakota counties are poised to go fully vote by mail in 2008. Story from the Grand Forks Herald.

(Thanks to DC for the catch on the state.)