Election Updates

Mike's on a conference call, Thad's recruiting a new colleague, and I'm standing in a line of twenty computers.

It's our annual gathering of blue blazers and khakis, otherwise known at the Annual Meeting of the American Political Science Association in Philadelphia PA. Here's where 10,000 political scientists get together to eat, drink, and exchange ideas and, increasingly, discuss election reform.

More updates as the weekend proceeds.

Speaking of opinion pieces, I authored one that came out this morning, focusing on the "Arizona Voter Rewards Act", which is on the ballot in Arizona this fall. The complete piece is here.

Here's my conclusion about this proposal --- the benefits are unclear, but the risks are real:

Combining some positive incentives to get people to the polls, with other reforms to make the process more competitive and interesting might induce qualified people to run for office, might lead to more exciting and more energized election campaigns and may give political parties and politicians more incentive to get people to the polls on election day.

But it is not necessarily the case that these wholesale changes in Arizona politics will be achieved by the voter lottery. And therein lies the risk of this measure - that it could become law and not achieve the laudable ends sought by the measure's proponents.

The Arizona voter lottery is an intriguing idea, but is also a risky reform attempt. There are so many open questions about it that the academic in me says we should study this idea further and not attempt this risky reform.

Dick Thornburgh and Richard Celeste, who co-chaired the recent National Academies of Sciences panel on e-voting, recently published an opinion piece, "Watch Out for Voting Day Bugs." The central arguments from their op-ed bear repeating here.

First, the problem:

For many jurisdictions, the 2006 elections will see the first large-scale use of electronic voting systems. Many organizations have learned the hard way that deployment and use of new technologies on a large scale virtually guarantee big surprises and unintended consequences: sudden system crashes, corrupted data or painfully slow systems. The usual remedies are to develop, test and evaluate small-scale prototypes before committing to organization-wide upgrades in technology, and to keep both old and new systems running for a while so that failures in the new system do not paralyze operations.

Second, the solution:

That's why we believe it will be essential this year that jurisdictions have backup and contingency plans that anticipate a wide range of possible failures in their electronic voting systems, including those that occur in the middle of the voting process on Election Day (or days).

And the conclusion:

Prudence and reasonable contingency planning should rule at this moment of truth for electronic voting, as election officials across the land work to retain public confidence in the face of new challenges.

By way of disclosure, I was a member of the same NAS panel, so it may come as no surprise I endorse the argument here. But dedicated readers of Election Updates also know that I've been arguing for some time now that election officials must do a better job with contingency planning and threat assessment.

The Brennan Center issued a study today about voting technology and residual votes. The key take-aways are that both DREs and precinct-count optical scan are systems that produce low residual vote rates and full-face machines, like those contemplated in New York, are terrible.

As the press release for the report states:

Significantly, several studies indicate that residual vote rates are higher in low income and minority communities. The Brennan Center study shows that improvements in voting equipment and ballot design produce substantial drops in residual vote rates in such communities. As a result, the failure of a voting system to protect against residual votes is likely to harm low-income and minority voters and their communities more severely than other communities. Among the report's key findings:

* Precinct Count Optical Scan (PCOS) and scrolling Direct Recording Electronic (DRE) voting systems are more accurate at recording voter intention than older voting systems. In 2004, residual vote rates were less than 1% for both technologies.

* Full-face DRE systems continue to be plagued with an unacceptably high residual vote rate. In 2000, 2002 and 2004, it exceeded that of either PCOS or scrolling DRE systems.

* Residual vote rates among voters earning less then $25,000 are higher on full-face DRE's ( 2.8%), than on either PCOS (1.4%) or scrolling DRE's (1.3%).

"The good news is that most states are selecting machines and designing ballots that will record more voters' choices accurately. The bad news is that major jurisdictions like Philadelphia, and perhaps New York City, plan to use voting technology that is known to have high error rates," said Lawrence Norden, Associate Counsel at the Brennan Center and lead author of the report.

The report can be downloaded here.

Thad already blogged this, but here is the early voting story from the Baltimore Sun. No more details than in the Post story.

Here is a Post story listing the dates and events in Maryland.

Here is the Perez decision (he was removed from the Attorney General's ballot, which may require reprogramming all voting machines statewide), which was issued at the same time.

This past week, Iowa joined a handful of U.S. states that allow overseas voters to receive and return their ballots electronically. Iowa currently has an estimated 1400 citizens overseas in the military, and they will be able to receive and send their ballot using email in this fall's elections. There are now 10 states that allow overseas voters to receive ballots electronically, and 7 that allow them to return their ballots electronically.

I've not been able to find much detail about how the email ballot transmission and return will work, only some brief reports in the media. For example, here is one explanation:

Brunelli says they have a webguard feature for safety and security with the e-mail and she says they also get the original voted ballot sent in as a check on the process. Brunelli says they'll use a common program to ensure the ballots are safe. She says the ballots that are sent back by e-mail will be scanned and sent back in the P-D-F format so they cannot be altered.

(Brunelli is the director of the Federal Voting Assistance Program). Once I find additional details, I'll pass them along.

At this point, there has been little research on overseas voting, and no studies that have focused on efforts like these to use new technologies to facilitate the participation of overseas citizens. Such studies need to be conducted, to determine the basic effectiveness of these efforts, along a number of important dimensions, like reliability, accuracy, privacy, security, and usability, just to name a few.