Election Updates

The new Ohio Secretary of State, Jennifer L. Brunner, is proposing that poll workers be conscripted, like jurors. She is asking the legislature to require residents to work at the polls the way they fulfill jury duty.

The entire proposal, which is reported in the NY Times, is interesting but you need to disentangle several aspects of her view, which the Times does not really do well. Here are the issues:

1. Should being a poll worker be a manditory duty of citizenship?
2. Should poll workers work from opening to closing?
3. Do you need partisan balance among poll workers?
4. Are poll workers old?

On the first point, it is important to remember that being a juror is required only inasmuch as you have to show up to be in a pool. The lawyers then throw out the "bad apples"--people with bias--to get what is a likely fair pool of jurors. Being a poll worker should be similar, if it is a requirement. For example, imagine a polling place with a couple of milquetoast poll workers and one flaming partisan who thinks provisional balloting is inherently wrong or that any ID requirement is wrong. You could easily get chaos as a result.

One solution is to provide incentives for county employees to be poll workers by allowing them the day off and paying them the poll worker salary as well. This is done in Los Angeles and other jurisdictions and may improve the quality of the poll worker pool.

Second, should poll workers work from opening to closing? Here, there are several competing interests. On the one hand, having one set of poll workers allows for clear chain of custody for everything in the precinct; if something screws up, we know where to point the figure. Having multiple crews screws this up, potentially. On the other hand, it is very inefficient to have one team all day; not only do they get tired and more likely to make mistakes at day's end, but they also over-staff polling places at mid-day when traffic is at its ebb so that there are enough workers at opening and closing. Balancing this may require having some people work all day and some work at the ends of the election day.

Third, there is the question of partisan balance. When I was growing up in Georgia, I assumed all poll workers were non-partisan because Georgia does not have party registration for voters. If I had grown up in Ohio, I would have thought all poll workers represented partisan interests. The point here is simple; poll workers don't have to be partisans. The problem with Ohio isn't that the poll workers may not have complete partisan balance in every precinct. The problem is that outside observers cannot observe polling place operations on election day! If the Secretary really wants to improve election operations, this would be one place to start.

Finally, the article makes claims about the age of poll workers. Here, we just don't have good data to generalize. I am aware of only 6 studies of poll workers that involve actual surveys of poll workers. Four were conducted by my friends at Brigham Young University and I have conducted the other two. Poll workers in Ohio are old, but there is a range of ages represented in the pool. Moreover, the poll workers in Utah are 10 years younger, on average, than those in Ohio. The point being, we should not generalize about poll workers without good data.

Finally, Doug Lewis is absolutely correct when he says to the NY Times that he "would prefer the Ohio initiative on a smaller scale along the lines of tests in small, medium and large counties. 'I think we need to keep an open mind but realize that some of these ideas may not be as convenient as they appear at first blush,' Mr. Lewis said."

Pilot testing of election ideas; now THAT is a novel concept!

See the following EAC website for the documents they have made available.

Yesterday there was this report from the Associated Press about the conviction of two election workers, who rigged the recount of ballots cast in the 2004 presidential election in Cuyahoga County.

While the recount was rigged, there isn't evidence this actually had any effect on the final result of the presidential election in Cuyahoga County:

Special prosecutor Kevin Baxter, who was brought in from Erie County to handle the case, did not claim the workers' actions affected the outcome of the election - Kerry gained 17 votes and Bush lost six in the county's recount.

But Baxter insisted the employees broke the law when they worked behind closed doors three days before the public Dec. 16, 2004, recount to pick ballots they knew would not cause discrepancies when checked by hand so they could avoid a lengthier, more expensive hand recount of all votes.

The following letter was sent yesterday from New York's State Board of Elections:

As you are aware, on January 4, 2007, the State Board of Elections learned that CIBER, our Independent Testing Authority had applied for and been denied interim accreditation by the EAC. The State Board of Elections made several requests (including two FOIL requests to the EAC) over the last three weeks for documents related to the EAC's denial of CIBER's application for certification.

This correspondence is to inform you that today (Thursday, January 25th,) CIBER forwarded relevant documents to the State Board of Elections. While these documents will assist the State Board in its analysis of the issues, this does not resolve the ongoing resistance by the EAC to release relevant documents in their possession, as well. We have every intention to pursue the disclosure of all relevant documents maintained by the EAC.

While we are reviewing the documents and making an assessment of the situation, in order to limit any potential liability, the State Board of Elections has requested that CIBER, and all subcontractors, stop all work on the contract with us pending further notice.

When we have completed our assessment of this situation, we will advise you of the next steps.

Peter S. Kosinski and Stanley L. Zalen
Co-Executive Directors

This follows other criticisms from the folks in New York about the EAC's unwillingness to provide information about their denial of accrediation to CIBER.

Hopefully this will help pry more information about the EAC's denial of accrediation to CIBER ...

The 2007 Electronic Voting Technology Workshop will be held on August 6, 2007 in Boston, MA. The event is sponsored by USENIX and ACCURATE. The event will be co-located with the 16th USENIX Security Symposium.
Submissions are due Sunday April 22, 2007, 11.59 pm PDT
Notification of acceptance will be Friday June 1, 2007
Final files are due Thursday June 28, 2007.

Program Chairs are Ray Martinez and David Wagner

More details on event website.

For those interested readers, I am going to post my notes from the "2008 and Beyond: The Future of Ethics and Election Reform" conference held in Columbus, OH last week.

These are just brief blurbs, and are completely my responsibility and not meant to be comprehensive summaries of the intent or conclusions of the authors. Appropriate comments aside ...

Panel 1:

Beth Rosenson presented a paper on ethics reforms and conflicts of interest. Unfortunately, I had to help set up my computer for the projection and was working on my own notes, so I did not take notes on Beth's paper. My apologies.

Ken Mayer: The impact of "clean" (publicly financed) elections on candidate emergence. Ken asked whether the candidate pools is increased--and would the reforms especially help challengers. He wanted to see if public financing reduces the influence of money.

He reported a mixed record of public funding. Some evidence of a more varied candidate pool (more challengers). Mainly needed in non-competitive system (but he did not say – these are precisely not the places where it would be adopted). As to who takes public $ -> Dems, Women, and more likely to take $ as races are less competitive.

Paul Gronke presented a paper on the relationship between early voting reforms and turnout. I summarized the state of early voting reforms in the United States, including a chart of the number of states with some significant amount of non-precinct level voting. We showed what evidence we could of non-precinct voting in 2006 (I gave a shout out here to Tokaji and Foley for writing about the necessity of getting higher quality election data out of the states). Finally, I showed some estimates of the impact of early voting reforms on turnout--increases that ranged from 1-4%, depending on the type of reform. These effects are significant by substantively modest (in line with most published research on the subject).

Thad Hall presented work from a number of pPoll workers surveys that were conducted in 2006. Hall argued that the efficacy of training methods is seldom studied. He stressed the gatekeeper role of poll workers. The world of elections is very dynamic, and thus training is important. He described what he called the "Spray and Pray" model vs. the "Small group, active learning" model of training (I've been to enough edu-talks to know where Thad is going with that one!). Finally he compared Cuyahuga County vs. Utah. This was a natural experiment because both locations adopted the same machines but different training methods (spray and pray, large groups in Cuyahoga; small groups and interactive experiences in Utah). Not surprisingly, Utah's pollworkers reported much more confidence in training and much higher job satisfaction.

[ED NOTE: I would say that the issue of poll worker training came out as one of the key issues throughout this conference.]

Lonna Atkeson & Kyle Saunders presented a paper that relied on voter surveys in two highly contested races, congressional races in NM 1 and CO 7. They wanted to relate how confident voters were that their ballots were being counted accurately and how satisfied they were with the voting experience to administrative variables like quality of poll workers and quality of election judges. Basically, the found a strong relationship between high quality poll workers and satisfied citizens. The machines don't matter as much as the people running them. This is really an important lesson, and one that I hope election reformers and election officials take to heart. They also found high confidence was positively related to easier to understand ballots, voter verified paper trails, and a longer period of time to fill out the ballot.

In what seems like a bizarre turn of events in the wake of Denver's problematic election this past November, stories are circulating today about a letter that apparently the USPS district manager in the Denver area sent to the Colorado Secretary of State's Office, criticizing both Denver and Sequoia Voting Systems for problems with absentee balloting in the January 2007. Here's a story from the Denver Post.

There is some interesting data in the Denver Post story, providing some statistics on the quality of the voter registration list used in Denver:

More than 10 percent of the nearly 300,000 ballots sent out for Denver's Jan. 30 mail-in election were returned, calling into question the accuracy of the city's voter registration list.

About 30,000 ballots sent to Denver voters have been returned as undeliverable, commission spokesman Alton Dillard said. An additional 3,500 ballots were sent out to wrong addresses - an error the commission blamed on a vendor.

Dillard said the delivery problems stem from the rush to schedule the mail-in election and the decision not to purge the rolls of inactive voters after the November election.

Given that little is known about the quality of voter registration lists, it is hard to evaluate a 10% inaccuracy rate, but it does seem high.

I suspect we'll have more to say about this, especially once Paul gets a chance to weigh in here and we get some additional information on what is happening in this election.

Here is the announcement and call for papers for the Second Annual Conference on Empirical Legal Studies, NYU, November 9-10, 2007.

A report coming out of Snohomish County, WA, which experimented with a fully by mail election in 2006, indicates that the election cost about the same as a comparable midterm election. The big costs for the county included printing, sorting, and mailing ballots. As county auditor Carolyn Dipenbrock notes "Elections are very labor intensive, and people do not understand how difficult it is to get the correct ballot to them."

Snohomish was hoping to save money, as they indicate in their vote by mail planning report. Their experience runs contrary to the much-reported cost savings that the state of Oregon experienced when they went fully vote by mail.

Story in the Seattle Herald. I tried, but can't find the Auditor's report online.

I just happened upon this survey report from NASS on state-by-state limitations on who can be in polling places. As you can easily see by skimming the list, among the states that responded to the NASS survey, there is considerable varation in who can be in polling places. This is especially true for independent election observers, as it is clear that they are either not allowed in number of the responding states, or they are required to obtain pre-election credentials to authorize their presence on election day. Enterprising students looking for research projects might think about quantifying the information in this report, and then correlating it with other measures, especially outcome measures, to ascertain whether easy polling place access by observers has any effect on outcome variables, for example, allegations of election fraud or reported incidents.

In an announcement late last week, the EAC received a recommendation from NIST that it accredit two new test labs:

NIST recommended that EAC accredit iBeta Quality Assurance and SysTest Labs to test voting systems against both the 2002 Voting System Standards and the 2005 Voluntary Voting System Guidelines. EAC will conduct further review of the recommended labs to address non-technical issues such as conflict of interest policies, organizational structure, and recordkeeping protocols. After the EAC review, the Commission will vote regarding full accreditation.

There's been some discussion about a report from Riverside County's (CA) election official, Barbara Dunmore, about some problems observed in that county's 2006 election, especially delays experienced by election day voters. There was recently a summary story in the Riverside Press-Enterprise regarding this report. However, the report itself hasn't appeared anywhere where I could find it.

I got wind of the report's availability in the minutes of a county board meeting, and you can obtain it from the county website here. Given that the county site was really slow, I've also put a copy on the Election Updates server, which should provide a quicker download for most readers. The report blames a long and complicated ballot, as well as an unplanned influx of late afternoon and evening voters, for most of the election-day voting delays.

There's a lot of interesting data in this short report, especially the county's analysis of where VVPAT's would likely need to have their paper changes (Appendices A and B), and the final table in the report, which gives some selected comments from what appear to be incident reports. It would be quite interesting to get all of the incident reports from the election officers ... that was very useful data to compile and analysis in Cuyahoga County (OH)!

As some of our readers have noted, since last Thursday evening, we have been experiencing some technical problems, mainly with the server that runs the VTP website (http://votingtechnologyproject.org), but some of these problems seem to have impacted the performance of the server that runs Election Updates. Sorry for these snafus, things should be running normally now.

A link to an upcoming talk by Matt Barreto of the University of Washington just came across my email:

WISER Brown Bag Series on Race, Ethnicity and Sexuality: Matt Barreto (UW Political Science), “The Impact of Vote-By-Mail in King County on Minority Voters,” Tuesday, February 6, 2007, 12:00-1:30pm, Gowen 1A.


This talk is at the University of Washington, Seattle. I'm not sure whether it's open to the public or not.