Election Updates

Rick Hasen posted links today to some interesting material from the February issue of the Havard Law Review on Election Law. Two of the essays caught my eye: "Detecting Illegal Votes in Contested Elections" and "Voter Identification Laws." Like Rick, I'm looking forward to reading this material soon.

There is an excellent essay on Charles Franklin's blog, Political Arithmetik, on the Palestinian exit polls. Charles is a leading expert on polling and statistics, and his blog is worth reading just to see what is currently on his mind.

But in his February 22 entry, "WSJ Numbers Guy on Palestinian Exit Polls", Charles digs a bit into the issue of polling errors in exit poll data. The issue at hand, just as was the case regarding the 2004 presidential election exit polls in the United States, is how can we know that survey response errors (in the case of the 2004 U.S. election and in the recent Palestinian election) are more likely to come from one political party relative to others. Here's what Charles wrote:

But here is the rub: How do you know the non-responses are overwhelmingly for Hamas? (Or for George Bush in 2004?) There may be good reason to think this is so, but what data support the inference? In the U.S. we have the advantage of sample precinct returns which provide a check on the exit polls. As data come in from sample precincts they are compared with exit results from the same precinct, which allows estimation of how much non-response might be affecting the exit results. Such an option doesn't exist in the Palestinian case, where the counting process is much slower. So a Palestinian exit pollster is faced with a dilemma: adjust the results based on substantive expectations (really, your best subjective judgement about non-respondents) and admit that your statistical results are shifted by a clearly non-data driven component (which could be wrong). OR, decline to introduce non-data driven elements into the calculations, with the clear risk that your results may be biased by selective non-response. That is a really tough decision. (A third option would be to compare current results with historical election returns, but in the Palestinian case there is very little past election data to use. That will, of course, improve over time if democratic elections continue to be the practice. Such a comparison can't account for across the board shifts, but might provide some leverage on non-response that is otherwise unavailable.)

Charles then notes that he is going to dig into this in more detail in the future, and gives links to two websites with data on the exit polls in the Palestinian elections: the Palestinian Center for Policy and Survey Research (PCPSR) and the Development Studies Programme of Birzeit University (DSP).

I'll keep watching Charles' blog for updates, but Election Updates readers might want to bookmark his blog for themselves!

There is a nice summary story in today's Electionline weekly summary on e-poll books. In addition to the positive comment in the story from me on how e-poll books if properly implemented should help resolve many problems that we have seen in past elections on voter registration (see the Caltech/MIT 2001 report for discussion), I believe that the quick mention of possible concerns about how e-poll books might be the source potential threats ("Some election watchdogs fear e-poll books could be the target for "denial of service" attacks on Election Day ...") comes from my paper on "Precinct Voting Denial of Service".

Though how I've gone from "election geek" to "election watchdog" in the eyes of our friends at Electionline is a mystery to me.

Thad and I wanted to welcome Melissa Slemin to Election Updates; Melissa is working with the Caltech/MIT Voting Technology Project (VTP) at Caltech, and we've invited her to join us and for her to use Election Updates to pass along information to our readers about what the VTP is doing, and other information that she comes across that might be of interest to our readers.

Welcome, Melissa!

There's a report from the University of Las Vegas, Nevada, of an interesting scheme to either facilitate vote fraud or to just undermine the credibility of a student government election there. Apparently an anonymous email was distributed warning of potential fraud in an upcoming election, and attached to the email was a file containing the last four digits of every student's identification number --- the same information that apparently students use as a password to vote.

Whether the anonymous email was really an attempt to faciliate fraud, or just to cast doubt about the credibility of an upcoming election, is unclear. This is an example of the sort of attack on a voting system that I wrote about in my "denial of service" threat analysis.

The response from the student government body in charge of running these elections, though, was unusual:

Zucco said CSUN's elections committee is working on a different system. In his e-mail, he stated, "We are currently working on a system that no one knows about and it will stay that way [until] the system is operational." He explained the secrecy is crucial until Elections can be certain the new system will be implemented.

It is not clear that a "security through obscurity" response here will either be effective, or whether it will have credibility, if potential voters do not know whether their votes are being counted as intended.

Paul Gronke has some useful summary information posted on his Earlyvoting blog. He received the same question that I did yesterday from a colleague, regarding whether or not there is any academic research as to whether by-mail voting might stimulate participate by lower income and minority voters.

Paul's answers are worth repeating here, as he is one of the experts in this field and has been doing some excellent research on early and by-mail voting:

I have not seen any studies of by mail's impact on primary elections. I have seen research on by mail in low intensity contests (most often, off cycle state and local elections), and the research shows that turnout is higher among regular voters--those who would go to the polls in most circumstances but may not if inconvenienced by the need to go to a precinct place.

There is no evidence of by-mail *expanding* the electorate to otherwise disempowered groups (racial and linguistic minorities).

For citations, Berinky et al. (200x POQ), Oliver (199x AJPS on absentee balloting, not the same as by mail voting), Magleby 199x article on by mail voting.


I think I have already answered this in my response to 1). No evidence of increased turnout. I should note that there was some evidence of increased turnout in the in-person early voting system in Florida in 2004 (see my most recent APSA paper) and in in-person systems in Texas in some elections (see Leighley and Stein APSA paper), but this is critically dependent upon the mobilization efforts of political parties and other GOTV organizations.

Yes, this is exactly what happens, but among those who are otherwise predisposed to vote. All-mail does not make voting "convenient" enough, apparently, to overcome the barriers that otherwise stand in the way of higher turnout in minority and disempowered communities.


So what I would say to the registrar is that, as far as turnout goes, by-mail is not a panacaea. It solves some problems (it increases turnout in local elections and probably would do the same in primaries), but may create others (undercutting the civic event that constitutes an election). We don't know much yet about the latter. The biggest hurdle to turnout is efficacy and feelings of disengagement--by mail voting does not address this.


By-mail voting may make it easier to voter among those facing linguistic or other barriers to turnout, but we simply don't know enough yet to say for sure.

This question has arisen because (as Paul correctly surmised) up in Sacramento there is a hearing going on this morning in the state Senate (Committee on Elections, Reapportionment and Constitutional Amendments) considering AB 707, which would authorize by-mail elections in California in time for the June 6, 2006 statewide primary elections.

And no, I am not testifying at these hearings ...

Okay, just when I thought things could not get much worse regarding the recent elections in Haiti, this odd little side-note was in the news today: Jacques Bernard, the newly-appointed head of Haiti's electoral council, has apparently fled the country. Here is summary from the Washington Post: "The head of Haiti's electoral council fled the country after opponents threatened his life and burned down his farmhouse nearly two weeks after disputed elections, an official said Monday."

One wonders what will happen next, as they are still in the midst of tabulating votes from the last election, and there is supposed to be a runoff election in about a month.

The National Commission on the Voting Rights Act issued a long (187 page) report titled "Protecting Minority Voters: The Voting Rights Act at Work, 1982-2005." For those interested in the general issue of the Voting Rights Act, it's effects on protecting minority voting rights, and the current debate over renewal of the Voting Rights Act (see my recent essay on Senator Obama's comments on VRA renewal), this report is an interesting read.

For readers of Election Updates interested in election administration issues, there is a very useful discussion within this report on minority language assistance programs. This discussion is buried in the report, from pages 67 through 75. In particular, there is a helpful discussion in the section titled "Examples of Effective Language Assistance Programs" on various efforts throughout the nation.

The following section, on "Language Assistance and Voter Turnout" is interesting, less compelling, but provocative. At this point in the report, they cite testimony at their hearings, as well as a study by James Tucker and Rodolfo Espino (which I'll try to find but haven't yet tracked down), which is asserted to provide evidence that language assistance programs have boosted minority voter turnout. The details provided about the Tucker and Espino study state that "the findings are based on a survey sent to 810 covered jurisdictions in 33 states. Over half responded, and complete responses were received from 361 jurisdictions in 31 states" (pages 67-68). What is difficult to understand is how one could focus only on covered jurisdictions and study the effect of language assistance programs on voter participation; rather the right methodology would seem to be to study all jurisdictions and to have variance from both jurisdictions without language assistance programs and those that do have language assistance programs to assess the effect of variance between program type and voter participation. In any case, I will try to locate this study and pass it along to readers of Election Updates, as it will be good to go to the source for details of their study. For enterprising students looking for research projects, here is a good one: look at how these VRA programs (and how well they are implemented) serve to enhance voter participation across the United States, since VRA implementation.