Personalized Voting

Yesterday and today I am attending an Accessible Voting conference and we have had an interesting experience. We were broken into 4 groups, discussing different aspects of the voting process — from the pre-voting registration and voter information component through voting modes (remote and in-person) and ballot design.

The fascinating part of the experience was this: several groups independently agreed on the need for two things in the voting process. First, there should be more options for voting — early, absentee, and vote centers — because it provides individuals with special needs options for voting that can accommodate their needs. Second, and more interesting, several groups wondered why each of us do not have the ability to create a voter profile that specifies the voting experience we want to have.

Imagine that, when you registered to vote that, after you registered, there was a survey you were given. (Something similar could be done at a kiosk for in-person voting or be included as a scannable document in an absentee voting packet). The survey asked you your preferences for voting. The survey would ask you several things, such as:

1. I would like to receive a voter guide before the election. (by mail, via email)
2. I would like to receive a voter guide in large print.
3. I would like to receive a ballot in large print. (in-person or absentee)
4. I would like to be a permanent absentee voter.
5. I would like to receive a list of early voting locations.
6. I have a Handicap Parking permit and will need that parking at my polling place.
7. I will be using the accessible voting machine at my polling place.
8. I would like an email reminder of my polling location.
9. I would like an email notification that my absentee ballot was received.

The idea here is simple: people should be able to express their needs and wants (not assumption-based because a person is identified as blind or disabled) and the system would provide voters with a preference-enhanced experience. Some people don’t need a sample ballot but others need one in large print or an audio version. Some people may want to be permanent absentee voters and others may want to get a list of early voting centers in a text message.

This system should be dynamic — I can update my preferences easily using the Internet, the phone, or some other mechanism — and should reflect my needs.

I realize that there are a million details here: cost, state laws, etc. However, the interesting thing here is that it could be scalable: states could provide this service across local jurisdictions and utilize new services like print on demand for helping make this all work.

How Jackson County, MO expanded their elections infrastructure

I’ve blogged a few times about the unanticipated infrastructure demands created by early voting. Most elections offices are designed to handle a few hundred citizens with questions about registration or disabled citizens needing use special access machines, not thousands or tens of thousands of voters showing up to cast a ballot. thousands of in-person voters.

This story from Jackson County, MO, just outside of Kansas City, illustrates the problem.

five election dates, new legislative districts thanks to the 2010 census and even seemingly simple things like generating new notification cards for every registered voter. And the November ballot – with a presidential race, several statewide races and initiatives, state legislative contests and possibly local ballot issues – is expected to be long.

The Democratic director of the board, Bob Nichols, noted “We had people lined up outside and in our office.” Tammy Brown, Nichols’s Republican counterpart, added “It is a crazy year.”

Adapt or die, as my colleague Doug Chapin often notes, and in this case, adaptation was easy. The story doesn’t note who saw the empty storefront across the street, but the Board has rented it, and just like that, more space for voting, shorter lines, and less stress on the elections staff.

Some Lessons from Iowa

Several people have written recently about the Iowa caucus results, including our good friend Doug Chapin.  I am going to pile on here but make a couple of broader points.

First, the Iowa results and the subsequent recertification of the results make a point that Lonna Atkeson, Mike Alvarez, and I make in our upcoming book on election audits:  election results are only meaningful if there is a chain of custody for the ballots.  You can audit, recount, and certify ballots after an election, but unless you can prove with a chain of custody that the ballots are not tampered with and are the same ballots that were cast on election day, any counting is rather pointless.  Elections are an end to end process, not a “count at the end of the night” process.  If people are not following procedures for handling ballots and there is not documentation of how ballots were handled — with signatures on forms documenting the hand off of ballots from location to location, for example, — then it is not possible to know what is being counted, let alone if the ballots are counted correctly.

Second, the process in Iowa also illustrates the problems associated with pure paper-based voting.  With either DRE electronic voting or voting with paper ballots that are scanned, there is a second, computer-based set of tabulated results.   If one set of results is lost, there is a backup.

How many votes have been cast in Florida?

A little mathematics and some web browsing skills, and it looks like somewhere between 8% and 17% of the ballots in Florida have already been cast.

Keep in mind that, as of today, four candidates (Bachmann, Cain, Huntsman, and Perry) were on these ballots, and three (Bachmann, Huntsman, and Perry) were still actively competing well into the absentee balloting period.

We’ll use Miami Dade County to make our calculations, but this same exercise can be done throughout Florida (and you can bet the Romney, Gingrich, and Santorum forces are making these calculations daily).

Miami Dade County, reports 42,149 absentee ballots returned as of close of business yesterday, totaling 32% of the 130,491 ballots mailed out.   The Miami Herald reports 143,000 ballots voted statewide, just over 30% of absentee ballot request.

According to GOP registration for the primary (the county code for Miami is “DAD”), 35% of Republicans requested an absentee ballot.

That gives us the turnout so far, and the percent of registered Republicans who requested an absentee ballot.  Now we need to estimate overall turnout.   In the 2008 presidential primary, turnout statewide was 38%, and in Miami Dade, turnout among Republicans in the 2008 primary was 40%.   Among Republicans in Miami Dade in 2008, 20.8% (33,548) of all ballots were cast absentee (another 24,744 were cast early in person).

Now we have almost all the pieces of our puzzle: registration, percent requesting an absentee ballot, percent voting overall, and percent voting absentee.

But wait!  We already have  42,149 absentee ballots returned in Miami-Dade, almost 10,000 more than the 2008 total.  What’s going on?

Maybe there are more Republicans?  No–according to the state, there are 7000 more GOP registrants in Miami-Dade in 2012 than in 2008.  Maybe Republican turnout will be higher?  Possibly, although the 2008 primary was still quite competitive on the Republican side.

But what if Republicans are voting absentee at a higher rate than in the past?  State party leaders said that absentee ballot requests are double the 2008 rate! The ground has shifted in Florida.

Now the final piece of the puzzle: we need to take into account absentee ballots that are never returned.  Charles Stewart, in his recent Election Law Journal article “Adding Up the Costs and Benefits of Vote By Mail”  has coined the term “ballot leakage” to describe absentee ballots that are but not counted, either because they are not returned (20% nationwide) or otherwise spoiled.  We’ll try to account for “leakage” below.

Let’s do the math:

32% of absentee ballots already returned x 1.25 (assuming only 80% of absentee ballots are actually voted) = 40% of all absentee ballots that will be voted have already been voted.

40% x 20.8% (percent of all ballots in Miami Dade that were absentee in 2008) = 8.32% of the GOP primary vote in Miami Dade has already been cast.

And if the state GOP if correct, and twice as many GOP primary ballots will be cast absentee, then almost 17% of the ballots are already in. That’s an awful lot of votes cast on a ballot containing four names of candidates who are no longer in the race.

GAO Report on Weekend Voting Released

http://gao.gov/products/GAO-12-69.  From the abstract:

For the 2010 general election, 35 states and the District provided voters at least one alternative to casting their ballot on Election Day through in-person early voting, no-excuse absentee voting, or voting by mail. Specifically, 33 states and the District provided in-person early voting, 29 states and the District provided no-excuse absentee voting, and 2 states provided voting by mail to all or most voters. Of the 9 states and the District where GAO conducted interviews, all but 2 states provided voters the option of in-person early voting in the 2010 general election, and 5 states and the District offered both early voting and no-excuse absentee voting. Implementation and characteristics of in-person early voting varied among the 7 states and, in some cases, among the jurisdictions within a state. For example, 5 states and the District required local jurisdictions to include at least one Saturday, and 2 states allowed for some jurisdiction discretion to include weekend days.

Continue reading →

Final Report on Russian Duma Elections Released

OSCE/ODIHR has released its final report on the Russian Duma elections today, which is available at http://www.osce.org/odihr/elections/86959. It contains a section on NVT and three recommendations in this regard.

Two types of new voting technologies were used during these elections. The first was a ballot scanning system called “KOIB”, introduced in 2004. The second was an electronic voting system called “KEG”, initially used in 2006. The later consisted of a touch screen to cast the vote and of an embedded printer to give voters the possibility to verify their choice whilst voting. The ballot scanners were used in 4,800 polling stations and 326 polling stations were equipped with the touch screen voting systems. (22)During election day, no serious problems or malfunction of either system were noted in polling stations visited by the OSCE/ODIHR observers. Poll workers seemed well trained and confident in managing the election process using the new voting technologies. They were supported by the designated staff that could assist voters when help was requested. OSCE/ODIHR observers noted that most voters needed such help. It was also noted that people providing assistance could easily see the contents of ballots to be scanned or the votes cast on the touch screen, violating the secrecy of the vote.

Although both systems provided a ‘paper trail’ (scanned ballots with KOIB and votes printed on a paper strip by an embedded printer with KEG), the absence of provisions for a mandatory random manual recount in a significant number of polling stations where new voting technologies are used is of concern. (23) In addition, the fact that votes in the KEG system are printed consecutively on one strip of paper could create the potential for the violation of the secrecy of the vote.

On the morning of election day, both systems were tested prior to actual voting, resulting in a printed test protocol, after which machines were reset and put in voting mode to start the voting process. Both types of new voting technology are based on ‘non-disclosed proprietary software’, not open to public scrutiny. Despite limited functional tests and certification of physical properties of the hardware, there has been no public independent evaluation or formal certification of these systems. (24) This can affect confidence of voters in both systems.

Recommendations

  • Mandatory recounts for a random significant sample of polling stations where new voting technologies are used should be carried out, as allowed for by current legislation. Such a measure can contribute to further enhancing trust in such systems.
  • To enhance the transparency and trust in new voting technologies used, evaluation and formal certification of the soft- and hardware by an independent public body against publicly available functional requirements could be considered, with the detailed evaluation report made public.
  • Either technical or procedural measures could be put into place to prevent poll workers from seeing the contents of ballots (for instance, through the use of privacy ballot covers) or votes being cast on touch screen machines (for instance, by using better privacy protection shields or proper voting booths) when helping voters.

Footnotes