Balkinization  

Monday, July 07, 2008

Popular Monitoring of Popular Elections

Heather K. Gerken

Archon Fung, one of the most interesting thinkers at Harvard's Kennedy School, has just come up with an intriguing idea for monitoring elections: a teched-up, wiki-based system for reporting problems on election day. It's modeled on the award-winning British site, fixmystreet.com, where people report maintenance problems (graffiti, potholes, broken street lights), locating the problem on a map and often attaching photographs to the entry. The site is interactive; it reports when a problem has been fixed and maps where current problems are so that you can figure out how things are working in your neighborhood. As you'll see from his introductory site, Fung envisions a much bigger version of this idea -- a national "weather map of election conditions" that would show you where the biggest problems are occurring based on real-time entries by trained election monitors and everyday citizens. You could then drill down into the map, figuring out exactly where problems were occurring in your state, city . . . even your polling place. The visuals would look something like this map of gas prices.

What makes Fung's idea promising is that it’s a "here to there" solution. It doesn't directly change how our elections are run. But it helps create an environment in which change is possible. As I explain in my forthcoming book, The Democracy Index: Getting from Here to There in Election Reform, the reason that it's hard to get election reform passed is that election problems are largely invisible to the average voter. Discarded ballots, long lines, machine breakdowns, registration problems -- these all occur routinely during the election process. But voters only become aware of these problems when a race is close enough for the problem to affect the outcome. Given that most races are not competitive, that's a bit like tracking annual rainfall by counting how often lightening strikes. Because voters learn about election administration problems in a haphazard, episodic fashion, politicians have no incentive to pay attention to them unless there’s what Rick Hasen calls an "electoral meltdown."

The magic of Fung's idea is that it makes election problems visible even in the absence of an electoral meltdown. If enough people participated so that coverage is thorough and consistent -- and that's a big "if," as Fung recognizes -- the site would be a great way to draw people's attention to routine election problems. Indeed, I suspect that the site would be highly addictive. Like many others, I spent an inordinate amount of time reading the updates on election problems that Talking Points Memo and Ben Smith's Politico blog provided during the primaries. Those blogs, however, could provide only piecemeal information to their readers. By harnessing the power of the wiki, Fung's "myfairelection" site could provide coverage that is both more systemic (giving you a sense of the big picture) and yet more personalized (letting you see what's going on in your own neighborhood).

While myfairelection.com would not pass muster with any political scientist (because it depends on reporting rather than random sampling, the gold standard of social science research), it would surely attract the attention of politicians as well. Consider, for example, the success that election reformers have had with the Election Incident Reporting System (EIRS), a web-based system that has allowed voter protection groups and individuals to report problems they've encountered in the last few elections. Because the data are available at the state and county level, reformers can tell state legislators and local council members that problems exist in their neighborhoods. As Tip O'Neil understood, one of the best ways to attract the attention of a politician is to document a problem in his district. That’s presumably why fixmystreet.com has had such a high success rate as well. It makes a problem visible and shames politicians into fixing it.

While Fung's idea is intriguing, there are definitely some kinks to work out before the proposal is ready for prime time.

Fung has already thought hard about two obstacles -- getting people to participate and avoiding the problem of spamming. Here's another: figuring out how to ensure that the site doesn't generate or amplify false rumors. Think about what happened in Arizona during this primary season. Contrary to usual practice, the primaries were closed, which meant that only registered Democrats could vote in the Democratic primary and only registered Republicans could vote in the Republican primary. Unsurprisingly, a number of independent voters were (quite properly) turned away. Because people didn't understand why voters were being turned away, rumors began to fly that thousands of voters had been wrongly purged off the registration lists. Relying on average people to report problems means that you risk a lot of misreporting.

The costs to misreporting are quite serious. Election administrators already do a hard job with few resources. I'm all for using data to hold election administrators and politicians accountable for problems in the system -- I've spent the last year writing on the subject -- but you have to have reliable, comparative data to do so. Otherwise, you'll just end up savaging the reputations of people who deserve better.

The wiki solution to this problem might be to ensure that election administrators have access to the site and an opportunity to tamp down rumors. I'm not sure this would be enough, as people tend to be suspicious of election administrators' motives. That’s because people see a problem, learn that the election official in charge has a partisan affiliation (most do), and find it all too easy to connect the dots and assume the source of the problem is partisan shenanigans. My research over the last year has convinced me, however, that most election problems are caused by a lack of resources. I've thus proposed a gentle version of the rule called "Hanlon’s Razor," which says that we should never attribute to malice that which can be explained by incompetence. In the context of election administration, we should never attribute to partisanship that which can be explained by a lack of resources. But how do we communicate that idea to the people logged in to myfairelection.com?

Here again, a wiki solution might help address the problem. First, the site might widen the lens for voters by showing them that the problems they see in their own polling place are happening throughout the country. Imagine, for instance, someone entered a report indicating that lines were long at her polling place. Next to that report the site might put a running total of how many other polling places nationwide were experiencing the same problem. Second, the site might ask experts to blog simultaneously to provide some perspective on these issues, or it might link to extant research. For instance, anyone who reported that she wasn't properly registered to vote might be asked if she wants to see the latest Pew Foundation report on the registration process.


A non-wiki solution to the problem of misreporting would be to allow only trained election monitors to post. Fung is halfway there, as he envisions an army of trained monitors assisting with the site. But I'm quite skeptical even as to the viability of Fung’s more limited proposal. It would take a mammoth amount of resources and legwork to get trained monitors in every polling place in the country (and some legal work, as some states ban election monitors in polling places). Fung astutely suggests piggybacking on the existing infrastructure -- relying on the monitors that the campaigns and good governance groups put in place each election cycle. Even then, it will be very tough to get a monitor in every polling place.

Finally, we should all be cognizant of the limits of the site. Fung is right to suggest that one important benefit of his proposal is that it would give us real-time feedback on how well the election is going, something that would be useful to campaigns and election administrators alike. But Fung is also right not to overclaim about the dependability of this data for evaluating election performance over the long haul. Social scientists are skeptical of reporting systems like these with good reason. We’d need more dependable, systemic data (that covers not just balloting on election day, but the registration and counting processes) in order to evaluate how well our election system is performing. That kind of data can be had, as I explain at length in my book on building a Democracy Index, but it will take more than one website to get it.

Nonetheless, Fung's proposal is quite intriguing. If he can figure out how to reap its benefits while mitigating its costs, he'll have accomplished a good deal in helping us get from "here to there" in election reform. Myfairelection.com might seem like a modest reform, but it's a modest reform that could make bigger, better reform possible.

Comments:

Just look for the problems being in minority communities as part of voter intimidation and suppression. You heard it first here.
Best,
Ben
 

Hastily blocklettered felt marker sign at polling place November 2008: "Voters not allowed to use camera equipped cellphones inside poll".

Then there was the picturesquely named BrooksBros Rebellion, in FL2000.

During one primary in a state which I shall leave unidentifed a freelance blogger documented uncertified chain of custody firmware printed circuit cards swapped into jammed voting machines to restore operationality.

In 2004 there was a large metropolis which sent firmware and vote machine on "sleepover" home with poll workers a day or two early, without any tamperproofing guarantee.

Then we have the stringent new laws in Indiana, the state that banned a community of elderly monasts from voting because of missing state certified identification.

I wonder where my birth certificate might be on election day, and whether the provisional ballot, if a voter receives such a ballot, actually is counted before the election is certified.

I think Fung has a good muse, but elections tend to elicit some primal urges. I like to reflect on the Ukraine outcome, people in the plaza, a recount corroborating a somewhat less grand victory for an incumbent in the orange Revolt, more camping out, the challenger forming the new government.

Though, as usual, BenD managed to frame the entirety of the problem in one direct clause.

I will have to read again what professor Hasen has elaborated as meltdown. Though, this year, if reports are accurate, registration increases will compensate for minor inaccuracies or unfairnesses.
 

This is a good enough idea that it's worth launching now. Right now the site has an explanatory video and a concept paper...viz., Harvard professors are no good at marketing or web development.

If this site doesn't get up and running, with some decent TV spots and a solid print-media strategy, it will fail...get the good professor some marketing advice right now!
 

Unfortunately, voting problems are typically most severe in areas where Wiki skills are low. The 800 number worked better, as long as people saw or heard the ads for it, although having looked through EIRS I can say that reports are by nature subjective and it's often hard to identify the general problem at work. Best was having trained volunteers at polling stations to talk to voters and poll workers.

But the biggest problem we found was not logistical or technological but financial. It's always difficult to get real interest in procedural issues, and it's longer and longer since Florida ...
 

Post a Comment

Older Posts
Newer Posts
Home