Wednesday, December 03, 2008

Will Obama Reduce the Chance That You Are Called for Jury Duty?

Ian Ayres

Crosspost from Freakonomics:
Photo: Tom Lemo

One of the changes that the “Yes We Can” movement has already wrought is a substantial increase in voter registration — particularly in swing states. In Virginia, for example, the number of registered voters increased by almost 10 percent.

Since voter-registration lists are also used to construct juror lists, a possible benefit of this registration boost is a reduction in the number of times the rest of Virginians are called for jury duty.

But the Obama bump is likely too muted. The construction of the juror rolls varies by state. States use a variety of databases, including not just voter registration lists, but utility bills, driver’s licenses, property tax rolls, welfare rolls, phone directories, and school lists. New York relies on “registered voters, licensed drivers, and the state’s mailing list of taxpayers.” So registering to vote does not necessarily mean that you are adding your name to your state’s juror rolls. (This is probably a good thing. It would be even harder to get citizens to overcome the voter’s paradox if registration subjected them to being called for jury duty.)

Any bump is also likely to be delayed, as the juror lists in many states are only updated periodically. Virginia tells us that newly registered voters would only hit the jury lists of the Commonwealth in 2010.



Nice post from the land of unintended consequences. I now have a new reason to work those voter registration drives.

Anecdotal evidence suggests that in my state any effect from increased votor registration would be swamped by other effects.

I'm called every year like clockwork. Others who are just as selectable (registered voters and drivers all) are called seldom or never. Clearly other factors are operating here, and their effect weighs in at a lot more than 10%.

So while it's fun to speculate and do math and project trends, the fact is probably that all the trend in question wouldn't be detectable above the noise. If you want to study likelihood of getting called, increase in pool size is the wrong place to look. The right place is probably selection bias.


Can you speak more on your thoughts about selection bias?

Many states get potential juror roles from DMV registration and from driver's licenses as well as voter rolls.

I've been asked to report for jury duty several times in CT, and once in CA so far....

So I'm not so sure that increased voter registration will lead to a much larger pool.


selection bias?

Sure. Over time, some are consistently called more than others. Basically the definition of selection bias.

I don't know why the bias exists, what maintains it. I don't attribute it to malice and very much doubt there is any. More likely just the mechanics of selection are flawed, and one result is some end up getting called more than others.

I have no inside knowledge of the mechanics. I will venture a guess: poor algorithm; random selections not equiprobable over time. A good algorithm demonstrates even coverage of the space. An excellent algorithm would select without replacement, thus satisfying a fairness condition "sooner or later everyone gets called".

Why doesn't the system do better? Look at what drives the system. Running out of potential jurors before a jury is assembled is immediate sharp pain. Being unfair to citizens is faint steady pain. The system responds by avoiding the immediate sharp pain but tolerates the faint steady pain. So selection bias creeps in, no one does the work to ferret it out, and the result is what we see.

So my guess is no one intends the selection bias, but no one is going to do anything any time soon to fix it.

Now: the specific ways in which selection bias creeps in are not random either. For instance, it's much easier on the system to select people who will show up. Evidence suggests that's exactly what the system does. It has been observed that an end result of that is juries filled with middle class white people judging defendants who are poor people of color. The accused leans over to his public defender and says "I don't see nobody on that jury who looks like me". The anecdotal evidence of my experience would agree.

So as I understand the situation, there is clear, longstanding, specific selection bias. The systems rolls along, juries get empanelled, courts seldom have to stop work for the day because they're out of potential jurors. But people like me get called more often. And juries filled with people like me are called upon to judge cases that turn on life experiences we couldn't possibly understand.

Another bonus is that it'll make it more likely that all the dead and fictional defendants can be seen by a jury of their peers.

Post a Comment

Older Posts
Newer Posts