« Buffetted | Home | Unhealthy »

May 2, 2005

Lethal lottery

What determines which death-row inmates eventually get executed? Not their crimes, according to criminologist Dee Wood Harper and computer scientist Stamos Karamouzis of Loyola University in New Orleans.

The Death Penalty Information Center says there are more than 3,400 inmates on death row in the U.S. On average, most remain there for at least a decade. Some 42% are black (compared with 12% of the U.S. population), while 46% are white (compared with 75%). In recent years, 60-70 death-row inmates have been executed each year—the largest number in Texas. Since 1976, when the death penalty was reintroduced, more than 960 prisoners have been executed across the U.S.

Harper and Karamouzis analyzed 28 years of death-penalty cases using an artificial neural network (ANN), a multiprocessor computing system that mimics the way the human nervous system processes information:

The main characteristic of such a computing system is the number of highly interconnected processing elements (neurons) working together to solve specific problems without being programmed with step-by-step instructions. Instead, ANNs are capable of learning on their own or by example through a learning process that involves adjustments to the connections that exist between the neurons.
“[W]e reconstructed the profiles of more than 1,300 death row inmates from a national population by using simple attributes such as the inmate’s sex, race, and highest year of education completed at the time of first imprisonment for capital offense. Then we performed various experiments in order to develop an ANN that is suitable to the profiles. We trained the network by letting it ‘witness’ 1,000 of the profiles more than 100 times each. Finally, we tested the ANN using 300 profiles that the network never witnessed before.

In all, the ANN was fed 19 different basic data points about each inmate—but no details about the crimes themselves, or about whether each defendant had been fairly represented. Despite this, the software was able to predict whether each inmate lived or died with greater than 90% accuracy.

Harper now finds himself aligned with those who have long opposed the death penalty on the grounds that it is, in the end, arbitrary. As he puts it:

“Predicting execution outcomes for prisoners under a sentence of death utilizing essential attributes that have no direct bearing on the judicial process has serious implications concerning the fairness of the death penalty.”

No kidding: clearly who (and where) you are matters a lot more than what you did.

Posted by Stephen at 12:30 AM in Legal issues | Permalink | TrackBack (0)

Trackback Pings

TrackBack URL for this entry:
http://www.disinterestedparty.com/cgi-bin/mt/mt-t.cgi/86