Even with all the usual caveats about such studies, it remains troubling that a new Urban Institute study shows between 8 and 15 percent of those convicted of serious sexual offenses in Virginia, from the early 1970's until the mid 1980's, were innocent. It suggests, among other things, what we already know from laboratory studies: people don't do a great job identifying perpetrators. Given the stakes in these cases - long prison bids or death, plus a lifetime of community notification for those who complete their sentences - this error rate is quite consequential. Then there is the other matter: in each case where an innocent person was convicted, a guilty person remained free.
Some people will surely argue that this is within the tolerable zone of error for a "beyond a reasonable doubt" paradigm. While I don't share that view, anyone should be troubled by the possible allocation of this error. Nobody would be surprised if erroneous convictions were more frequent in cases involving defendants of color as well as defendants with limited financial resources.
Others may dismiss this study as an artifact of old DNA testing technology. I hope this is true. But this only means that as a society, we may have a duty to double check our old convictions in light of new technology. And it should also sound a cautionary note: when these convictions occurred, we were quite confident of the accuracy of then-modern science. Perhaps new technology is moving us closer and closer to "accuracy" in verdicts. Indeed, this is the promise of our new high-tech surveillance society. For now, however, you'll just have to color me a skeptic.