How often does it happen in America that innocent people are convicted of crimes they did not commit? Is it really as common as the estimates we so often see cited, or is the real frequency different?

The Arizona Law Review has published new estimates by two authors giving different answers to this question: a lot lower and drastically lower.

Paul Cassell of the University of Utah gives us the "drastically lower" answer in the lead article. His estimate of the frequency of wrongful convictions is a range of 0.016% to 0.062%, in contrast to the commonly cited estimates ranging from 1% to 4%. George Thomas of Rutgers gives us the "a lot lower" answer in his "partial response" article, with a range of 0.125% to 0.5%. Prof. Cassell also has a reply article and a post at the Volokh Conspiracy.

We should not, of course, minimize the problem of wrongful convictions. A single one is a miscarriage of justice and devastating to the people involved. That is why, even as we limit repeated reviews of convictions on innocence-irrelevant grounds, we should maintain or widen avenues of review for cases of genuine "got the wrong guy" innocence. I insisted on that in the drafting of what became California's Proposition 66.

We also should not uncritically accept dubious estimates from notorious exaggerators. See Ward Campbell, Exoneration Inflation (2008). I welcome these new contributions to the field and look forward to reading them more closely.