Just How Many Wrongfully Convicted Remain Incarcerated?

Co-Editor Brian Howe of the Ohio Innocence Project discuss a recent study on the rate of false convictions.  Part two of this discussion will post tomorrow.  Brian writes:

 There has been a lot of press coverage recently about a new study published by the National Academy of Sciences estimating the number of wrongful convictions:  Rate of false conviction of criminal defendants who are sentenced to death.  SR Gross et al., Proceedings of the Nat'l Academy of Sciences.

First, the punchline:

"We present a conservative estimate of the proportion of erroneous convictions of defendants sentenced to death in the United States from 1973 through 2004, 4.1%."

This is tricky business for a number of reasons.  The raw figures on capital cases, executions and exonerations (depending on the definition) are widely available. There are so many variables involved in each of these, however, that simply dividing exonerations by the number of inmates on death row to calculate the rate of wrongful convictions would be of little use.  The benefit of this study is that it attempts to incorporate the numbers into a general model that approximates the actual workings of our legal system, and can therefore be used to provide a broadly applicable number. 

This requires plenty of assumptions, of course, all of which are fascinating and deserve posts in and of themselves. 

First, the authors do note the possibility that capital cases may perversely be more likely to produce wrongful convictions, since police and prosecutors are under more pressure to convict for sensationalist murders than minor felonies.  Unfortunately, without data, there is no way to test for or incorporate this hypothesis into the conclusion.

In addition, and more importantly, the authors acknowledge the relative frequency at which death sentences are commuted to life in prison. This practice has two effects on the estimate of wrongful incarceration.  First, the authors assume that the rate of innocence of those who receive commutations is slightly higher than those who do not-- the thinking goes that if a reviewing court has any doubt about a person's actual guilt, but not enough doubt to actually exonerate, the easiest solution is to void the sentence or commute it directly to life in prison.  Second, the authors assume that the heavy amount of litigation required for exoneration stays steady for those under a current threat of execution, but declines rapidly if the inmate does manage to successfully reduce his or her sentence. 

As a result, under this analysis, the authors assume that most exonerations are likely to happen for those currently under threat of death, even if a higher proportion of wrongfully convicted inmates will have their sentences commuted and subsequently be forgotten.

In my opinion, the most interesting assumptions-- and they are big ones-- weigh in favor of a more conservative estimate of the actually innocent.  First, the authors assume that of the 1320 defendants executed since 1977, none of them have been actually innocent.  The authors also assume a perfect success rate for exonerations of innocent defendants who have been on death row for at least 21 years.  In other words, the estimate of wrongful convictions is only as low as 4% insofar as we assume our post-conviction system has been 100% infallible at spotting and exonerating the innocent.  

Tomorrow's post will continue the analysis.