A report by University of Michigan staffers and law students _ Exonerations in the United States: 1989 through 2003 _ was released to the public on April 23, 2004. The report analyzed data from 328 cases during that 15 year period in which the defendant was officially declared, "not guilty of a crime for which he or she had previously been convicted."

The report concentrated on rape and murder convictions, since 319 of the 328 cases studied involved a defendant convicted of one or both those crimes. One of two areas the researchers focused on, was how often several factors known to contribute to a wrongful conviction - eyewitness misidentification, perjury and a false confession - were present in those cases. It was found that 64% of the people exonerated of rape and/or murder had been misidentified, 15% had falsely confessed, and a prosecution witness had committed perjury in 44% of the cases.

The other area reported on is how race relates to exonerations. It was found that people of various races are exonerated at about the same rate as they are convicted _ unless the person was under 18 at the time of arrest. Almost eight times as many blacks wrongly convicted of a crime as a juvenile are exonerated than are whites. That finding is consistent with the phenomena of erroneous cross-racial identifications, since many of those cases involved the mistaken ID of a black person by a white.

So in general the report's findings tend to be a fine tuning of what is already known about wrongful convictions. However there is one area where the report engages in speculation, and that is about how many non-death row exonerations there would have been if all cases "were reviewed with the same level of care that we devote to death sentences." That figure is estimated in the report to be 28,642 cases, which it refers to as "a shocking prospect." While appearing numerically impressive, it is put into perspective by considering that based on Bureau of Justice Statistics data, it is estimated that from 1989 to 2003 there were 14,295,000 felony convictions in state and federal courts. So the report's estimate that there should have been 28,642 exonerations from 1989 to 2003 amounts to 2/10th of 1%, or .2% of the felony convictions during those 15 years. Furthermore, the report's estimate amounts to a projected average of 1,909 exonerations a year, compared with 953,000 convictions yearly _ or 1 out of 500.

The Supreme Court inferred in Schlup v. Delo, 115 S. Ct. 115 (1995) that the legal system may only need to ascertain guilt to an accuracy rate of 99% (99 out of 100). Thus the reports contention that 499 out of 500 convictions are of a guilty person provides powerful support to the contention that while not 100% accurate, the United States has a discerning legal system. Hence if the report's estimate of wrongful convictions is to be believed, then this countries legal system is neither broken nor in need of significant reforms. After all, the average new car has an average of over one significant problem within 90 days of its purchase, so with a correct conviction rate of 99.8%, car manufacturers ought to be consulting with judges and prosecutors on how to manufacture a more reliable product. Consequently the report plays directly into the hands of prosecutors, judges, police and corrections officials who contend the system works remarkably well at weeding out the innocent from the guilty.

However that assessment stands in stark contrast with the much different conclusion that can be drawn from the findings of a much more extensive study of criminal convictions published in June 2000. That study - A Broken System: Error Rates in Capital Cases _ examined every capital cases finalized from 1973 to 1995. It found that 68% of those 4,578 cases were reversed on appeal due to prejudicial error; that "7% of capital cases nationwide are reversed because the condemned person was found to be innocent"; and that on retrial, the defendant was given a lesser sentence in 82% of those reversed cases. So based on the findings of that extensive multi-year study that was overseen by the esteemed Professor James Liebman (co-author of Federal Habeas Corpus Practice and Procedure), if every one of the 14,295,000 criminal conviction in this country from 1989 through 2003 had been subjected to the same degree of appellate review as is a capital case, then it can be estimated that 9,720,600 of those convictions (68%) would have been reversed, with the result that 680,442 of the defendants (7%) would have been exonerated, and 7,970,892 of the defendants (82%) would have been re-sentenced to a lesser punishment. An important indication that although they may seem high at first glance, those estimates are actually conservative, is that since the study only included cases through 1995, the number of exonerations it reported wasn't skewed upward by the many defendants convicted of capital crimes who have been cleared by DNA in the past nine years.

Consequently, unlike the report of April 2004, the findings reported in A Broken System (and its follow-up report, A Broken System, Part II, February 2002) are consistent with the proposition that there are serious systemic errors in the ability of this country's trial courts to accurately distinguish the innocent from the guilty. In fact the report concluded that in regards to capital cases, the "system is collapsing under the weight of that error, and the risk of executing the innocent is high." The problems are no less acute in cases that don't involve as severe of a penalty.

So while the analysis in the April 2004 report of various factors that can contribute to a wrongful conviction is valuable, there are sound reasons to take with a grain of salt, its downplaying of how often they occur.

The 37 page report, Exonerations in the United States: 1989 through 2003, can be purchased from PLN for the $5 cost of printing, handling and mailing. The 175 page report, A Broken System: Error Rates in Capital Cases (June 2000) can be purchased from PLN for the $7 cost of printing, handling and mailing. Or both reports can be ordered for $10. Stamps OK.

Subscribe today

Already a subscriber? Login