In April, Jerry Miller, an Illinois man who served 24 years for a rape he did not commit, became the 200th American prisoner cleared by DNA evidence. His case, like the 199 others, represented a catastrophic failure of the criminal justice system.
When an airplane crashes, investigators pore over the wreckage to discover what went wrong and to learn from the experience. The justice system has not done anything similar.
But a new study does. Brandon L. Garrett, a law professor at the University of Virginia, has, for the first time, systematically examined the 200 cases, in which innocent people served an average of 12 years in prison. In each case, of course, the evidence used to convict them was at least flawed and often false — yet juries, trial judges and appellate courts failed to notice.
"A few types of unreliable trial evidence predictably supported wrongful convictions," Professor Garrett concluded in his study, "Judging Innocence," to be published in The Columbia Law Review in January.
The leading cause of the wrongful convictions was erroneous identification by eyewitnesses, which occurred 79 percent of the time. In a quarter of the cases, such testimony was the only direct evidence against the defendant.
Faulty forensic evidence was next, present in 55 percent of the cases. In some of those cases, courts put undue weight on evidence with limited value, as when a defendant's blood type matched evidence from the crime scene. In others, prosecution experts exaggerated, made honest mistakes or committed outright fraud.
Most of the forensic evidence involved problems with the analysis of blood or semen. Forty-two cases featured expert testimony about hair, an area that is, Professor Garrett wrote, "notoriously unreliable."
Informants testified against the defendants in 18 percent of the cases. (In three cases, it turned out they had an unusually powerful motive for their false testimony, as DNA evidence proved they were in fact guilty of the crime they had pinned on the defendant.)
There were false confessions in 16 percent of the cases, with two-thirds of those involving defendants who were juveniles, mentally retarded or both.
The 200 cases examined in the study are a distinctive subset of criminal cases. More than 90 percent of those exonerated by DNA were convicted of rape, or of both rape and murder, rape being the classic crime in which DNA can categorically prove innocence. For other crimes, there is often no biological evidence or, if there is, it can give only circumstantial hints about guilt or innocence.
Only 14 of those exonerated had been sentenced to death, 13 in rape-murders. There is a widespread misconception that DNA evidence has freed many inmates from death row, but it is actually a rare murder not involving rape in which biological evidence can provide categorical proof of innocence.
"DNA testing is available in fewer than 10 percent of violent crimes," said Peter Neufeld, a founder of the Innocence Project at Cardozo Law School, which was instrumental in securing many exonerations. "But the same causes of wrongful convictions exist in cases with DNA evidence as in those cases that don't."
Professor Garrett's study strongly suggests, then, that there are thousands of people serving long sentences for crimes they did not commit but who have no hope that DNA can clear them.
In a second forthcoming study of false convictions, this one focused on capital cases, two law professors — Samuel R. Gross of the University of Michigan and Barbara O'Brien of Michigan State — cautioned that "exonerations are highly unrepresentative of wrongful convictions in general."
"The main thing we can safely conclude from exonerations is that there are many other false convictions that we have not discovered," the Michigan study said. "In addition, a couple of strong demographic patterns appear to be reliable: black men accused of raping white women face a greater risk of false conviction than other rape defendants; and young suspects, those under 18, are at greater risk of false confession than other suspects." Professor Garrett also found that exonerated convicts were more apt to be members of minority groups than was the prison population generally. For instance, 73 percent of the convicts cleared of rape charges were black or Hispanic, compared with 37 percent of all rape convicts.
The courts performed miserably in ferreting out the innocent. Thirty-one of the 200 exonerated prisoners, for instance, had appealed to the United States Supreme Court, but the justices refused to hear 30 of the cases. In the one case they did hear, they ruled against the inmate. Of course, appeals courts do not typically reconsider a jury's factual findings, focusing instead on asserted procedural errors. Only 20 of the 200 even appealed on the ground that they were innocent; none of those claims were granted.
Perhaps the most troubling finding in Professor Garrett's study was how reluctant the criminal justice system was to allow DNA testing in the first place. Prosecutors often opposed it, and 16 courts initially denied requests for testing.
Yet DNA evidence can do more than free the innocent. In many cases, it also identified the person who actually committed the crime.
In 40 percent of the cases handled by the Innocence Project, Mr. Neufeld said, DNA not only exonerated the innocent prisoner but also provided evidence that helped identify the person who committed the crime. "In every single one of those cases that perpetrator had committed violent crimes in the intervening years," he said.
The era of DNA exonerations should be a finite one. These days, DNA testing is common on the front end of prosecutions, meaning that in a few years, the window that the 200 exonerations has opened on the justice system will close. We should look carefully through that window while we can.