Innocents in Prison

Many thousands of wrongly convicted people are rotting in prisons and jails around the country.

As recently as 20 years ago, it was extraordinarily rare for a convicted prisoner to establish his or her innocence conclusively enough to get public attention. That changed with breakthroughs in DNA science.

The 205th DNA exoneration since 1989 was recorded earlier this month by the Innocence Project, a group of crack defense lawyers who have made such cases their mission. The exonerated prisoners—including 15 who had been sentenced to death—have been found innocent by courts, prosecutors, or governors based on post-conviction DNA testing.

But America has been too slow to appreciate that the DNA exonerations, and other evidence, suggest that many thousands of other wrongly convicted people are rotting in prisons and jails around the country. And our federal, state, and local governments and courts have done far too little to adopt proposed criminal justice reforms that could reduce the number of innocent people convicted while nailing more of the real criminals.

The case of the most recent DNA exonoree, Byron Halsey, was typical: Based on a confession full of obviously false details, extracted by high-pressure interrogation, he spent 19 years in prison in New Jersey for two heinous child murders committed by another man in 1985. Halsey was able to prove his innocence only after a 2002 New Jersey law forced reluctant prosecutors to give his counsel access to DNA evidence. In Halsey's and some 70 other DNA-exoneration cases, DNA also helped to establish the guilt of the real perpetrators. All or almost all had committed other violent crimes before being caught.

The kind of DNA evidence that can conclusively prove innocence or guilt is available only in a small fraction of cases, mainly rapes and rape-murders in which sperm is recovered. But these cases are the tip of the iceberg. Extrapolation from DNA and other exonerations suggests that the overall error rate in violent-crime prosecutions is probably much higher than the estimates of less than 1 percent offered by many criminal-justice officials.

In a stunning non-DNA case that shows how low law enforcement can go, a federal District judge in Boston ordered the government on July 26 to pay $102 million in damages to victims of the FBI's "outrageous" role in framing four men for a 1965 gangland murder. The four spent a combined total of 109 years in prison. Two died there.

"FBI officials up the line allowed their employees to break laws, violate rules, and ruin lives," wrote the judge, Nancy Gertner. She found that the FBI had known all along that a mob hit man named Joseph Barboza was lying when he claimed he saw the four men kill a mobster named Edward Deegan. Barboza and the FBI protected the real killer, Vincent Flemmi, because Barboza and Flemmi were FBI informants.

Although that was decades ago, the FBI (like most police agencies) still insists on a policy—its virtual prohibition of tape-recording interviews—that allows agents to conceal or distort what suspects and witnesses say. The usual motive is not to frame innocent people but to fit evidence to preconceived (and often mistaken) theories of the case.

The FBI has defended its policy on the grounds that taping might deter suspects from talking and that jurors might be offended to hear police using lawful but deceptive interrogation tricks such as lying to suspects and feigning sympathy. But the first concern seems far-fetched in most cases and the second is illegitimate. Refusing to tape interviews hides important evidence from juries as well as lending itself to abuse.

Police and/or prosecutorial misconduct appears to figure in more than half of the 205 convictions that DNA has proven false. Specific reasons for those convictions are catalogued in "Judging Innocence," a study by Brandon Garrett, a law professor at the University of Virginia, slated for publication in January in the Columbia Law Review. Mistaken eyewitness identifications—often due to police subtly pointing witnesses toward the people the cops suspect—figured in 79 percent of these false convictions. Flawed or corrupt testimony by scientific "experts" (about hair, blood types, and the like) figured in 55 percent. False confessions, mostly by juvenile defendants, figured in 16 percent.

What percentage of the 2.2 million men and women locked up in state and federal prisons and local jails are actually innocent of the charged crimes? This is not knowable with any confidence. But we can extrapolate from a few data points.

Samuel Gross, a University of Michigan law professor, has calculated that 2.3 percent of all prisoners sentenced to death between 1973 and 1989 have been exonerated and freed. His research suggests that the vast majority in fact did not commit the crimes. And an unknown number of innocents have not been exonerated.

Does this suggest that we may have 50,000 innocent prisoners (2.3 percent times 2.2 million) in this country? Possibly, but not necessarily. Gross and Barbara O'Brien of Michigan State Law School stress in another forthcoming article that capital cases (and the DNA exonerations) are not representative of false convictions in general.

On the one hand, for example, the percentage of actual innocents may be higher among capital murder defendants—virtually all of whom go to trial—than among the 90 percent of violent-crime defendants whose cases end in plea bargains.

On the other hand, a much higher percentage of, say, robbery convictions than of capital murder convictions are based on unreliable eyewitness identifications. So perhaps the robbery error rate is much higher than 2.3 percent. And it's clear that most people wrongly convicted of lesser crimes are far less likely to win exoneration than are death-row inmates (who get much better post-conviction legal representation) and rape convicts (when DNA evidence is available).

Gross and others have also found that a very disproportionate number of exonerated prisoners are black or Hispanic. Discrimination no doubt accounts for some of this. So does the especially high error rate in cross-racial identifications, which creates special risks for black men wrongly accused of raping or robbing white people.

But well-off white men are not exempt from wrongful prosecution. This was spectacularly illustrated by the fabricated rape charges against three innocent Duke lacrosse players. Durham District Attorney Mike Nifong falsely demonized them as rapists, racists, and "hooligans," thereby gaining enough black votes to win what had been an uphill election battle. Nifong (who is white) also rigged a photo-identification process to frame the three for a nonexistent crime, hid DNA proof of innocence, and lied to the public and the court for many months before North Carolina Attorney General Roy Cooper took over the case and declared the defendants innocent.

Nifong has been disbarred and still faces a contempt-of-court charge. But his richly deserved fate is almost unheard-of. Most state bars and judges have given passes even to prosecutors who have hidden or falsified evidence to put innocent men and women on death row.

The DNA exonerations have spurred reforms by some states, but most still use shockingly unreliable police practices.

The good news, according to the Innocence Project, is that 42 states and the District of Columbia now require that inmates be given access to any DNA evidence; 22 (plus D.C. and the federal government) require preservation of DNA evidence; and 22 (plus D.C. and the federal government) compensate victims of false convictions.

But only three states (and some cities) have adopted the reforms to eyewitness procedures—such as excluding officers who know which person is the suspect, among other safeguards—that experts say are necessary to prevent false identifications. And only the District of Columbia, nine states, and some 500 localities keep police honest by requiring them to tape interrogations.

The courts, meanwhile, have "performed miserably in ferreting out the innocent" convicts, as Adam Liptak observed in a July 23 New York Times column. The Garrett study shows that the Supreme Court refused to hear the appeals of 30 of the first 200 prisoners who were later exonerated and rejected the one appeal it did hear.

The justices, and Congress, have also made it difficult for state prisoners to seek relief in lower federal courts. Most recently, on June 15 the conservative high court majority slammed the door on a murder convict because his lawyer had filed his appeal three days late—on the mistaken advice of a federal District judge. This decision seems perverse, especially in an era of DNA exonerations when, as a Justice Department commission observed in 1999, "the strong presumption that verdicts are correct, one of the underpinnings of restrictions on post-conviction relief, has been weakened."

Indeed, so formidable are the procedural obstacles facing falsely convicted innocents that only 10 percent of the 205 DNA exonorees made their innocence the basis of their initial appeals—and none was successful. This reflects the appeals courts' almost exclusive focus on whether the prosecutor or judge made procedural errors and almost complete deference to findings of guilt by juries and trial judges. They should defer less.