One in 25 Inmates Sentenced to Death Are Likely Innocent

The death sentence is the lie society tells itself about criminal justice. It's not punishment or mending or making the world more "right," so much as it is validation for every other act of justice or would-be justice a society engages in, from parking tickets to years in solitary confinement. To take the whole mess seriously we need to be able to say that our judgments are this correct, correct enough to tie someone to a bed and fill them with poison. Meanwhile, the death penalty fills us full of rightness and infallibility, and we go on meting justice in something like comfort.

The death sentence used to be the lie society told itself about justice, I should say. It used to be we could convict someone in a court of law, lock them away, and eventually do the deed. That person would be guilty because they were found guilty and that's just all there was. Some of us called the lie a lie, but the system could ignore those calls because, well, that's all there was. Now, there's more to guilt and innocence. DNA has invalidated the lie, exposed more of the innocent occupying death row and the innocent already beyond death row. It's a truly fucked up time for crime and punishment.

Despite the outing of numerous wrongful convictions, there's a stalwart crew still dedicated to the lie of rightful conviction, and the notion that putting people to death has any place in a world where guilt has become provisional and science's power to exonerate is only now becoming realized. Maybe statistics would help, a probable body count for justice gone awry. A study out today in the Proceedings of the National Academy of Sciences offers just such a number for "false conviction among death-sentenced criminal defendants in the United States."

The notion of rightful conviction persists in large part because of the idea that that it's not possible to know in the very first place what the rate of false conviction even is over the entirety of criminal incarceration. This is the "that's all there was" above, and it's summarized in the paper like this: "There is no systematic method to determine the accuracy of a criminal conviction; if there were, these errors would not occur in the first place." This statistic is what's known as a "dark figure;" it exists, but is unknowable. As such, the lie would not appear to be irrational, at least until we narrow the field to the death penalty by itself.

“We start with the problem: you have a type of error here that is not itself observed," said the study's lead author Samuel Gross, in an interview. "So how can we estimate how frequently wrongful conviction happens? What is the search process that is best at detecting these errors? The answer is that the most effective process—the most efficient machine—is the one that applies to defendants while they're under the threat of execution."

The death penalty is different from other sentences because it has a much higher exoneration rate than any other sort of conviction. In fact, while death penalty cases rank below one percent of all criminal cases, they make up 12 percent of the total exoneration rate. The disproportion is 130 to 1. According to the paper, the reason is simple: "far more attention and resources are devoted to death penalty cases than to other criminal prosecutions, before and after conviction." Death, after all, is the most serious business there is. All death sentences are reviewed on automatic appeal after trial, most more than once. Death row defendants are typically represented by lawyers all the way up to their eventual execution—or exoneration.

"The high exoneration rate for death sentences suggests that a substantial proportion of innocent defendants who are sentenced to death are ultimately exonerated, perhaps a majority," write the study's authors. "If so, we can use capital exoneration as a basis for estimating a lower bound for the false conviction rate among death sentences." False conviction rates have been estimated for death row inmates in the past—the two major studies finding rates between 2.3 and 3.3 percent—but they face an interesting limitation. Basically, a large number of death row cases that leave death row don't do so through exoneration; instead, they're renegotiated to life in prison, where exoneration rates fade back into the "background rate" for the criminal justice system as a whole.

Today's study is the first to attempt to fix that error. It does so by assuming that cases dropped from death row actually remain on death row with death row resources and death row scrutiny until either exoneration or execution. The statistical method used is what's known as survival analysis: what percentage of a given population will survive beyond a set period of time. "What we end up doing is estimating what the rate would be if everyone had the benefit of that very exacting process," said Gross.

The resulting figure, "a conservative estimate of the proportion of erroneous convictions of defendants sentenced to death in the United States from 1973 through 2004": 4.1 percent. That's probably a low number, but it extrapolates out to 50 innocent deaths (though the authors argue the actual death figure should be lower because of commutations to life sentences).

That's the disquieting flip side: those who weren't exonerated, executed, or still on death row, are often now (still) serving life sentences. " One of the conclusions of the study is that in addition to the 138 people who were sentenced to death since 1973 who have been exonerated, there are 200 who haven't been identified, and we still don't know who they are. We know that there are such people and we have a good estimate of the number," Gross said. In fact, the closer an inmate is to innocence, the more likely he or she is to be resentenced to life in prison.

The study has implications not only death row—where inmates receive the utmost scrutiny—but also across the criminal justice system. "What can we say, from this, about other cases? The answer is nothing directly." Gross said. "There are lots of reasons to believe that capital cases are different, even from other murders. You can't generalize from this to other crimes. But it does tell you something"

Inmates under the threat of execution receive the most scrutiny, and even if the other cases are different, the rate of wrongful conviction is likely similar, Gross said. "Here we can come up with a reliable estimate for one group and that suggests that the rate for other violent similar felonies is something similar. It does suggest that it's not hugely higher or hugely lower than this," he said. "Beyond that we can't say."

Judge Antonin Scalia famously gave his own estimate for wrongful capital convictions in 2007: .027 percent. Which brings us back to the lie we started with nicely enough, that of justice perfect enough to wield death. Judge Learned Hand, in 1923, put it like this (as quoted in the paper): "Our [criminal] procedure has always been haunted by the ghost of the innocent man convicted. It is an unreal dream.” Well, the ghost is here, looking at all of us.