The Tortured Statistical Debate Over the Death Penalty

A productive policy conversation after a botched Oklahoma execution depends on rigorously supported claims. Those can be tough to come by.

The botched execution of Clayton Lockett in Oklahoma a few weeks ago has prompted a new round of debate over the value and morality of capital punishment, and this is a very good thing. Candid talk about the death penalty does two important things. It provides the ignorant with the opportunity to discover facts they had not before considered. And it entices the learned to think anew, and aloud, about old patterns and practices.

In this second category, for example, my colleague Ta-Nehisi Coates chimed in recently to remind everyone of one immutable truth about capital punishment: "In America, the history of the criminal justice—and the death penalty—is utterly inseparable from white supremacy." Combine that with prosecutorial misconduct, and a meaningless right to counsel, and cowardly trial judges, and empty appellate review, you get the death-penalty regime we have.

Along this higher plane, away from the prejudice and bigotry and zeal over the torture of another human being, the conversation I am following most closely in the wake of Lockett's grim death is between Radley Balko at The Washington Post and John Lott, the academic who founded the right-wing Crime Prevention Research Center. It's instructive not just for what it says about America's capital punishment structure but for what it says about science and statistics as well.

The debate started two weeks ago, on May 1, right after Lockett, a convicted murderer, was tortured to death by Oklahoma officials. Matt Lewis wrote a reasonable piece at The Week laying out "The conservative case for capital punishment." He wrote: "I believe in second chances. I believe in reform and rehabilitation. But I also believe in evil."

Balko responded to Lewis's piece with a case for "Why conservatives should oppose the death penalty," in which he argued that conservative skepticism of government authority and competence—like the inability to execute someone in a manner that isn't torture—should never be more intense than in the context of capital punishment, when life is on the line.

I am a fan of both Lewis and Balko and I found the exchange smart, even refreshing. Lewis is right that there could be a role for the justice system's ultimate punishment. And Balko is right to point out that capital punishment in America today is meted out arbitrarily, and capriciously, often after dubious trials and appellate procedures, bad lawyering, and poor judging.

Here is where Lott enters the story. Last week, at National Review, he wrote: "There is overwhelming evidence that the death penalty deters murder and saves lives. Combine that with the fact that errors, few to begin with, are becoming ever less common, and objections to the death penalty are basically eviscerated."

I recoiled, not just because I don't believe for a second that there is "overwhelming evidence that the death penalty deters murder and saves lives" but also because I have spent the past five years chronicling the many ways in which legal and factual errors, along with intentional misconduct on the part of public officials, pervade death-penalty regimes around the nation.

Balko must have had the same reaction, because he responded a few days later. His piece is vintage Balko, a blend of sense and sensibility, and at its core is this line: "It’s when Lott starts citing statistics that his piece begins to fall apart." Balko asserted that Lott applied a "facile" way of interpreting statistics about the racial component behind capital punishment and that Lott unjustifiably downplayed the significance of exonerations in capital cases.

That, in turn, must have made Lott recoil. On Tuesday, Lott responded to Balko's column, reiterating at the end of his piece an old argument he has made many times before: "Most peer-reviewed studies by economists, as I show in my book Freedomnomics, find that each execution saves roughly 15 to 18 potential murder victims."

Following this conversation day-to-day, with its array of statistics and facts and studies and evidence, made my head spin. The debate can be boiled down, I think, to two primary components: 1) the racial disparity in capital punishment, and 2) the deterrent effect of the death penalty relative to the error rate of executing innocent people.

For help understanding (and answering) these questions, I asked John H. Blume, a professor at Cornell Law School who directs the school's respected Death Penalty Project. In essence, I asked who was right and why. Here's Blume's response:

Every credible study has found a statistically significant race of victim effect (i.e., more death sentences are imposed in white victim cases), and the effect is most pronounced in black defendant/white victim cases. I have written about this, as have many others, but the effect is real, and it does not disappear when other factors are variables are taken into account.  

[Lott] is right about whites being sentenced to death at a higher rate in the aggregate than African-Americans but this is a result of the same phenomena. Whites overwhelmingly kill whites, and black overwhelmingly kill blacks. So, since the overall death sentencing rate is higher in white victim cases, more death sentences in the aggregate will be imposed on white defendants.

But, the most death sentences (again the effect is statistically significant) is found in white victim/black defendant cases, and this effect is most pronounced in the Southern states that were part of the Old Confederacy (but is present in all states that retain the death penalty).

Blume also debunked Lott's claim about the deterrent effect:

As for the claim that each execution saves 15-18 potential murder victims, it is, as my recently deceased colleague and the grandfather of empirical legal studies Ted Eisenberg said after reading it, “preposterous.” There have been numerous studies over decades attempting to establish whether there is (or is not) any deterrent effect from sentencing persons to death and executing persons sentenced to death. 

No reliable study by credible researchers has ever found any deterrent effect. And the idea that you could pinpoint a relatively precise number of persons whose lives were saved by an individual execution is sophistry of the highest order.

Both Balko and Lott cited the work of Samuel Gross, a professor at the University of Michigan Law School, who just helped publish an important study estimating "that if all death-sentenced defendants remained under sentence of death indefinitely at least 4.1 percent would be exonerated. We conclude that this is a conservative estimate of the proportion of false conviction among death sentences in the United States."

I asked Gross essentially the same question I asked Blume. On deterrence, he told me, we have Lott, who argues that "innocent people’s lives are saved thanks to the death penalty," and then we have an April 2012 report from the National Academy of Science's National Research Council (my emphasis):

Research to date on the effect of capital punishment on homicide rates is not useful in determining whether the death penalty increases, decreases, or has no effect on these rates, says a new report from the National Research Council. The committee that wrote the report evaluated studies conducted since a four-year moratorium on the death penalty was lifted in 1976, and it found that the studies do not provide evidence for or against the proposition that the death penalty affects homicide rates. 

These studies should not be used to inform judgments about the effect of the death penalty on homicide, and should not serve as a basis for policy decisions about capital punishment, the committee said.

The lack of evidence about the deterrent effect of capital punishment—whether it is positive, negative, or zero—should not be construed as favoring one argument over another, the report stresses. "Fundamental flaws in the research we reviewed make it of no use in answering the question of whether the death penalty affects homicide rates," said Daniel S. Nagin, Teresa and H. John Heinz III University Professor of Public Policy and Statistics at Carnegie Mellon University Pittsburgh, and chair of the committee that wrote the report.

"We recognize that this conclusion may be controversial to some, but no one is well-served by unsupportable claims about the effect of the death penalty, regardless of whether the claim is that the death penalty deters homicides, has no effect on homicide rates or actually increases homicides."

Gross also disputed Lott's characterization of his new National Academies of Science paper on exonerations and the death penalty. "I wonder if he bothered to read it," Gross said in an email Tuesday. He continued:

Lott says: "Many times guilty people go free because evidence that was used at trial is later determined during an appeal to have been improperly obtained. The paper co-authored by Gross defined 'exoneration' as occurring when someone was removed from death row by a 'legal action by courts or executive officials.'"

This is flawed, Gross alleges, for the following reasons:

1) Many—most—death sentenced prisoners have their convictions, or more often their sentences, reversed for all sorts of technical reasons. But few are cleared and freed.

(2) We do NOT define an exoneration as a "removed from death row." That's just false. We say specifically and repeatedly that most death sentences inmates are removed from death row but not exonerated—in fact our methodology depends on that difference.

(3) We do define an exoneration for this study as a case in which someone who was sentenced to death has been completely cleared of the offense for which s/he was sentenced to death and freed (except for a few who had died). We claim that in those cases (as opposed to "was removed from death row") the defendants are overwhelmingly factually innocent. But we also have a section on sensitivity analyses in which we address the consequences of mis-classifications, and we conclude that we have produced a conservative estimate.

It is clear that the debate over the death penalty in America, which preceded the Declaration of Independence and the drafting of the Constitution, will continue long after Balko and Lott and Blume and Gross (and I) are gone. But no one now or later benefits from a debate based in whole or part on what the NAS calls "unsupportable claims about the death penalty."

I'll leave the fighting over stats to the academics. I know what I cover every day on this beat: The death penalty as currently implemented in America is both unconstitutional and immoral, because too many capital trials are unfair, too few appellate reviews are meaningful, and too few judges, prosecutors, witnesses, and politicians care enough to change that.