"Long before it's in the papers"
April 28, 2014


At least one in 25 death-sentenced people are innocent, study claims

April 28, 2014
Special to World Science  

A new study es­ti­mates that at least one in 25 peo­ple sen­tenced to death in the Un­ited States are in­no­cent.

Re­searcher Sam­u­el Gross of Uni­vers­ity of Mich­i­gan Law School and col­leagues pub­lished the anal­y­sis in this week’s early on­line edi­tion of the re­search jour­nal Pro­ceed­ings of the Na­tio­n­al Aca­de­my of Sci­en­ces.

Some re­search­ers have held that the rate of false con­vic­tions is un­know­a­ble since, if there were some sys­tem­atic way to de­termine con­vic­tion accuracy, errors wouldn’t occur in the first place. In the absence of hard data, some have claimed the false-con­viction rate is al­most zero.

Gross and col­leagues fo­cused on death sen­tences, which have a far high­er ex­oner­a­t­ion rate than oth­er sen­tences. “A high proportion of false con­victions that do come to light and produce exon­er­a­tions are con­cen­trated among the tiny min­ority of cases in which de­fend­ants are sentenced to death,” they wrote. “This makes it poss­ible to use data on death row exon­er­ations to esti­mate the over­all rate of false con­viction among death sen­tences.”

Not­ing that most death-sen­tenced de­fen­dants are re-sen­tenced to life in pris­on and don’t stay on death row, the re­search­ers an­a­lyzed ex­oner­a­t­ions from 1973, when the Un­ited States es­tab­lished the death pen­al­ty in its mod­ern form, to 2004.

Their da­ta came from the De­part­ment of Jus­tice and the Death Pen­al­ty In­form­a­t­ion Cen­ter, a non­prof­it that tracks ex­oner­a­t­ions among cap­i­tal de­fen­dants.

In terms of actual exon­erations, “2.3% of all death sen­tences im­posed from 1973 through 1989 resulted in exon­er­ation by the end of 2004,” the re­search­ers wrote. But the exon­er­ation rate be­comes much lower for de­fend­ants re-sen­tenced to life in pri­son, they noted. Us­ing “sur­vival anal­y­sis,” a tech­nique from ep­i­de­mi­ology, they es­ti­mat­ed that if all death-sen­tenced de­fen­dants had stayed un­der ex­e­cu­tion threat for 21 years, at least 4.1 per­cent would have been ex­onerated.

Of course, they added, the study must con­tain er­rors—such as by not ac­count­ing for the exon­erations, un­known in num­ber, that must also be erron­eous. But Gross and col­leagues claimed their work makes “con­serv­ative as­sump­tions” over­all.

* * *

Send us a comment on this story, or send it to a friend

Sign up for


On Home Page         


  • W­eird oc­ean sound fi­nally ex­plained: a whale

  • Peo­ple found to re­member atro­cities in ways that fa­vor their group


  • Was black­mail essen­tial for marr­iage to evolve?

  • Plu­to has even cold­er “twin” of sim­ilar size, studies find

  • Could simple an­ger have taught people to coop­erate?

  • Diff­erent cul­tures’ mu­sic matches their spe­ech styles, study finds


  • F­rog said to de­scribe its home through song

  • Even r­ats will lend a help­ing paw: study

  • D­rug may undo aging-assoc­iated brain changes in ani­mals

A new study estimates that at least one in 25 people sentenced to death in the Un ited States are innocent. Resarcher Samuel Gross of Uni vers ity of Michigan Law School and colleagues published the analysis in this week’s ear ly online edition of the research journal PNAS. Some researchers have held that the rate of false convictions is un knowable since such errors often remain un noticed forever. Gross and colleagues focused on death sentences, which are associated with a far higher exoner ation rate than other sentences. Noting that most death-sentenced defendants are resentenced to life in prison and don’t stay on death row, the researchers analyzed exoner ations from 1973, when the modern form of the death penalty was established in the Un ited States, to 2004. Their data came from the Department of Justice and the Death Penalty Inform ation Center, a nonprofit that tracks exoner ations among capital defendants. The authors found that the high exoner ation rate applies on ly while the defendants remain un der threat of execution. Using “survival analysis,” a technique in epidemiology, they estimated that if all death-sentenced defendants were to stay un der execution threat indefinitely, at least 4.1% would be exonerated.