Overview

The Death Penalty Information Center has found that a siz­able por­tion of all death sen­tences that have been imposed since 1972 have been over­turned because of pros­e­cu­to­r­i­al mis­con­duct. Yet, many instances of mis­con­duct go unad­dressed and most pros­e­cu­tors who com­mit mis­con­duct are not held accountable.

DPIC cat­a­logued every known death sen­tence that was reversed because of pros­e­cu­to­r­i­al mis­con­duct and every exon­er­a­tion in which pros­e­cu­to­r­i­al mis­con­duct was present (here­inafter ​“mis­con­duct exon­er­a­tion”). Building upon DPIC’s past research and data from a vari­ety of oth­er sources, DPIC has iden­ti­fied more than 550 cap­i­tal pros­e­cu­to­r­i­al mis­con­duct rever­sals and mis­con­duct exon­er­a­tions. This means that more than 5.6% of all death sen­tences imposed since 1972 were reversed because of pros­e­cu­to­r­i­al mis­con­duct or result­ed in a misconduct exoneration.

The dataset high­lights the preva­lence of pros­e­cu­to­r­i­al mis­con­duct in exon­er­a­tion cas­es. More than 100 of the sen­tences in the dataset are exon­er­a­tions in which the wrong­ful con­vic­tions were affect­ed by pros­e­cu­to­r­i­al mis­con­duct. This rein­forces the find­ings in DPIC’s 2021 Special Report: The Innocence Epidemic, which found that 69% of death-row exon­er­a­tions have involved offi­cial mis­con­duct by police, pros­e­cu­tors, or oth­er government officials.

Scope

DPIC attempt­ed to iden­ti­fy every death sen­tence over­turned because of pros­e­cu­to­r­i­al mis­con­duct or that result­ed in a mis­con­duct exon­er­a­tion. Because death sen­tences are the rel­e­vant met­ric, some peo­ple who were vic­tims of ser­i­al mis­con­duct at mul­ti­ple stages of their cas­es appear more than once. For exam­ple, if a defen­dan­t’s ini­tial cap­i­tal con­vic­tion was reversed because of pros­e­cu­to­r­i­al mis­con­duct, and then was sen­tenced to death again in a retri­al or resen­tenc­ing pro­ceed­ing that was reversed because of new mis­con­duct, we count­ed that as two death sen­tence rever­sals. Most famous­ly, Curtis Flowers was tried six times by the same pros­e­cu­tor which result­ed in two mis­tri­als and four rever­sals on pros­e­cu­to­r­i­al mis­con­duct grounds. Flowers is includ­ed in the dataset four times — once for each pros­e­cu­to­r­i­al misconduct reversal.

The dataset is nar­row­ly focused, doc­u­ment­ing only cap­i­tal-case rever­sals and exon­er­a­tions, but not every instance in which pros­e­cu­to­r­i­al mis­con­duct occurred. This dataset includes cas­es in which the cap­i­tal con­vic­tion was reversed, lead­ing to a com­plete­ly new tri­al, and cas­es in which just the death sen­tence was reversed, lead­ing to a new sen­tenc­ing pro­ceed­ing. Cases in which mis­con­duct was found in both the guilt and penal­ty phas­es of a sin­gle tri­al are count­ed only once. Cases in which a court found pros­e­cu­to­r­i­al mis­con­duct but dis­missed it as ​“harm­less error” were not includ­ed in the dataset, nor were cas­es in which a defen­dant alleged pros­e­cu­to­r­i­al mis­con­duct but the court grant­ed relief on oth­er grounds with­out reach­ing the misconduct claims.

The nature of pros­e­cu­to­r­i­al mis­con­duct cas­es makes it dif­fi­cult to rank one coun­ty or state as worse than anoth­er. The num­ber of mis­con­duct rever­sals or exon­er­a­tions in a state or coun­ty depends on a com­bi­na­tion of fac­tors, includ­ing the preva­lence of mis­con­duct, how much mis­con­duct is dis­cov­ered, and how will­ing courts and oth­er deci­sion­mak­ers are to reverse tri­al out­comes or exon­er­ate indi­vid­u­als in cas­es involv­ing pros­e­cu­to­r­i­al mis­con­duct. For exam­ple, Philadelphia County, Pennsylvania has the most instances of pros­e­cu­to­r­i­al mis­con­duct rever­sals and exon­er­a­tions, but while this doc­u­ments a high rate of mis­con­duct in Philadelphia, it does not nec­es­sar­i­ly sig­ni­fy that Philadelphia pros­e­cu­tors engaged in more mis­con­duct than pros­e­cu­tors in any oth­er coun­ty. The large num­ber of doc­u­ment­ed mis­con­duct rever­sals could be a reflec­tion of the high qual­i­ty work of the insti­tu­tion­al cap­i­tal defense orga­ni­za­tions in Philadelphia in inves­ti­gat­ing and expos­ing mis­con­duct, the cre­ation of a con­vic­tion integri­ty unit in the Philadelphia District Attorney’s Office devot­ed to cor­rect­ing past injus­tices, the recep­tive­ness of coun­ty, state, and fed­er­al courts gov­ern­ing Philadelphia to address mis­con­duct claims, and the thor­ough­ness of DPIC’s infor­ma­tion regard­ing Philadelphia cap­i­tal tri­als, includ­ing access to unpub­lished tri­al court opin­ions. In short, the absence of doc­u­ment­ed cas­es of cap­i­tal con­vic­tions or death sen­tences reversed for pros­e­cu­to­r­i­al mis­con­duct does not equate to the absence of mis­con­duct in capital prosecutions. 

Methods

When col­lect­ing data for this study, we start­ed by search­ing for poten­tial cas­es that met our cri­te­ria for a pros­e­cu­to­r­i­al mis­con­duct rever­sal. We gath­ered data from var­i­ous datasets, legal resources, and stud­ies ana­lyz­ing pros­e­cu­to­r­i­al mis­con­duct. DPIC staff and interns con­duct­ed on-line research of legal data­bas­es to dis­cov­er court opin­ions in which cap­i­tal appel­lants alleged pros­e­cu­to­r­i­al mis­con­duct, as well as inde­pen­dent legal research to find addi­tion­al appel­late rever­sals for pros­e­cu­to­r­i­al mis­con­duct in high-use death penal­ty states. External infor­ma­tion sources includ­ed the Open File archive, the Broken System project at Columbia Law School, case lists main­tained by the Habeas Assistance and Training (HAT) Project, and data from researchers and lit­i­ga­tors across the country. 

Categorization

Despite eth­i­cal and con­sti­tu­tion­al oblig­a­tions to pur­sue jus­tice while pro­tect­ing con­sti­tu­tion­al rights, pros­e­cu­tors have engaged in a wide vari­ety of mis­con­duct in cap­i­tal cas­es. The study clas­si­fies mis­con­duct into sev­en preva­lent cat­e­gories and a catchall cat­e­go­ry of ​“oth­er” that encom­pass­es a num­ber of less com­mon types of mis­con­duct rever­sals. These cat­e­gories are: with­hold­ing favor­able evi­dence; jury dis­crim­i­na­tion; improp­er argu­ment; improp­er ques­tion­ing; false evi­dence; improp­er evi­dence; and vio­la­tions of the right to coun­sel. Some cas­es involved mul­ti­ple types of pros­e­cu­to­r­i­al mis­con­duct, so a sin­gle mis­con­duct rever­sal or exon­er­a­tion may be list­ed under sev­er­al different categories.

Withholding favor­able evi­dence: This cat­e­go­ry includes cas­es in which the gov­ern­ment improp­er­ly with­held evi­dence favor­able to the defen­dant. This includes vio­la­tions of the U.S. Supreme Court’s 1963 rul­ing in Brady v. Maryland. A Brady vio­la­tion occurs when­ev­er the gov­ern­ment with­holds evi­dence that is favor­able to the defense, but it results in a rever­sal only if a court deems the vio­la­tion to have been mate­r­i­al to guilt and/​or pun­ish­ment. Materiality is mea­sured by whether there is a rea­son­able prob­a­bil­i­ty that the out­come of the tri­al or penal­ty phase would have been dif­fer­ent if the evi­dence had been dis­closed to the defense. This cat­e­go­ry also includes cas­es in which the pros­e­cu­tion failed to cor­rect false evi­dence pre­sent­ed by state wit­ness­es. Joe D’Ambrosio​’s case pro­vides a notable exam­ple of a Brady vio­la­tion. D’Ambrosio was exon­er­at­ed in 2012 after spend­ing 23 years on Ohio’s death row. The pros­e­cu­tion admit­ted to with­hold­ing mul­ti­ple pieces of evi­dence includ­ing phys­i­cal evi­dence that could be test­ed, a motive for a pros­e­cu­tion wit­ness to have com­mit­ted the mur­der, and wit­ness state­ments about the date of the mur­der and who com­mit­ted it. A fed­er­al dis­trict court barred the state from retry­ing D’Ambrosio because the state action of “[hold­ing] D’Ambrosio on death row, despite wrong­ful­ly with­hold­ing evi­dence that ​‘would have sub­stan­tial­ly increased a rea­son­able juror’s doubt of D’Ambrosio’s guilt’” was of an ​“extra­or­di­nary” nature.

Jury Discrimination: This includes cas­es in which the pros­e­cu­tion excludes jurors on the basis of race, sex, or reli­gion in vio­la­tion of Batson v. Kentucky and its prog­e­ny. Timothy Foster​’s case illus­trates jury dis­crim­i­na­tion in death penal­ty tri­als. Foster was sen­tenced to death in Floyd County, Georgia in 1987. Foster is a Black man who was tried by an all-white jury. His con­vic­tion was over­turned by the U.S. Supreme Court in 2016 because pros­e­cu­tors improp­er­ly exer­cised their dis­cre­tionary jury strikes on the basis of race to exclude African American jurors. Foster obtained the pros­e­cu­tors’ jury selec­tion notes, which showed that pros­e­cu­tors had high­light­ed the names of each of the Black prospec­tive jurors in green on four dif­fer­ent copies of the jury list; cir­cled the word ​“BLACK” next to the ​“Race” ques­tion on the juror ques­tion­naires of five Black prospec­tive jurors; iden­ti­fied three Black prospec­tive jurors as ​“B#1,” ​“B#2,” and ​“B#3”; and ranked the Black prospec­tive jurors against one anoth­er in case ​“it comes down to hav­ing to pick one of the black jurors.” The Supreme Court point­ed out that, “[e]vidence that a prosecutor’s rea­sons for strik­ing a black prospec­tive juror apply equal­ly to an oth­er­wise sim­i­lar non­black prospec­tive juror who is allowed to serve tends to sug­gest pur­pose­ful dis­crim­i­na­tion.” The rea­sons giv­en by pros­e­cu­tors for strik­ing one Black juror, Marilyn Garrett, includ­ed her age and the fact that she was divorced, but they allowed three out of four divorced white jurors and white jurors of sim­i­lar age to Garrett to serve.

Improper Argument: This is a broad cat­e­go­ry that encom­pass­es sit­u­a­tions where the pros­e­cu­tion makes inflam­ma­to­ry, inap­pro­pri­ate, or uncon­sti­tu­tion­al state­ments dur­ing the guilt or sen­tenc­ing phase of the tri­al. A notable exam­ple of a death sen­tence being over­turned because of an improp­er argu­ment is the case of Dale Edward Flanagan in Clark County, Nevada. Flanagan’s death sen­tence was over­turned twice because the pros­e­cu­tor made prej­u­di­cial argu­ments. In 1988, the Supreme Court of Nevada high­light­ed six instances of improp­er argu­ments by the pros­e­cu­tor war­rant­i­ng rever­sal, includ­ing that the pros­e­cu­tor ​“weave[d] in an inap­pro­pri­ate appeal to pas­sion” when the pros­e­cu­tor stat­ed sweet things he remem­bered about his grand­moth­er before dis­cussing how Flanagan killed an elder­ly cou­ple. Flanagan was again sen­tenced to death, but his sen­tence was again over­turned because the pros­e­cu­tor argued that Flanagan’s occult beliefs and mem­ber­ship in a coven were exam­ples of his bad char­ac­ter. The Supreme Court of Nevada ruled that the prosecutor’s use of Flanagan’s occult beliefs vio­lat­ed his First Amendment right to Free Exercise.

Improper Questioning: Like Improper Argument, this cat­e­go­ry is also broad and encom­pass­es instances where the pros­e­cu­tion asks inflam­ma­to­ry, inap­pro­pri­ate, or uncon­sti­tu­tion­al ques­tions of a tes­ti­fy­ing wit­ness. Joseph Green​’s case pro­vides a notable exam­ple of improp­er ques­tion­ing. Green was sen­tenced to death in Bradford County, Florida in 1993. In 1996, the Supreme Court of Florida over­turned Green’s con­vic­tion due in part to the prosecutor’s improp­er ques­tion­ing regard­ing a wit­ness’ his­to­ry of alco­hol use. On direct exam­i­na­tion, the wit­ness tes­ti­fied that she had not been drink­ing when she wit­nessed the mur­der. On cross exam­i­na­tion, the pros­e­cu­tion inquired about her his­to­ry of alco­hol use and elicit­ed from the wit­ness that she was a recov­er­ing alco­holic and had been com­ing from an Alcoholics Anonymous meet­ing when she wit­nessed the mur­der. The Court held that evi­dence of wit­ness drug or alco­hol use for impeach­ment pur­pos­es is imper­mis­si­ble unless it is used to demon­strate that the wit­ness was under the influ­ence of drugs/​alcohol at the time they wit­nessed the event in ques­tion, or such evi­dence of pri­or use demon­strates a wit­ness’ inabil­i­ty to cred­i­bly tes­ti­fy about what they saw. As there was no show­ing that the wit­ness’ pri­or alco­hol use served either of these pur­pos­es, this line of ques­tion­ing was imper­mis­si­ble. Green was lat­er exon­er­at­ed when, on retri­al, the court found that the state’s star wit­ness was incom­pe­tent to tes­ti­fy. This dis­missal of tes­ti­mo­ny, cou­pled with the fact that the pros­e­cu­tion could not present any phys­i­cal evi­dence link­ing Green to the crime, led the cir­cuit court judge who had orig­i­nal­ly sen­tenced Green to die to dis­miss all charges, say­ing that there was no evi­dence tying Green to the murder.

Improper Evidence: This cat­e­go­ry involves instances in which the pros­e­cu­tion presents inad­mis­si­ble evi­dence and/​or expert evi­dence not sup­port­ed by sci­en­tif­ic rig­or, such as the now-dis­cred­it­ed use of bite mark evi­dence. Robert DuBoise​’s case pro­vides a notable exam­ple of the use of improp­er evi­dence. DuBoise was exon­er­at­ed in 2020 after being in prison for 37 years for a Florida mur­der and rape con­vic­tion that was based on junk-sci­ence bite-mark evi­dence, and false jail­house wit­ness tes­ti­mo­ny. The foren­sic odon­tol­o­gist who tes­ti­fied at tri­al, Dr. Souviron, had pre­vi­ous­ly said, ​“If you tell me that is the guy that did it, I will go into court and say that is the guy that did it,” at the International Association of Chiefs of Police. The ​“bite mark” that led to DuBoise’s con­vic­tion was not even a bite mark at all. Further, the police alleged­ly con­spired with the jail­house infor­mant to pro­duce false tes­ti­mo­ny at trial.

False Evidence: This cat­e­go­ry includes cas­es in which, dur­ing tri­al, the pros­e­cu­tor presents or fails to cor­rect evi­dence they know to be false. A notable exam­ple of false evi­dence being pre­sent­ed by pros­e­cu­tors is the case of Kenneth Simmons who was sen­tenced to death in Richland County, South Carolina in 1996. Simmons was grant­ed a new tri­al in 2017 because pros­e­cu­tors had pre­sent­ed false DNA tes­ti­mo­ny that ​“severe­ly deprived” Simmons of his due process rights. During the tri­al, the pros­e­cu­tion called an expert wit­ness that offered false and mis­lead­ing DNA tes­ti­mo­ny that pur­port­ed to link Simmons to the mur­der. The judge that grant­ed Simmons a new tri­al said that the prosecution’s ​“mis­rep­re­sen­ta­tion of the strength of the DNA evi­dence to the jury” was ​“over­whelm­ing.” The judge also wrote that the pros­e­cu­tion had pre­sent­ed the jury with ​“con­fus­ing, mis­lead­ing, and inac­cu­rate” infor­ma­tion about the DNA evi­dence, includ­ing show­ing the jury a chart that con­tained fab­ri­cat­ed DNA results, using the chart to make addi­tion­al incor­rect claims about the DNA evi­dence dur­ing clos­ing argu­ments, and false­ly argu­ing that Simmons was the only pos­si­ble source of the DNA.

Violation of Right to Counsel: This includes cas­es in which the pros­e­cu­tion intro­duces evi­dence that the pros­e­cu­tion helped elic­it in vio­la­tion of the defendant’s right to coun­sel or that the pros­e­cu­tor knew was gained in vio­la­tion of the defendant’s right to coun­sel. For exam­ple, John Satterwhite’s Texas death sen­tence was over­turned by the United States Supreme Court in 1988 because Satterwhite’s Sixth Amendment right to coun­sel was vio­lat­ed. The pros­e­cu­tors did not inform Satterwhite’s attor­neys that they were hav­ing a psy­chi­a­trist eval­u­ate Satterwhite for com­pe­ten­cy to stand tri­al, san­i­ty at the time of the crime, and future dan­ger­ous­ness. The pros­e­cu­tor then called the psy­chi­a­trist to tes­ti­fy dur­ing the penal­ty phase hear­ing. The Supreme Court ruled that the psy­chi­a­trist was a key wit­ness for deter­min­ing future dan­ger­ous­ness as the only licensed physi­cian to tes­ti­fy, and the pros­e­cu­tor placed sig­nif­i­cant weight on his tes­ti­mo­ny dur­ing clos­ing argu­ments to the jury. Consequently, the Supreme Court found it impos­si­ble to say beyond a rea­son­able doubt that the psychiatrist’s tes­ti­mo­ny on future dan­ger­ous­ness did not influ­ence the jury.

Other: This cat­e­go­ry includes a vari­ety of cas­es such as dou­ble jeop­ardy vio­la­tions, con­flicts of inter­est, pros­e­cu­to­r­i­al involve­ment in coerced con­fes­sions, and oth­er cas­es where the rever­sal is due to improp­er acts of the pros­e­cu­tor. Clinton Young​’s case is an exam­ple of a rever­sal due to a con­flict of inter­est. Young’s con­vic­tion was vacat­ed in 2021 because the pros­e­cu­tor on his case also was work­ing as a law clerk for the judge pre­sid­ing over the case. Assistant District Attorney Ralph Petty was one of the pros­e­cu­tors on Young’s case, while also work­ing as a clerk, con­duct­ing legal research, and mak­ing rec­om­men­da­tions to the pre­sid­ing judge. The Texas Court of Criminal Appeals ruled this was a clear con­flict of inter­est. Petty sur­ren­dered his law license and, in an April 13, 2021, order pro­hibit­ing Petty from prac­tic­ing law in Texas, the Texas Supreme Court said the ter­mi­na­tion of Petty’s license was ​“in the best inter­est of the pub­lic [and] the profession.”