Skip navigation
The Habeas Citebook: Prosecutorial Misconduct - Header
× You have 2 more free articles available this month. Subscribe today.

Deliberately Convicting the Innocent: Exonerations Expose the Criminal Justice System’s Callous Indifference Toward Official Misconduct

Decades later, Morton’s attorneys would discover that district attorney Ken Anderson had failed to disclose that a few days after the murder—when Morton was already in custody—someone attempted to use a credit card belonging to Christine and cashed a $20 check missing from her purse. Anderson had also failed to disclose that the Mortons’ toddler, Eric, had witnessed the murder and told his grandmother that “a monster” had killed his Mommy when “Daddy was not there.” 

But perhaps most significant was a bloody bandana that had been found near the Mortons’ home. In 2005, Morton’s attorneys began seeking DNA testing of the bandana. But the new district attorney, John Bradley, fought the request, promising that it would fail to identify any “mystery killer.” Five years later in 2010, when testing was finally ordered, that is exactly what happened: genetic material was found on the bandana from Christine Morton and another man who was not her husband.

The next year, a DNA match was made to Mark Alan Norwood, a 57-year-old convicted felon who was also tied to the 1988 bludgeoning murder of another young Austin-area mother in her bed, 34-year-old Debra Baker. A few months following the DNA match, after serving more than 24 years in prison, Morton was released on October 4, 2011, and exonerated eight days later. He was 56 years old.

Norwood was convicted of murdering Christine Morton in 2013 and Debra Baker in 2016. He received a life sentence and could be eligible for parole in 2031.

In 2013, former D.A. Anderson, who by then had become a state judge, was arrested and charged with evidence tampering. He avoided that charge by pleading guilty to contempt of court, resigning his job and surrendering his law license. He spent 96 hours in jail. Yet he is also the only prosecutor ever incarcerated for misconduct leading to a wrongful conviction.

Michael Morton is but one of the nearly 1,300 factually innocent people who are known to have been convicted due to the deliberate deception of government officials. In a report titled “Government Misconduct and Convicting the Innocent: The Role of Prosecutors, Police and Other Law Enforcement” (2020) (hereinafter referred to as the “Report”), authors Samuel R. Gross, Maurice J. Possley, Kaitlin Jackson Roll, and Klara Huber Stephens examined the first 2,400 postings to the National Registry of Exonerations (“Registry”).

The postings examined include exonerations from 1989 to February 27, 2019. This article provides a brief review of some of their general findings. It will also review the authors’ classification of the five primary types of misconduct as well as examine the limitations of the Report. And it will consider the authors’ suggested reforms to reduce or prevent future misconduct. Finally, it will address the abject failure in holding government offenders accountable.

This undertaking requires explanation of a couple of terms used in the Report. “Official misconduct” was defined by the authors to occur when “a prosecutor, police officer ... forensic analyst or child welfare worker violated an official duty in the investigation or prosecution of a criminal case, and the violation contributed to the conviction of a defendant who was later exonerated” and the “misconduct in investigations and prosecutions ... produced unreliable, misleading or false evidence of guilt, or ... conceal[ed], distort[ed], or undercut true evidence of innocence.”

To qualify as an exoneration, it had to be reported as such in the Registry. “Exoneration” was defined as occurring when a convicted defendant, following a post-conviction reexamination of the evidence against him or her, was either: (1) declared factually innocent by a government official/agency with authority to make that declaration or (2) relieved of all consequences of the criminal conviction by an authorized government official/agency. In the latter circumstance, that action had to be the result, at least in part, of evidence of innocence that either was not presented at the defendant’s trial or was not known to the defense at the time the defendant pleaded guilty.

General Findings

Seventy percent of official misconduct the Report found was committed by a combination of prosecutors (30%), law enforcement officers (35%), forensic scientists/analysts (3%), and child welfare workers (2%). The Report didn’t include misconduct by judges or by defense attorneys (although many instances of official misconduct came to light as a result of pursuing claims of ineffective assistance of counsel).

Official misconduct occurred in 1,296 (54%) of the 2,400 exonerations examined. The overwhelming majority of detected misconduct occurred in prosecutions of violent crimes. This is due, at least in part, because most exonerations (78%) were for violent felonies (including both those with misconduct and those without misconduct.) The following table shows the number of exonerations for each type of violent crime where official misconduct occurred and the rate of occurrence of misconduct for all exonerations of that crime:

The table reveals that the more serious offenses have higher rates of misconduct. Most likely, this is due to two factors.

First, there is increased pressure on public officials to solve serious crimes (a community demands to a greater extent to know who committed a murder compared to knowing who broke a few windows in parked cars). That increased pressure may lead to a rushed effort to solve the crime that is highly susceptible to misconduct.

Second, conviction for a serious crime has a greater chance of a more thorough post-conviction investigation that will uncover misconduct. For example, the amount of time and other resources spent on post-conviction investigations of murder convictions (especially those resulting in a death sentence) simply aren’t spent on convictions for lesser offenses that resulted in a few years in prison or probation.

Examining the Five General Categories of Official Misconduct

Almost all of the official misconduct identified fell into five general categories: (1) witness tampering, (2) misconduct in interrogations, (3) fabricating evidence, (4) concealing exculpatory evidence, and (5) misconduct at trial.

Witness Tampering

Harold Franks, 59, was shot and killed in a 1975 robbery in Cleveland, Ohio, that also left bystander Ann Robinson, 58, seriously wounded. Police quickly identified two suspects, Paul Gardenshire,16, and Ishmael Hixon, 23. However, police suspended investigating them when 12-year-old Eddie Vernon came forward to say he had seen the shooter, whom he identified as Ricky Jackson, 18, flee the scene in a car with brothers Ronnie and Wiley Bridgeman, who were 17 and 20, respectively.

No physical evidence tied the men to the crimes, and none had a criminal record. In fact, the car seen speeding away from the seen was actually traced to Hixon. Nevertheless, they were convicted and sentenced to death—later commuted to life in prison—based almost entirely on Vernon’s eyewitness testimony.

After a journalist investigating the story tracked him down, Vernon recanted, eventually testifying at a 2014 hearing that he had not witnessed the crime. He said he had heard a rumor that the three men were involved, so he invented the lie because he wanted to help the police. Vernon further testified that he tried to recant before the trials, but police threatened to arrest his parents for perjury if he changed his story. The police never revealed any of this to the court or the defendants.

At that same hearing, evidence was finally presented that Hixon, one of the original suspects, pleaded guilty in 1976 to more than a dozen aggravated robberies. Gardenshire, the other original suspect, had been convicted of the same crime in 1978.

Later in 2014, the three men sent to prison for Franks’ murder were exonerated. By that time, Ronnie Bridgeman had been paroled after serving 29 years; the other two were still imprisoned, having served 39 years—all of them as innocent men.

Witness tampering happens when witnesses are induced to provide false evidence or to withhold accurate evidence. In 325 of the 409 cases in which witness tampering was identified, the defendant was wrongly convicted of murder, sexual assault, or child sexual abuse. In 80% of those cases, the tampering was carried out by police officers, with the remainder carried out by prosecutors and child welfare workers.

Witness tampering was generally accomplished by threats (occurring in 130 exonerations) and manipulation (occurring in 305 exonerations). Threats included telling witnesses that if they did not cooperate, they or their close relatives would be charged, sentenced to prison or death, lose custody of their children, etc. Manipulation generally involved paying witnesses with money or other valuables, promising leniency in their own criminal cases or release from custody, as well as other benefits. But manipulation also included convincing witnesses that defendants were guilty based on evidence that didn’t exist and then convincing the witnesses that the defendants would go free without the witnesses’ false testimony.

Three common types of witness tampering that may be accomplished, either by threats, manipulation, or both, are: (1) procuring false testimony, which is the practice of inducing a witness to testify to facts the officer or prosecutor knows the witness did not perceive; (2) tainted identification, which occurs when police deliberately induce a witness to identify a suspect during a lineup or other identification procedure, whether the witness recognizes the suspect or not; and (3) improper questioning of a child victim, which occurs when the child is repeatedly, insistently, and suggestively questioned by a government official who will not allow the child to deny that he or she was a victim of sex abuse. The following are illustrative examples:

When Randall Dale Adams’ car ran out of gas in Dallas on November 27, 1976, he was given a ride by 16-year-old David Harris. Later that night, when the 37-year-old Adams was no longer with him, Harris shot and killed a police officer. Once the teenager became a suspect, he claimed that Adams had been driving and shot the officer with Harris’ gun. Because Harris was a suspect himself, police needed a “credible” witness to make the case. An eyewitness described the shooter as Black or Mexican (Adams is white), failing to identify Adams in the lineup. But an officer promised to drop pending robbery charges against the witness’ daughter if she would change her statement and make the identification. All of this was concealed from the defense—but not the prosecutor— and the witness testified at trial that she had identified Adams unassisted. Adams was convicted and sentenced to death in 1977. Eight years later, a documentary filmmaker dug into the case and discovered that Harris had subsequently received a death sentence for another murder—a sentence that never would have occurred but for the false testimony given by the witness at Adams’ trial. In 1988, Harris recanted his testimony against Adams, who was exonerated the following year. Harris was executed for the second murder in 2004.

In the investigation that led to Timothy Cole’s 1985 rape conviction in Lubbock, Texas, police showed the 20-year-old victim six photographs. Five were mugshots of “fillers”—other Black men who were not actual suspects—all photographed in profile and in black and white. Cole’s photo, however, was a Polaroid color picture with him facing the camera. The victim, who was white, immediately chose the Polaroid as the photo of her Black assailant. Cole, a 24-year-old U.S. Army veteran and Texas Tech student, spent 13 years in prison, where he died—10 years before he was exonerated, posthumously, by DNA testing and a confession from the actual rapist. He was also posthumously awarded a degree in law and social justice by Texas Tech in 2015. The victim, Michelle Murray Mallin, maintains she was pressured by Lubbock police to misidentify Cole.

Misconduct in Interrogations

In 1986, officers from the Chicago Police Department (“CPD”) arrested Aaron Patterson and Eric Caine in connection with the fatal stabbing of Vincent Sanchez, 73, and his 62-year-old wife, Rafaela. During interrogation by Lieutenant Jon Burge, Caine was told that Patterson had confessed that he and Caine killed the two victims. Caine denied being involved and refused to confess. Detectives then beat him and took him to see Patterson, who also had been so savagely beaten he could barely speak. Caine then signed a confession.

Patterson was left alone in a room with a confession to sign. He used a paperclip to scratch a message on a metal bench: “Police threaten me with violence. Slapped and suffocated me with plastic. No lawyer or dad. Sign false statement to murders.”

Patterson and Caine were both convicted of murder in 1989. Patterson, 25, was sentenced to death. He received a pardon based on innocence in 2003. Caine, 24, was sentenced to life in prison; he was exonerated in 2011 on the same day Burge—by then promoted to Commander—went to prison for perjury and obstruction of justice related to his torture of suspects.

For centuries, a confession was considered “regina probationum”— “the queen of evidence.” Studies reveal that jurors today are still convinced of guilt by a defendant’s confession almost as much as by DNA evidence.

There were false confessions in 292 of the exonerations in the Report. Misconduct during interrogations produced 165 (57%) of those false confessions. But what is considered misconduct in an interrogation?

The U.S. Supreme Court ruled in Rogers v. Richmond, 365 U.S. 534 (1961), that police behavior must be so “coercive” that it makes a confession “involuntary” in order to violate a defendant’s constitutional right to due process.

However, in expanding on this ruling in Haynes v. Washington, 373 U.S. 503 (1963), the Court confirmed that the use of violence or threats of violence automatically requires exclusion of a confession and that other types of misconduct also require courts to weigh the “totality of the circumstances” in order to determine if a confession was involuntary.

Threatening to charge a suspect’s relatives if he or she doesn’t confess is misconduct. So is “sham plea bargaining,” which occurs when police promise to obtain lenient sentencing or dismissal of charges. Police have no authority to determine which charges to bring to trial or to determine sentencing, meaning sham plea bargaining is “lying about the law” and is prohibited.

Nearly one-third of all exonerations from Chicago—including a whopping 54% of the Windy City’s murder exonerations—involved false confessions. That’s far higher than the rates for the rest of the country, where one-tenth of all exonerations and 18% of murder exonerations involved false confessions. How those false confessions were obtained holds a clue to the difference: 77% of Chicago’s false confessions were obtained by misconduct compared to 49% elsewhere in the U.S.

What makes Chicago such an outlier is attributable, at least in part, to the CPD’s once-systemic practice of torturing suspects. The most notorious offenders were Burge and his crew who, throughout the 1970s and 1980s, used beatings, suffocation, electrical shocks, and threats of death—even pulling the trigger on guns near suspects’ heads—to force confessions, mostly from Black men. Burge wasn’t responsible for all of the torture. Thirty-three of the 52 exonerations in which the CPD coerced suspects with violence didn’t involve him, and in 13 of those, the torture during interrogations happened after he was fired.

Of course, police misconduct during interrogations isn’t limited to Chicago. In August 2006, Herbert Fields, 70, was found shot dead near his car in New Haven, Connecticut, in an apparent robbery. Police Detective Clarence Willoughby interrogated a suspect, 16-year-old Bobby Johnson, without his parents present. In response to Johnson’s repeated claims of innocence, Willoughby threatened him with the death penalty unless he confessed, promising he would then get probation. Willoughby also lied to him that police had physical evidence which connected him to the crime. Johnson eventually tape-recorded a confession, which police helped him rehearse and even revise after details of ballistics testing contradicted it, with Willoughby warning Johnson that his “deal” for probation would be revoked if he didn’t make his story match the evidence. Johnson pleaded guilty to murder and was sentenced to 38 years in prison. He was exonerated in 2015, after Willoughby was caught extracting other false confessions and new evidence implicated another man in Fields’ murder.

In 43% of exonerations involving false confessions, no official misconduct was detected. However, in America, police are permitted to engage in conduct that, while not legally defined as misconduct, is known to produce false confessions. Police may not lie about the law, but they are permitted to lie about the facts and/or evidence. False confessions are often produced when police lie to suspects about failing polygraphs or about evidence that matches their DNA or fingerprints. Police may feed details of the crime to suspects for use in their confessions—and prosecutors then argue to the jury that the only way the defendant could know those details is if he committed the crime. Police may also make certain promises or threats, such as promising not to arrest a suspect if he confesses, even if they then arrest him after his confession.

In 2004, a DNA match found the man who had raped 16-year-old Fancy Figueroa after she got home from school in May 1997 in Queens, New York. But by then, she had a seven-year-old conviction for filing a false police report about the rape. Police who had investigated it had taken her to a hospital for a rape kit, where it was discovered she was pregnant. Believing she made up the rape to conceal her pregnancy, police promised to help find her rapist if she would confess she was lying. She complied, was charged, and convicted, and police dropped the case because she had “admitted” making up the charge. Yet the officers’ actions were not legally considered misconduct because they hadn’t promised anything out of their power to do and because that power included the prerogative to do nothing.

Fabricating Evidence

At the Chicago rape trial of Calvin and Larry Ollins, Omar Saunders, and Marcellius Bradford, in 1988, forensic analyst Pamela Fish testified that semen found on the victim’s body could have come from three of the four defendants, who ranged in age from 14 to 18. But Fish knew from blood testing she had conducted and hidden that none of the defendants could actually have been the source of the semen.

This is just one example of police or other officials inventing incriminating evidence by lying about their own observations—a factor the Report found in about 10% of exonerations. In 75 exonerations, police officers or forensic analysts working for police agencies presented false evidence by lying about results of forensic testing of trace evidence. In 5% of exonerations, officers falsely testified that they witnessed defendants commit crimes—most commonly by lying that defendants had possessed drugs or assaulted an officer. And in 36 exonerations, police fabricated confessions by defendants.

False or misleading forensic evidence contributed to 590 (25%) of the exonerations. But in most of those cases, forensic experts didn’t engage in misconduct. The faulty evidence was a result of once-accepted testing procedures that have since been shown to be unreliable. Said another way, the experts accurately followed the testing methods and honestly reported the results of those tests, but the methods were based on junk science.

Forensic fraud, on the other hand, occurs when a “state actor” engages in the “deliberate falsification of forensic evidence in order to help obtain a conviction,” the Report says. This occurred in 75 exonerations.

In 10 of them, experts claimed false matches of a defendant’s hair with hair recovered from the crime scene. Three months before testifying at a 1987 rape trial, West Virginia State Trooper Fred Zain wrote in a report that hair recovered from the victim was pubic hair. At trial, however, he concealed the report and testified it was “high unlikely” that the hair didn’t come from the blond beard of defendant Glen Woodall. Woodall’s subsequent conviction wasn’t overturned until DNA testing years later led to his 1992 exoneration.

In four of the remaining 65 exonerations, police planted the evidence that was then subjected to forensic analysis. At Shirley Kinge’s 1990 arson trial in Tompkins County, New York, State Police Trooper David Harding testified he found the 54-year-old’s fingerprints on a gasoline can recovered after the fire. Kinge was exonerated in 1992 after Harding admitted he had planted the evidence and pleaded guilty to perjury.

Two more exonerations later resulted from the discovery of fingerprint evidence planted by New York State Troopers: William Labolt, Jr., whose prints were planted by Lt. Craig Harvey, was exonerated in 1994, and Mark Prentice, whose fingerprints Harding also planted, was exonerated in 1995.

The 2017 exoneration of Desmond Ricks, by then age 50, resulted from the discovery that the bullets forensic analysts examined before his 1992 trial had almost certainly not come from the corpse of the victim, Gerry Bennett, but had instead been swapped by Detroit police for bullets taken from a gun belonging to Ricks’ mother.

Sometimes, forensic analysts changed their conclusions simply to satisfy the desire of a police officer or prosecutor to convict a particular defendant. In Adolph Munson’s 1984 murder trial in Oklahoma, Medical Examiner Ralph Erdmann first testified that the victim was killed by a large caliber bullet. After prosecutors told him they needed it to be a small caliber bullet, he revised his testimony and said maggots had enlarged the .22 caliber holes.

Twenty-five of the 75 cases of forensic fraud—one-third—involved notorious repeat offenders. Zain presented false evidence in seven trials that led to convictions of defendants later exonerated. Fish gave false testimony in 10 cases. Erdmann lost his medical license after it was uncovered that he filed reports in thousands of autopsies he never conducted.

The Report found that in five percent of exonerations, police fabricated evidence in crimes that never occurred. In August 2006, George Almond, 28, was visiting friends at a Chicago public housing apartment when CPD Sgt. Ronald Watts led a drug raid on the building. Familiar with Watts’ reputation for planting drugs on people, Almond waited until he thought police had left before he stepped into the hallway. Watts was there. He stopped Almond and demanded money. When Almond produced too little, Watts arrested him. At the police station, another officer placed 20 bags of heroin on a table that Watts then claimed he’d confiscated from Almond. Almond served 16 months of a two-year sentence after pleading guilty to possession of a controlled substance.

Four years later, Watts and another officer under his command, Kallatt Mohammed, were captured on surveillance video taking money from a drug courier—who was actually an FBI informant. An investigation then exposed a host of people from whom they had extorted money, planting drugs on those who didn’t pay up. By February 2019, there had been 66 exonerations—including Almond’s—which resulted from Watts’ misconduct.

Police sometimes resort to lying even in misdemeanor cases. In July 2014, a 27-year-old police officer in Bethel, Alaska, Andrew Reid, severely beat 48-year-old Wassillie Gregory, claiming Gregory had “clawed” at the officer’s hand. Without benefit of counsel, Gregory pleaded guilty to harassment of a police officer, a misdemeanor. He was exonerated a year later after a tip from an eyewitness led Reid’s superiors to surveillance video showing the officer handcuff a compliant Gregory and then repeatedly slam him into the pavement.

In 31 exonerations, police made up confessions that they falsely testified they had heard defendants make. In other cases, police tricked suspects into signing confessions unknowingly. After CPD officers interrogated Omar Aguirre about a 1997 murder, they told him a document printed in English was his release paper and instructed him to sign it. Aguirre, who neither spoke nor read English, signed what was actually a confession to the murder that had been prepared by police. His 1999 conviction was overturned in 2002 when the FBI tied the murder to a street gang.

Concealing Exculpatory Evidence

In 1984, John Thompson was arrested for a notorious robbery-murder in New Orleans. His picture was published in a newspaper, and based on that, he was identified as the perpetrator in an earlier carjacking. District Attorney Harry Connick, Sr., opted to try the carjacking case first, using that conviction then to obtain a death sentence against Thompson at the murder trial. Fourteen years later, a month before Thompson’s scheduled 1999 execution, a report surfaced from a forensic test on a bloodstain left by the robber on one of the carjacking victims, which prosecutors had ordered. The results— which proved the stain didn’t come from Thompson—had been hidden from the defense in his trial. In fact, prosecutors had even gone so far as to remove the piece of cloth with the stain and destroy it.

Thompson’s carjacking conviction was dismissed. At his retrial on the murder charge in 2003, evidence was presented of a $15,000 reward paid to the prosecution’s primary witness, Richard Perkins, by the family of the victim, Raymond Liuzza. The jury deliberated for 35 minutes before acquitting Thompson.

By far, the most frequent official misconduct in the exonerations the Report examined was the concealment of exculpatory evidence. In 44% of all exonerations and in 61% of exonerations for murder convictions, state actors had deliberately concealed material evidence.

As Alex Kozinski, former chief judge of the U.S. Court of Appeals for the Ninth Circuit, wrote in United States v. Olsen, 737 F.3d 625 (9th Cir. 2013), concealment of exculpatory evidence by prosecutors has “reached epidemic proportions in recent years”—and courts, the judge chided, were responsible because judges failed to take action.

Writing in the Preface to Criminal Law 2.0, 44 Geo. L.J. Ann. Rev. Crim. Proc. (2015), Kozinski elaborated on this concern, worrying that the high number of exonerations based on DNA testing exhibits that “there are disturbing indications that a non-trivial number of prosecutors—and sometimes entire prosecutorial offices—engage in misconduct that seriously undermines the fairness of criminal trials.”

Almost 60 years ago, the U.S. Supreme Court held in Brady v. Maryland, 373 U.S. 83 (1963), that “suppression by the prosecution of evidence favorable to the accused ... violate[s] due process where the evidence is material either to guilt or to punishment.”

Yet in “Trial & Error,” a study published by the Chicago Tribune (Jan. 11, 1999), Ken Armstrong and Maurice Possley found at least 381 homicide convictions across the country had been reversed between 1963 and 1999 “because prosecutors concealed evidence suggesting innocence or presented evidence they knew to be false.” Sixty-seven of those defendants had been sentenced to death, 24 of whom were exonerated before 1999.

Orleans Parish in Louisiana—where the Thompson case came from—has the highest per capita exoneration rate of any county in the U.S. In 78% of those exonerations, exculpatory evidence was concealed. Three of those cases, involving innocent men sentenced to death, reached the U.S. Supreme Court: Kyles v. Whitley, 514 U.S. 419 (1995); Connick v. Thompson, 563 U.S. 51 (2011); and Smith v. Cain, 565 U.S. 73 (2012).

All three of these cases were prosecuted by the Orleans Parish District Attorney’s Office headed by Harry Connick, Sr. In her concurring opinion in Thompson’s case, Justice Ruth Bader Ginsberg noted that “Connick created a tinderbox in Orleans Parish in which Brady violations were nigh inevitable. And when they did occur, Connick insisted there was no need to change anything.”

Exonerations for concealed evidence that were reviewed in the Report involved two broad types. The first, “substantive evidence,” includes objects or information that bears directly on a defendant’s guilt or innocence regardless of whose testimony introduces or provides it. So in some exonerations examined in the Report, there were objects concealed such as murder weapons, clothing, surveillance videos, and biological evidence.

Examples of concealed information included results of DNA and other forensic testing that excluded the defendant, eyewitnesses who identified someone other than the defendant, and ballistic evidence showing the deadly bullets were not fired from the defendant’s gun.

The second broad type of concealed evidence, “impeachment material,” challenges the credibility of a prosecution witness. It has no direct bearing on a defendant’s guilt except as it goes to the credibility or trustworthiness of the evidence of guilt that the prosecution is putting forth.

Examples of concealed impeachment evidence from the Report included witnesses receiving favorable plea bargains or dismissal of charges in exchange for their testimony; witnesses receiving cash rewards for their testimony; prior inconsistent statements made by witnesses; and histories of witnesses’ criminal conduct. In one case, the prosecution hid a surveillance tape that showed a purported eyewitness had left the scene an hour before the alleged crime occurred.

In the 520 exonerations where exculpatory evidence was concealed, impeachment evidence was hidden in 416 (80%); substantive evidence was hidden in 325 (63%); and both types were hidden in 262 (50%). Prosecutors were responsible for concealing impeachment evidence in 286 of the cases and for concealing substantive evidence in 218. Police were responsible in 172 cases, and forensic analysts were to blame in 29 cases.

Misconduct at Trial

The most common form of official misconduct at trial was perjury by law enforcement officers (304 or 13% of the 2,400 exonerations). In 256 (84%) of those cases, police officers committed the perjury, forensic examiners committed perjury in 18%, and a handful of cases included perjury by both.

Police officers sometimes lied to make their conduct appear consistent with witnesses’ statements. For instance, at the 1992 murder trial of Steven Dewitt in Washington, D.C., a police officer testified that he stopped Dewitt’s car because it had a temporary license plates ending in “818,” matching an eyewitness description, he said. Ten years after Dewitt’s conviction, defense attorneys discovered that the actual eyewitness report included the numbers “829.” In fact, police had stopped a car with those plates shortly after the murder but let the driver—apparently their real suspect—go. Dewitt was exonerated in 2004.

Most often, however, police lied about the conduct of the investigation. After 25-year-old Keith Price’s fatal 2006 shooting in Allen Epstein’s car in Georgia, Epstein told a police detective that Price had pulled a plastic pellet gun on Christopher Roesser, 25, before Roesser shot and killed him with a handgun. Afterwards, Epstein threw the pellet gun outside the car, where it was recovered.

But at Roesser’s 2008 murder trial, Epstein changed his story, apparently unwilling to admit that he and Price had been involved in a drug deal with Roesser. He denied ever saying anything about the pellet gun, and the detective also denied it. Roesser was convicted. At his 2011 retrial, an investigator working with the medical examiner told the jury that the detective had told him exactly what Epstein originally told police, leading to Roesser’s exoneration.

In 186 (8%) of the exonerations, it was prosecutors who committed misconduct at trial by failing to disclose that a witness had committed perjury. In Napue v. Illinois, 360 U.S. 264 (1959), the U.S. Supreme Court held that “a conviction obtained through use of false evidence, known to be such by representatives of the State,” violates the Fourteenth Amendment and cannot stand “when the State, although not soliciting the false evidence, allows it to go uncorrected when it appears.” This holds true whether the permitted perjury pertains to substantive or impeachment evidence.

Additionally, this misconduct almost always results in concealing exculpatory evidence as well. For example, at Ulysses Charles’ 1984 trial in Boston for a triple rape, a forensic analyst testified that no seminal fluid was detected on the vaginal swab. But both the analyst and the prosecutor knew that fluid had been found on the swab and that it came from a person with a blood type different than Charles’. The prosecutor allowed the testimony to go uncorrected, and Charles spent 18 years in prison before he was exonerated in 2002.

The most frequent type of uncorrected perjury—occurring in 20% of exonerations of this type— occurred when prosecutors knowingly let witnesses lie about incentives they received in exchange for their testimony. At Jabbar Collins’ murder trial in Brooklyn in 1994, three witnesses testified that they made no deals with the prosecution in exchange for their testimony against Collins. The prosecutor—who had in fact made a deal with each of the three—ridiculed to the jury the idea that any such deal had been made. In 2010, a federal judge compelled the prosecution to produce documents that described the deals that had been made as well as how the prosecution had threatened the witnesses to force them to give their testimony.

Prosecutors dropped their objections to Collins’ release, and the judge vacated the convictions. But the Brooklyn District Attorney’s Office reiterated its “position, then and now ... that we believe in this defendant’s guilt.” The judge responded, “It’s really sad that the D.A.’s office persists in ... saying they did nothing wrong here.”

Prosecutors themselves are not permitted to lie in court, regardless of whether or not they are under oath. The American Bar Association’s Model Rules of Professional Conduct, Rule 3.3 states that a lawyer shall not knowingly ... make a false statement of fact or law to a tribunal.” Yet lying in court by prosecutors is known to have occurred in 94 (4%) of the exonerations.

In 1997, in Hampton, Virginia, Ricky Cullipher was convicted of a shooting that left his friend in a coma. Cullipher’s attorney asked for a trial delay because a witness failed to appear, but the prosecutor told the judge that the witness would add nothing to the testimony from previous witnesses. However, the prosecutor knew the missing witness was an eyewitness who would have testified the victim had shot himself.

Prosecutors may not argue purported facts to a jury that the prosecutor knows to be false. Yet in 2009, Timothy Parkes was convicted of bank fraud in Tennessee after the prosecutor argued to the jury that Parkes would get to “keep the $4 million” if the jury acquitted him—even though the prosecutor had earlier convinced the trial judge to prohibit the defense from presenting evidence that showed the money had already been repaid.

In cross-examination of witnesses, prosecutors may ask “leading questions” which, while not technically statements, can leave a jury with a false impression. Asking, “You committed this particular crime, didn’t you?” will almost certainly receive an answer of “no” from a defendant—while also leaving the jury with the idea that he did what the veiled assertion suggested.

Discipline of Officials

Some form of discipline was imposed in only 219 (17%) of the exonerations where misconduct occurred. But in many instances, it was a single punishment meted out to one official for repeated misconduct in several cases. And in 70% of the cases—154 in all—where discipline was imposed, it was for behavior other than the specific misconduct that led to the exoneration.

In the 381 homicide cases between 1963 and 1999 that Armstrong and Prossley found were reversed, only two of the hundreds of prosecutors involved in the misconduct were disciplined: one was fired but reinstated and the other was suspended for 30 days. Forty-two of the other prosecutors who were from the Chicago area went on to become judges.

A 2003 study by the Center for Public Integrity of over 11,000 criminal cases from state courts where prosecutorial misconduct was alleged found that judges reduced sentences, vacated convictions, or dismissed charges in 2,000 of those cases. Yet bar complaints were filed against just 44 prosecutors, with only two of them disbarred and 12 suspended.

In the Report, prosecutors were disciplined in only 4% of the 2,400 exonerations. Eleven prosecutors were disciplined by the offices that employed them (though just two were fired), and 14 were disciplined by bar authorities or courts (with only three disbarred).

In only two exonerations were prosecutors convicted of a crime—for which they received only minor punishments. Former Williamson County, Texas, District Attorney Ken Anderson served four days in jail on a criminal contempt conviction after concealing evidence that would have prevented Michael Morton from serving 24 years in prison on a false murder conviction.

And former Durham County, North Carolina, District Attorney Michael Nifong served one day in jail for criminal contempt after concealing DNA evidence in a nationally publicized case involving three white members of the Duke University Lacrosse team falsely accused of raping a Black exotic dancer. The charges in that case were dropped, but the negative publicity led to an investigation of Nifong that revealed he had hidden exculpatory evidence in obtaining a murder conviction against Darryl Howard in 1991. Howard was exonerated in 2016.

Police officers were disciplined in 160 cases examined in the Report. Officers were convicted of crimes in 127 cases, while in 33 cases they were disciplined by the police force that employed them. Though this number is significantly higher than it is for prosecutors, it is considerably lower than the total number (843) of exonerations where misconduct by police officers helped secure false convictions. Punishments other than criminal convictions included suspensions, assignment to desk duty or other less-desired duties, and reprimands.

Having 127 exonerations where an officer was convicted of criminal misconduct does not mean there were 127 police officers convicted because some officers were repeat offenders. Lieutenant Burge, Sergeant Watts, and Officer Mohammed each received a single conviction, but together, they were responsible for the misconduct in 85 of the 127 cases where an officer was convicted. (Due to the nature of the available records, the exact number of officers convicted is not known, but it appears to be about 30.)

Even those criminal convictions handed down were usually not based on the officers’ underlying misconduct. For example, Burge wasn’t convicted for his acts of torture; he received 54 months in prison for lying under oath. Watts wasn’t convicted for extorting money or falsely claiming drugs belonged to arrestees; he was convicted for stealing money from an undercover officer in a federal sting and sentenced to 22 months in federal prison.

Forensic analysts were disciplined in 62 cases—47% of those in which their misconduct led to an exoneration. Three analysts were convicted of crimes; one received 30 days in jail (for writing false drug-analysis reports); one was placed on home confinement for nine months (for lying about credentials); and one received probation (also for lying about credentials). Other forms of discipline included loss of official certification, loss of employment, and suspensions. Only 13 different analysts who were disciplined accounted for 35 of those cases (and just six analysts for 28 of those).


After Michael Morton was exonerated, the state of Texas passed the Michael Morton Act, which, among other things, requires prosecutors to provide defendants access to “any offense reports, any designated documents, papers, written or recorded statements of the defendant or a witness, including witness statements of law enforcement ... or any designated books, accounts, letters, photographs, or objects or other tangible things ... that constitute or contain evidence material to any matter involved in the action and that are in the possession, custody, or control of the state....” (Texas Code of Criminal Procedure, Art. 39.14).

But broadly speaking, all of that was required before Morton was tried. Such laws or rules would help—if they were enforced. It seems likely that the reason Ken Anderson withheld exculpatory evidence in Morton’s case is that such evidence was routinely concealed by prosecutors in Williamson County. And this pattern of behavior existed because rules against it were seldom enforced. Strict enforcement of existing rules is one “reform” that would reduce wrongful convictions in general and wrongful convictions stemming from misconduct in particular.

Reforms in the manner of gathering evidence, on the other hand, have been shown to be very effective in reducing the likelihood of misconduct. Recording the witness identification process helps to reveal any taint. It also reveals whether or not the witness actually identified the suspect. Other identification reforms the Report points out include: “having the lineup administered by a person who doesn’t know the identity of the suspect, and therefore can’t communicate the identity to the witness, purposefully or inadvertently; selecting ‘fillers’ for the lineup who resemble the witnesses’ description of the criminal; telling each witness that the criminal may or may not be in the lineup; recording the level of confidence of any witness who makes an identification; and never permitting multiple witnesses to view a lineup together.”

Likewise, recording interrogations sharply reduces misconduct for the obvious reason that abusing a suspect—or making unlawful threats or promises and other misconduct—would be exposed. Recording or videotaping interrogations has been the law in Illinois since 2003, a reform that was passed in direct response to the abuses of Burge and his underlings.

No U.S. jurisdiction requires that all witness interviews be recorded. But this would provide a record of what was said, enable a jury to examine the witnesses and the circumstances where the statement was made, prevent officers from lying about or rephrasing what was said, and prevent witnesses from being threatened or manipulated.

Another desperately needed reform arises from the U.S. Supreme Court’s modified interpretation of its Brady decision with its subsequent ruling in United States v. Bagley, 473 U.S. 667 (1985). Brady held that a prosecutor is required to disclose material exculpatory evidence, and when it was decided, the universal understanding of “material” was “any evidence that tends to prove the probability or improbability of a fact in issue.”

But in Bagley, the Supreme Court created a new definition of “materiality” for Brady purposes. Under Bagley, evidence is “material” only if “there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different.” This is akin to the “no harm, no foul” rule the Court has established with its “ineffective assistance of counsel” test announced the previous year in Strickland v. Washington, 466 U.S. 668 (1984). But it’s simply unworkable for Brady purposes.

For example, prior to even taking the case to trial, how can a prosecutor know the impact of any evidence on a finder of fact? Only after the trial can that be known when a verdict is issued. Again, the role of the prosecutor is not to weigh the evidence before trial and then make a determination that the exculpatory evidence in his or her possession isn’t persuasive and therefore conclude there is no duty to disclose it.

The original Brady test was simpler and more workable. Once a prosecutor decides on which charges to bring and determines the facts at issue, it is possible to make an informed decision as to whether any piece of evidence in possession tends to disprove any fact at issue. Unless the prosecutor personally witnessed the crime, humility would dictate, recognizing that he or she doesn’t truly “know” whether the defendant is guilty, that the prosecutor’s appropriate role is to see that all material evidence is put before the finders of fact so that they can make an informed and good-faith determination of whether the prosecution has carried its burden of proving “guilt beyond a reasonable doubt.”

Without needing the Court to revisit Bagley, though, an additional reform suggested by this author would be having someone in the prosecutor’s office to play the role of “devil’s advocate” when determining whether there is any exculpatory evidence in the prosecutor’s possession. Skilled prosecutors know how to play the role of a skilled defense attorney. The assigned devil’s advocate would not be on the team prosecuting the case but would take the role of “if I were the defendant’s attorney, would this evidence be of value to me?”


The number is staggering: 1,296 factually innocent people convicted due to the deliberate misconduct of government officials. Considering the Report’s limitations, the actual number of cases is likely far higher. For example, the Report acknowledged 149 exonerations that occurred in Houston, Texas, but did not count them among those resulting from official misconduct because the identifiable evidence supported, at best, a conclusion of negligence.

Those 149 defendants all pleaded guilty to drug crimes. But later laboratory testing proved the material evidence with which—and for which—they were arrested contained no drugs at all. After apprehending these 149 people, did officers conduct “field tests” on materials in their possession and falsely claim a positive result as grounds for probable cause to arrest? We don’t know. The case records don’t reveal how these suspected drug materials came to be in the possession of those arrested or why anyone was charged with drug offenses for possessing the materials. So the Report doesn’t include these and many other known exonerations in its list of those that resulted from official misconduct.

It is highly probable that a tremendous amount of misconduct occurring in minor offenses goes undetected. For example, Sergeant Watts was the culprit in 66 of the 2,400 exonerations, yet his misconduct was uncovered only after he was arrested for an unrelated offense. How often does similar misconduct occur in minor crimes by similar bad actors operating under color of authority? How would a defendant prove it when it’s the defendant’s word against the word of two or three officers?

Misconduct also goes undetected in more serious cases. If not for Thompson’s death sentence for murder, the intensive—and expensive—investigation that uncovered the misconduct in his carjacking conviction wouldn’t have been undertaken. How many others have been falsely convicted of robbery, rape, and other serious crimes due to misconduct that was never detected because no one had the time, money, or interest in investigating?

Official misconduct often occurs as well in convictions obtained with ineffective defense counsel—since defense attorneys, like prosecutors, are officers of the court. But it isn’t often detected. Unless a defendant sues his attorney, no one goes looking for any misconduct.

Finally, officials commit misconduct with the intention of not being caught. If a detective destroys evidence of innocence, e.g., a surveillance video or DNA showing someone other than the defendant committed the crime, how would we ever know? Many of the instances of official misconduct only come to light through a series of random events, i.e., pure luck.

The 1,296 exonerations uncovered by the Report are only the “known” cases where official misconduct resulted in a conviction of the factually innocent. The actual number is most likely substantially higher. Ten times higher? One hundred times? No one knows.

But two things are known with absolute certainty: (1) official misconduct is a major cause of convicting the factually innocent, and (2) the gravity of the misconduct and its consequences are not being addressed.

In 1,076 (83%) of the 1,296 cases where official misconduct caused an innocent person to be unlawfully deprived of liberty, no one was disciplined. No one. How serious is that offense? Before you attempt an answer, consider the statutory penalties for kidnapping and abduction.

In these cases, innocent people were sentenced to prison—many to life-long terms, and 93 of them were sentenced to death. On average, these innocents spent 11 years in prison (13.9 years for those wrongly convicted of murder).

It is rational to argue that these issues need to be addressed as the serious criminal offenses that they are. Is there really much difference between one who kidnaps a victim and one who falsely and knowingly convinces a court to sentence someone to prison? Or between one who shoots another with intent to kill and one who hides evidence with the intent to convince a jury to send an innocent person to the death chamber? 


As a digital subscriber to Criminal Legal News, you can access full text and downloads for this and other premium content.

Subscribe today

Already a subscriber? Login



PLN Subscribe Now Ad
Advertise Here 4th Ad
Disciplinary Self-Help Litigation Manual - Side