When a prisoner is granted their freedom because they were wrongly convicted of a crime, the focus turns to the years — or decades — they spent behind bars, their feelings upon release and their hopes for the future. But a new study digs into the reasons people are wrongly convicted, and it has found that 54 percent of those defendants are victimized by official misconduct, with police involved in 34 percent of cases, prosecutors in 30 percent, and some cases involving both police and prosecutors.

The study by the National Registry of Exonerations reviewed 2,400 exonerations it has logged between 1989 and 2019, nearly 80 percent of which were for violent felonies. Of the 2,400, 93 innocent defendants were sentenced to death and later cleared before they were executed.

The study also found that police and prosecutors are rarely disciplined for actions that lead to a wrongful conviction. Researchers found that 4 percent of prosecutors involved in those convictions were disciplined, but the penalties were “comparatively mild” and only three were disbarred. Police officers were disciplined in 19 percent of cases leading to wrongful convictions, and in 80 percent of those cases officers were convicted of crimes, such as Chicago police Sgt. Ronald Watts, who led a group of officers who planted drug or gun evidence leading to 66 false convictions.

“Misconduct by police, prosecutors and other law enforcement officials is a regular problem,” said Samuel R. Gross, an emeritus professor at the University of Michigan Law School and a co-founder of the registry, “and it produces a steady stream of convictions of innocent people.” And because the data for exonerations is gathered randomly, from news reports and legal rulings, with no central repository of court statistics, “it is clear to us that misconduct occurs in many more cases” than the registry tracks.

The study cites five types of misconduct that lead to wrongful convictions: witness tampering, misconduct in interrogations, fabricating evidence, concealing exculpatory evidence and misconduct at trial. It found that Black defendants were slightly more likely than White defendants to be victims of misconduct, 57 percent to 52 percent, but that the racial difference was much larger for drug crimes — 47 to 22 percent — and in murder cases, 78 to 64 percent. In some instances, authorities purposefully sought to improperly influence a case, other times their missteps were unintentional or based on flawed techniques, Gross said.

The report notes that President Trump’s administration has reduced or eliminated measures that address wrongful convictions, such as shutting down an office focused on forensic science and limiting federal oversight of local police departments. And the report found that prosecutorial misconduct in federal white collar cases has increased over the last 18 years. But Gross pointed out that there are positive trends as well, with misconduct in interrogations and police lineups greatly reduced and reliance on faulty science such as hair analysis, arson evidence or bite marks no longer leading to improper convictions.

“In policing, if you don’t reform, somebody’s going to reform you,” said William G. Brooks III, the chief of the Norwood, Mass., police and a widely recognized proponent of “blind administration” photo lineups for crime witnesses. That process uses an officer completely unconnected with a case to show six photos to a witness, giving no hint to the witness of which one might be the suspect while video-recording the process.

Brooks was dubious initially but wound up getting blind lineups into mandatory statewide training in Massachusetts in 2004, and he has trained police on the process in 29 states and Canada. Massachusetts also has required video-recording of suspect interrogations since 2013. “I think great strides have been made,” Brooks said, “but we’re not finished. There are still states where these reforms are not in place.” The study said misconduct in police interrogations had dropped drastically since 2003, particularly in cases involving abusive questioning of supposed victims of child sex abuse.

Flawed scientific conclusions helped lead to the 46 percent of cases with wrongful convictions that weren’t the result of official misconduct. Gross said about a quarter of all exonerations emanated from bad forensic science, as well as witnesses who overstated the findings of such evidence. Ineffective work by defense lawyers was another cause and “probably as much of a problem as misconduct of prosecutors, but we just don’t know,” because many such cases go unreported or un-exonerated, Gross said.

And then there are cases where people plead guilty simply to get out of jail rather than wait months for trial, although the evidence later clears them. “My guess is,” Gross said, “the most common cause of false convictions is pretrial detention,” often in low-level drug or theft cases. In Harris County, Tex., 149 people who pleaded guilty to drug possession before the lab could even test their drugs were later exonerated when the lab determined the seized substances weren’t illegal.

The study is a detailed analysis of 30 years of exonerations, but Gross and others are careful to emphasize that the 2,400 cases are far from a comprehensive count, since there is no centralized national database of criminal cases at the state and local levels. So an estimate of how often wrongful convictions occur, as a percentage of overall cases, is not possible.

But to say that only 2,400 wrongful convictions occurred across 30 years would be wrong, Gross said. He noted that a 2014 study found that 4.1 percent of all death sentences between 1973 and 2004 were imposed on innocent defendants, in cases presumably with much more legal work involved. “What does that tell you about non-death cases? We don’t know,” Gross said. “Maybe it’s not 4.1 percent (rate of exoneration), but maybe it’s 2 percent. … It is clear to us that misconduct occurs in many more cases, but we can’t say what percentage.”

The study notes that 60 county prosecutors across the country have formed conviction integrity units to review old cases for possible wrongful convictions. Nelson Bunn, executive director of the National District Attorneys Association, noted that many exonerations over the past five years have been the result of these units and prosecutors taking “a proactive role in correcting injustices of the past.”

Prosecutors from a conviction integrity unit in Baltimore last year helped free Alfred Chestnut, Ransom Watkins and Andrew Stewart who were wrongly convicted as teenage friends in a 1983 murder. Police at the time had ignored evidence pointing to another person as the killer, the prosecutors said.

Bunn also noted that prosecutors accused of “official misconduct” often refers to a failure to turn over something in the state’s files to the defense, but that “does not mean there was intentional misconduct, but something that may have just been missed in good faith.” Bunn also said some wrongful convictions occurred because enhanced technology and processes such as DNA testing weren’t available at the time, but now enable testing to exonerate the wrongly convicted. This happened recently in Florida with the release last month of Robert DuBoise, after recently discovered DNA showed he hadn’t killed a Tampa woman in 1983. The conviction was based largely on now-discredited bite-mark analysis.

The study acknowledges there are other areas to examine, including quantifying ineffective assistance by defense attorneys. Miriam Krinsky of Fair and Just Prosecution also noted there is limited information on misdemeanor exonerations, although misdemeanors “are the doorway for so many into the justice system and all that follows.”

Fair and Just Prosecution, which supports liberal prosecutors across the country, has called on local district attorneys to create a “Misdemeanor Post-Conviction Integrity” process, saying that “integrity issues in misdemeanor cases are as, if not more, profound as concerns arising in felony cases, for numerous reasons.” But Krinsky said she was not aware of any prosecutors who have tried that approach.

The lack of discipline for police and prosecutors involved in wrongful convictions is not new, Gross said, largely because both professions do most of their regulation themselves. “The impulse not to punish one’s own is strong,” Gross said. For prosecutors, “it’s a disciplinary system put in place by lawyers, run by lawyers, and professional discipline is often very lax. Police officers are more likely to be disciplined, but if they are, that discipline is hidden from view, including from prosecutors.” The study found that forensic analysts involved in wrongful convictions were disciplined in 47 percent of the cases.

The National Registry of Exonerations is a project that was founded in 2012 in conjunction with the Center on Wrongful Convictions at Northwestern University School of Law, and is a joint project of the University of Michigan Law School, the Michigan State University College of Law and the University of California, Irvine Newkirk Center for Science & Society.