Reading Time: 33 minutes

On January 2, 1983, in the early morning hours, James A. Buckley died at a service station in St. Louis County. The 19-year-old white male attendant had been shot seven times, with robbery as the apparent motive.

When 24-year-old Ellen Reasonover heard about the murder on the television news later that day, she mentioned to her mother that she had stopped at the service station after midnight to seek change for the laundromat. As she approached the cashier’s cage, Reasonover recalled, she saw a black male inside, walking to the rear of the building. She assumed he was the attendant, but he didn’t respond to her knocking. Reasonover glimpsed two other black males in the shadows of the service station lot, but thought nothing of it. She then drove to a nearby convenience store for the change she still needed.

Reasonover’s mother urged her to contact the police, who’d asked the community for tips that might help solve the homicide. Ellen Reasonover did as her mother suggested. When she called the police, she used a made-up name, but used her true name as soon as she arrived at the police station.

When police asked Reasonover why she had used a false name over the telephone, she explained that, as a black woman dealing with detectives under pressure to solve a murder with a young white male victim, she had to overcome a lifetime of generalized distrust. Furthermore, her half-brothers were criminals known to local police, causing Reasonover concern about guilt by association. But she understood her civic duty, she said, so she decided to come forward.

Imagine Ellen Reasonover’s surprise when, later that year, a jury convicted her of Buckley’s murder. Sixteen years later, a federal judge—a Republican appointee who had once served as a prosecutor—released Reasonover from prison. Reasonover, the judge concluded, was almost certainly innocent, and without question had been a victim of prosecutorial misconduct.

Steven Goldman, the prosecutor who put Reasonover in prison and whose misconduct, the federal judge ruled, contributed to the miscarriage of justice, benefited from the high-profile conviction; the voters of St. Louis County elected him to a state court judgeship.

Reasonover’s case shows what can go wrong as prosecutors collaborate year in, year out with the same group of police officers, forensic scientists, expert witnesses and judges. They come to trust each other, developing bonds that sometimes lead to shortcuts. If the evidence is weak, it becomes easier to ignore those weaknesses or paper them over. At each step, errors are ratified rather than exposed.

What follows is a catalog of lessons about prosecutors and their allies in the criminal justice system, revolving around an intensive study of Reasonover’s conviction. Though her case is extreme, the kinds of misconduct that occurred can be found in hundreds of other cases.

Lesson One: Premature conclusions can ensnare the innocent. Police detectives assumed Reasonover’s guilt within a few days of the murder, seeming to disregard evidence that pointed away from her, according to available documents. Prosecutor Goldman apparently adopted those assumptions in preparing his case against Reasonover.

Police asked Reasonover to view mug shots of possible perpetrators. They told her not to worry about a perfect match. Rather, police told her to look for features similar to the men she reported glimpsing from a distance in the shadows of the service station. When Reasonover identified pictures of two men who turned out to be incarcerated for other offenses, police arrived at the conclusion that she must be trying to deflect suspicion from herself. They put Reasonover in a jail cell, with no access to a lawyer. By contrast, when another potential witness, a white male, pointed to one of the same mug shots that Reasonover had chosen, the police did not put him in a cell.

Within a couple of days, the prosecution team allowed Reasonover to leave jail without filing charges against her. But, according to available documents, it appears investigators were so fixated on Reasonover that they did little to pursue other leads, such as service station visitors who reportedly had purchased marijuana from the victim at work; an individual with whom the victim fought at a party the month before; and area robbers with a modus operandi similar to what occurred the night of the murder.

Lesson Two: Lack of solid evidence does not prevent charges from being filed. Charging a suspect despite lacking physical evidence (no weapon, no blood, no saliva, no semen, no skin, no hair); no reliable eyewitness testimony; no credible confession; no clear motive; and no coherent theory of the crime is often a bad move. Police and prosecutors lacked all those elements in the case of Ellen Reasonover. Before the arrest, authorities possessed no hint of Reasonover’s involvement other than their reaction to her perceived behavior after she voluntarily called the police with information.

A prosecutor’s authority to charge someone with a crime is an awesome power, without any realistic, meaningful checks. Conscientious prosecutors understand the danger. New York City prosecutor Mark D. Cohen states it plainly: Charges should not be filed to use as leverage in plea bargaining, or to enhance a prosecutor’s political standing, he says. Other prosecutors may use the National District Attorneys Association standards, which make clear that charges should be filed only if they “can be substantiated by admissible evidence at trial.” Some elected district attorneys have tried to limit abuse by making the decisions themselves in questionable cases.

In Brooklyn, N.Y., elected district attorney Charles Hynes told his staff prosecutors that he would personally decide whether to charge a suspect based on a single eyewitness account. Hynes instituted the policy because Brooklyn had for several years suffered embarrassment due to wrongful convictions. He says that during a recent year-long stretch, he rejected 15 of 70 proposed prosecutions because the sole eyewitness in the rejected cases lacked credibility.

Of course, when the elected prosecutor or a deputy district attorney makes the final decision there’s always the danger that pressures related to electoral politics will play a role. E. Michael McCann, the elected prosecutor in Milwaukee, Wis., since 1968, has stood for re-election several times and understands the temptations. “The discretion of the district attorney is broad and subject to almost no control or judicial review,” McCann said. “The burden of proving that a district attorney has abused his or her discretion … because of political ambition or community pressure will be extremely difficult, if not impossible, to prove.” Take for example a prosecutor who files a charge in a weak but sensitive case, figuring “a closing argument artfully orchestrated to stoke the passions of the jury but not so fevered as to trigger reversal will surely tempt those jurors to convict, overriding doubts that ought to persist due to deficiencies in the evidence. In such cases, only the district attorney will know what the defense attorney, the judge and the investigating officers may suspect, and the jury never perceives: that the district attorney has compromised the precious integrity of his or her office.”

Whether prosecutors are seeking higher office or not, McCann worries about how many never reduce or dismiss a charge as new evidence becomes available. Too many prosecutors have a mistaken notion that the original decision is “garbed with a pristine quality, and any reduction therefrom appears to cast the district attorney in an adverse light and to compromise the integrity of the criminal justice system. … In some cases, an elected district attorney, not wishing to reach a politically unpalatable but proper decision, justifies inaction by saying, ‘Let the jury decide.’ While this may appear to be graced with a patina of deference to the jury system, it is in fact an abdication of the district attorney’s responsibilities. The jury will rarely, if ever, know as much about the case as the prosecutor.”

Lesson Three: Prosecutors can tilt the system of checks and balances. Theoretically, a grand jury of citizens called from the registered voter rolls serves as a check on a prosecutor. So should the judge at the preliminary hearing. In practice, grand juries and judges rarely serve as a brake on prosecutors. But the system of checks and balances can fail, as it did in the Reasonover case. A grand jury heard the prosecutor’s evidence and indicted an innocent woman.

Traditionally, grand juries hear only the prosecution’s case, and the anecdotal evidence is overwhelming that grand juries hardly ever fail to follow the prosecutor’s lead. Mark Cohen of New York says that in many jurisdictions his colleagues have no obligation to tell grand juries about exculpatory or mitigating evidence. Such behavior is almost never discussed publicly. Journalists have no access to grand juries. Grand jurors themselves are supposed to remain silent. Defendants and their lawyers worry that they will be unable to negotiate with prosecutors later if they make grand jury proceedings the battleground for allegations of state error or misconduct.

Still, there are appellate court rulings that shed light on how a prosecutor can improperly manipulate a grand jury. A New Jersey ruling, for example, threw out the sexual assault indictment of James H. Gaughran in Hudson County (Jersey City). The victim, a 17-year- old woman, made the allegations against the 19-year-old Gaughran, whom she’d known for about nine months. The alleged assault occurred during an evening of smoking marijuana and drinking beer. Gaughran maintained that they’d had consensual sex. An examination of the woman at a hospital emergency room tended to support his version. Here is what the grand jurors heard about the medical examination:

Question by the prosecutor: “And were you taken to the Jersey City Medical Center?” Answer: “Yes, I was.” Question: “And were you given a gynecological exam while you were there?” Answer: “Yes.”

In overturning the conviction, the appellate court commented: “The results of the exam were not given to the grand jurors. They were not given the defendant’s statement. Nor were they advised of the drug use by the victim … The defendant contends the medical examination directly supports the defendant’s version and directly contradicts the victim’s version. The state concedes the report is exculpatory but argues it has no duty to present exculpatory evidence.”

The appellate judges in the Gaughran case made it clear they disagreed with the state’s contention. “It is essential that the jurors be informed of the relevant facts,” they wrote. “The evidential impact of the medical report should not be underestimated. It directly contradicts the victim’s claim of anal and vaginal penetration and does not support her claim of a one and a half hour struggle. The Grand Jurors could not have been expected to ask for the results of the medical exam. They were skillfully misled by omission into believing it had corroborated the victim’s testimony. By withholding relevant and highly exculpatory evidence in its possession, the State treated this Grand Jury as a rubber stamp, its playtoy, and clearly infringed upon this Grand Jury’s decision-making function. In fact, given the nature of the medical report and its devastating impact on the presentation, the failure to present it can be termed an intentional subversion of the process.”

Another “subversive” tactic prosecutors may use at the grand jury stage involves what is referred to as “saving back” one or more charges arising from the conduct leading to trial. If something goes wrong for the prosecution, those withheld criminal charges might be used to pressure the defendant into accepting a plea bargain. Saved back charges are supposed to be barred from future prosecution if known to the district attorney at the time of the original prosecution.

At the preliminary hearing stage, a judge can serve as a check on a prosecutor. Again, however, the anecdotal evidence is abundant that few weak cases are thrown out at preliminary hearings. Individual prosecutors see the same judges over and over, day after day. Banter can turn into professional friendships and sometimes social friendships. Ex parte communications, without defense counsel present, might start to seem normal.

Robert N. Kepple of the Texas District and County Attorneys Association said it might unfold like this: “The court has a lot of cases set on the docket Monday. The judge is a former prosecutor, the type who still asks the prosecutor ‘Are we ready?’ The judge, over coffee before the docket call, asks you about the cases set for trial that day. You know that what the judge wants to do is evaluate the one…he feels is worth trying.” Kepple says the ethical prosecutor has a simple solution available—common sense: “…The key to avoiding problems with ex parte communications is to remember the very definition of ex parte, that is, a communication [involving] fewer than all of the people legally entitled to participate. If prosecutors are always mindful of who should be or would want to be involved in a communication, they can avoid trouble.”

Lesson Four: Prosecutors should cautiously evaluate testimony and confessions. Police and prosecutors in various jurisdictions place too much value on apparent confessions, jailhouse informants and eyewitness testimony. Confessions sometimes turn out to be wrong, be it because they were coerced, the suspect was mentally ill, or the suspect confessed after a jailhouse informant concocted a story about hearing an admission of guilt. Eyewitness testimony may also turn out to be wrong: DNA evidence has shown that even a rape victim in the closest proximity imaginable to the perpetrator has misidentified the attacker. In the Reasonover case, authorities lacked a confession, eyewitness testimony, physical evidence or motive—raising serious questions about the decision to charge her. They did, however, have the testimony of two jailhouse informants.

Goldman made deals with two inmates who briefly shared the same cell with Reasonover at different times. Both inmates suffered from prior credibility problems; both told stories of alleged Reasonover confessions that violated common sense; both received future considerations from Goldman that were never disclosed to the defense or the jury. Years later, when Reasonover was represented by highly competent, persistent lawyers, one of the two informants refused to repeat her allegations under oath. The other could not be interviewed again because she had committed suicide.

Reasonover supposedly confessed spontaneously to an inmate named Rose Jolliff. Another inmate in the same cell, Marquita Butler, despite being pressured to confirm Jolliff’s account, told police and prosecutors that Jolliff was lying. Jolliff had not worn a recording device; the prosecution team possessed only her verbal account, which both Butler and Reasonover denied.

Five days later, after Reasonover had been released from police custody, the prosecution team arranged for Jolliff to call their suspect, hoping to entrap Reasonover. During the conversation, Reasonover mentioned her innocence eight times—to the same woman who supposedly heard an unprovoked confession five days earlier. Reasonover and her trial lawyers never knew about the recording of that conversation, which suggested innocence, before facing the jury that voted for a murder conviction.

The confidence prosecutors place on informants is surprising given how easy it is for an experienced convict to manipulate the system. Problem confessions, informants and eyewitness testimony have backfired again and again. But the learning curve is either temporary or non-existent in numerous jurisdictions. Confessions sometimes turn out to be false, maybe because of coercion by the interrogators, maybe because of the suspect’s mental illness, maybe because a jailhouse informant concocted a story about hearing an admission of guilt. The informants rarely provide information without expecting a payoff. As for eyewitness testimony by informants, victims or bystanders, it is sometimes filled with error for all sorts of well-intended or nefarious reasons.

Milwaukee District Attorney E. Michael McCann says that decades of working closely with police have led him to comprehend the false confessions phenomenon. The veteran prosecutor learns the degree of rectitude demonstrated by each officer, the biases of some. “The prosecutor may see the officers interrogate prisoners and hear their testimony, as well as that of the interrogated prisoners, during motions to suppress confessions. This experience can materially aid the conscientious prosecutor in a hard-nosed assessment as to whether police overreaching was involved in securing a confession and how much reliance should be placed on inculpatory statements reported but challenged by the accused. However, while long-term working relationships with officers can provide significant information as to their credibility or suspected mendacity, such contacts can also give rise to friendships that may impair the prosecutor’s capacity for critical assessments of claims of police overreaching in security confessions…Other prosecutors, being of unconscionably timorous bent, may be unwilling to confront police hostility occasionally triggered by a decision not to use a confession with the adverse implications that such a determination has for a conviction.”

McCann says that while a confession might not be out-and-out fabricated by police, it might still be untrue because of the suspect’s mental instability combined in some way with the tactics used to extract the confession. He worries about lack of candor by police “critical to the jury’s determination as to how trustworthy the confessions are.” Based on his experience, McCann says “On occasion, either in response to the district attorney’s question or by a volunteered comment, an officer or commander of integrity will tell the district attorney that he or she has doubts about the veracity of a confession.”

More and more courts are allowing the defense to present expert testimony about the prevalence of false confessions, including the reasons behind the surprisingly large numbers. In such circumstances, prosecutors often labor mightily to discredit the defense expert. That is no surprise—the testimony should be scrutinized in front of a jury because the generalizations might be inapplicable to the particular case.

Just as more and more courts are allowing defense experts to testify about the false confession phenomenon, more and more courts are allowing defense experts to explain why eyewitness testimony is sometimes unreliable. The eyewitness testimony used by prosecutors sometimes turns out to be well intentioned but mistaken, sometimes intentionally misleading. Police who question eyewitnesses can exercise a huge influence. McCann in Milwaukee worries about the “overzealous officer who inappropriately suggests information to a witness and who pushes an identification witness for more certainty than the witness has. Some officers carefully record both inculpatory and exculpatory information. Other officers tend to brush aside exculpatory statements and record principally inculpatory ones.”

DNA evidence is beginning to persuade previously strong advocates of relying on confessions and eyewitness testimony that they are mistaken more frequently than the conventional wisdom has allowed. When DNA testing demonstrates that even a rape victim in point-blank proximity to the perpetrator misidentified the attacker, all eyewitness testimony must be evaluated with care.

The pervasiveness of prosecutors using informants, and informants using prosecutors, came to the attention of many laypeople in 1989, in large part because the Los Angeles Times outed professional informant Leslie Vernon White. Reporters Ted Rohrlich and Robert W. Stewart opened their expose like this:

“When veteran jailhouse informant Leslie Vernon White picked up a telephone last fall and showed authorities how easily he could fake the confession of another inmate, he cracked open a window on a secret world. At any given time, Los Angeles County jails hold between 50 and 100 informants—many of them career criminals like White—who have engaged in relentless campaigns to implicate their fellow prisoners in crimes, and thus earn the ultimate favor from authorities: early release from custody.”

Before the White expose, prosecutors repeated their mantra to journalists, defense attorneys and other skeptics over and over: Yes, informants can be risky to use. But we have a two-pronged test we apply. Did the informant tell us details known only to the criminal? And did the informant and the defendant spend time together in jail?

White demonstrated that the two-pronged test could be compromised. As the Times reporters explained, “Equipped with only a telephone and the last name of the inmate he did not know, White impersonated police officers and prosecutors and squeezed enough information from law enforcement officials to fabricate a plausible confession. Then he created a phony record showing that he and the accused had been together in jail.”

The Times reporters studied “dozens of criminal cases involving alleged inmate confessions.” Three involved confessions faked by inmates and used against other suspects. In “many other” cases, the Times reporters could not tell for sure whether inmates had lied about confessions by others. Furthermore, threats by police and prosecutors that jailhouse informants would testify if a case proceeded to trial caused defendants to plead guilty in exchange for reduced sentences. The Times reporters determined that when defendants did go to trial, the prosecution had used jailhouse informants on the witness stand in hundreds of cases. Although it is usually impossible to prove what goes through a prosecutor’s mind, the Times reporters found credible evidence that at least some Los Angeles County prosecutors knew their jailhouse informants were lying, but called them as witnesses anyway.

Lesson Five: Prosecutors should not rely uncritically on their scientific and forensic experts. Reasonover was given a stress test—a modified lie detector test—which was ruled inadmissible in court. A prosecution witness was hypnotized to recover a memory of what he’d seen on the night of the murder. His hypnosis-induced recollection and subsequent testimony implicating Reasonover’s boyfriend was allowed to stand.

When understood and used properly, forensic and other technological evidence can be powerful tools. But the evidentiary value of such forensic material can be diminished when it is mishandled, improperly labeled, or poorly inventoried by police officers. Sometimes, even when the evidence is collected properly, the analysis may be faulty or the forensic scientist untrustworthy. Fred S. Zain in West Virginia and Joyce Gilchrist in Oklahoma City are among the forensic technicians employed by police departments who are best-known for their alleged incompetence or their willingness to doctor evidence.

Zain is known to have provided misleading evidence in numerous trials that helped the prosecution and hurt the defense. It is unknown whether prosecutors using his evidence knew about the problems; it is known, however, that even after Zain came under suspicion, some West Virginia prosecutors continued to ask that he specifically conduct forensic tests because they disliked the results from other scientists.

Zain, a law enforcement officer, began directing the serology division of the West Virginia State Police Crime Laboratory in 1979. Some of his staff suspected his competence and honesty at an early stage, but either said nothing or had their suspicions ignored by police superiors. All that began to change in 1987, with the trial of Glen Dale Woodall for assault. As one of the prosecution’s experts, Zain testified that his scientific examination of semen recovered from the victims matched Woodall’s blood type. Statistically, Zain testified, the chance of such a match was six in 10,000.

Appellate judges upheld the conviction. But Woodall and his lawyers refused to give up, and in a habeas corpus proceeding they convinced the state Supreme Court to order DNA testing. The results proved Woodall’s innocence. The trial court freed him from prison five years into his maximum 335-year term. Woodall sued for damages. The state’s insurer began an investigation that included a review of Zain’s work. The findings led to a $1 million payment for Woodall.

A subsequent investigation into Zain’s work led to a West Virginia Supreme Court ruling that all convictions involving Zain’s findings would be analyzed to determine if the remaining evidence supported those convictions. The state Supreme Court commented that the details of Zain’s conduct “are shocking, and represent egregious violations of the right of a defendant to a fair trial. They stain our judicial system and mock the ideal of justice under law.” The judges directed Kanawha County prosecutor William C. Forbes and the U.S. attorney to determine whether Zain should be charged with a crime.

A Kanawha County grand jury indicted Zain in 1998 for providing false testimony in the prosecutions of James E. Richardson, John Earl McLaurin and Jimmy C. Gardner. At Zain’s trial, jurors could not reach a unanimous verdict; Zain died of cancer at age 52 in December 2002, before he could be retried. At the time of his death, a dozen prisoners had been released or scheduled for retrial because of his testimony; the state had paid at least $6.5 million in damages to inmates.

While Zain was being prosecuted, an August 2001 ruling by the U.S. Court of Appeals for the Tenth Circuit in the case of Alfred Brian Mitchell was casting serious doubt on the work of Oklahoma City forensics expert Joyce Gilchrist. The case wended its way to the federal circuit after an Oklahoma City trial court found Mitchell guilty of murdering a college student volunteering at a community center for disadvantaged juveniles during 1991. The court also found Mitchell guilty of rape and sodomy.

The conviction and death sentence rested partly on the testimony of Gilchrist, a forensic chemist with the Oklahoma City Police Department. Gilchrist’s testimony about her examination of the physical evidence, coupled with other testimony that the victim and her boyfriend last had sexual relations eight days before the murder, convicted Mitchell.

Before trial, Gilchrist sent the evidence to Michael Vick at the Federal Bureau of Investigation laboratory’s DNA unit for testing. At trial, Gilchrist characterized Vick’s report as inconclusive. It was much later, during an evidentiary proceeding supervised by a federal judge as part of Mitchell’s appeals, that the defense saw Gilchrist’s notes of her conversations with Vick. Those notes suggested that the FBI’s tests cleared Mitchell of committing rape either before or after murdering the victim. One of Gilchrist’s own tests had the same result.

In reversing the death penalty, the Tenth Circuit judges commented not only on Gilchrist, but also on the trial prosecutor: Gilchrist “provided the jury with evidence implicating Mr. Mitchell in the sexual assault of the victim which she knew was rendered false … by evidence withheld from the defense. Compounding this improper conduct was that of the prosecutor, whom the [federal] district court found had ‘labored extensively at trial to obscure the true DNA test results and to highlight Gilchrist’s test results,’ and whose characterization of the FBI report in his closing argument was ‘entirely unsupported by evidence and … misleading.’”

Not all questionable forensic evidence stems from lies, tied to pro-prosecution bias. So-called expert witnesses sometimes have weak credentials, or rely on hokum posing as science. Prosecutors around the nation used to retain the services of a University of North Carolina-Greensboro anthropology professor named Louise Robbins, who said she could match crime-scene footprints to the footwear of perpetrators. Few other forensic scientists endorsed the validity of Robbins’ techniques. But prosecutors called on Robbins over and over, banking on the good will of the trial judge to certify her as an expert. Robbins helped convict defendants across the nation until her technique was shown to yield results that were no better than chance would have produced.

Robbins probably would have remained below the radar of most defense lawyers, legal scholars, journalists and general readers except for her involvement in an especially egregious case of prosecutorial zealotry. The case began in 1983, with the abduction of a 10-year-old girl from her comfortable home in DuPage County, part of the Chicago suburbs. When searchers found her body two days later, they saw she had been assaulted and murdered. Finally, more than a year later, prosecutors—under pressure from the girl’s parents, elected officials and the citizenry at large—charged three men based on very questionable evidence. About a year after the arrests of Rolando Cruz, Alejandro Hernandez and Stephen Buckley, the trial began. Jurors found Cruz and Hernandez guilty, but deadlocked on Buckley. By the end of 1985, a career criminal named Brian Dugan had confessed to murdering the girl, acting by himself. He passed a lie detector test, and convinced numerous observers, including seasoned law enforcement officers, that he indeed abducted and murdered the victim. The prosecutors and the trial judge refused to believe Dugan, however. So Cruz and Hernandez sat on death row while their appeals proceeded. Buckley, meanwhile, wondered whether and when prosecutors would re-try him, based largely on Robbins’ testimony about his shoes.

Chicago Sun-Times reporter Thomas Frisbie watched the case unfold. Like other observers, he thought Robbins’ testimony against Buckley violated common sense, not to mention scientific principles. So he started contacting other anthropologists and forensic scientists in related fields. Every scientist Frisbie contacted demeaned her work. Robbins herself would not discuss the validity of her techniques in the Buckley case with Frisbie. A law professor who taught a course on scientific evidence told Frisbie it seemed the only standard used by prosecutors and judges allowing Robbins to testify as an expert “is that it be incriminating to the defendant.”

DuPage County State’s Attorney Jim Ryan finally announced during early 1987 that his office would not re-try Buckley, in large part because Robbins was suffering from a serious illness. “Apparently,” Frisbie said, “Ryan could find no other expert who would support Robbins’ conclusions.” Robbins died later that year. Eventually, both Hernandez and Cruz would also be cleared.

The next year, an Illinois appellate court reversed the murder and sexual assault conviction of Dennis J. Ferguson. Robbins had been a prosecution witness. The appellate judges ordered a new trial for Ferguson largely because of their finding that Robbins’ testimony was worthless. The judges added, however, that the prosecutor aggravated the situation by frequently misstating the evidence. The prosecutor iced that impermissible behavior with this passage from the closing argument: “You have to believe that they [prosecution witnesses] are all liars or fools, every one of them. And for you to find the defendant not guilty … you have to believe he told you the truth. You have to feel his brothers told you the truth. And that all of the persons I just named are liars and fools. It is your decision.” The appellate judges noted “For a prosecutor to inform a jury that in order to believe the defense witnesses the jury must find that each of the State’s witnesses was lying is such a misstatement of law as to prejudice the defendant and deny him a fair trial.”

Lesson Six: Some prosecutors still withhold evidence. In Ellen Reasonover’s case, prosecutor Goldman withheld a tape recording suggesting her innocence, a recording made without Reasonover’s knowledge while she was in a jail cell. It is possible that police gave it to him without commenting on its contents. But it is unlikely that Goldman never listened to the tape, especially given his reputation for thoroughness. The tape captures a 56-minute conversation between Reasonover and her former boyfriend as they sat in separate, nearby holding cells secretly rigged by police with recording devices. Police and prosecutors suspected the former boyfriend of collaborating with Reasonover and another man to murder the service station attendant. The tape captures two people sounding befuddled about why they are in jail, about why they are suspects. The conversation by itself does not prove innocence, but if introduced at trial could have pushed one or more jurors into the realm of reasonable doubt.

The recording came into evidence during a post-conviction proceeding because of a private investigator’s question to a prosecution witness many years after trial. At that point, the state could have professed ignorance, could have claimed the tape had been destroyed, or otherwise frustrated Reasonover’s post-conviction lawyers. Fortunately for Reasonover, a prosecutor other than Goldman, upon request, looked for the recording, located it, and provided it to Reasonover’s last-ditch appellate lawyers. Without it, Reasonover would still be serving her prison term.

Most prosecutors do not destroy evidence that could establish innocence. It does happen, though, and probably more often than prosecutors themselves acknowledge publicly. For example, while compiling a primer on prosecutorial behavior in Kentucky, defense lawyer Jerry J. Cox found multiple appellate rulings revolving around destruction of evidence.

Withholding evidence that might be favorable to the defendant is called a Brady violation, after a 1963 U.S. Supreme Court decision. That ruling involved a 1958 homicide in Anne Arundel County, Md., that demonstrated the ugly consequences of non-disclosure.

John Leo Brady, a 25-year-old abandoned by his parents while a baby, a poor student, an Air Force enlistee with a spotty record leading to premature discharge, an indifferent employee drifting from menial job to menial job, found himself broke while his girlfriend was pregnant. When he noticed the new car of a long-time acquaintance, physically slight, hard-working loner William Brooks, age 53, Brady decided to steal the Ford Fairlane. Brady planned the theft with Donald Boblit, the developmentally disabled brother of his girlfriend. The available evidence suggests that Brady had no intention of killing Brooks, who had treated him generously in the past. But, that same evidence suggests, Boblit panicked during the car theft, beating Brooks and then strangling him to death. After their arrests, Brady and Boblit told conflicting stories. Among his various versions, Boblit told authorities that he, not Brady, was responsible for the homicide.

Prosecutor C. Osborne Duvall apparently never considered telling Brady’s defense lawyer about that confession, nor did Duvall have to do so by law. Brady and Boblit both ended up on Death Row. A prison chaplain who believed Brady’s account that Brooks died at the hands of Boblit convinced E. Clinton Bamberger Jr., a former Catholic school student of his, to consider helping with an appeal. Bamberger, a former prosecutor with the state attorney general’s office, had come to believe in a level playing field while working for the government.

Bamberger obtained the transcripts from the separate trials of Brady and Boblit. The Brady trial transcript yielded no grounds for appeal. But in the Boblit transcript, the lawyer noticed a vague reference to a July 9, 1958, confession, called the “fifth statement.” Examining the exhibits attached to the transcript, Bamberger found four statements, but not the July 9 document. When he obtained that document, which included Boblit’s confession, he used it as the basis for Brady’s appeal.

A Maryland court accepted the appeal, and, later, so did the U.S. Supreme Court. Writing for the majority, Justice William O. Douglas stated, “the suppression by the prosecution of evidence favorable to an accused upon request violates due process where the evidence is material either to guilt or to punishment, irrespective of the good faith or bad faith of the prosecution … Society wins not only when the guilty are convicted but when criminal trials are fair; our system of the administration of justice suffers when any accused is treated unfairly.”

It might seem simplest for prosecutors automatically to provide everything—except the identities of confidential informants—in their files to the defense. Some prosecutors indeed operate like that. Some do not; it is human nature to want to win, and turning over everything sometimes makes obtaining a conviction more difficult.

The gray areas can be immense. The timing of the disclosure is one of those gray areas. If a prosecutor hangs onto Brady material long enough, it might lose its value for the defense. For example, if the prosecutor delays disclosure of significant evidence with exculpatory or impeachment potential, the defendant will be disadvantaged before the trial when deciding whether to accept a plea agreement rather than risk an uncertain fate in front of a jury.

Douglas R. Roth, a Sedgwick County (Wichita), Kan., prosecutor, instructs his colleagues in a chapter of The Prosecutor’s Deskbook to “avoid the temptation of making a self-determination that the possible exculpatory evidence is not credible or admissible and therefore does not need to be disclosed to the defense. The test under Brady is not whether the evidence is admissible. …Given a prosecutor’s duty to seek the truth and to not convict innocent persons, the firm belief that a particular defendant is guilty should always be subject to reasonable review and re-evaluation.”

Admonitions such as Roth’s seem to be ignored more frequently than previously thought. “For a while after the Brady decision, every occurrence of withholding by a prosecutor meant a risk for the prosecution,” Katherine Goldwasser, a former federal prosecutor in Chicago who now teaches law at Washington University in St. Louis, told the Center. Her first boss taught her “If it hurts, turn it over to the defense.” Law and practice have moved a long way from that maxim, Goldwasser says, as courts have provided prosecutors new loopholes. To demonstrate a Brady violation today, the defense must prove non-disclosure of normally secret information, then demonstrate that the non-disclosure contributed to a guilty verdict.

“Legal standards will never stop diabolical prosecutors,” Goldwasser said. “But going back to the original Brady standard could stop a basically decent prosecutor from withholding information useful to the defense. Most prosecutors aren’t out to break the rules.”

Many prosecutors emphasize that turning over everything is simplest. Joshua Marquis, the elected district attorney in Clatsop County, Ore., says it is his policy: “I say give the defense everything. What’s the downside, as long as you don’t jeopardize an informant’s life? It’s an easy call for me.” A few prosecutors mentioned that sometimes the discovery material helps defense attorneys see their clients’ lies, making a plea bargain easier to reach.

Lesson Seven: A pattern of excluding potential jurors by race or gender should raise a warning. A jury of all whites tried Reasonover, despite a substantial minority population in St. Louis County. The use of peremptory challenges by the prosecution (or the defense) to remove a certain type of juror from the panel might be grounded in impermissible racism or sexism.

Proving a violation, however, is usually impossible, absent a leaked smoking-gun memo or an obvious historical pattern, because it requires reading the mind of the prosecutor. If the prosecutor can offer a reason for each challenge that sounds race-neutral or gender-neutral, there is nothing a trial judge or appellate judges can do but nod their heads.

Defense attorneys talk about extreme instances where they believe the prosecutor dropped the criminal charges in the middle of jury selection because of an unfavorable looking panel, only to refile the charges later in hopes of drawing a more favorable panel.

Just as racial or gender bias continues to be an issue for American society at large, the fairness of jury selection will likely be questioned in some cases. During 2002, theDallas Morning News demonstrated the continuing nature of the problem with a front-page story containing this opening paragraph: “A Supreme Court ruling in the case of Texas death row inmate Thomas Miller-El, convicted of killing an Irving hotel clerk, could set a standard for proving racial discrimination by prosecutors in jury selection, legal experts say.” The defendant’s lawyers first raised the jury fairness issue during the 1986 trial. The trial judge found no evidence of wrongdoing, but the appellate process kept the allegations alive until the U.S. Supreme Court agreed to hear oral arguments in October 2002.

In an eight-to-one decision, the U.S. Supreme Court said Miller-El should receive a substantive hearing from a lower court on his allegations about jury selection. The court majority said, “A comparative analysis of the venire members demonstrates that African-Americans were excluded from [Miller-El’s] jury in a ratio significantly higher than Caucasians were. Of the 108 possible jurors reviewed by the prosecution and defense, 20 were African-American. Nine of them were excused for cause or by agreement of [both] parties. Of the 11 African-American jurors remaining, however, all but one were excluded by peremptory strikes exercised by the prosecutors. On this basis, 91 percent of the eligible black jurors were removed by peremptory strikes. In contrast, the prosecutors used their peremptory strikes against just 13 percent (four out of 31) of the eligible non-black prospective jurors qualified to serve … ” Furthermore, the court majority ruled, during the questioning of potential jurors, prosecutors used different tactics based on race when asking about willingness to invoke the death penalty.

Such conduct at trial, combined with persuasive evidence about the Dallas County District Attorney’s office historical practice of discriminating against African-American potential jurors, led to the court majority’s lopsided vote in Miller-El’s favor.

Lesson Eight: Juries empowered to impose the death penalty might be more likely to convict. Anecdotal evidence and social science research suggest that excluding all potential jurors who oppose the death penalty changes the tenor of deliberations. It is apparently human nature to think that when the death penalty is on the table, the defendant must be guilty, or the prosecutor would request a lesser punishment. Reasonover faced a jury who could impose the death penalty. After the jurors found her guilty, 11 of the 12 voted to execute her, despite a case based solely on police suspicions and the testimony of two jailhouse informants. Only one juror stood between Reasonover’s death at the hands of the state and her later release from prison by a federal judge.

In the Dallas case of Miller-El heard by the U.S. Supreme Court, his lawyers argued that the prosecution used peremptory challenges to eliminate 10 of 11 possible black jurors while retaining whites with similar backgrounds. The state claimed that the blacks stricken through peremptory challenges expressed unwillingness to impose the death penalty, which meant their exclusion from the jury would be justified by law. “The vast majority of non-minority panelists favored the death penalty and were willing to impose it, while the majority of African-American panelists were opposed to the death penalty or unwilling to impose it,” the state’s brief argued. Some supporters of Miller-El’s position argued that the state’s line of reasoning promotes unfairness because it suggests that African-Americans in general are less desirable as jurors than whites because of differing views about imposing a death sentence. Miller-El’s petition referred to a Dallas Morning News series based on 100 randomly selected felony trials. It showed that 86 percent of blacks eligible for jury duty ended up not serving because of prosecutors’ peremptory challenges.

Lesson Nine: Improper opening statements and closing arguments and direct and cross-examinations can infect the fairness of a trial. An appellate court in the Reasonover case said the prosecutor “was consistently blatant in his use of leading questions” posed to the jailhouse informants during trial. But because Reasonover’s trial lawyers never objected to the leading nature of the questions, the appellate court would do nothing about the problem.

Defense lawyers are generally vocal about prosecutors’ conduct during trials—that is part of their job, after all. But sometimes judges, from the bench, issue some of the strongest condemnations of prosecutorial tactics.

Paul E. Peifer, an Ohio Supreme Court justice, explained in a commentary article his dissent in an appeal from death row inhabitant Angelo Fears: “A prosecutor’s ‘overzealous’ remarks may not seem that important. After all, why should prosecutors restrain themselves when all they are doing is trying to convict a guilty person? In fact, it is extremely important. Our system of justice is … founded on the principle that no matter how guilty a defendant may appear, the accused shall receive an impartial trial; everyone according to the same rules. But in this case, Angelo Fears’ prosecutors time and again made statements during the trial that may have unfairly influenced the jury. For example, the prosecutors defamed a psychologist who testified as an expert witness. The prosecutors called the doctor the ‘mouth piece’ of the defense, insinuating that the defense paid the expert simply to have him parrot their opinions. The prosecutors also made repeated references to notes that the doctor took during an interview with Fears. The judge had already ruled that the state couldn’t view the notes, but the prosecutors kept bringing them up, making it appear to the jury that important information was hidden in those notes.” Peifer warned that if prosecutors in the Fears case and other cases “don’t reign themselves in, a guilty criminal will wriggle out of a conviction because of prosecutorial misconduct.”

In North Carolina, the state Supreme Court lectured prosecutors on improper closings in an opinion that vacated a death sentence: ” … The issue of improper closing arguments has become a mainstay, if not a troublesome refrain, in cases before this court. In virtually every capital case, many other criminal cases, and a growing number of civil cases, this issue is being vigorously advocated as grounds for reversible error. Therefore, we take this opportunity to revisit in some detail the limits of proper closing argument; the professional and ethical responsibility of attorneys making such arguments; the duty of our trial judges to be diligent in overseeing closing arguments; and the possible ramifications for failing to keep such arguments in line with existing law.” The court added, “If attorneys were to scrupulously comply with these seemingly simple requirements, then the issue of alleging improper argument on appeal would prove an exception instead of the rule. Regrettably, such has not been the case; in fact, it appears to this court that some attorneys intentionally ‘push the envelope’ with their jury arguments in the belief that there will be no consequences for doing so.”

Amie L. Clifford, assistant director of the National College of District Attorneys, warns her colleagues about the types of misconduct that can occur during closing arguments, such as commenting about the defendant’s failure to testify at trial; characterizing prosecution evidence as uncontradicted when the only person able to contradict it with certainty is the defendant, who chose not to testify; making direct statements about the defendant’s refusal to participate in a post-arrest interview with law enforcement officers; issuing derogatory comments about the defendant that could be considered name calling (“animal,” “monster,” “pervert,” etc.); injecting personal opinion in a manner that amounts to vouching for a witness (“I submit to you that I think the victim is telling the truth about the details of the rape, despite the lack of semen.”); appealing to the jury’s prejudices or passions, such as comparing the heinous nature of the crime at issue to the Columbine High School mass murder; stating to the jury that an acquittal will lead to dire consequences for the community, such as heightened gang crime sprees; asserting personal knowledge concerning facts not in evidence.

Defense lawyers Daniel E. Monnat of Wichita and Paige A. Nichols of Lawrence, Kan., have identified 10 kinds of prosecutorial misconduct during closing arguments. The first five fall under their label “elementary misconduct”: referring to facts not in evidence; commenting about the defendant’s failure to testify or present evidence; name calling or scare tactics meant to inflame jurors; begging for sympathy, such as asking jurors to place themselves in the victim’s shoes; and stating personal opinions, including bolstering the credibility of state’s witnesses as well as impugning the credibility of the defendant, defense counsel or defense witnesses.

The next two fall under the label “distorting the burden of proof”: arguing that the jury cannot acquit unless it finds state’s witnesses are mistaken or lying; and arguing that the jury must convict if it finds the defendant or defense witnesses lied. Next is “denigrating defense tactics”: suggesting that the defense theory of the case is a sham. Finally, come two points under the label “misstating the jury’s responsibility”: minimizing jurors’ obligations by suggesting appellate courts can correct mistakes; and overstating jurors’ obligations by asking them to support law enforcement or send a message to the criminal element.

Prosecutors, in various professional publications, show they are aware of the closing argument and cross-examination pitfalls. Jean G. Sturtridge, a St. Clair County, Mich., prosecutor, advises colleagues that the state “should not use the power of cross-examination to discredit or undermine a witness if the prosecutor knows the witness is testifying truthfully.” Another precept: Prosecutors “should not ask a question that implies the existence of a fact that he either knows to be untrue or has no reasonable, objective basis for believing is true.”

Al M. Dominguez, the elected prosecutor for Colorado’s 19th District, based in Greeley, admonishes colleagues to stay away from all kinds of tricks during cross-examination, including some that might be legal. Dominguez singled out a technique called “pitting.” “During direct examination, the defendant testifies to a story that totally contradicts … prosecution witnesses,” he wrote. “During cross-examination, the prosecutor asks the defendant whether all the other witnesses are lying. … Because testimonial inconsistencies between witnesses can be the product of mistake or another cause, it is misleading and unfair to give the jury the impression that the witness is lying. It is the province of the jury to weigh the credibility of the witnesses … “

Lesson Ten: Prosecutors should not interfere with defense access to prosecution witnesses or tamper with witnesses for either side. In the Reasonover case, defense counsel experienced difficulties trying to talk to the two jailhouse informants before trial.

Douglas R. Roth, writing in The Prosecutor’s Deskbook, warns his colleagues to refrain from obstructing defense access to witnesses. “The best practice for prosecutors is to instruct witnesses, preferably in writing, that they may talk with defense representatives if they choose to but that they are not obligated to.”

Misconduct can also occur when a prosecutor needs to contact a defense witness. “Just visiting a defense witness and interviewing [him] does not raise ethical concerns,” says Robert N. Kepple of the Texas District and County Attorneys Association. “But just how far can a prosecutor go when it comes to telling the witness [he or she] is less than impressed with the story? The key may be intimidation. Did the prosecutor act in a way that intimidated a witness into not testifying?” If the question ends up before a judge, the issue usually revolves around allegations that a prosecutor threatened a witness with perjury. “A prosecutor cannot engage in deliberate and badgering threats designed to quash significant testimony,” he adds. Prosecutors are also limited in what they can say to those testifying for the state. If a witness contradicts an earlier account, the prosecutor can warn the witness to be truthful, but cannot emphasize such warnings to the point that intimidation seems to be involved. A refusal to grant immunity from prosecution for the prior inconsistent testimony can be a form of intimidation, in that it is meant to bind the witness to one particular version, whether true or not. Even when the state’s witness is credible at first and remains consistent in her account, there is what Kepple calls “a balancing act that prosecutors must perform every time they prepare a witness for trial.” That balancing act allows detailed discussion of the law and the facts, as well as the importance of certain facts. But untruthfulness or even shading the facts to fit a particular scenario cannot be encouraged.

Former prosecutor Bennett L. Gershman is pessimistic that witness-coaching can be monitored, given the private settings in which it occurs. Calling witness-coaching the “dark secret” of prosecutorial conduct, Gershman says wrongful convictions and other documented errors of the criminal justice system can often be linked to “techniques used by prosecutors to prepare, shape and polish the testimony of their witnesses.” Based on his own experience as a prosecutor as well as his research since becoming a law professor, Gershman says it is “indisputable that some prosecutors coach witnesses with the deliberate objective of promoting false or misleading testimony. Prosecutors do this primarily to eliminate inconsistencies between a witness’s earlier statements and her present testimony; avoid details that might embarrass the witness and weaken her testimony; and conceal information that might reveal … the prosecutor has suppressed evidence.”

Lesson Eleven: Appellate courts sometimes ignore exculpatory evidence withheld due to misconduct. The year after Reasonover’s murder conviction, a team of Washington Post reporters began investigating the conduct of police and prosecutors, based on a tip. The Post story referred briefly to the secretly recorded jailhouse conversation between Reasonover and her former boyfriend, but the reporters did not know the precise contents of the tape. Reasonover’s first team of appellate lawyers requested the tape, but the state replied that the two individuals “did not discuss anything of substance concerning the killing,” adding that the prosecution “should not be faulted for not disclosing a conversation which contained nothing of evidentiary value to either party.”

In other words, the state cast doubt on Reasonover’s credibility because she never discussed details of a murder she said she never committed. The appellate judges neither asked for nor heard the tape. In their denial of Reasonover’s original appeal, the judges said “Although it would appear that [the tape] should have been disclosed to the defense, absence of any indication in the record of the content of the conversation prevents our determination of whether the state’s failure in this regard amounts to prosecutorial misconduct warranting reversal.”

The appellate judges also brushed aside the prosecution’s failure to fully disclose deals with the two jailhouse informants, because, the judges said, they could not decipher the criminal-history printouts of the informants, so chose not to bother with that evidence.

It’s not uncommon for appellate judges to give prosecutors a pass, excusing their missteps with the term “harmless error.” If a prosecutor bends or breaks a rule, even if the violation appears to be intentional, appellate judges can say they fail to see how a change in conduct would have altered the verdict. It is the equivalent of the vernacular “No harm, no foul.” Most of the time, appellate judges refuse even to identify the offending prosecutor by name.

“The harmless error rule, once an appellate mechanism to prevent technical violations from upsetting a verdict, has evolved into the most powerful judicial weapon to preserve convictions despite serious errors or misconduct,” Gershman writes. “And whereas habeas corpus historically was invoked to vindicate fundamental constitutional rights, the Great Writ has been rendered almost nugatory by procedural obstacles to the bringing of claims and by increased deference to state court interpretations of federal constitutional guarantees.”

Former Kansas prosecutor and trial judge Robert L. Gernon agrees, as he views allegations of misconduct from his perch on the appellate bench. Back in 1967, Gernon notes, U.S. Supreme Court Justice Potter Stewart recognized the quandary of a harmless error rule, saying it committed the court “to a case-by-case examination to determine the extent to which we think unconstitutional comment on a defendant’s failure to testify influenced the outcome of a particular trial. This burdensome obligation is one that we here are hardly qualified to discharge.”

Gernon said the contradiction is inherent: ” … Appellate courts will not reweigh the evidence. However, at times that is precisely what reliance on the harmless error rule requires.” The inherent contradictions result “in a sliding-scale approach to some cases. The exact same conduct will result in a reversal if the case is close, but an affirmance if the evidence of guilt is strong.”

With habeas corpus relief from the federal courts severely restricted by a 1996 Congressional act, sound appellate decisions by state courts are more vital than ever to the functioning of the criminal justice system. Reasonover’s habeas petition made it to the federal courts just before the new law mandated a cutoff. As Barry Scheck, Peter Neufeld and Jim Dwyer write in their book, the Antiterrorism and Effective Death Penalty Act of 1996 closed a relief valve. The law “gives condemned prisoners six months after their state appeals to ask for federal intervention, and sets a one-year time limit for all other cases.”

Lesson Twelve: Individuals from outside the criminal justice system are often the only post-conviction hope of those denied a fair hearing. Judges do not often revisit closed cases, relying on the concepts of harmless error and finality. Defense attorneys may not care, or they may be overwhelmed with other cases, or they may not be competent enough to investigate. In Reasonover’s case, an investigation by the tiny, not-for-profit Centurion Ministries of Princeton, N.J., followed by Centurion’s hiring of post-conviction lawyers from Kansas City, Mo., made the difference.

Other outsiders made a difference on Reasonover’s behalf. On Christmas Day 1998, for example, a private investigator who had helped on an early Reasonover appeal saw an article in the St. Louis Post-Dispatch about the convicted murderer’s attempt to clear her name. The investigator called one of Reasonover’s last-ditch appellate lawyers, offering to turn over his old files. In those files, the lawyer noticed evidence of a previously undisclosed leniency deal between the prosecutor and one of the jailhouse informants. The information appeared in a memo from the informant’s public defender to an office colleague. The memo said the informant “is going to be a witness in a capital murder case that Steve Goldman is trying … After she testifies, she is going to plead guilty [to a charge unrelated to Reasonover] … and be given probation. The details of the plea can be worked [out] after she testifies. The state does not want to allow … defense attorneys to bring up any kind of deal that might have been made … I have been assured by Steve Goldman that the state isn’t going to burn her, that she will receive probation.”

Lesson Thirteen: Police and prosecutors sometimes do little to search for the actual perpetrators of a crime after learning the original suspect is innocent. Although police and prosecutors in the Reasonover case said they had identified her two alleged accomplices, they never charged either man with murder. Furthermore, when the Centurion Ministries investigation identified a plausible murderer, St. Louis County authorities did nothing, at least publicly. With Reasonover freed, the killing of the service station attendant remains unpunished and unsolved. As many as three murderers might be at liberty to kill again. But as far as police and prosecutors are concerned, they cleared the case from the books.

Your support is crucial!

Our newsroom needs to raise $121,000 by end of the year so we can hold the power accountable and strengthen our democracy in 2024. Public Integrity doesn’t have paywalls and doesn’t accept advertising. We depend on individuals like you to sustain quality journalism.