The National Registry of Exonerations has added a code that identifies cases in which the evidence against the exoneree included testimony by a jailhouse informant—a witness who was in custody with the exonerated defendant and who testified that the defendant confessed to him.

“Jailhouse snitch” testimony, as it is commonly known, is notoriously unreliable because the incarcerated witnesses are strongly motivated to say what the prosecution wants, usually because they get substantial reductions in their own sentences in return. This problem has been in the news lately.  Cameron Todd Willingham  was executed for murder by arson in Texas in 2004 based on forensic evidence that was later discredited, and testimony from a jailhouse informant. That informant recently came forward to say that his testimony was false and was procured by a secret deal with the prosecutor. And in Orange County, California, the District Attorney is under fire because an aggressive, illegal and secret program of using jailhouse informants in murder prosecutions has come to light.

To find the jailhouse informant cases in the Registry, go to the Detailed View  under the Browse Cases tab, place the cursor over the heading “Tags” in the heading bar above the list of cases, click on the down arrow arrow.jpg that appears and choose “JI” (“Jailhouse Informant”) from the drop-down menu.

Eight percent of all exonerees in the Registry were convicted in part by testimony from jailhouse informants, 119 out of 1,567.  They are concentrated among the worst crimes: 102 out of 119 are murder cases, 15%, of all murder exonerations, compared to 2% of all other exonerations (17/900). Among murders, the more extreme the punishment, the more likely we are to see a jailhouse informant, ranging from 23% of exonerations with death sentences to 10% of murder cases in which the defendant received a sentence less than life in prison. See Table 1.


This pattern makes sense.  There is more at stake in prosecutions for murder than for lesser crimes—and more yet among the minority of highly aggravated murders that get sentences of life without parole or death—so the authorities appear to be more willing to cut deals or do other favors for prisoners in return for testimony, which can easily lead to perjury.

There is one noteworthy pattern among the few exonerations in which jailhouse informants helped convict innocent defendants for lesser crimes than murder: they are much more likely to testify in federal than in state cases. About 10% of federal exonerations in non-murder cases included jailhouse informant testimony (8/81), compared to 1% of non-murder exonerations in state courts (9/819). We can’t generalize from this small number, especially since four of the eight federal informant exonerations were in a single four-defendant drug prosecution, but it suggests that trials for crimes other than murder are more likely to include jailhouse informants if they take place in federal court.


- Samuel Gross & Kaitlin Jackson


National Academy Releases Report on Eyewitness Identification

On October 2, 2014, the National Research Council of the National Academy of Sciences released a report entitled Identifying the Culprit: Assessing Eyewitness Identification. This is an important document, but not because it breaks new ground. Instead, it does an excellent job of summarizing what we do and do not know about eyewitness identification testimony, and it offers recommendations on how to collect and use eyewitness identification evidence in cases in which a criminal’s identity is in dispute.

 peep 2.jpg 

The care and detail of the Report, the distinction and the wide range of experience of the members of the committee that wrote it, together with the prestige of the National Academy, may make it uniquely influential.

The Report recommends that police use procedures designed to protect eyewitness memories from contamination, and to collect information about the reliability of witness identifications. Specifically, the Report recommends that police departments: 

  • Train all law enforcement officers in eyewitness identification, focusing specifically on factors that affect vision and memory, and on procedures to prevent contamination.
  • Implement “double-blind” identification procedures in which the officer who administers an in-person or photographic lineup does not know which of the people is the suspect, and thus cannot deliberately or unintentionally influence the witness.
  • Develop and use standardized witness instructions at identification procedures.
  • Document witnesses’ level of confidence in their judgments at the time of the identifications.
  • Videotape all identification procedures.

The Report also recommends that judges take steps to assess the value – and possible dangers – of eyewitness identifications by conducting pretrial inquiries about eyewitness identifications regardless of whether one of the parties raises an objection to the admissibility of the evidence.  In the same vein, the Report recommends that judges make juries aware of the pitfalls of eyewitness identification by admitting expert witness testimony, requiring detailed evidence of the circumstances of identifications, and giving jury instructions on how to evaluate eyewitness testimony.

Finally, the Report calls for better data collection and more scientific research. In particular, the Report finds that existing research does not support a conclusion one way or the other on the comparative value of sequential lineups – in which the people or pictures are presented one at a time – as compared to the older and more common practice of simultaneous lineups, in which a witness is shown a line of several people or an array of photographs all at once. The Report recommends additional and better studies comparing these two procedures. The Report also recommends the creation of a National Research Initiative on Eyewitness Identification to coordinate future research and help make use of its findings.

As of October 14, 2014, 35% of the exonerations in the Registry include mistaken eyewitness identifications, 509 cases out of 1446. The distribution of all known exonerations that include mistaken eyewitness identifications by category of crime is displayed in these graphs, and was discussed in a recent Registry newsletter.


-  Kaitlin Jackson & Samuel Gross




Female Exonerees:

Trends and Patterns


Only 8% of the exonerees we know about are women (119/1,432), about the same proportion as female prison inmates in the United States. 


There are two major, interrelated differences between female exonerees, as a group, and male exonerees.

Child victims: 40% of the women in the Registry (47/119) were exonerated of crimes with child victims, compared to only 22% of male exonerees (293/1,313). Most of these women were convicted of child sex abuse – 22%, compared to 11% of male exonerees – but none were convicted of sexual assault on an adult. In addition, nearly 30% of female homicide exonerees were convicted of killing children (15/51), but only 17% of male homicide exonerees (105/601).

Nonetheless, men account for 86% of exonerations with child victims because male exonerees outnumber females by 11 to 1.

No-crime cases: 63% of female exonerees were convicted of crimes that never occurred (75/119), three times the rate for men (21%, 271/1313). Seventy-five percent of female no-crime exonerations involved violent crimes (56/75); and in three-quarters of those cases the supposed victims were children (42/56).

Nearly 90% of women exonerated of crimes against children were convicted of violent crimes that never happened (42/47). Nearly 70% of these cases were in two comparatively uncommon categories:

  • Child sex abuse hysteria cases. None of the 26 women exonerated in a child sex abuse case was convicted of a crime that actually occurred. (By comparison, 27% of men exonerated in child sex abuse cases were convicted of actual crimes that other men committed.)  The major reason is that 80% of these women (21/26) were exonerated in "child sex abuse hysteria" cases.

From the mid-1980s to the mid-1990s, some prosecutors and child welfare workers around the country became convinced that people caring for children, often day care workers and school employees, were sexually abusing those children on a massive scale. More than a hundred defendants were convicted as a result, often based on accusations including bizarre, unsubstantiated and highly improbable satanic rituals. The Registry lists 51 child sex abuse hysteria exonerees, 21 women and 30 men; all were convicted of crimes that never happened. They make up 0.2% of exonerations of men, but 28% of exonerations of women.

The most recent child sex abuse hysteria conviction in the Registry occurred in 1995. Since then, only one woman has been exonerated for any crime involving child sex abuse.

  • Shaken Baby Syndrome cases. Twelve exonerations involve "Shaken Baby Syndrome" or SBS (now called "Abusive Head Trauma"). In almost all, the exonerees were convicted on a now widely discredited theory that violent shaking of infants can produce immediate and extreme neurological damage or death without external or skeletal injuries. It now appears that these deaths and injuries were caused by unrelated accidents or undiagnosed pathologies.

Two-thirds of SBS exonerees are women (8/12), which is hardly surprising. One implication of the theory of SBS is that the infant was injured or killed by shaking by the last adult who took care of the child before medical personnel intervened, and women do the vast majority of the work of caring for infants.

Five additional women were exonerated after convictions for killing children who died by accidental means unrelated to claims of shaking.

A terrible consequence of these trends is that some innocent mothers –Teresa Engberg-Lehmer and Nicole Harris, for example –must grieve for children who died of accidental or natural causes, while in prison for killing those children themselves.

In sum, while most female exonerees were falsely convicted of violent crimes against adult victims, or of non-violent crimes, a substantial and disproportionate minority were convicted of violent crimes against children, usually violent crimes that never happened. Judging from known exonerations, women are the likely victims of false convictions for violent crimes that are believed to have been committed by care-takers in roles that are overwhelmingly filled by women – as parents and other family care givers, and as day care workers and teachers of young children.


- Kaitlin Jackson & Samuel Gross





​Paul Casteleiro has worked to free wrongly convicted defendants for three decades, primarily with Centurion Ministries in New Jersey. "I love doing these cases,” Casteleiro says. “You get to hold the system accountable." (Third from the left with the blue tie)


​Three years after our inauguration, the Registry has added its 1,600th exoneration: that of Michael McAlister, who was sentenced to 35 years in prison for attempted rape in Virginia in 1986, and exonerated last week.

Mr. McAlister was misidentified as the perpetrator of one of many sexual assaults committed by serial rapist Norman Derr, to whom he bore a striking resemblance. Police and prosecutors had expressed doubts about McAlister’s guilt since before his conviction, but it took 29 years and the threat of life imprisonment as sexual predator to secure his exoneration.

Norman Derr is also the real criminal behind another exoneration, that of Jerry Lee Jenkins, who was falsely convicted of a rape Derr committed in Maryland in 1987. Jenkins was exonerated by DNA in 2013. Derr is one of three criminals we know of who committed crimes that led to false convictions and eventual exonerations of multiple defendants in unrelated cases. The other two are Eddie Mosley (see Frank Lee Smith and Jerry Frank Townsend), and Aaron Doxie III (see Julius Ruffin and Arthur Whitfield).

Michael McAlister is on the left, Norman Derr on the right.


​The Chinese criminal justice system continues to abuse suspects and convict innocent defendants, despite President Xi Jinping's public commitment to make the "rule of law" a high priority for political reform. A recent article in The New York Times describes the routine use of torture to force suspects to confess. Many of them are permanently crippled and disfigured. Even when the true culprits are identified and convicted, and the innocent defendants eventually released, the authorities rarely admit their errors.


​From 1972 to 1991, Chicago Detective Commander Jon Burge and his detectives systematically tortured African American defendants until they confessed to crimes, often ones they did not commit. On May 6, in an unprecedented move, the Chicago City Council—drawing on the United Nations Convention against Torture—passed a reparations ordinance to recognize and compensate survivors of Burge's torture from a 5.5 million dollar city fund. In addition to compensation, the city will offer psychological counseling, create a public monument for the torture victims, and teach Chicago Public Schools students about this “ugly chapter” in the city’s history.


​The FBI and the Department of Justice have acknowledged that 26 out of 28 examiners from the FBI Laboratory Microscopic Hair Analysis Unit gave inaccurate testimony that favored prosecutors in more than 95% of 268 trials that have been reviewed from a period lasting more than two decades. Defendants were sentenced to death in 32 of the trials; 14 have been executed or have since died in prison. University of Virginia law professor Brandon L. Garrett called the forensic failure a “mass disaster." The Washington Post reports: "The findings likely scratch the surface. The FBI said as of mid-April that reviews of about 350 trial testimonies and 900 lab reports are nearly complete, with about 1,200 cases remaining."


​In response to growing criticism of the reliability of bite mark identification, the American Board of Forensic Odontology (ABFO) recently conducted a study to try to demonstrate that human bite marks are unique, and that human skin is capable of accurately capturing those unique indentations. The study, however, showed no such thing. Thirty-nine experts in Forensic Odontology were asked to review pictures that might or might not show human bite marks, and they disagreed in their interpretations in the vast majority of the cases. The ABFO is redesigning the study and plans to redo it a second time with a different decision tree. Experts in forensic evidence have criticized the ABFO's decision not to publish the results of the initial study. For example, Case Western Law Professor Paul Gianelli said, “If this were truly a science-based organization, I would not only expect them to be extremely troubled by the results of this study, I would expect them to want to publish the results.”


​In an unusual turn of events Lane Tolbert, who prosecuted exoneree Frank Sealie, reopened the investigation shortly after Sealie was sentenced to life in prison, and secured the exoneration within ten months. According to Tolbert, "There was something about this one that stuck with me," and made him worry about the conviction. Tolbert, who describes himself as tough on violent crime, explained his doubts:  "If I know in other cases [that] inadmissible evidence points to the defendant I can know I was doing the right thing... Nothing else [except one eyewitness] pointed to Frank Sealie."


​In recent years, Shaken Baby Syndrome (SBS), now called “Abusive Head Trauma,” has come under fire as an invalid medical diagnosis—and yet it remains the basis of many prosecutions and convictions for child abuse, felony assault and homicide. The Washington Post, in an in-depth multi-part series examines the deep divide within the medical and criminal justice communities on the validity of this diagnosis.

The series includes an explanation and a history of this diagnosis, including visual aids; a survey of biomechanical testing on crash dummies designed to simulate babies; case studies, including interviews with parents and other caregivers who were prosecuted for supposedly injuring  or killing infants in their care; interviews with doctors and forensic scientists from both sides of the aisle, including some who have changed their minds over time in light of new evidence; and national data on almost 2,000 SBS prosecutions since 2001, including more than 200 that ended in dismissal, acquittal or exoneration after conviction.
The Washington Post series is a must-read for anybody who is interested in the issue. In addition, PBS recently did a segment on Shaken Baby Syndrome that is also useful for people hoping to learn more.

Currently there are 13 shaken baby exoneration cases in the Registry.


​A.M. "Marty" Stroud III,  the lead prosecutor who sent exoneree Glenn Ford to death row in Louisiana, publicly apologized in a long letter to the editor of the Shreveport Times on March 8. He explains, "I did not hide evidence, I simply did not seriously consider that sufficient information may have been out there that could have led to a different conclusion. And that omission is on me." Stroud urges Louisiana to pay Ford compensation for the 30 years he spent wrongfully imprisoned.

Other prosecutors have apologized for false convictions. Several weeks ago, for example Mark Larson, the Chief Criminal Deputy of the King County Prosecutor's Office in Washington state, wrote an impressive description of the exoneration of Brandon Olebar. But Mr.Straud's letter, coming from the very prosecutor who obtained Ford's death sentence, is uniquely powerful and moving.


A recent series in the Washington Post discusses the dangers of bite mark identification, a pseudo-scientific procedure that contributed to the false convictions of at least 17 exonerees. 

The scientific community has been highly critical of bite mark analysis. In 2009, the National Academy of Sciences found that it lacked a "scientific basis.” Nonetheless, practitioners of bite mark analysis cling to their practices, and their professional organization – American Board of Forensic Ondontology – has responded to research pointing out the flaws in the field by trying to discredit individual researchers. 

Worse, bite mark identification is still embraced as reliable science in courtrooms across the country. Some courts go so far as to say that bite mark analysis has reached a sufficient level of "general acceptability" that it should be admitted in evidence without a Frye or Daubert hearing to test its reliability. 

Unlike fingerprint or tool-mark identification, the question is not how to make bite mark analysis more accurate, but whether it should be used at all. Science points to no, but many judges say otherwise.

1 - 10  more Next