Doubts Over Shaken Baby Syndrome Often Don’t Delay Death Sentences
Since 1992, courts in at least 17 states and the Army have exonerated 32 people convicted in shaken baby syndrome cases based on new studies about the syndrome, according to the National Registry of Exonerations.
One was 41-year-old Mary Weaver from Marshalltown, Iowa, who was babysitting on the morning of Jan. 22,1993, for 11-month-old Melissa Mathes. At one point, the child stopped breathing and was taken to a hospital where she died on the next day. Weaver was convicted of murder in 1994 and was sentenced to life in prison without parole. Weaver was exonerated in 1997 after testimony was presented debunking shaken baby syndrome as the cause of death.
Another case involved Julie Baumer, 27, of White, Mich. In October 2003, Baumer brought her 6-week-old nephew to the hospital after she noticed that he was lethargic, fussy, and unwilling to eat. The child died at the hospital, allegedly because Baumer had shaken the baby enough to cause lethal injuries. Baumer was sentenced to 10 to 15 years in prison for child abuse. In 2010, the conviction was overturned on the grounds that the results of shaken baby syndrome are disputed. Baumer was set free.
The findings in the 32 cases, in part, led the state of Texas in 2013 to enact Article 11.073, commonly called the “junk science writ.” California, Connecticut, Michigan, Nevada and Wyoming have since passed similar “junk science” statutes.
The law was designed to open a path for someone to challenge their conviction if there is new scientific evidence that was unavailable at the time of the trial. At the time the Texas law was enacted, proponents said it was a solution to wrongful convictions based on faulty science.
In practice, the Texas Common Court of Appeals has not granted relief under the junk science writ to a single person sentenced to death. The appeals court has decided 25 junk science applications filed by people facing the death penalty since September 2013, which constitutes 34 percent of all applications filed.
Of the 25 applications filed by people sentenced to death, 18 were dismissed or denied without review because the court determined they were procedurally barred. Most, 64 percent, were dismissed or denied by a page long order with no substantive information.
“The deadly consequences of this pattern are clear: People may be executed following convictions that rest on faulty science because they are unable to obtain relief under 11.073 (junk science writ). This is especially concerning because the rate of wrongful convictions of death-sentenced people is quite high,” according to a report from the non-profit, Texas Defender Service.
Since 1973, 200 people have been exonerated from U.S. death rows including 18 from Texas. A 2014 study found that only seven applications filed by death-sentenced applicants were ever reviewed on the merits in Texas.
Texas Defender reported that the junk science writ “recognizes that many scientific disciplines, at one time regularly used to secure convictions, have been found to be riddled with errors.” The group said disciplines include fields as varied as blood spatter evidence, forensic pathology, fire debris investigation, shoeprints, bitemark evidence, firearms identification, gunshot residue, and DNA. Despite advances in various areas of forensics, new DNA evidence is the only consistently reliable type of evidence upon which the Texas appellate court has granted relief. Of the 3,479 exonerations in the U.S. since 1989, only 594 involved DNA evidence.
“This is problematic because DNA evidence is only a tiny subset of potentially false forensic evidence, and in fact, most wrongful convictions are not based on DNA,” Texas Defender reported.
The junk science writ was meant to expand access for people wrongfully convicted on the basis of false or unreliable forensic evidence. However, Texas Defender found that the Texas Court of Criminal Appeals has shown “a disregard for discredited scientific methods…meaning that potentially innocent people will be executed.”
Other methods that have been greatly improved include advances in proteomics, the study of proteins to gain insights that DNA alone can’t provide, such as cluses about the time of death and the age of the deceased.
Another area of advancement has been in blood pattern analysis software that helps reconstruct crime scenes by analyzing bloodstains, providing information into the sequence of events.
Advanced systems like the Integrated Ballistic Identification System (IBIS) use 3D imaging and comparison algorithms to identify firearms and their components.
Researchers have developed fluorescent carbon dot powders that make latent fingerprints visible under UV light, improving the detection and analysis of fingerprints.
Hand-held chemical analysis devices like the MasSpec Pen can quickly detect and measure compounds to determine the presence of drugs and explosives.
Portable, paper foldscopes allow for immediate analysis of forensic samples. Artificial intelligence and machine learning also are being used to analyze and interpret forensice data faster and more efficiently.
Now, the life of Texas death row prisoner Robert Roberson hangs in the balance as lawmakers consider his junk science writ defense because of new research that has added doubt about the validity of shaken baby syndrome allegations.
At the same time, 1,700 miles away, the N.J. Jersey Supreme Court has been hearing testimony over whether it should be the first state high court to bar trial evidence from doctors that parents killed or injured their children through shaking. Last year the state’s Superior Court Appellate Division upheld a groundbreaking ruling that evidence about shaken baby syndrome was inadmissible because of its unreliability as a diagnosis.
In July 2005, the Court of Appeal of England and Wales heard four appeals of shaken baby syndrome convictions. One case was dropped, the sentence was reduced for one, and two convictions were upheld. The court found that the existence of accepted symptoms, including retinal bleeding, subdural hematoma, and acute encephalopathy are not 100 percent diagnostic of shaken baby syndrome and that clinical history is also important.
Cody Mason is a lawyer with the New Jersey Public Defender Office and the lawyer of a father jailed for years while he fought the testimony of a doctor who said he abused his child. He argued against doctors lumping various kinds of abuse into shaking baby syndrome. The engineer who created the theory said that decades of research has been unable to prove that shaking alone can cause death.
Roberson, who is on the autism spectrum, was scheduled to be executed last Thursday for the 2002 killing of his 2-year-old daughter Nikki Curtis. Roberson has lost appeals to the state Board of Pardons and Paroles and the U.S. Supreme Court.
The U.S. Supreme Court declined to take the case on legal grounds. Justice Sonia Sotomayor said that she believed there were significant questions about the evidence about shaken baby syndrome in Roberson’s case. She said the execution should be halted to prevent a “miscarriage of justice from occurring.”
Roberson’s execution has been temporarily delayed by the Texas Supreme Court while the Legislature hears testimony over whether outdated scientific evidence led to a faulty conviction and that the child actually died from complications with severe pneumonia.
Texas Attorney General Ken Paxton and Gov. Greg Abbott have both argued the execution should not be delayed because only the attorney general and not the Legislature can offer even temporary clemency.
The night before Roberson brought her to a Palestine, Texas, emergency room, Nikki had fallen off a bed. The cause of death was determined to have been shaken baby syndrome. Roberson’s lawyer has claimed that the child’s symptoms were misdiagnosed and were more likely the result of days of illness leading up to her death, as well as hospital treatment that may have left her with bruises, such as intubation and emergency resuscitation.
Roberson’s lawyer said the child had double pneumonia that had progressed to sepsis and she had been prescribed two medications now seen as inappropriate for children that would have further hindered her ability to breathe.
Shaken baby syndrome, more commonly referred to as a type of “abusive head trauma,” is accepted as a valid diagnosis by the American Academy of Pediatrics. Abusive head trauma generally occurs when a caregiver violently shakes a child. The U.S. Centers for Disease Control and Prevention (CDC) says it is the leading cause of child abuse deaths in children younger than 5.
Natalie Montfort, a psychologist who works with adults who have autism, reviewed the case and testified that “So much of the evidence that the state presented against Mr. Roberson at trial was about manifestations of his then-undiagnosed autism.”
Shaken baby syndrome was the only evidence presented to the jury. Witnesses in the emergency room had testified that Roberson was odd, uncaring and unemotional when he brought his daughter to the hospital. Such behavior is often typical of people on the autism spectrum and do not necessarily indicate a lack of caring.
“His reactions were misconstrued as evidence of guilt. The jury was told to believe he was callous and remorseless,” Montfort said.
Texas Defender reported that the junk science writ is implemented when the court finds that “had the scientific evidence been presented at trial, on the preponderance of the evidence, the person would not have been convicted.”
The court however has required the evidence to be strong enough to eliminate any rational basis for a conviction, which is the legal innocence standard.
“This places much too high a burden on the people it was designed to help. Innocent people often cannot definitively prove that there is no rational basis for their conviction years or decades after it occurred,” Texas Defender wrote.
“In practice, people seeking relief under 11.073 might need to go beyond proving the State’s reliance on flawed science — they might need to provide evidence affirmatively showing innocence,” Texas Defender wrote. “For most innocent people who were convicted based on problematic forensic evidence, this standard is impossible to meet.”
As a result, the appellate court has not done enough “to consistently protect all people who have been convicted on false and discredited scientific evidence.”
The National Registry of Exonerations listed others who have had their sentences reversed over objections to the shaken baby claims, including:
Derrick Allen of Durham, N,C., was convicted in 1999 of murdering a toddler in his care, 2-year-old Adesha Artis. Allen was sentenced to a term of 43 to 54 years in prison. In 2016, after serving 17 years in prison, Allen’s sentence was exonerated.
A trial court judge ruled that the state had used the threat of the death penalty as leverage to secure a plea from Allen, while at the same time withholding evidence to which he was entitled. The judge also called the lab results regarding Adesha as “deceptively written” and “designed to obscure the fact that confirmatory testing was performed … and yielded negative results.”
Vincente Benavides, 42, of Kern, Calif., was released from California’s death row on April 19, 2018, 25 years after he was sentenced to death for the rape, sodomy, and murder of a 21-month-old girl in Delano, Calif. His convictions were vacated and the charges were dismissed in 2018 after experts concluded that the girl had never been raped or sodomized, and that the medical evidence used to convict Benavides was wrong.
Michael Griffin, 19, of Genessee, Mich., was sentenced to life in prison in 2010 for killing his infant daughter, Naviah Griffin. The conviction was reversed and Griffin was free after serving 13 years in prison.
At 6:58 p.m. on September 30, 2009, Griffin brought his daughter to the emergency room at Hurley Medical Center in Flint, Michigan. Griffin said that Naviah, then seven months old, had fallen out of her motorized swing in an upstairs bedroom. Griffin said he had heard a thud, rushed to the room, and found his daughter on the floor, unresponsive and bleeding from her lip. She died the next morning.
At the time of Griffin’s trial, the science behind Shaken Baby Syndrome was already being questioned, as a growing body of research showed that the markers for the diagnosis — retinal hemorrhaging, brain bleeding, and brain injuries — could be due to illnesses and accidental injuries. Sheldon Siegel, Griffin’s court appointed lawyer, whose law license had been suspended in 2013, failed to present an expert witness for the defense.
Christopher Lyman, 29, of Geary, Kans., was sentenced to life in prison in 2015 for the death of his nephew, Johnathan Swan. The conviction was reversed in 2023, after Lyman had been in prison for eight years.
Lyman had checked on Johnathan and found the child limp and unresponsive. Lyman said later that he tried CPR and heard some gurgling in Johnathan’s lungs. Lyman said that bruises found on the child’s body might have happened when Johnathan and another boy were playing, because sometimes the boys collided. Lyman also said that two weeks earlier, he was carrying Johnathan when he fell down the stairs, but he didn’t think the boy had been injured because Lyman took the brunt of the fall.
Dr. Janet Ophoven, a pediatric forensic pathologist, wrote, “My review indicates that there was a combination of misdiagnosis and a rush to judgment regarding the child’s condition.” Ophoven said the findings that Johnathan’s retinal hemorrhaging, brain swelling, and brain bruising were signs of abuse is a “a controversial determination based on today’s medical science.”
She noted that there are 25 medical conditions that can cause retinal hemorrhaging.
On January 16, 2023, Judge Courtney Boehm of Geary County District Court granted Lyman a new trial, based in part on ineffective assistance of counsel. On July 12, 2023, Lyman’s 39th birthday, the state dismissed the charges with no explanation.