On a frigid evening in January 2018, Brian Wharton was sitting at home in Hawkins, a small town in the backcountry of east Texas, when there was a knock on the door. A woman introduced herself as a lawyer working to spare the life of Robert Roberson, a death row inmate who had come within four days of execution.
Wharton immediately understood the significance of this visit. He was a retired detective from the police department in Palestine, another small town about 80 miles away, and in 2002 he had been involved in a case that stuck in his memory.
On 31 January that year, Wharton was called to Palestine hospital after receiving reports of a mortally ill child. Her father, Roberson, had brought his two-year-old daughter Nikki into the ER in a comatose state, the little girl lying limp in his arms and turning blue. She died the following day.
Wharton led the investigation. He arrested Roberson hours after the girl died, even before an autopsy had been performed, having been advised by doctors that this was a case of “shaken baby syndrome” – the theory that the child had died from violent shaking that caused fatal brain injuries.
Wharton went on to testify at Roberson’s capital trial. The jury convicted Roberson in February 2003 and sent him – now branded Prisoner 999442 – to death row.
Wharton had been troubled by key aspects of the case from the start. Why had the doctors and nurses seized on shaken baby syndrome as the cause of Nikki’s injuries almost immediately, and why with such certainty? Why had a nurse told him she had found indicators of sexual assault on Nikki’s body when he could see no such evidence?
Why, when he escorted Roberson back to his house on the day of Nikki’s collapse, could he find no signs of a violent altercation? Why were there no marks or indentations on the walls, no holes in the sheetrock, nothing – and no blood anywhere except a few drops on a washcloth Roberson gave him?
Above all, why had Roberson himself acted so strangely? As a police officer, Wharton had dealt with both perpetrators and victims of violent crimes, but Roberson’s behavior fitted the pattern of neither criminal nor victim. He came across as flat, emotionless, matter-of-fact, as though nothing was registering with him.
All these things bugged the former detective to the point that he felt haunted by them. So when the doorbell rang, and Roberson’s lawyer stood before him, Wharton knew exactly what to do.
“Come in,” he said. “I’ve been waiting for you.”
•••
On Tuesday, the US supreme court will convene in Washington for what is known as a “long conference”, preceding the start of its 2023-24 term. The justices will meet in private to sift through the backlog of cases that have amassed over the summer, tossing out those they are disinclined to consider.
The lawyer Gretchen Sween will be waiting on tenterhooks for news of which petitions the court has discarded. Among the huge pile of requests awaiting the justices’ attention is one she prepared for her client: Roberson.
“We’ll be watching anxiously to see if we make it through that gauntlet,” Sween told the Guardian.
Roberson’s supreme court appeal could be his last shot at judicial review. Should the justices reject his approaches, the state of Texas would be free to set a new execution date.
Roberson has asked the court to review the decision of the Texas state courts to deny a retrial despite new evidence that has emerged casting doubt on his conviction. Since Sween took over as his attorney in 2016 she has discovered that:
Nikki, who shortly before she died had been placed in Roberson’s custody, was chronically sick almost from birth. She suffered bouts of breathing apnea which caused her suddenly to stop breathing and collapse. Shortly before her father rushed her to hospital, she had had diarrhea for five days and had a fever of 104.5F (40.3C).
Nikki also had severe undiagnosed pneumonia.
She had been prescribed medications, including an opioid, that are no longer considered safe for children as they can cause fatal breathing problems and oxygen deprivation.
The sexual assault allegation was unsupported by any evidence, and the nurse who raised it was unqualified to identify it.
Then there is the overarching question of shaken baby syndrome. The petition asks the supreme court to weigh whether a death sentence based largely on a scientific hunch presented to the jury as fact should be allowed to stand when, in the intervening years, that theory has been largely discredited.
In other words: should Texas be allowed to execute a man on the basis of junk science?
“This is a shocking, social failure,” Sween said. “This is not simply a matter of having the wrong guy, it’s a matter of there having been no crime committed in the first place.”
•••
Texas houses its 179 male death row inmates in the Allen B Polunsky Unit, a complex of windowless grey boxes resembling a Walmart supercenter shrouded in barbed wire.
Roberson, 56, was waiting in the secure metal cubicle assigned for media visits. He was wiping smudges off the bulletproof glass, an attention to detail that seemed incongruous given the circumstances.
He was dressed in the all-white onesie that is death row uniform in Texas. From the first exchange, his highly unusual way of speaking was evident.
“Thank you for for for for for, ah ah, this interview, sir sir,” he said, blinking fast behind his glasses, in time with his stutter.
That was a clue as to why medical staff were so quick to suspect Roberson when they encountered him, and why Wharton found his behaviour so singular. Long after his capital trial – too late for the jury to factor it in – Roberson was diagnosed with autism.
Roberson described his life of poverty at the time he began looking after Nikki. Separated from the girl’s mother, and living with a new partner, he worked three newspaper delivery routes to try to support them.
“After gas and stuff, you don’t make too much from a paper route,” he said.
Asked what he hoped for Nikki when he took custody of her, he said: “I was hoping to raise her, you know, to be a father to her, to raise her to have a good life and stuff, and for us all to be happy. To provide for her best I can, to nurse her, to love her, to pray for her.”
Roberson said Nikki was sick from the day she joined him, recalling that they had to visit the doctor’s surgery because she was so ill. On the morning of 31 January 2002, he said he found she had fallen out of bed and was lying face down on the floor.
“She had a little bit of blood on her lip, I wiped it off on a towel,” he said.
He lifted her back on the bed and they went back to sleep. Then he described what happened after he awoke a second time.
“When I got back up she wasn’t breathing and I listen to her heart, weren’t breathing. I tried to wake her up, her heart was still beating but she wasn’t breathing.”
He added: “Only thing I could think was she might have fell off and might have hit her head, but I didn’t see it, so I don’t know.”
He drove Nikki to the ER and handed her to the medical staff.
“They lifted the little girl, and she’s limp and started turning blue, and stuff.”
He said he could tell that the nurses and doctors suspected him straight away of doing something bad to Nikki. During the trial, the prosecution said he came across as cold and uncaring because he didn’t cry, but he said that was because he had trained himself not to cry during his difficult childhood.
“It’s hard for me to cry because I got my behind whooped by my father who would say, ‘Shut up or dry it up or we’ll give you some more.’ My dad was kind of abusive, you know.”
Roberson was taken into custody within hours of Nikki’s death. He remains there more than 21 years later.
I asked him how he felt about being convicted of shaking his daughter to death.
“I would never think about shaking her,” he replied. “And that’s God’s honest truth. I don’t know what happened to her, Mr Pilkington. I wouldn’t want that to be on nobody: to lose a child, especially if you tried to do right and you loved her and tried to get to know her, then to be accused.”
Roberson said he knew nothing about the medical hypothesis that fueled his prosecution until long after he was arrested. He read about it in a brochure someone gave his mother.
“I’d never heard about shaken baby in my life,” he said.
•••
Shaken baby syndrome (SBS) began life in the early 1970s as a medical hypothesis for why some children become severely ill, some fatally, showing signs of internal head trauma but little or no signs of external trauma. One of the first to float the idea was a British pediatric neurosurgeon, Norman Guthkelch, who in 1971 suggested the cause might be violent shaking.
By the 1980s, SBS had developed into a detailed theory focusing on three symptoms, known as the triad, that it was believed could indicate abusive shaking: bleeding between the tissue layers covering the brain, swelling on the brain, and bleeding at the back of the eyes. When that triad was seen, even without any outward sign of force, child abuse might be behind it.
By the 1990s, a form of groupthink had crept in, medical practitioners firming up what had begun as a mere hunch into a formal diagnosis. The collective thinking now was that where those three symptoms were seen in young children, not only could abusive shaking be presumed, it had to be taken as the cause.
In 2001, the American Academy of Pediatrics, the world’s leading pediatric medical organization, issued official guidance. It said that where a young child had internal head injuries there was “the need for a presumption of child abuse” – a categoric finding that rapidly became medical, and then forensic dogma.
So it was that a year later, when Roberson brought Nikki to ER, he was met by a cordon of medical staff trained in this thinking, with detectives and prosecutors following their lead.
“When we got to Palestine hospital, the conversation was about shaken baby,” Wharton recalled, “so for me and my detectives that’s the direction we were going.”
The problem was that the presumption that the triad must be caused by violent shaking had no scientific foundations. As a new book of essays, Shaken Baby Syndrome, published by Cambridge University Press and edited by prominent scientists and law professors in the US, UK, Europe and Japan, points out, shaking is essentially a biomechanical issue pediatric doctors know little about.
One of the first biomechanical studies to be carried out into SBS was conducted in 1987, almost 20 years after the hypothesis emerged. It found that violent shaking produces a force inside the head well below that involved in impact injuries such as car crashes, and much weaker than child abuse experts had assumed.
The new book’s conclusion is that 50 years after Guthkelch’s hunch, “this hypothesis still lacks a scientific evidence base”.
In 1997, shaken baby syndrome became a global sensation with the televised trial of a British nanny, Louise Woodward, in Massachusetts. She was prosecuted following the death of the eight-month-old in her care, Matthew Eappen.
Woodward was convicted of second-degree murder and sentenced to life, later reduced to involuntary manslaughter and time served. Her trial was so high-profile that it brought to public attention serious doubts about the credibility of SBS raised by the nanny’s defense attorney, the co-founder of the Innocence Project, Barry Scheck.
Since Roberson was sentenced to death in 2002, those doubts have steadily grown. Dr Patrick Barnes, a key prosecution witness in the Woodward trial, retracted his testimony and became a vocal critic of the syndrome.
Even Guthkelch, its creator, came out in 2012 and said his work had been distorted. He never intended, he said, to suggest that the three symptoms could be assumed to be shaken baby syndrome in the absence of any other evidence of abuse.
“There was not a vestige of proof when the name was suggested that shaking, and nothing else, causes the triad,” he said.
Over time those medical doubts have filtered through to the courtroom. Since 1993 there have been 32 exonerations in the US following SBS convictions, according to the National Registry of Exonerations.
This week, an appeals court in New Jersey ruled that SBS was “junk science” and “scientifically unreliable” and could not be used by prosecutors at trial.
In Ohio, Alan Butts was released in December after 19 years in prison having been granted a new trial. His story is uncannily similar to Roberson’s: he was accused of the “shaking baby” death of his girlfriend’s two-year-old son, Jadyn, in 2002, the year Nikki died.
New evidence showed Jadyn had pneumonia at the time of his death – just like Nikki. The toddler was on medication no longer given to children because it can cause sudden death – just like Nikki.
Katherine Judson is a lawyer who has represented Butts for several years. She is also executive director of the Center for Integrity in Forensic Sciences, which has filed a brief supporting Roberson in his supreme court petition.
In Judson’s view, the Butts and Roberson narratives are so remarkably akin that they provide an excellent example of “flukey justice”.
“You can have the exact same facts in a different jurisdiction,” she said, “and one person gets released while the other person languishes on death row.”
•••
In recent years, the Innocence Project has become increasingly involved in exposing junk science. The group has focused on a range of forensic techniques that in their application have put innocent people at risk.
They include bitemark analysis, under which it was believed erroneously that a bite on a corpse’s skin could be matched to a suspect’s dentition; hair microscopy, which claimed falsely that a single hair from a crime scene could be linked visually to an individual; and fire science, which suggested that analysis of a burn pattern could determine whether a fire was arson – until it was discovered that intentional and accidental fires leave the same imprints.
The Innocence Project has joined Roberson’s legal team. Vanessa Potkin, director of special litigation, said the group had decided to fight for the prisoner’s life because “an innocent man is at serious risk of being executed”.
She added: “Like so many faulty forensics that have played a role in the nation’s exonerations, the evidence upon which his conviction and death sentence rest has been disproven since the time of his trial.”
Paradoxically, Texas is a leader in countering junk science. In 2013, the state introduced a first-in-the-nation “junk science writ” that allowed prisoners – especially those on death row – to challenge sentences on grounds of misused forensic science. It was under this law that in 2016 Sween saved Roberson from imminent death by securing a stay of execution four days before his scheduled lethal injection.
But the hope generated by the new junk science law in Texas has proven a chimera. There have been about 70 attempts by death row inmates to utilize the law and of those the number that have obtained relief is zero.
Kosoul Chanthakoummane was one of those who appealed through the junk science law. He had been put on death row on the back of three different types of junk science: hypnosis of a witness to obtain identification, bitemark analysis and a discredited form of DNA testimony.
In August 2022, Texas executed him anyway.
There is a real possibility that that fate now awaits Roberson. Having secured a stay of execution, he was granted an evidentiary hearing at his trial court. Sween presented a mass of new evidence. Expert witnesses were called to speak about the growing consensus that SBS is junk science.
The court heard that Nikki’s symptoms could all be explained by natural rather than violent causes, including her pneumonia and chronic illness and the life-threatening medication. It was also told about Roberson’s autism, and the unfounded nature of the sexual assault allegation.
Despite the voluminous testimony, a state judge denied the appeal. She said shaken baby syndrome remained a reasonable diagnosis and that Roberson “failed to provide any newly discovered evidence” undermining his conviction.
In January, the Texas court of criminal appeals affirmed that decision.
The state is now actively lobbying the supreme court justices to persuade them not to take Roberson’s case. Its argument cites testimony from Teddie, Roberson’s estranged girlfriend, who said he had a bad temper and would yell at Nikki and hit her.
Roberson’s legal team counter that Teddie was not a credible witness. Her own sister testified for Roberson, insisting she had never seen him mistreat Nikki and telling the jury Teddie had a problem with telling the truth.
Brian Wharton, who after retiring from law enforcement retrained as a cleric and is now pastor of a United Methodist church situated 10 miles away from death row, testified at the junk science hearing. He told the court that, drawing on his 14 years’ experience as a police officer, he had come to the view that shaken baby syndrome was a fallacy and that without it there was no evidence supporting Roberson’s death sentence.
“What case would we have been able to make?” Wharton told the Guardian. “There was no crime scene, no forensic evidence. It was just three words: shaken baby syndrome. Without them, he would be a free man today.”
Roberson said the worst part of his fight to stay alive was not knowing. Will the supreme court take his petition, or will they not?
He has had so many knock-backs that he knows the worst could happen. But he says he’s ready for whatever comes. He recently participated in the first faith-based program on Texas’s death row and said it had deepened his religious belief.
Roberson has a tattoo on his right arm which he inked himself when he was still a child, aged 17. That was many years before he grew up and became a father to Nikki.
“F T W,” it says.
The Guardian asked what it stood for. He replied: “Then? Or now?”
“Then it was ‘F The World’. That’s how I felt when I was a kid. But I changed, and now it’s ‘Follow The Way’.”
That will be his guiding star for whatever lies ahead.
“I hope and pray that God gives them the knowledge for the people to make a righteous decision,” Roberson said. “I know I didn’t do it. I’m not guilty. So I’m at peace with the Lord.”
