Robert Roberson Is Going to Die Because Courts Aren't Built for Science Questions
For Robert Roberson, a sea change in the science hasn't set him free.
Much like the Melissa Lucio case, Robert Roberson awaits execution for a crime that may have never happened. Roberson’s case differs in two key respects: the state is actively trying to set an execution date, and his innocence case centers not on subsequent medical opinion but on a collapse in the theory used to convict him.
“Shaken Baby Syndrome:” Intuitive But Wrong
When pediatricians first named “shaken baby syndrome,“ they ensured a long life for the diagnosis the same way emergency physicians did with “excited delirium:“ a vague name seemingly describing an intuitive pathology. Babies and toddlers famously wield comparatively large heads on comparatively weak necks and spines; of course throttling one can cause serious whiplash injuries.
The syndrome, however, refers to a more specific pattern of brain injury. Without the neck or spinal strength to lessen the force at which they’re being throttled, shaken baby syndrome posits that this whiplash leads to veins shearing off the skull, leading to subdural hematoma (a bleed between the brain and skull). What’s more, according to early proponents of the diagnosis, certain anatomical findings could conclusively suggest the syndrome, notwithstanding any other evidence. According to the original theory, if a patient between less than two years old shows retinal hemorrhage, subdural hematoma, and hypoxemic encephalopathy,1 investigators can be completely confident that they died by violent shaking.
This was the medico-legal consensus when Robert Roberson brought his daughter to an ER in 2002. Though investigators noted “no sign of violence“ at the home, an MRI revealed the supposedly definitive signs, and Roberson was arrested for her murder before pathologists could even complete an autopsy. Roberson was convicted and sentenced to death in 2003.
Changing Science Doesn’t Always Change Sentences
Since then, however, medical understanding of the diagnosis has completely shifted. Biomechanical analyses found that a throttling sufficient to cause injuries attributed to shaken baby syndrome would almost certainly cause neck injury, a symptom rarely found in the diagnosis. And while proponents relied on the idea that ground-level falls were almost never lethal, pathologist John Plunkett, found 18 cases in an 11-year period. 12 of these cases exhibited a “lucid interval” between the accident and onset of symptoms, undermining prosecutors’ reliance on witness testimony that shaken baby victims acted normally beforehand (as Roberson’s in-laws had done). Meanwhile, scientists found an increased role for prenatal Vitamin D deficiency in babies’ bone formation, suggesting that nutritional shortfalls (as a British pathologist found more than 30 cases) could be responsible for a number of misdiagnoses. While The American Academy of Pediatrics stuck by its guns2 throughout most of these developments, it finally relented (somewhat) in 2017 with a new position renaming the diagnosis “abusive head trauma,“ a have-your-cake-and-eat-it-too phrase allowing them to incorporate their critics’ findings without retracting their original claims. The evolving science—which contributed to 31 exonerations3—was enough to grant Roberson a stay of execution just five days before his original scheduled date. Indeed, the state of Texas itself conceded the faults in the science presented at Roberson’s trial in a separate action.
“Shaken Baby“ Cases Prove the State’s Unfitness for Judging Medical Questions
Once convicted, however, defendants in cases of faulty science run up against an anti-scientific defect in the system.
Legal principles dictate a certain deference to the trial court’s decision. For “evergreen“ questions like, say, Brady violations, that’s pretty reasonable; the court system would be hopelessly bogged down (more than usual) if the losers were entitled to endless rehearings of the facts. With scientific and medical questions, however, it means a reversal of the logic by which science moves.
Scientific reasoning runs on the idea that, given the deeper evidence base from which it develops, new research tends to be more reliable. Accordingly, when a Conviction Integrity Unit (a team of prosecutors that looks for flaws in their predecessors’ work) in Tennessee reevaluated Russell Maze’s life sentence for the 2004 supposedly shaken-baby death of his son Alex, they determined that the cause of death would likely not hold up to contemporary scientific scrutiny. The judge, however, held that this new understanding—testified to by seven doctors and the case’s lead detective—was no more than a “different opinion,” upholding Maze’s conviction and sentence. The theory that prevailed at trial effectively holds more weight because it was given first—the exact opposite of the logic by which that theory is supposed to operate.
Roberson ran into a similar problem. In appeals, his team compiled a 17-page brief challenging the science by which he was convicted; the state responded with a few instances in which the possibility of “blunt force trauma” was raised at trial—with the state’s pathologist making clear that the shaken baby syndrome theory would qualify as such—to argue that the diagnosis was not the basis of the conviction. The Texas Court of Criminal Appeals “rubberstamped“ the prosecution’s theory, holding that since the retooled theory is still “a recognized diagnosis,“ the sea change in its understanding since the trial amounts to merely a “battle of the experts.”
With the Supreme Court denying his final petition, Roberson’s last hope is with Governor Greg Abbott and the state’s Board of Pardons and Paroles. Though he has some unexpected support (the supervising detective on the investigation is pleading for his life), neither the Board nor the governor are known for generous views on clemency. Roberson could die as early as October because Texas can’t admit it was wrong, and the system is disinclined to make them.
Respectively: bleeding from the eyes, the aforementioned brain bleed, and brain dysfunction from a lack of oxygen.
The 2014 documentary The Syndrome tracks the capture of the AAP’s child abuse committee by “shaken baby syndrome“ proponents.
31 cases in the National Registry of Exonerations use the “SBS“ tag; I did not delve into what role new science on the topic played in each case.
I understand the perspective of the court re: putting more weight on the first analysis instead of opening the system up to endless appeals. But you're so right that science (and opinions that flow from it) requires the latitude to self-correct. Peer review and evolution of theories over time what makes science research strong. Just like a politician whose views evolve over time with better information is better than one who digs in on a position despite its foundation being disproved, placing being right the first time over accepting that they might have been wrong.
The death penalty is cruel and barbaric. It's also way more expensive than housing someone for life.