Shaken baby syndrome has been discredited. Why is Robert Roberson still on death row?

By Jeff Kukucka and David Faigman

In a last minute effort to save the life of a man on death rowa bipartisan group of Texas lawmakers just did something extraordinary: They did it unanimously quoted Robert Robersonconvicted in 2003 of murdering his daughter based on the now-discredited theory of shaken baby syndrome, to testify before them for five days after he was scheduled to be executed, effectively forcing the state to keep him alive.

Roberson is one of several people who have been jailed for injuries to a child that prosecutors say resulted from violent shaking. But research has revealed serious flaws in those determinations and dozens of other defendants who were wrongfully convicted under that theory. they were exonerated. Yet Roberson remains on death row, even as politicians, scientists and others — including the lead detective who investigated him and one of the jurors who convicted him — have spoken out on his behalf. If his execution goes ahead, they and many others believe Texas will be killing an innocent man for a “crime” that never happened.

As our scientific understanding of shaken baby syndrome has evolved over the past 20 years, justice is demanding that courts reexamine old convictions. in the light of new findings. This is especially true for Roberson, who would be the first person in the US to be executed for a conviction based on shaken baby syndrome. Regardless of one’s view of the death penalty, the ultimate punishment must be held to the ultimate standard of proof — and Roberson’s case falls short of that standard.

About supporting scientific journalism

If you like this article, please consider supporting our award-winning journalism by subscription. By purchasing a subscription, you help ensure the future of impactful stories about the discoveries and ideas that shape our world today.

The theory behind shaken baby syndrome dates back to the early 1970s, when two medical researchers – Norman Guthkelch and John Caffey – separately published the first scientific papers explaining that shaking an infant can cause fatal internal injuries even in the absence of external injuries. Over time, doctors and law enforcement officials, among others, began to widely rely on a triad of symptoms—cerebral bleeding, brain swelling, and retinal bleeding—as definitive proof that someone had abused a trembling child. To support this theory, the researchers cited cases where a child showed these symptoms and a caretaker testified to shaking the child, which apparently confirmed the triad as a reliable way to diagnose abuse.

There is no doubt that shaking a baby can cause injuries, including those that comprise the triad of shaken baby syndrome. Newer research, however, has shown that shaking is not only way of causing those injuries: they can also result from an accidental “short fall” (eg falling off the bed) as well as other medical causes (eg pneumonia, wrong medication) – all of which are valid for Roberson’s daughter. In fact, a 2024 study found that the lesions historically used to diagnose tremors are actually More likely to result from accidents than from shaking. In short, modern science understands that the presence of these symptoms does not necessarily mean that a child has been abused, nor does their absence mean that they have not.

Why have clinicians mistakenly trusted this triad of symptoms for so long? The short answer is that correcting misconceptions requires a feedback loop MEAN often lacking in child abuse investigations. When a physician diagnoses a living adult and prescribes a treatment, the effectiveness of that treatment provides feedback on the correctness of his diagnosis; if treatment proves ineffective, doctors can learn from this misdiagnosis and adjust future diagnoses accordingly. However, such feedback is not always sufficient; for example, doctors practiced bloodletting for centuries because it was generally accepted and seemed to work for some patients, even though it was an illusory correlation. As for shaking, doctors rarely learn if a child has been actually shaken because the child is usually deceased or unable to articulate what happened, and thus doctors rarely receive feedback that the triad has led to an incorrect diagnosis.

In terms of studies that used a caregiver’s confession to establish that abuse had occurred, it is now known that innocent people sometimes confess to the crimes they did not commit, so the confession are is not synonymous with truth. Some scholars have even argued that the unique circumstances of suspected shaking cases (eg, the emotional state of the suspects) create a particularly high risk of false confession.

Complicating matters further, child abuse determinations are subject cognitive biaswhere extraneous information leads experts to interpret the same injury in different ways—at least one of which must be incorrect. In one study, for example, medical professionals rated the same childhood injury more often as abuse rather than an accident if she is told that the child’s parents are unmarried or drug users — both appear to be true for Roberson. Another study found that the same extraneous factors drove emergency room doctors misdiagnose accidental injuries as abuse in 83 percent of cases.

Even just knowing about a criminal charge can affect how a doctor evaluates a child’s injuries. In one study, independent experts reviewed medical records from cases where, unknowingly, a fellow expert had testified that the child had been shaken. In 94 percent of these cases, independent experts concluded that “the child’s head injuries … possibly, or even probably, had a non-traumatic cause.”

Autopsy decisions are also unreliable. In a 2021 study, opinions of medical examiners whether a child’s death was an accident or homicide was strongly influenced by the child’s race and who brought the child to the hospital, even if the child’s injuries and history were otherwise identical. In response, leading medical examiners explained that manner of death “is not a ‘scientific’ determination” and “often does not hold up well in court.” However, jurors—including some at Roberson’s trial—often hear and trust these weak views, which has led some scholars to support that mode of death testimony. should not be admissible in US courts, as is the case in almost every other country.

As research debunking shaken baby syndrome has grown, so have successful legal challenges to criminal convictions based on it, including another case from Texas where — just eight days before Roberson’s scheduled execution — a man was granted a new trial on the grounds that “scientific knowledge has evolved” since his 2004 trial and would “probably result in an acquittal” in 2024. Before his death in 2016, even Guthkelch—one of the theory’s architects—lamented that “his friendly suggestion to avoid harming children became an excuse for imprisoning innocent parents.” Roberson is one of those innocent parents.

Science is constantly evolving, and when it reveals a mistake of the past, we don’t just resign ourselves to it; we are taking corrective action. Our legal system should be no different. When Robert Roberson was convicted, the triad of wounds was widely accepted as evidence of shaking, but as science has progressed, this is no longer the case. The legal guarantee of due process must accommodate such progress, especially when a person’s life literally depends on it. For the law to ignore evolving scientific knowledge is not only unfair; it is criminal.

This is an opinion and analysis article, and the views expressed by the author or authors are not necessarily theirs american scientist