The clinical diagnostic process invokes unvalidated general-causation theory (shaking) as an explanation for clinical findings in infants. These medical findings (subdural haemorrhage, retinal haemorrhage, and encephalopathy) are non-specific and develop in natural diseases and accidents. Yet child protection teams associate these findings with abuse. Such ‘diagnosis’ of abuse, triggers social service and law enforcement intervention. Outside the clinical system, which errs on the side of child safety, the SBS/AHT general-causation theories have been challenged. Biomechanical, neuropathological, and forensic pathology research disputes the validity of the shaking theory. Medical ethicists and epidemiologists question the clinical reliance upon data and studies limited by circular reasoning and case selection bias. While ‘child abuse’ may be a ‘valid diagnosis’ for triggering social service intervention, it is not a scientifically sound diagnosis. Lacking foundational validity and support in the relevant scientific fields, SBS/AHT lacks reliability and general acceptance. Expert opinion of SBS/AHT general causation theory is inadmissible under a Daubert or Frye analysis.

This is a brief presentation of the evidence from a systematic literature review of the diagnostic accuracy in suspected traumatic shaking. The national and international reaction to this systematic literature review is also addressed, along with rebuttal of the criticism and an interpretation of the hostile reception of the review. We argue that despite the fact that a scientific controversy often includes competing theories about mechanisms, the shaken baby controversy also includes a controversy about correlation knowledge, because its function is to corroborate (or falsify) the applied theories about mechanisms. Moreover, we argue that long personal experience and groupthink within child protection teams have influenced the development of biased gold standards, resulting in turn in circular reasoning: hence most of the shaken baby literature is flawed.