Shaken baby syndrome/abusive head trauma struggles to find a solid scientific foundation, largely because of a circularity confound, the same diagnostic features under study serve as the basis for categorising cases for research purposes. In an attempt to overcome this circularity, researchers have turned to confessions as a sorting criterion in the research, under the belief that confessions are independent of the diagnostic features and hence not subject to circularity. However, none of the research examines the nature and reliability of the confessions, or the interrogations that produce them; they simply accept the confessions as true, reliable, and independent. Research on interrogations and false confessions, however, along with extensive and wholly consistent anecdotal evidence, strongly suggest that SBS/AHT confessions are largely if not entirely produced by interrogator reliance on the diagnostic findings. That reliance undermines both the independence of the confessions, and hence their ability to break free from circularity, and the reliability of these confessions as a group. On the current state of knowledge, confessions cannot be relied upon to substitute for science to support the SBS/AHT hypothesis.