In this paper we describe a custody case that centred on allegations of child sexual abuse. A pair of preschool-aged sisters accused their biological father of various sexual acts, though their allegations were made under problematic conditions and were contradicted by other statements they made. In an affidavit written by one of us (MB), we describe the relevant memory development research in the course of presenting the court with a scientific analysis. We find compelling evidence of multiple risk factors in the way the daughters' recollections were elicited. Although the “ultimate question” of guilt is beyond our purview, our identification of risks was instrumental in the legal system's decision that the children's allegations were not valid. We put this analysis forward as an example of evidence-based testimony in which scientific findings from the memory literature can be used to frame an expert's analysis.