--- - |2+ Purpose Prosecutors working with child sexual abuse (CSA) cases involving young children have raised concerns that reliability criteria from the Supreme Court of Sweden are holding children's testimony to impossible standards (e.g., expecting the child's testimony to be long, rich in detail and spontaneous). This study aimed to address these concerns by investigating how District Courts and Courts of Appeal employ said criteria in their testimonial assessments of young child complainants. Methods Court documents from District Courts (n = 100) and Courts of Appeal (n = 45) in CSA cases involving 100 children age 7 years and under were analysed with respect to the courts’ testimonial assessments. Results Testimonial assessments were more frequently referenced in acquitting verdicts and in cases with evidence of low corroborative value. Richness in detail was the most frequently used reliability criterion, followed by spontaneity. Most criteria were used in favour of the children's testimony. However, the length criterion was typically used against the reliability of the children's testimony. Conclusions Our findings confirm prosecutors’ concerns that criteria from the Supreme Court are frequently used in evaluations of young children's testimony. This is troublesome, as some criteria do not correspond to current research on young children's witness abilities. For example, compared to testimony given by older children or adults, testimony provided by a young child is typically not long or rich in detail. We encourage prosecutors to extend their own knowledge on young children's capability as witnesses and present this to the court. - 'Legal and Criminological Psychology, Volume 23, Issue 2, Page 176-191, September 2018. '