This article provides new findings from a national study involving 18 forensic interview sites of 137 children who were randomly assigned to a four or eight session extended evaluation. Cases assigned to the eight session protocol were significantly more likely to be classified "credible disclosure" of sexual abuse (56.6%) than cases assigned to the four session protocol (29.5%) and significantly less likely to be classified "credible nondisclosure" of sexual abuse (9.2%) than cases in the four session protocol (24.6%). When four versus eight sessions, demographic variables, and case characteristics were entered into a regression, variables that predicted likelihood of sexual abuse were eight session protocol, older victim age, and caretaker belief the child had been sexually abused. When new disclosures were examined by session in the eight session protocol, 95% of new disclosures occurred by the sixth session. (Contains 3 notes and 7 tables.)