Best-practice forensic interview protocols recommend that interviews with children should include a practice narrative. While the benefits of practice narratives have been consistently demonstrated in non-Aboriginal populations, particular interaction features in many Aboriginal communities may affect how the interview technique applies to Aboriginal children. Our study aimed to examine the effectiveness of practice narratives on the memory reports of 64 Aboriginal children (6 to 15 years) from three remote Australian communities. Children participated in a staged 30-minute innocuous event, and were interviewed one to two days later by experienced interviewers (half with a practice narrative and half without). Logistic and multiple linear regressions demonstrated that practice narratives did not predict the accuracy or informativeness of Aboriginal children’s subsequent accounts. Unexpectedly, results revealed that girls in our sample produced more words and target details, as well as fewer confabulations compared to the boys. The implications of these findings for forensic interviews with Aboriginal children are discussed.