--- - |2 Summary Children aged 6–8 (N = 84) were interviewed 1 week after participating in a repeated event. Half received training in labeling episodes of a repeated autobiographical event (Label Training); remaining children practiced talking about the same without label training (Standard Practice). Subsequently, children recalled the target event in two recall order conditions: script for the events followed by a specific instance (Generic‐first) or the reverse (Episodic‐first). Training effects were modest, but the research has important implications for interviewers' elicitation of children's labels for instances of repeated events because 98% of the labels generated were unique. The study provides additional support for the notion that recalling the script first can be beneficial. Children in the Generic‐first condition were more accurate for some types of details, and reported more information in the first half of the interview about details that changed across instances, than children in the Episodic‐first condition. - Applied Cognitive Psychology, Volume 32, Issue 5, Page 600-609, September/October 2018.