The participants were 44 children ages 4-6 who were assigned as participants or observers for a modified version of the game Simon Says, in which they took part with a confederate. The children were told before the game began that it was a memory game. The children's memory was assessed afterwards by (1) the Step-wise Interview; (2) the Step-wise Interview and big dolls; or (3) the Step-wise Interview and small, detailed dolls and props. The interviews included three leading questions. Results revealed no main effect of interview type on the overall amount or accuracy of the children's accounts. No main effects for the interview type or the status as participant versus observer were observed for the leading questions. Five-year-olds remembered more details and were more accurate in their accounts than were the 4-year-olds. Females were more accurate than males in their accounts with the small detailed toys and props. Findings revealed that the dolls did not compromise the memory of children overall and that the use of anatomically neutral dolls may be a viable alternative in situations that do not support the use of anatomically detailed dolls. However, the study was exploratory and further research is needed. Tables and 34 references (Author abstract modified)