Tell Me Everything You Discussed: Children's Memory for Dyadic Conversations after a 1-Week or a 3-Week Delay

In child abuse investigations, children are often asked to recount previous conversations related to the allegations (i.e., "conversational testimony"). To explore children's ability to provide conversational testimony, we staged a semi-structured novel dyadic conversation between an adult researcher and 8-year-old children (n = 90). Children's gist recall and recognition memory for their own statements, their conversational partner's statements, and question-answer pairs were tested after either a 1-week or a 3-week delay. The results revealed that children recounted a minority of the conversation, although children recalled more after a short delay (7%) than after a long delay (4%). A majority of children's free recall statements were accurate (68%); however, approximately one-third of their free recall statements were incorrect. Children almost exclusively recounted their own statements, and rarely recalled any of the adult's statements or the question-answer pairs during free recall. Reports of the adult's statements and question-answer pairs increased with cued recall questioning, but remained minimal. During recognition testing, children were able to distinguish between true and false recognition items for their own statements and the adult's statements, but performed at chance level on recognition items concerning question-answer pairs. Forensic implications of the results are discussed.