A research assistant blind to the hypothesis transcribed each trial starting at the experimenter’s target questions (e.g., ‘‘Who is afraid of raccoons?”) and ending when the experimenter started the next trial. This removed any reference to the condition from the transcript so that coders would be blind to the condition. An independent coder and the first author coded children’s responses from all trials, and disagreements were resolved through discussion. Recordings from 15 individual trials were unclear and could not be transcribed or coded, although the remaining trials for the affected chil-dren were retained and coded. An additional 3 trials were also excluded because the experimenter made errors in the follow-up questions, and again the remaining trials from these children were retained and coded. Thus, of 340 total trials, responses from 322 were coded.