By using these undirected prompts, we hoped to experimentally alter the proportion of children
providing explanations while also generating enough variability in the content of children’s
responses within each condition to permit a comparison of children who did and did not explain
regardless of experimental prompt.