This study examined the effect of social influence on children's witness reports, with respect to a number of details varying in centrality. Children (N = 115; age = 10 years, 4 months to 13 years, 8 months) were interviewed about a personally experienced event. Half of the children were interviewed together with a confederate who answered the interview questions before the child did, while the other half were interviewed alone. Children were influenced by the confederate's answers to withhold some critical details observed (omission errors), but not to add details not observed (commission errors). When the children were asked to follow up on their reports, truthful reports contained more information than did false reports.