One hundred twenty-eight 5- to 7-year-old children were interviewed using the National Institute of Child Health and Human Development (NICHD) Investigative Interview Protocol about an event staged 4 to 6 weeks earlier. Children were prepared for talking about the investigated event using either an invitational or directive style of prompting, with or without additional practice describing experienced events. The open invitation prompts (including those using children’s words to encourage further reporting) elicited more detailed responses than the more focused directive prompts without reducing accuracy. Children were most responsive when they had received preparation that included practice describing experienced events in response to invitation prompts. Overall, children were highly accurate regardless of prompt type. Errors mostly related to peripheral rather than central information and were more likely to be elicited by directive or yes/no questions than by invitations. Children who provided accounts when asked about a false event were less accurate when describing the true event. Children who received preparation that included practice recalling a recent event in response to directive and yes/no questions were least accurate when questioned about the false event first. The data provide the first direct evaluation of the accuracy of information elicited using different prompt types in the course of NICHD Protocol interviews, and underscore the importance of how children are prepared for subsequent reporting.