Courts in many Western countries are increasingly willing to accept children as competent witnesses, thus increasing the number of children providing testimony. It is thus essential that forensic investigators and others responsible for eliciting information from children have the skills and knowledge necessary to draw out the most accurate and detailed accounts possible. The current study probed the effects of 4 variations on the original Step-Wise Interview protocol on the quality of 141 young children’s eyewitness recall in 2 age groups: those 3 to 5.5 years and those 5.5 to 8 years. The four variations tested the effects of direct versus indirect probes and rapport-building using past versus general events. Participants observed a performance of a story in the classroom or day care setting and were subsequently interviewed about the story 1 to 2 weeks later. It was hypothesized that more accurate statements would be elicited using direct as opposed to indirect probes and that rapport-building talk of previous events would increase the accuracy of information provided by the children. It was also expected that the older children would provide more accurate eyewitness accounts. Results indicated that older children reported more information about the event and provided more accurate information than did the younger children. When direct probes were used, younger children provided as much information as did older children, but provided less information when indirect probes were used. The lowest eyewitness accuracy provided by younger children was observed when they were asked direct probes and when they had practice recalling a past event through rapport-building. Future research should probe the current findings regarding rapport-building.