Results indicate that simulated practice interviews involving trained postgraduate students were more effective compared to fellow participation in reinforcing interviewers' use of open-ended questions. Although the performance of all participants improved with practice, the beneficial effect of having trained actors play the role of a child was robust. It was evident irrespective of whether performance was measured using the absolute number of proportion of open-ended questions, or the number of open-ended questions asked prior to the first specific question. The benefit of using trained actors was also recognized by the participants, whose self-ratings of their ability to maintain open-ended questions was higher in the actor when compared to the fellow-participant practice conditions. The importance of finding a superior practice effect in the trained-actor condition at the 12-week followup speaks volumes; innovations are successful when participants are tested immediately after training, however, the effect often dissipates when performance is assessed outside the training context, and after a long post-training retention interval. In the current study, the actor-trained participants maintained better performance over time, even though they had not engaged in post-training practice. The differences between the proportions of open-ended questions across the practice sessions were numerically larger at the 12-week followup than the immediate post-test suggesting there may have been greater positive transfer of learning to the new post-criterion interview context among the actor-trained interviewers. Participants were 50 professionals who were employed by a State child protection service in Australia. Tables and references