How can teacher education programs prepare novices who are ready to provide effective instruction on their first day in the classroom? One potentially promising strategy is focusing teacher preparation on practical classroom skills, rather than educational theory and disciplinary content. In particular, thousands of prospective teachers are using virtual reality simulations as a platform for practicing teaching techniques prior to using them in classrooms. As a result, teacher educators are able to observe prospective teachers in ways that are otherwise difficult to replicate in a university classroom or in student teaching, while facilitating the provision of feedback, with opportunities for “do-overs.” This study employs a randomized control trial to examine to what extent teaching improves in such simulated classrooms, contingent on the structure of feedback prospective teachers receive. A sample of 120 prospective teachers were randomly assigned to one of three modes of feedback: (1) self-reflection without external feedback, (2) traditional coaching following a simulator session, and (3) “bug in the ear” coaching throughout the simulator session. Growth in practice was assessed using multiple methods, both qualitative and quantitative. The talk will highlight the potential value of simulated classrooms for both instruction and research, and detail implications for teacher preparation design. Avoiding the trap of technology for technology’s sake necessitates systematic empirical research into how these technologies are being used, and the degree to which and ways in which the technology supports the purported goals of improving teaching in real classrooms with real children.