This research study examined whether and how professional development can help teachers design sequences of instruction that lead to improved science learning. The efficacy of three professional development programs and a control condition was compared in a cluster randomized trial involving 53 middle school science teachers from a single school district. The four conditions varied along two dimensions: (a) the extent to which the programs guided teachers' selection of curriculum materials and (b) whether or not teachers received explicit instruction in models of teaching associated with particular methods for designing instruction. Results indicated that the two programs most effective at improving students' science learning were the ones in which teachers received explicit instruction in models of teaching. (Contains 4 notes, 8 tables, and 1 figure.)