Abstract:In physics education research (PER), for many years now we have focused our attentions on finding ways to improve our instruction and have achieved some notable successes. In this paper, we suggest that the time has come to embed this activity in a more complete and scientific view of PER, one that builds a coherent understanding of the system of teaching and learning in addition to improving the practice of our instruction. We outline five broad topics of interest for PER and discuss questions that need to be addressed in each topic over the next few years. The topics are: the model of the participants, the model of the contexts, the model of the content, the engineering of instruction, and the overall epistemology of PER -How do we decide when we think we know something?