Rebound effect studies are useful for policy making, as they indicate the extent to which increases in energy efficiency lead to lower energy savings than those predicted by engineering calculations. Existing rebound studies assume energy consumption changes arise from an economically rational response, by consumers, to cheaper energy services. This misses the possibility that energy efficiency increases could have effects on wider society, organization and material-human relationships which lead to increases in energy consumption for reasons other than economic utility. This paper argues that setting rebound studies within a properly worked out social theory would give a fuller picture of these causal routes. Noting that ‘practice theories’ drawn from Schatzki are being used increasingly in energy consumption studies, the paper critically expounds Schatzki’s development of practice theory and human-material ‘arrangements’, offering this as a framework for rebound studies. It then uses a small case study, of growth in computer use in a university research cluster, to compare rebound effect results obtained by the traditional approach and the Schatzkian approach. The latter gives rebounds up to 100% higher than the former. The paper recommends that policy take these wider causal determinants into account when estimating likely effects of energy efficiency increases.