The exposure and diffusion of energy consumption information in building occupant peer networks has been shown to influence an individual's energy consumption decisions. In this paper, we develop an agent-based computational model for individual energy consumption behavior based on data collected during an experiment on residential energy use. We simulate the building occupants' decision making and the information transmission process. By comparing the impact of several parameters in the network level computational model and validating the parameters in a second experimental setting, our research serves to clarify how network relations can be leveraged for modifying energy consumption behavior. Network degree and weight were identified as the major structural parameters that impact building occupants' conservation decisions, while network size was found to have no significant impact. These findings have important implications for the design and effectiveness of residential energy feedback systems designed to promote energy conservation in residential buildings. (C) 2011 Elsevier B.V. All rights reserved.