Modeling and simulation can be powerful tools for analyzing multi-agent systems, such as networked robotic systems and sensor networks. In this paper, it is shown concretely how instances of both these elements fit into a general methodology for multi-level modeling, providing insight into system dynamics. Use of the resulting general framework is illustrated through application to a specific sample case study involving a robotic wireless sensor network engaged in an acoustic detection task. We then compare and contrast the resulting family of models, highlighting explicitly the trade-off between realism and simplicity.