Model Transparency in Educational System Dynamics

Model transparency is increasingly identified as a positive or even necessary characteristic of system-dynamics-based learning environments, where model transparency is usually identified as providing modified causal-loop diagrams, equations, or verbal descriptions of a model. The theses of this presentation are: (1) Model transparency may be beneficial for some educational goals and conditions, but model opacity may be beneficial for others. (2) Model transparency is a continuum (from transparent to opaque) and is multidimensional (for different aspects of a model, such as its variables, stock-flow combinations, and cause-effect relationships). (3) There are many methods of providing information about a model, and these too will depend on the goals and other characteristics of a learning environment. Rather than seeking to prove that model transparency is valuable, system dynamics researchers should be elaborating on how goals and other conditions determine optimal levels and methods of transparency.