Anders Bouwer, Vania Bessa Machado and Bert Bredeweg. 2002.
Interactive Model Building Environments.
(In book) The Role of Communication in Learning to Model,
(eds.) P. Brna, M. Baker, K. Stenning and A Tiberghien,
pages 155-182, Lawrence Erlbaum Associates, London, UK.
(Final draft to publisher: PDF)
Abstract
Interactive model building environments are educational software systems which support scientific investigation by allowing students to build models and experiment with them. The basic idea is that students use the model building environment to construct their own model of a scientific domain, which the computer can use to run simulations, thereby providing predictions and causal explanations. The models are qualitative models, which explicitly represent the domain knowledge required for causal reasoning about system structure and behaviour. If a valid model has been constructed or assembled, a simulation can be started, and its progress is controlled by the student. To communicate the contents of the model and the simulation results, these are represented graphically as state-transition diagrams, causal networks, structural hierarchies, tables and graphs. Students can navigate through these interactive visualisations to get an overview, check specific predictions, search for explanations, investigate trends and look into details of the simulation. In this way, model building environments can support two learning goals: learning how to model scientific phenomena, and learning by making and checking predictions and explanations with the simulation model. Since the approach taken is domain-independent, the same model-building environment can be used for domains varying from ecology to physics. This article discusses two parts of such a model building environment that have been implemented as separate prototypes. One prototype is VISIGARP, a model inspection tool that supports learners in controlling and investigating simulation models. The second prototype is MOBUM, a model building tool that support learners in constructing simulation models. Integration of both types of systems will lead towards fully interactive model building environments, allowing students to participate in the process of scientific investigation to a larger extent than traditionally possible.