Abstract:

System dynamics modeling is one way to apprehend complex problems and design effective decision policies. The first task of the modeling process is conceptualization. This requires modelers to develop an understanding of what the problem is, find out which aspects may be relevant, gather information like time series data and natural language explanations, classify behavior modes of variables and construct a tentative endogenous (causal) explanation of these modes. It is a complex task, but poorly supported by learning activities and materials.

However, there is a lack of educational materials. Available textbooks suppose readers to have a sufficient background in quantitative methods, and university courses are mainly at the graduate level. Many self-guided learners use textbooks and downloadable materials, and then elaborate flawed personal conceptualization tactics, leading to flawed models.

To make steps towards an appropriate learner support, the author has developed a series of tasks descriptions, procedures, decision rules, data formats and explanations based on the available literature and personal experience. Tasks are decomposed into procedures, and procedures call up decision rules. Information needed or generated is stored according to the data formats, and any modeling concept referenced in a task, procedure, rule, or data definition, is based on an explanation. This allows a novice to find guidance just at the level of detail where her own skills are not yet sufficient: a complete novice needs the rules and explanations to carry out the procedures which implement the tasks. An advanced beginner may carry out model conceptualization by just following the procedures, and then advance to only follow the task declarations.

Each of these elements is like a node in a conceptual network. The resulting network has been implemented as interactive web platform. The platform supports navigation and search in a highly usable format and tracks users/students’ navigation. Students can therefore obtain feedback on their own learning progress: if the frequency of visits to a certain type of node (say, rules) decreases, this is a trace of learning. This will hopefully reinforce a sense of autonomy in learners. Similarly, lecturers can monitor the progress of their class.

This practical modeling guide "MoNo-guide" (for "Modeling Novice" or "Modelador Novato" in Spanish) has been launched in November 2022 in Spanish (for students in Chile, but it is designed to operate in English, too), and it is free to use. At the current stage of development, there are many limitations. However, SD modeling is one of several styles of computational modeling and as such, it also plays a role in teaching mathematical modeling. At the conference, the network of nodes and the guide will be presented, so that interested lecturers and researchers can engage in an ongoing dialogue, and hopefully further this line of work.

Keywords:

Problem solving, Computational modelling, Model conceptualization, Self-guided learning.