System Dynamics Modeling: Validation for Quality Assurance

The etymological root of "valid" is in the Latin word "validus", which denotes attributes such as strong, powerful and firm. A valid model, then, is well-founded and difficult to reject because it accurately represents the perceived real system which it is sup-posed to reflect. This system can be either one that already exists or one that is be-ing constructed, or even anticipated, by a modeler or a group of modelers. The validation standards in System Dynamics are more rigorous than those of many other methodologies. Let us distinguish between two types of mathematical models, which are fundamentally different: Causal, theory-like models and non-causal, statistical (correlational) models [Barlas & Carpenter 1990]. The former are explanatory, i.e., they embody theory about the functioning of a real system. The latter are descriptive and express observed associations among different elements of a real system. System dynamics models are causal models. Non-causal models are tested globally, in that the statistical fit between model and data series from the real system under study is assessed. If the fit is satisfactory, the model is considered to be accurate ("valid", "true"). In contrast, system dynamicists postulate that models be not only right, but right for the right reasons. As the models are made up of causal interdependencies, accuracy is required for each and every variable and relationship. The following principle applies: In case only one component of the model is shown to be wrong, the whole model is rejected even if the over-all model output fits the data [Barlas & Carpenter 1990]. This strict standard is conducive to high-quality modeling practice. A model is an abstract version of a perceived reality. Simulation is a way of experimenting with mathematical models to gain insights and to employ these to improve the real system under study. It is often said that System Dynamics models should portray problems or issues, not systems. This statement must be interpreted in the sense that one should not try to set the boundaries of the model too wide, but rather give the model a focus by concentrating on an object in accordance with the specific purpose of the model. In a narrower definition, even an issue or problem can be conceived of as a "system", i.e., "a portion of the world sufficiently well defined to be the subject of study" [Rapoport 1954]. Validity then consists in a stringent correspondence between model system and real system. We will treat the issue of model validation as a means of assuring high-quality models. We interject that validity is not the only criterion of model quality, the other criteria being parsimony, ease-of-use, practicality, importance, etc. [Schwaninger & Groesser 2008]. In the following, the epistemological foundations of model validity are reviewed (Chapter II). Then, an overview of the methods for assuring model validity is given (Chapter III). Further, the survey includes an overview of the validation process (Chapter IV) and our final conclusions (Chapter V).