Interpreting model results — examples from an energy model

With the increasing use of complex computer models for high-level policy decisionmaking, the problem of correctly interpreting and communicating model results becomes a more general concern. This paper traces misconceptions about the use of models to the existence of different conceptions of the term “model.” Policy models are quite often less theory-based than models in the traditional disciplines, especially in cases where the policy models deal with the long-term developments of sociotechnical systems. The authors examine the use of an example of one such model. Generalising from the authors' experiences in other fields of application, e.g., global modeling, the problems of interpreting model results are discussed. The proper use of future-oriented policy models is clarified by the introduction of typologies implying distinctions, e.g., between forecasting, “what-if,” and learning models, and between different “levels” of results, viz. model outcomes, model inferences and policy-issue oriented interpretations.