The role of models in measurement outside the laboratory

For measurement outside the laboratory, uncertainty evaluations have to be model-based. Therefore models should not only represent the measurand but also its environment. For measurement in social science these models will be too complex, therefore instead of white-box models, grey-box (modular-designed) models are preferred. They are much simpler. Uncertainty evaluations of measurements outside the laboratory are more of the kind of Type B evaluations. In social science, due to the lack of authoritative scientific institutions, Type B evaluations will be of a different character. In case of measurements based on grey-box models they are the so-called structure-oriented behaviour tests. (c) 2013 Elsevier Ltd. All rights reserved.