In my own approach to modelling, I try to bear three questions in mind: A) Does the approach generate sufficient confidence to lead to application?; B) Does the application produce improvement in planning/decision making?; C) Does the gain from this improvement justify the costs of the modelling exercise?. These are essentially practical questions in which the idea of validity looms large and I am grateful to Faivre & Sanchez for developing their notion of validity. I think they have captured the essence of validity in their description of it as "the adequacy of the model as a mimic of the system which it is intended to represent".
- This could also be of interest