Abstract

Much of the scientific process involves “selective ignorance”: we include certain aspects of the systems we are considering and ignore others. This is inherent in the models that we utilize as proxies for biological systems. Our goal usually is to isolate components of these systems and consider them at only certain temporal and spatial scales. The scales and questions induce different metrics for what might be considered a “good” model. The study of mathematical and computational models is replete with differing views of the terms verification, validation, corroboration, and so on. I have often argued that criteria for determination of model utility should be established prior to model construction, but this is rarely done in the application of models in biology. The question I address is whether it is feasible to develop a general approach to model evaluation, that includes all the forms of models typically applied in biology—animal and cell/tissue culture ones as well as mathematical and computational ones