Full Program »
On Goodness of FitWhat does it mean for a model to fit some data, and what does it mean for a model to make accurate predictions? In both cases, we face the same problem: to provide reasons for assessing the distance between theory and evidence. Call this the goodness of fit problem. In this paper, I argue that solving the goodness of fit problem requires an appeal to pragmatic criteria, and therefore, pragmatic encroachment in estimation and model selection is unavoidable. I then briefly argue for the claim that science is pragmatically encroached.
University of Illinois at Urbana-Champaign