In classical statistics we trust current data to give us a trustworthy model, with certain assumptions often built in like the normality of data. The model parameters are generated by the data, and once we have the parameters we often assume that they will apply to future samples. Say, for example, that we use a regression-type model to explain ...