I have the strong impression that many of these problems can be avoided by using bayesian statistics and including reasonable prior information (possibly hierarchical). See Gelman on this topic. I suppose that this illustrates that prior information is especially important in more complex fields.
I have the strong impression that many of these problems can be avoided by using bayesian statistics and including reasonable prior information (possibly hierarchical). See Gelman on this topic. I suppose that this illustrates that prior information is especially important in more complex fields.