johannes.buchner.acad [ät] gmx.com
http://arxiv.org/abs/1402.0004
Is this interpretation allowed?
Problems with
true model $\rightarrow$ assigned model
red: wabs $\rightarrow$ powerlaw
red: powerlaw $\rightarrow$ wabs
If you answer these questions, you are doing calibrated Bayes.
Using non-Bayesian methods (simulations, visualisation, p-values).