Open Access
Translator Disclaimer
November 2016 Contextuality of Misspecification and Data-Dependent Losses
Peter Grünwald
Statist. Sci. 31(4): 495-498 (November 2016). DOI: 10.1214/16-STS561


We elaborate on Watson and Holmes’ observation that misspecification is contextual: a model that is wrong can still be adequate in one prediction context, yet grossly inadequate in another. One can incorporate such phenomena by adopting a generalized posterior, in which the likelihood is multiplied by an exponentiated loss. We argue that Watson and Holmes’ characterization of such generalized posteriors does not really explain their good practical performance, and we provide an alternative explanation which suggests a further extension of the method.


Download Citation

Peter Grünwald. "Contextuality of Misspecification and Data-Dependent Losses." Statist. Sci. 31 (4) 495 - 498, November 2016.


Published: November 2016
First available in Project Euclid: 19 January 2017

zbMATH: 06946239
MathSciNet: MR3598727
Digital Object Identifier: 10.1214/16-STS561

Rights: Copyright © 2016 Institute of Mathematical Statistics


Vol.31 • No. 4 • November 2016
Back to Top