I study how to measure and rank information transmission when agents have misspecified priors, misinterpret signals, or update beliefs in a non-Bayesian way. I axiomatize and characterize measures of meaningful information transmission that account for both the objective features of a signal structure and agents’ (mis)interpretations of it. Meaningful information measures coincide with canonical ones under correct Bayesian beliefs, but otherwise extend them in a disciplined way grounded in the same underlying decision geometry. I then connect this class of measures to the robust order over misinterpreted experiments from Morris (1991) and derive a new equivalent portfolio representation of that order. Finally, in an empirical application to political correctness on college campuses that leverages the experiment in Braghieri (2024), I show that canonical and meaningful information measures and orders can differ sharply.

For futher information please contact: erika.somma@unibocconi.it