We learn the importance of stories through theory and practice. We focus on cases where a news outlet observes the realization of a global situation and must decide whether to report the realization to a consumer who pays the opportunity cost to use the report. The buyer’s best reporting probability is monotone to the fair value rule, a statistical measure of the amount of “news” in the buyer’s relative past fulfillment. We show that a specific scoring rule derived from the statistical literature unambiguously captures important patterns in possible reporting across several US television news sites. We argue that the batting rule can serve as an alternative control in settings where a researcher wishes to examine bias in news reporting. Scoring significantly reduces the appearance of bias in our applications.
That’s a new paper from Luis Armona, Matthew Gentzkow, Emir Kamenica, and Jesse M. Shapiro. I take this to mean that the real bias is towards surprising news rather than bad news per se? By Paul Novosad.
Source link