Abstract
Modern statistics, building on decision theory, does not explicitly take into account incentive problems and strategic behavior in the collection and analysis of data. In this talk I overview recent game-theoretic models that explicitly incorporate conflicts of interest between researchers and decision makers. I also discuss implications for the organization and regulation of data collection and analysis. The talk is largely based on the following recent papers:
Strategic Sample Selection with A. Di Tillio and P.N. Sorensen (2021 Econometrica),
Research and Approval Process: The Organization of Persuasion with E. Henry (2019, American Economic Review), and
P-Hacking in Clinical Trials and How Incentives Shape the Distribution of Results across Phases with J. Adda and C. Decker (2000 PNAS).