Francesco Bogliacino (Università Cattolica del Sacro Cuore)

Measuring Norms: A Comparison of the Predictive and Descriptive Power of Three Methods

Abstract: We analyze and compare three methods of measuring norms: the Krupka and Weber (KW) coordination game, the two-step approach by Bicchieri and Xiao (BX), and a novel Binarized Scoring Method (BSM) we introduce that elicits the full distribution of normative beliefs. We test their effectiveness in two distinctive ways. First, we compare the fit and predictive power of the norms elicited by each method in 3 versions of the dictator game, which differ in how the pie is initially allocated. Then we use vignettes to assess the extent to which the methods can recover existing norms for various naturally occurring settings. We find that the KW method yields better predictive power within a norm-dependent utility model. All 3 methods effectively recover norms in field settings, although KW is more robust to false positives.