Decision Making versus Testing: A Changing of the Guards in Empirical Research?

Abstract

      In this paper we will elaborate and give support to Manski's central message  in “Statistical decision theory should supplant hypothesis tesing" (Manksi [15]; see also Manski [16]). Decisions should take the place of testing. This review should benefit practioners in empirical research of econometrics and in statistics in general. We emphasize thatWald's Statistical Decision Theory should be used in empirical research instead of “traditional" hypothesis testing (both frequentist and Bayesian testing), since hypothesis testing is only a one-sized-fits-all way to make decisions. The goal of empirical research is to make optimal decisions, and it is possible to make decisions  without using testing.