Computational Issues in the Statistical Design and Analysis of Experimental Games

Publisher Website
Google Scholar
Abstract
One goal of experimental economics is to provide data to identify models that best describe the behavior of experimental subjects and, more generally, human economic behavior. We discuss here what we think are the three main steps required to make experimen tal investigations of economic games as statistically informative as possible: finding the solution of the ex perimental game under the postulated equilibrium or other economic models, selecting from a potential class of experimental designs the optimal one for dis criminating between those models, and choosing an optimal stopping rule that indicates when to stop sam pling data and accept one model as the best explana tion of the data. Each step can be computationally in tensive. We offer an algorithmic presentation of the necessary computations in each of the three steps and illustrate these procedures by examples from our re search on learning models in experimental games with incomplete information. These three steps of experi mental design and analysis are not limited to experi mental games, but the computational burden of imple menting these algorithms in other experimental envi ronments—for example, market experiments—requires further considerations with which we have not dealt.

This publication has 8 references indexed in Scilit: