Replication Package for: The Experimenters' Dilemma: Inferential Preferences over Populations

This repository contains the replication package for “The Experimenters’ Dilemma: Inferential Preferences over Populations” by Luca Rigotti, Alistair J. Wilson, and Neeraja Gupta, conditionally accepted at Experimental Economics. The paper compares the statistical power per research dollar of five participant pools (lab undergraduates, a virtual lab, Amazon Mechanical Turk, CloudResearch, and Prolific) using one-shot prisoner’s dilemma games. It weighs the lower cost per observation of online samples against their attenuated responses to incentives. The package includes the de-identified raw session exports from oTree (lab, October 2019; virtual lab, September 2021) and Qualtrics (MTurk, August 2020; Prolific, February–March and June 2021; CloudResearch, June 2022) with session timestamps. It also includes the experimental materials (oTree source code, Qualtrics survey files, and experimenter protocols), a Stata cleaning script that builds the analysis datasets, and a Jupyter notebook (Python, calling Stata) that reproduces the tables, figures, and numbers reported in the main text and the online appendix. Literature benchmarks from Charness, Rigotti, and Rustichini (2016) and Gächter, Lee, Sefton, and Weber (2024) enter the notebook as published treatment-level counts.