The competitive advantage of sanctioning institutions revisited: A multilab replication

Sergio Lo Iacono; Wojtek Przepiorka; Vincent Buskens; Rense Corten; Marcel van Assen; Arnout van de Rijt

doi:10.1093/pnasnexus/pgad091

The competitive advantage of sanctioning institutions revisited: A multilab replication

PNAS Nexus. 2023 May 2;2(5):pgad091. doi: 10.1093/pnasnexus/pgad091. eCollection 2023 May.

Authors

Sergio Lo Iacono^{1

2}, Wojtek Przepiorka², Vincent Buskens², Rense Corten², Marcel van Assen^{2

3}, Arnout van de Rijt⁴

Affiliations

¹ Department of Sociology, University of Essex, Colchester CO4 3SQ, UK.
² Department of Sociology/ICS, Utrecht University, Utrecht 3584 CH, The Netherlands.
³ Department of Methodology and Statistics, Tilburg University, Tilburg 5037 AB, The Netherlands.
⁴ Department of Political and Social Sciences, European University Institute, Fiesole FI 50014, Italy.

Abstract

Is peer sanctioning a sustainable solution to the problem of human cooperation? We conducted an exact multilab replication (N = 1,008; 7 labs × 12 groups × 12 participants) of an experiment by Gürerk, Irlenbusch, and Rockenbach published in Science in 2006 (Gürerk Ö, Irlenbusch B, Rockenbach B. The competitive advantage of sanctioning institutions. 2006. Science. 312(5770):108-111). In GIR2006 (N = 84; 1 lab × 7 groups × 12 participants), groups that allowed members to reward cooperators and punish defectors were found to outgrow and outperform groups without a peer-sanctioning institution. We find GIR2006 replicated in accordance with all preregistered replication criteria in five of the seven labs we sampled. There, the majority of participants joined groups with a sanctioning institution, and participants cooperated and profited more on average than in groups without a sanctioning institution. In the two other labs, results were weaker but still favored sanctioning institutions. These findings establish the competitive advantage of sanctioning institutions as a robust phenomenon within the European context.