A Critical Analysis of Punishment in Public Goods Games

Garrison W. Greenwood, Hussein Abbass, Eleni Petraki

Research output: A Conference proceeding or a Chapter in BookConference contributionpeer-review

2 Citations (Scopus)

Abstract

Social dilemmas arise whenever individuals must choose between their own self-interests or the welfare of a group. Economic games such as Public Goods Games (PGG) provide a framework for studying human behavior in social dilemmas. Cooperators put their self-interests aside for the group benefit while defectors free ride by putting their self-interests first. Punishment has been shown to be an effective mechanism for countering free riding in both model-based and human PGG experiments. But researchers always assume, since this punishment is costly to the punisher, it must be altruistic. In this study we show costly punishment in a PGG has nothing to do with altruism. Replicator dynamics are used to evolve strategies in a PGG. Our results show even a minority of punishers can improve cooperation levels in a population if the cooperators who punish are trustworthy. Finally, we argue punishment as a strategy in social dilemmas is never altruistic.

Original languageEnglish
Title of host publicationProceedings of the 2018 IEEE Conference on Computational Intelligence and Games, CIG 2018
PublisherIEEE, Institute of Electrical and Electronics Engineers
Pages1-5
Number of pages5
ISBN (Electronic)9781538643594
ISBN (Print)9781538643594
DOIs
Publication statusPublished - 2018
Event14th IEEE Conference on Computational Intelligence and Games - Maastricht, Netherlands
Duration: 14 Aug 201817 Aug 2018

Publication series

NameIEEE Conference on Computatonal Intelligence and Games, CIG
Volume2018-August

Conference

Conference14th IEEE Conference on Computational Intelligence and Games
Abbreviated titleCIG 2018
Country/TerritoryNetherlands
CityMaastricht
Period14/08/1817/08/18

Fingerprint

Dive into the research topics of 'A Critical Analysis of Punishment in Public Goods Games'. Together they form a unique fingerprint.

Cite this