Abstract
Strong reciprocity explains prosocial cooperation by the presence of individuals who incur costs to help those who helped them (‘strong positive reciprocity’) and to punish those who wronged them (‘strong negative reciprocity’). Theories of social preferences predict that in contrast to ‘strong reciprocators’, self-regarding people cooperate and punish only if there are sufficient future benefits. Here, we test this prediction in a two-stage design. First, participants are classified according to their disposition towards strong positive reciprocity as either dispositional conditional cooperators (DCC) or dispositional free riders (DFR). Participants then play a one-shot public goods game, either with or without punishment. As expected, DFR cooperate only when punishment is possible, whereas DCC cooperate without punishment. Surprisingly, dispositions towards strong positive reciprocity are unrelated to strong negative reciprocity: punishment by DCC and DFR is practically identical. The ‘burden of cooperation’ is thus carried by a larger set of individuals than previously assumed.
Highlights
MethodsWe conducted the experiments between November 2012 and February 2014
The aim of study 1 was to provide a direct test of the often-invoked assumption that strong positive reciprocity and strong negative reciprocity are linked: strong reciprocators are necessarily DCC and may punish if others contribute less than them; DFR never punish, and cooperate only if there is a threat of punishment
The main result of study 1 is that the individual tendency to punish defectors is independent of dispositions towards strong positive reciprocity: the punishment behaviour of DCC and DFR is virtually indistinguishable; both punish those who contribute less than them to a similar degree
Summary
We conducted the experiments between November 2012 and February 2014. The sample size was determined in expectation of heterogeneity in cooperative dispositions and to ensure a minimum number of participants from each type. We recruited at least 90 participants in each treatment to ensure a minimum expected number of 20 DFR. The experiment was approved by the University of Nottingham School of Economics Ethics Committee and informed consent was obtained from all participants. The experiment was computerised with z-Tree[66]. Each session consisted of reading the instructions, computerised control questions, two experimental games and a questionnaire. The control questions were designed to check participants’ understanding of the games’ payoff functions. Participants had to correctly answer all control questions before the start of the experimental games.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
