Shutting down social preferences in the lab: an experimenter demand approach
How to cite
Semb, Joakim. 2019. "Shutting down social preferences in the lab: an experimenter demand approach." AEA RCT Registry. AEARCTR-0004647. https://doi.org/10.1257/rct.4647-1.1
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
Social preferences pose a challenge for any lab testing of theoretical predictions that are based on unobservable vNM utilities. Depending on the theory being tested, the unobservable mapping of laboratory payoffs to vNM utilities will, at best, introduce considerable noise, and at worst, bias the results. I investigate the efficacy of a very straightforward method for reducing the influence of social preferences on decision-making in the lab: instructing subjects to behave selfishly. The question is a simple one: does maximization of one's own payoff better predict subjects' behavior in experiments when they are explicitly instructed to achieve this goal?
To answer this, a first experiment will be conducted in which subjects play a one-shot prisoner's dilemma with one of three treatments: (1) social framing of actions, (2) neutral framing of actions, and (3) neutral framing of actions with explicit instructions to behave selfishly.
Experimental design
Pen-and-paper experiment in which subjects play a one-shot prisoner's dilemma with one of three treatments: (1) social framing of actions, (2) neutral framing of actions, and (3) neutral framing of actions with explicit instructions to behave selfishly (see enclosed instructions for exact phrasing). After choosing their own action, subjects state their beliefs about the distribution of actions among other participants. Payment is based on payoff in the PD and on the accuracy of beliefs.
Subjects also answer an exit survey composed mainly of likert scale items about motives behind action choice.
Design details
Cited by
No citation data available for this registration yet.