Social (P)references: A Study Of Peer-Dependent Reference Points
How to cite
Eisfeld, Thomas. 2023. "Social (P)references: A Study Of Peer-Dependent Reference Points." AEA RCT Registry. AEARCTR-0011107. https://doi.org/10.1257/rct.11107-1.2
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
In contrast to standard economic theory, many economic experiments report peer effects on individual choices under risk, suggesting individuals have peer-dependent risk preferences. At the same time, tools to discuss peer-dependent risk preferences are limited and often arbitrary. Based on a model of peer-dependent reference points, we conduct two experiments -- a peer experiment and a control experiment. Building on a within-subjects design, participants will face Multiple Price Lists (MPLs) while merely observing what another participant (Peer) earns. We test peer effects on individual choice under risk in settings where ranks can and cannot reverse. We conduct another experiment to control for other confounding factors that could drive participant behavior.
Experimental design
There will be a peer experiment and a control experiment. In the peer experiment, participants will be randomly assigned a role: either a Decision Maker (DM) or a Peer. DMs will then be randomly paired with a peer and will face Multiple Price Lists (MPLs) where they can choose between a risky prospect and a safe amount of money. While completing these MPLs, DMs will see the earnings of their paired Peer on the screen. Peers will always earn a fixed amount of money. Hence, peer earnings act as a mere cue. Since we plan to test for differences in individual risk preferences between situations where social ranks can change and situations where social ranks are fixed, we will implement three treatment rounds in the experiment by varying Peer earnings: one, where DMs are ensured to be better off than their Peer, one where ranks can reverse (i.e., where Peer earnings fall in between what DMs can earn with the risky prospects), and one where DMs are ensured to be always worse off than their Peer. The order of the treatment rounds will be randomized. At the end, earnings from one round of the treatment stage will be picked at random to determine the final payoff that participants will received (together with a show-up fee). To reduce costs while allowing for sufficient flexibility in terms of design for each experimental session, two DMs will be randomly paired with one peer.
To account for other confounding factors (such as anchoring effects), we plan to run a control experiment. The control experiment will replicate the peer experiment, with the exception that Peers will be absent. To implement this, we will tell participant that at the beginning of the control experiment, they could have earned counterfactual earnings. We design these counterfactual earnings to match the peer earnings from the peer experiment and show them to participants in combination with the same MPLs from the peer experiment. Comparing the peer experiment with the control experiment then allows us to isolate peer effects on individual risk preferences.
Design details
Cited by
No citation data available for this registration yet.