Using Equivalent Offsets to Test Reference Dependence: Evidence from Three Experimental Paradigms

How to cite

Wang, Ao. 2025. "Using Equivalent Offsets to Test Reference Dependence: Evidence from Three Experimental Paradigms." AEA RCT Registry. AEARCTR-0015748. https://doi.org/10.1257/rct.15748-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Ao Wang*

* Primary investigator

Registry ID
AEARCTR-0015748
DOI
https://doi.org/10.1257/rct.15748-1.0
First registered
Last updated
Status
In development
Trial dates
2025-06-02 – 2025-07-31
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/15748

Abstract

This paper tests reference dependence by examining whether the effect of a payoff increase on choice behavior is offset by an equivalent increase in the reference point, following the framework of Rees-Jones and Wang (2022). We apply this approach across three experimental paradigms: an effort task, an investment game, and a binary lottery choice experiment.

Experimental design

Below we summarize the main features of the planned experiments. For detailed experimental design, screenshots, data-generating processes, theoretical analysis, and proposed statistical methods, please refer to the analysis plan.

There are three experimental settings: an investment game, an effort task, and a binary lottery choice experiment.

In the Investment Game, subjects make repeated choices between safe and risky options to accumulate earnings. Across-subject variation comes from random assignment to either [Baseline], where earnings depend solely on the subject’s choices, or [AutoInvest], where earnings also reflect automatic exposure to an index fund. Within-subject variation arises from random variation in payoff magnitudes across rounds.

In the Effort Task, subjects first complete transcription tasks and then answer 50 questions involving a choice between a low-paying lottery requiring no further effort and a high-paying lottery requiring additional effort. Variation is entirely within-subject, as both payments and effort requirements vary randomly across questions.

In the Binary Lottery Choice experiment, subjects make 50 choices between two lotteries. Across-subject variation comes from assignment to either [Plain], where lotteries are presented in a standard format, or [Contingent], where the same lotteries are structured to isolate the contingency of the common consequence. Within-subject variation is introduced through random variation in lottery payoffs.

Design details

Randomization method
oTree, Prolific
Randomization unit
Individual Level and Question Level
Clusters
NA
Observations
1,400 subjects
Arms
Investment Game: 500 subjects - 250 for "Baseline" - 250 for "AutoInvest" Effort Task: 500 subjects Binary Lottery Choice: 400 subjects - 200 for "Plain" - 200 for "Contingent"
Primary outcomes
How choices—such as selecting a risky investment, supplying more labor, or opting for a tri-outcome lottery—respond to experimental variation in payoff levels and hypothesized reference points.

Cited by

No citation data available for this registration yet.

Keywords