A Nonparametric Test for Cumulative Prospect Theory

How to cite

He, Songyu. 2025. "A Nonparametric Test for Cumulative Prospect Theory." AEA RCT Registry. AEARCTR-0015722. https://doi.org/10.1257/rct.15722-1.0

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

Songyu He*

* Primary investigator

Registry ID
AEARCTR-0015722
DOI
https://doi.org/10.1257/rct.15722-1.0
First registered
Last updated
Status
In development
Trial dates
2025-06-18 – 2025-07-01
Countries
Private
JEL codes
D90
Registry record
http://www.socialscienceregistry.org/trials/15722

Abstract

As a leading non-expected utility model, Cumulative Prospect Theory (CPT) (Tversky and Kahneman, 1992) rationalizes a wide range of deviations from expected utility theory while keeping most of its theoretical axioms. This project proposes an experiment to jointly analyze and test the two major factors in CPT: probability weighting function and rank dependence. Specifically, under a nonparametric framework, we test the validity of rank dependence and estimate the first-order derivative of probability weighting functions.

Intervention

Our experiment has two treatments – gain and loss treatment. Every subject only participates in a single treatment. In the gain treatment, subjects chooses among monetary lotteries. In contrast, in the lost treatment, subjects chooses among lotteries with outcomes being potential workloads they need to complete at the end of the experiment. Each treatment contains nineteen multiple price lists (Holt and Laury, 2002), which can be divided into two parts: (p,Z) lists and (q,X) lists. In each treatment, there are ten (p,Z) lists and nine (q,X) lists. The experimental interface is programmed using oTree (Chen et al., 2016).

Experimental design

This experiments recruit 550 subjects (275 male and 275 female) from Prolific with the following inclusion criteria: Inclusion criteria:
Approval rate ≥ 98%. Fluent in English. Located in US or UK. Minimum high school education. Age 18–50.
Participants will be excluded from analysis if: their choices lead to 3 or more first-order stochastic dominance (FOSD) violations.
They have 7 or more total mistakes in three comprehension questions (the order of answers in questions are randomized)

Design details

Randomization method
Randomization done by computer.
Randomization unit
Individual
Clusters
550 individual
Observations
550 individual
Arms
275 in gain treatment. 275 in loss treatment.
Primary outcomes
The primary outcomes are the nineteen choices of subjects for each price list. Each choice indicates a pair of lotteries that are indifferent to the subjects.

Cited by

No citation data available for this registration yet.

Keywords