Allais paradox with different elicitation methods

Last registered on October 31, 2025
View Trial History

Pre-Trial

Trial Information

General Information

Title
Allais paradox with different elicitation methods
RCT ID
AEARCTR-0016765
Initial registration date
October 24, 2025

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
October 31, 2025, 8:02 AM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Primary Investigator

Name
Zhihua Li
Affiliation
Contact Primary Investigator

Other Primary Investigator(s)

Additional Trial Information

Status
In development
Start date
2025-11-01
End date
2026-12-31
Keywords
Behavior, Lab
Additional Keywords
JEL code(s)
C91, D01, D81, D83
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
This paper re-examines the classical Allais paradox by exploring how different elicitation methods and choice modes influence its manifestation. Recent research has suggested that observed violations may reflect decision errors or deeper departures from fundamental principles rather than failures of the independence axiom. Building on this line of inquiry, we argue that the implementation of choice tasks themselves may play a critical role in generating paradoxical behavior. Our study investigates whether the paradox is confined to binary choices or whether it extends to alternative elicitation frameworks. We further examine the relationship between task comparability and the consistency of preferences, offering new insights into the mechanisms underlying the paradox and its implications for models of decision-making under risk.
External Link(s)

Registration Citation

Citation
Li, Zhihua. 2025. "Allais paradox with different elicitation methods." AEA RCT Registry. October 31. https://doi.org/10.1257/rct.16765-1.0
Sponsors & Partners
Experimental Details

Interventions

Intervention(s)
Intervention Start Date
2025-11-01
Intervention End Date
2026-12-31

Primary Outcomes

Primary Outcomes (end points)
The Allais paradox is most likely to emerge in binary choice tasks, but becomes less pronounced or disappears when preferences are elicited through alternative methods (e.g., valuation tasks or broader choice formats). This would suggest that the paradox is partly an artifact of the elicitation mode rather than a robust violation of the independence axiom.
Primary Outcomes (explanation)
The valuation tasks (such as certainty equivalent elicitation) or broader portfolio-style formats require participants to assess each option more holistically and broadly, often in relation to a personal reference value rather than to another nearly identical lottery. This shift reduces the salience of small probability differences and discourages narrow attribute-wise comparisons, leading to more internally consistent responses. In other words, alternative elicitation methods provide a decision context in which biases triggered by the framing of binary choices play a smaller role.

Secondary Outcomes

Secondary Outcomes (end points)
According to the comparability hypothesis, preferences will be more consistent in low-comparability tasks (e.g., certainty equivalents where options differ in structure) than in high-comparability tasks (e.g., probability equivalents where options share a similar structure). This would highlight how task comparability shapes decision biases and consistency.
Secondary Outcomes (explanation)
The second objective is to test the comparability hypothesis proposed by Li (2025). This hypothesis suggests that when individuals evaluate two options with a similar structure high comparability) – for example, two two-outcome lotteries – they are more likely to engage in attribute-by-attribute trade-offs, making them more susceptible to biases such as salience effects. Conversely, when evaluating options with different structures (low comparability) – for example, a sure outcome versus a risky lottery – individuals are more likely to engage in holistic valuation, thereby reducing bias. To test this, we also elicit probability equivalents (PE) of the lotteries. If the comparability hypothesis holds, we expect more consistent behavior in CE tasks (low comparability) than in PE tasks (high comparability).

Experimental Design

Experimental Design
The study investigates whether the Allais paradox is specific to binary choice by comparing behavior across different elicitation methods. Participants make decisions both through portfolio-style and valuation-based tasks, allowing us to test the robustness of the paradox beyond standard binary comparisons. In addition, we vary the comparability of options to examine whether preference consistency depends on the structural similarity of the lotteries being evaluated.
Experimental Design Details
Not available
Randomization Method
Randomisation of questions
The lotteries and decision tasks will be presented in the program in random order.

Randomisation of the risky decisions in the experiment
When implementing the real payment of the lottery, the physical device will be used to realise the probability, such as drawing a marble from an opaque bag, or throwing two ten-sided dice.

Randomisation of participants
There will not be any randomisation of the participants. The two experiments are within-subject, so all participants will get the same set of decision questions.

For each participant, the only difference is how the questions are presented, e.g., the order of the questions.

Randomization Unit
The experiment is conducted at the individual level, with no clustering or additional levels of randomization.
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
The plan is to conduct the experiment at the Behavioural and Decision Research Lab at Beijing Normal - Hong Kong Baptist University with the student sample and then online via the Prolific platform with a more representative sample.

Sample size: planned number of observations
500 - 1000 participants in total.
Sample size (or number of clusters) by treatment arms
Not applicable.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Not applicable.
Supporting Documents and Materials

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
IRB

Institutional Review Boards (IRBs)

IRB Name
Beijing Normal Hong Kong Baptist University Research Ethics Committee
IRB Approval Date
2025-10-14
IRB Approval Number
REC-2025-48
Analysis Plan