Emotion revelation and trust: valence, beliefs and ambiguity attitudes

How to cite

XIAO, Fantine. 2026. "Emotion revelation and trust: valence, beliefs and ambiguity attitudes." AEA RCT Registry. AEARCTR-0018300. https://doi.org/10.1257/rct.18300-2.1

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

Fantine XIAO*

* Primary investigator

Registry ID
AEARCTR-0018300
DOI
https://doi.org/10.1257/rct.18300-2.1
First registered
Last updated
Status
On going
Trial dates
2025-12-01 – 2026-08-31
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/18300

Abstract

This project studies how the opportunity to reveal emotion affects trust behavior, including trust decisions, beliefs, and ambiguity attitudes. I implement a modified trust game in which trustors may reveal positive or negative emotions, with varying degrees of flexibility over valence and intensity to the trustees, before the trustees decide how much to return. The study identifies the mechanisms through which access to emotion revelation shapes trusting behavior within a rigorous economic framework. It also examines patterns of emotion revelation and the motivations behind them.

Intervention

Subjects are randomly assigned to one of four conditions that vary access to emotion revelation toward the trustee: a control condition with no message, a positive-valence condition in which they can indicate how happy they would feel if trust is reciprocated, a negative-valence condition in which they can indicate how disappointed they would feel if trust is not reciprocated, and a free-valence condition in which they can choose the valence and intensity of the emotional message sent to the trustee. After answering treatment-specific questions, they all make a trust decision and then choose between lottery pairs to elicit beliefs and ambiguity attitudes.

Experimental design

The experiment is based on a modified trust game in which all participants act as trustors and are matched with separately recruited trustees. Each trustor chooses between a no-trust option, in which they keep 5 dollars and allocate 5 dollars to the trustee, and a trust option, in which they keep 2 dollars and send 8 dollars to the trustee. The 8 dollars is tripled to 24 dollars, after which the trustee decides how much to return.

Before making the trust decision, participants report how happy they would feel if trust is reciprocated and how disappointed they would feel if trust is not reciprocated, each at one of three intensity levels: slightly, moderately, or very. Participants are then randomly assigned to one of four conditions. In the control condition, no emotional message is sent to the trustee. In the negative-valence treatment, participants are told that if they choose trust, the trustee will see a message stating how disappointed they would feel if the trustee does not reciprocate their trust, before the trustee decides how much to return. Participants are asked to choose the intensity of that message. In the positive-valence treatment, participants are told that if they choose trust, the trustee will see a message stating how happy they would feel if the trustee reciprocates their trust, and participants choose the intensity. In the free-valence treatment, participants are allowed to choose both whether to reveal happiness or disappointment and the intensity of the chosen emotion.

After the trust decision, all participants complete a bi-section choice task, based on the belief-hedge method introduced by Baillon et al. (2018, 2021). Participants make decisions between ambiguous and risky options. The probability in the risky option is varied until indifference is reached within a pre-determined error margin. These choice situations are designed to elicit the matching probability associated with events concerning the trustee's return amount. The resulting matching probabilities allow us to construct three indexes: (i) an ambiguity-aversion index, capturing aversion to ambiguity about the trustee's return amount; (ii) an ambiguity-insensitivity index, capturing how precisely or imprecisely participants perceive this uncertainty; and (iii) an expected return amount, reflecting participants’ beliefs about how much they expect to receive from the trustee.

Design details

Randomization method
Randomization is implemented automatically through Qualtrics’ built-in randomizer.
Randomization unit
individual
Clusters
No clustering
Observations
400 online participants -2026 May followup: 100 online participants
Arms
100 online participants per treatment
Primary outcomes
-trust decision -ambiguity aversion index -a-insensitivity index -expected return
Secondary outcomes
-emotion revelation gap

Cited by

No citation data available for this registration yet.

Keywords