An Experimental Study of Dynamic Disclosure Games

How to cite

Huang, Lingbo; Dong, Lu. 2026. "An Experimental Study of Dynamic Disclosure Games." AEA RCT Registry. AEARCTR-0018182. https://doi.org/10.1257/rct.18182-1.0

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

Lingbo Huang* , Lu Dong

* Primary investigator

Registry ID
AEARCTR-0018182
DOI
https://doi.org/10.1257/rct.18182-1.0
First registered
Last updated
Status
In development
Trial dates
2026-03-30 – 2026-04-30
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/18182

Abstract

This experiment investigates how the \emph{structure} of evidence accumulation affects disclosure strategies in dynamic voluntary disclosure games. We study both static and dynamic settings in which an informed sender seeks to influence the actions of an uninformed receiver by disclosing verifiable evidence that arrives over time. Theory predicts that early disclosure is strategically harmful for the sender. We test these predictions experimentally and examine whether the mere presence of an intermediate disclosure opportunity affects behavior even when theory predicts it should not.

Intervention

The experiment employs a between-subjects design with six treatments. In each treatment, a sender observes private evidence about an uncertain state and chooses what to disclose to a receiver, who then reports a guess about the state value.

Each session runs a single treatment. Subjects are randomly assigned to fixed roles (Sender or Receiver) and re-matched with a new partner each round (strangers protocol). Subjects complete 30 paid rounds preceded by practice rounds.

Experimental design

The six treatments are:

Static Treatments (Baseline). Two single-stage Dye-type disclosure games varying the probability of evidence:
• Static-Low: Urn contains 15 Green, 15 Red, and 70 White balls.
• Static-High: Urn contains 35 Green, 35 Red, and 30 White balls.

Dynamic Treatments (Main). Two two-stage games varying the evidence structure:
• Sequential: Evidence follows a unique path. Urn 1 contains 70 Green and 30 White balls; Urn 2 (drawn only if Urn 1 is Green) contains 50 Green and 50 Red balls.
The sender makes disclosure decisions in two steps.
• Independent: Two independent sources. Urn 1 contains 70 Green and 30 White balls; Urn 2 contains 50 Green and 50 Red balls. Both are drawn independently. The sender makes disclosure decisions in two steps.

Timing Controls (Robustness). Two single-stage games using the same two-urn structure as the dynamic treatments but collapsing the two periods into a single disclosure decision:
• Static-Sequential: Same urn structure as Sequential, single disclosure period.
• Static-Independent: Same urn structure as Independent, single disclosure period.

Design details

Randomization method
Randomization done by the software
Randomization unit
Treatment randomization is employed at session level. Pairing is randomized in each round with roles fixed throughout the session.
Clusters
Each treatment has 6 sessions of 10 subjects each. 36 sessions in total
Observations
36 sessionsm, 360 participants in total.
Arms
60 participants per treatment.
Primary outcomes
Early disclosure rate, Receiver guess, Disclosure rate by type
Secondary outcomes
Timing effect, Overall informativeness, Sender’s informativeness

Cited by

No citation data available for this registration yet.

Keywords