Effectiveness of Information Manipulation - An Experimental Investigation of Informational Autocracies

How to cite

Höhne, Christian; Gallo, Edoardo. 2024. "Effectiveness of Information Manipulation - An Experimental Investigation of Informational Autocracies." AEA RCT Registry. AEARCTR-0014479. https://doi.org/10.1257/rct.14479-1.0

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

Christian Höhne* , Edoardo Gallo

* Primary investigator

Registry ID
AEARCTR-0014479
DOI
https://doi.org/10.1257/rct.14479-1.0
First registered
Last updated
Status
In development
Trial dates
2024-10-08 – 2024-10-19
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/14479

Abstract

The purpose of this project is to investigate if communication by a group leader can increase contributions to a public good by the other group members if the communicated information is unverifiable and incentives between the leader and members are misaligned. This is to experimentally approximate the environment of informational autocracy (Guriev and Treisman, 2019, 2020) while abstracting from the existence of motivated beliefs and identity.

Experimental design

To model this decision environment, we use a game design combining dynamics of public good (PG) and trust games. Four participants -- of Type A -- can contribute to a public good, while one player -- Participant B -- can either add to or embezzle from the public good. Participant B thus plays a version of TG with multiple investors, while Participants of Type A play a PG with uncertain return.
The payoff for Participants of Type A depends on three components: their individual contribution to the public good, contributions by other Type A's and Participant B's action. Type As know their own contribution, the final size of the group account, and their payoff for each round. They may also receive information about the composition of contributions leading to the final size of the public good, that is overall contributions by all Participants of Type A and Participant B's action choice.
We experimentally vary whether the provided information is verifiable to Participants of Type A and whether it can be manipulated by Participant B. This way we can identify differences in contribution and embezzlement behavior under different informational regimes and evaluate the effectiveness of information manipulation.

Design details

Randomization method
Algorithmic randomisation to experimental sessions based on availability and algorithmic randomisation to groups and roles within the session.
Randomization unit
We randomize the treatment by experimental session. In each session, we randomly assign 5 participants to one group. Within this group, we randomly assign one participant the role 'B' and all the remaining participants play role "A".
Clusters
100 groups
Observations
500 participants
Arms
25 groups or 125 participants by treatment arm
Primary outcomes
- contribution to public good - beliefs about public good contribution - embezzlement behavior - choice of communication

Cited by

No citation data available for this registration yet.

Keywords