Complexity and Higher-order Rationality: an Experimental Study

How to cite

Li, Jiaying. 2024. "Complexity and Higher-order Rationality: an Experimental Study." AEA RCT Registry. AEARCTR-0013275. https://doi.org/10.1257/rct.13275-1.0

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

Jiaying Li*

* Primary investigator

Registry ID
AEARCTR-0013275
DOI
https://doi.org/10.1257/rct.13275-1.0
First registered
Last updated
Status
On going
Trial dates
2024-03-28 – 2024-12-31
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/13275

Abstract

We study how the complexity of games influences subjects' higher-order rationality. Our approach involves employing the ring game to gauge an individual's rationality level, with the game’s complexity varied to examine the potential causal relationship between game complexity and subjects' higher-order rationality. This study is particularly focused on three main sources of complexity: the structure of the game, the number of zero payoffs within the payoff matrices, and the salience of iterative strategies and dominant strategies in the ring game. Instead of playing with other human players, participants will play with computer-controlled players, which are pre-programmed to always maximize their achievable payoffs and act as equilibrium players. This setup ensures that any reflection on the depth of subjects' thinking is primarily influenced by the game's inherent complexity, rather than their perceptions of how other players might respond to varying levels of complexity. As the study progresses, we aim to compare different methods of measuring complexity, including ex-ante approaches like the subjective rating of games, as well as ex-post methods that consider reaction times and error rates.

Experimental design

We vary the complexity across ring games. We will consider the following factors that affect complexity:
1. The structure of the game, including variations in the game grid (e.g., 3x3 and 5x5 designs).
2. The number of zero payoffs within the payoff matrices.
3. The salience of iteration and dominant strategy in the ring game.

During the experiment, subjects first read the instructions and must pass a comprehension test to continue. This ensures they understand how to interpret the payoff matrices of the ring games. They will then play 18 different ring games followed by 4 cognitive reflection tests. Finally, they are asked to complete an anonymized questionnaire that gathers demographic information, including age, gender, and field of study.

Design details

Randomization method
Subjects will be randomly assigned to different order of treatment.
Randomization unit
We randomize at the individual level.
Clusters
120 individuals (or more, as budget allows)
Observations
120 individuals (or more, as budget allows)
Arms
120 individuals (or more, as budget allows)
Primary outcomes
Our outcome variables are subjects' identified levels in each game, reaction times, and cognitive reflection test scores.

Cited by

No citation data available for this registration yet.

Keywords