Collective Rationality: An experimental Study
How to cite
QU, Xiangyu; Liang, Xinyu; Cai, Chun; Li, Xiaoxi. 2025. "Collective Rationality: An experimental Study." AEA RCT Registry. AEARCTR-0017015. https://doi.org/10.1257/rct.17015-1.0
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
This experiment investigates the behavioral foundations of collective rationality in group decision-making under uncertainty. Participants make choice over risky options, both individually and after group deliberation. The design tests whether collective choices satisfy standard rationality. By comparing individual and group-level choices, the experiment identifies how discussion, information exchange, and cognitive constraints shape collective preferences. The findings aim to provide empirical foundations for theories of collective decision-making and preference aggregation in economics and behavioral social science.
Experimental design
Our experimental design includes two treatments, with group sizes of 3 and 5 participants, respectively. Each experiment consists of 2 stages: the first stage involves individual decisions, and the second stage involves collective decisions (in groups of 3 or 5).
Baseline Condition (Group-size-3 condition)
In the first stage, using the methods of Choi et al. (2007, 2014), the decision maker’s (DM’s) task is to independently distribute points between two accounts, labeled BLUE(x) and RED(y). We can caculate the exchange rates(p) between BLUE and RED accounts. Thus, a decision obtains a tuple (x, p) whereby a DM selects a bundle (x, y) under the exchange rates (px, py).
Since measuring rationality requires a collection of such decisions, we include 25 tasks with randomly generated lines (Choi et al., 2014). After that, we measure the economic rationality of these 25 decisions, (xi, pi) (i=1,…, 25).
The second stage involves collective decision-making, including 25 rounds as well. At the beginning of each round, each group member makes an independent allocation choice following the method from Stage 1. They then proceed to an “deliberation” phase, such as structured discussion, silent voting, during which a collective decision is reached. Collective rationality for each group is measured using the same method as in the first stage. The design enables identification of whether collective choices satisfy rationality.
Group-size-5 condition
The Group-size-5 condition proceeds exactly as the Group-size-3 condition, except that each group consists of five members instead of three in the collective decision stage.
Design details
Cited by
No citation data available for this registration yet.