Individual and committee sequential search in the laboratory

How to cite

Herman, Clément. 2023. "Individual and committee sequential search in the laboratory." AEA RCT Registry. AEARCTR-0011115. https://doi.org/10.1257/rct.11115-1.0

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

Clément Herman*

* Primary investigator

Registry ID
AEARCTR-0011115
DOI
https://doi.org/10.1257/rct.11115-1.0
First registered
Last updated
Status
In development
Trial dates
2023-03-26 – 2023-10-28
Countries
Private
JEL codes
D83, D72
Registry record
http://www.socialscienceregistry.org/trials/11115

Abstract

The objective of this study is to analyze the behavior and explore the effectiveness of individual and collective sequential search in a laboratory experiment. The design of our experiment is based on the work of Albrecht, Anderson, and Vroman (JET, 2010). Our aim is to evaluate the predictions of that model and identify factors, including behavioral, that were not accounted for. We are particularly interested in understanding the impact of committee size, voting rules, and offer distributions on search duration, strategies adopted by participants, and ultimately, their welfare.

Intervention

Subjects will take part in a search experiment, either individually (individual treatment) or in groups of 3 (committee treatment). The group or individual will receive offers until either acceptance or random termination of the game. Each offer corresponds to a round. When an offer is accepted, the search ends, and each subject receives the value of their own offer (or the value of the common offer in the common value treatment) as experimental points, to be converted into monetary payments at the end of the session.
Offers are drawn independently each round, from a commonly known distribution. We will elicit subjects’ reservation values using an incentive-compatible BDM mechanism (after Brown, Flynn and Schotter, AER, 2011): subjects will enter their “minimum acceptance value” (i.e., their reservation value) before their offer is drawn. Subjects will set their reservation value each round before an offer is made. Subjects may set different reservation values in different rounds.
In the individual search, an individual receives one offer every round. A subject will automatically accept the offer when it is equal to or above that value, otherwise she will reject.
In the committee search case, a vote will be cast automatically, to accept the offer (common value case) or set of offers (private value case) if the offer has a value that is equal to or above the minimum acceptance value, or to reject otherwise.

In all treatments: time preferences will be experimentally induced with random termination. The experimental discount rate will be set at 95%, which corresponds to a 5% termination probability per round. We will implement a block random termination à la Frechette and Yuksel (EE 2017), with blocks of 5 rounds.

In the committee treatments, subjects will not observe the reservation value set by other players for that round, nor will they observe the vote cast by each player or the total precise number of votes cast in favor of acceptance. They will only observe if their group decided to accept or reject the offer or set of offers. They will also have access to the history of the reservation values, offers drawn, and own vote, for the previous rounds of the current game.

For the committee treatments, we will also elicit the subjects’ beliefs about the reservation values set by a random other group member, in later rounds of the game. Subjects will be incentivized for accuracy.

In the committee with private value treatments, we will also introduce games where subjects play against group members who are computers with pre-defined strategies that are known to the subject. The goal of these treatments is to allow the analysis of best-response behavior when agents know the strategies played by other committee members. We will vary these pre-defined strategies.

Experimental design

Participants will be randomly assigned to 10 (5 x 2) treatments :
- Search setup (5): individual, committee with common value (unanimity and majority), committee with private value (unanimity and majority). In the committee search with common value, the committee receives a common offer every round. In the committee search with private value, each member of the committee receives an individual offer every round, drawn separately for each group member from the same common distribution. For each committee treatment, we will vary the voting rule for the group to decide to stop the search: majority (group stops searching when at least two group members vote to stop) or unanimity (group stops searching when all members vote to stop).
- Distribution of offers (2): distribution A and distribution B, where distribution B is a mean-preserving spread of A. Distribution A: with probability 90%, a random integer number between 1 and 30, with probability 10%, a random integer number between 31 and 100. Distribution B: with probability 70%, a random integer number between 1 and 10, with probability 30%, a random integer number between 11 and 100.

At the end of the session, subjects will play two versions of the dictator game to assess their altruism as well as two versions of the risky project, where we vary the game parameters. Subjects will also complete a short survey to better understand the strategies that they played, as well as their feedback on the experiment itself.

Design details

Randomization method
Between-subjects design : one treatment is played in each session.
Randomization unit
All subjects within a session will be assigned to the same treatment, different sessions will receive different treatments.
Clusters
Around 210 experimental subjects (maybe more, maybe less, depending on how recruitment goes) in 10 treatments.
Observations
Around 210 experimental subjects
Arms
21
Primary outcomes
Game duration (number of rounds played) Reservation values set by committee members / individuals in every round of the game Welfare (final payoffs) Beliefs about other group member’s reservation values.

Cited by

No citation data available for this registration yet.

Keywords