A Field Experiment on the Role of Reference Points and Peer Information in Contest Reward Design

How to cite

Lou, Hanlin; Xu, Hui; Chen, Yanlin. 2025. "A Field Experiment on the Role of Reference Points and Peer Information in Contest Reward Design." AEA RCT Registry. AEARCTR-0015943. https://doi.org/10.1257/rct.15943-1.0

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

Hanlin Lou* , Hui Xu , Yanlin Chen

* Primary investigator

Registry ID
AEARCTR-0015943
DOI
https://doi.org/10.1257/rct.15943-1.0
First registered
Last updated
Status
In development
Trial dates
2025-05-08 – 2025-05-23
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/15943

Abstract

This study investigates how upfront rewards and peer group composition influence individual performance and belief formation in contests with a rank-based reward system. We conduct a field experiment with Year 10 and Year 11 students from a Chinese high school, randomly assigning participants to one of six experimental groups. Within each class, students are assigned to groups of five. They participate in a vocabulary contest, where they have two days to memorize 100 English words and compete against their group members. Rewards are based on their performance ranking within the group. The experimental groups vary along two dimensions: whether students receive upfront tokens before the contest (none, the average reward, or the highest prize), and whether groups are formed randomly or based on prior performance in a related task. Beliefs about expected rankings are elicited both at the beginning and after the test preparation period. This research provides new insights into how reward structures and peer reference points shape motivation, performance, and belief updating in competitive settings.

Intervention

We implement a 3 × 2 between-subject design, with two intervention dimensions described as follows:

Initial Reward Manipulation (Pre-allocated Rewards):

1, Zero initial reward: No rewards are provided prior to the contest.

2, Moderate initial reward (Average prize): An initial reward equal to the average prize amount is allocated to each participant before the contest.

3,High initial reward (Highest prize): An initial reward equal to the highest available prize is allocated to each participant before the contest.

Group Formation:

1,Random group assignment

2,Similar-ability grouping (based on participants' previous performance)

Experimental design

We conduct the experiment with Year 10 and Year 11 high school students from a Chinese high school. At the beginning of the experiment, participants are randomly assigned at the class level to one of six experimental groups. Within each experimental group, participants are further divided into subgroups consisting of five students each. Participants receive a vocabulary list containing 100 English words and have two days to memorize them. They are informed about an upcoming vocabulary memory contest scheduled for two days later. The contest consists of a vocabulary test, and participants' rewards depend on their rank within their subgroup, determined by their test scores. The reward scheme is identical for all participants; however, reward distribution methods and subgroup formation rules vary across experimental groups.

Participants receive an information sheet (A4 size) detailing the reward scheme, subgroup formation method, and the score distribution of their subgroup members from the most recent English exam.

Participants earn tokens as contest rewards, which they can exchange for tangible goods at the end of the experiment. Rewards are distributed based on rankings within each five-person subgroup: 25 tokens for 1st place, 20 tokens for 2nd place, 15 tokens for 3rd place, 10 tokens for 4th place, and 5 tokens for 5th place.

Treatment Groups:

Treatment Group 1: No tokens are provided before the vocabulary test. Participants are randomly assigned to subgroups of five within each class.

Treatment Group 2: No tokens are provided before the vocabulary test. Participants within each class are ranked based on their recent English test scores, and every five participants with similar scores are grouped together.

Treatment Group 3: Participants receive 15 tokens before the vocabulary test. Rewards are adjusted according to the contest results. Participants are randomly assigned to subgroups of five within each class.

Treatment Group 4: Participants receive 15 tokens before the vocabulary test. Rewards are adjusted according to the contest results. Participants within each class are ranked based on their recent English test scores, and every five participants with similar scores are grouped together.

Treatment Group 5: Participants receive 25 tokens before the vocabulary test. Rewards are adjusted according to the contest results. Participants are randomly assigned to subgroups of five within each class.

Treatment Group 6: Participants receive 25 tokens before the vocabulary test. Rewards are adjusted according to the contest results. Participants within each class are ranked based on their recent English test scores, and every five participants with similar scores are grouped together.

Additionally, we elicit participants' subjective expectations about the likelihood of ranking in the top 1, top 2, top 3, and top 4 positions at two separate times: at the beginning of the experiment and immediately before the vocabulary test. We will use a follow-up survey after participants receive their rewards to elicit their satisfaction with the contest outcomes and the reward system.

Design details

Randomization method
Randomization done in office by a computer
Randomization unit
Class level randomization
Clusters
36 classes
Observations
1800 students
Arms
300 students per experimental group
Primary outcomes
Contest performance, measured by participants' test scores on the English vocabulary test.

Cited by

No citation data available for this registration yet.

Keywords