Promises, Companionship and Dazi

Last registered on January 06, 2026
View Trial History

Pre-Trial

Trial Information

General Information

Title
Promises, Companionship and Dazi
RCT ID
AEARCTR-0017500
Initial registration date
December 23, 2025

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
January 06, 2026, 6:52 AM EST

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

Country
China
Region
Changsha, Hunan Province

Primary Investigator

Name
Qian Zhang
Affiliation
Central South University
Contact Primary Investigator

Other Primary Investigator(s)

PI Name
Jianxin Wang
PI Affiliation
Business School, Central South University
Contact Investigator
PI Name
Yulei Rao
PI Affiliation
Business School, Central South University
Contact Investigator
PI Name
Daniel Houser
PI Affiliation
Interdisciplinary Center for Economic Science, George Mason University
Contact Investigator

Additional Trial Information

Status
In development
Start date
2025-12-25
End date
2026-04-30
Keywords
Behavior
Additional Keywords
Promises, Companionship, Dazi culture
JEL code(s)
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
A lengthy literature suggests that promises play a critical role in creating business partnerships, however, an important open question remains: whether promises promote the formation of companionships that do not include a pecuniary component? We address this question by conducting a randomized controlled trial using the social-media platform RedNote.
We post on RedNote that we are looking for one or more companions to take part in an activity (e.g. visiting a museum). People express interest via direct message (DM). In half of our responses we explicitly promise to try to make the trip enjoyable, and in the other half (randomly determined) we indicate the same but do not make an explicit promise. Our interest is in the frequency with which people agree to take part in the activity, and whether it varies according to the presence of a promise. Our study sheds new light on the role of communication and promises in the formation of non-business partnerships.
External Link(s)

Registration Citation

Citation
Houser, Daniel et al. 2026. "Promises, Companionship and Dazi." AEA RCT Registry. January 06. https://doi.org/10.1257/rct.17500-1.0
Sponsors & Partners
Experimental Details

Interventions

Intervention(s)
We randomly assign our responses to DMs we receive to either the promise or no-promise treatment.
In the promise treatment our response includes an explicit promise that we will do our best to enhance the enjoyment of the activity.
In the no-promise treatment, our message is the same except that it does not include an explicit promise.
Intervention Start Date
2025-12-25
Intervention End Date
2026-04-15

Primary Outcomes

Primary Outcomes (end points)
Whether people indicate a willingness to take part in the activity with us.
Primary Outcomes (explanation)
We use the first response we receive to the promise/no-promise intervention message to measure interest in joining the activity. Any subsequent message will not be used to measure interest (though we will save any additional messages and may use them for subsequent exploratory analyses).
We say people are willing to join if they reply to our message within 48 hours, and in their reply includes a pre-specified code we provided to them in our intervention message, or an indication of their available time. We say people are not willing to join if they do not reply to our intervention message within 48 hours.
Some messages may not clearly indicate interest or disinterest (e.g., they neither reply with the code nor their available time). We will classify unclear messages as either showing or not showing interest by using either human (XH algorithm, Xiao and Houser 2005 PNAS) or computer (LLM) coders.

Secondary Outcomes

Secondary Outcomes (end points)
Whether people make an effort to join the activity (e.g., reservations at the location of the activity).
Whether people actually participate in the activity with us.
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
We post on RedNote that we are looking for one or more companions to take part in an activity (e.g. visiting a museum). People express interest via direct message (DM). In half of our responses we explicitly promise to try to make the trip enjoyable, and in the other half (randomly determined) we indicate the same but do not make an explicit promise. Our interest is in the frequency with which people agree to take part in the activity, and whether it varies according to the presence of a promise.
Experimental Design Details
Not available
Randomization Method
Using the order in which participant’s messages are received, we will alternate between sending promise and no promise messages. Whether the first message contains a promise will be randomly determined.
Randomization Unit
Individual level
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
No clusters
Sample size: planned number of observations
Based on the data of a lab experiment which test the effects of promises on companionship, we calculate the Cohen’s d of 0.48. We calculated the minimum sample size needed at this effect size by G Power, as shown below: t tests - Means: Difference between two independent means (two groups) Analysis: A priori: Compute required sample size Input: Tail(s) = Two Effect size d = 0.48 α err prob = 0.05 Power (1-β err prob) = 0.8 Allocation ratio N2/N1 = 1 Output: Non-centrality parameter δ = 2.8397183 Critical t = 1.9773035 Df = 138 Sample size group 1 = 70 Sample size group 2 = 70 Total sample size = 140 Actual power = 0.8050897 As shown above, the minimum sample size needed is 140, and 70 in each group.
Sample size (or number of clusters) by treatment arms
About 100 observations in promise group
About 100 observations in no promise group
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Given the sample size, we calculate the minimum detectable effect size using the G*Power, as shown below: t tests - Means: Difference between two independent means (two groups) Analysis: Sensitivity: compute required effect size. Input: Tail(s) = Two α err prob = 0.05 Power (1-β err prob) = 0.8 Sample size group 1 = 100 Sample size group 2 = 100 Output: Non-centrality parameter δ = 2.8152618 Critical t = 1.9720175 Df = 198 Effect size d = 0.398 Based on the sample size of 200, the study has 80% statistical power at a 5% significance level to detect the difference in mean values between treatments.
Supporting Documents and Materials
IRB

Institutional Review Boards (IRBs)

IRB Name
A study of partnership decision making
IRB Approval Date
2025-12-19
IRB Approval Number
STUDY00000936
Analysis Plan