Reducing Single-Use Cutlery in Food Delivery Industry by Social Norm Nudges

How to cite

Pan, Yuhang; He, Guojun; Sawada, Yasuyuki; Tan, Elaine; Park, Albert. 2025. "Reducing Single-Use Cutlery in Food Delivery Industry by Social Norm Nudges." AEA RCT Registry. AEARCTR-0016613. https://doi.org/10.1257/rct.16613-1.0

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

Yuhang Pan* , Guojun He , Yasuyuki Sawada , Elaine Tan , Albert Park

* Primary investigator

Registry ID
AEARCTR-0016613
DOI
https://doi.org/10.1257/rct.16613-1.0
First registered
Last updated
Status
Completed
Trial dates
2024-01-01 – 2025-08-31
Countries
China ()
Registry record
http://www.socialscienceregistry.org/trials/16613

Abstract

The rapid expansion of online food delivery has contributed to the excessive use of single-use cutlery (SUC), posing notable environmental challenges. Traditional policy instruments, such as bans or price-based disincentives, remain difficult to implement in decentralized digital platforms, motivating interest in behavioral interventions as a scalable and low-cost alternative. Beginning in 2024, we started a large-scale field experiment in collaboration with a leading food delivery platform to test whether social norm–based nudges can reduce SUC usage. During the intervention phase, the platform embedded norm reminders into the ordering interface, either highlighting the share of consumers who opt out of SUC (descriptive norm) or appealing to responsible consumption (injunctive norm). After this treatment period, the intervention will be withdrawn, allowing us to examine whether changes in ordering behavior persist once the nudges are removed. This experimental design will provide novel evidence on whether digital norm-based nudges can generate lasting behavioral change through habit formation. The study aims to advance the understanding of pro-environmental behavior in high-frequency digital consumption contexts and to inform the design of scalable, ESG-aligned interventions for sustainable consumption.

Intervention

We refer to the first treatment arm as Personalized Historical Feedback (T1). In this condition, the platform provided users with individualized information that reminded them of their own past choices and placed those choices in relation to a broader behavioral benchmark. The intervention therefore combined a personal feedback component with a normative reference, encouraging reflection on one’s own prior behavior in light of social information. The second treatment arm is labeled Descriptive Norms & Identity Labeling (T2). Here, the platform communicated how other consumers had behaved with respect to cutlery usage and simultaneously categorized the focal user as either part of this majority or not.

Experimental design

We conducted a randomized controlled field experiment on a major online food delivery platform beginning in 2024. The purpose of the study is to test whether digital, norm-based messages can influence consumers’ choice to forgo single-use cutlery.
Consumers were randomly assigned into different conditions. The control group will see the standard ordering interface without additional prompts. The first treatment arm, Personalized Historical Feedback (T1), introduces an informational cue that combines feedback on users’ own past ordering choices with a reference to broader behavioral benchmarks. The second treatment arm, Descriptive Norms & Identity Labeling (T2), provides information about the prevalence of cutlery avoidance among other consumers and highlights whether the focal user aligns with this group. After an intervention period, the messages will be removed, and behavior will continue to be tracked to assess whether effects persist in the absence of active nudges, thereby allowing us to study potential habit formation.

Design details

Randomization method
Randomization done in office by a computer
Randomization unit
Consumer
Clusters
2% consumers in selected cities.
Observations
In the round of analysis, we will randomly select about 200,000 consumers.
Arms
1% of consumers in selected cities for each treatment arm.
Primary outcomes
Consumer-level orders and expenditures at the food delivery platform, and their related pro-environmental behaviors reflected via other records.

Cited by

No citation data available for this registration yet.

Keywords