How to Combat Cyber Scams? A Randomized Control Trial on a Major Online Payment Platform in China

How to cite

Meng, Juanjuan; Li, Wei. 2023. "How to Combat Cyber Scams? A Randomized Control Trial on a Major Online Payment Platform in China." AEA RCT Registry. AEARCTR-0011899. https://doi.org/10.1257/rct.11899-1.0

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

Juanjuan Meng* , Wei Li

* Primary investigator

Registry ID
AEARCTR-0011899
DOI
https://doi.org/10.1257/rct.11899-1.0
First registered
Last updated
Status
In development
Trial dates
2023-08-13 – 2023-08-18
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/11899

Abstract

In the rapid evolution of the digital economy, the escalating severity of online fraud has spurred widespread concern, prompting the exploration of effective countermeasures. This paper plans to collaborate with a major online payment platform in China and run a randomized control trial. We will explore in the context of online payment, what is the most effective intervention in aiding users to recognize scams and reducing the incidence rate of online fraud. We will also analyse the heterogeneous effects of various interventions on different population.

Intervention

1. General risk reminder: presents a uniform interface and standardized messaging to all users.
2. Personalized risk reminder: provides tailored warnings based on user-selected transaction purposes and recipient information, explicitly highlighting potential fraud scenarios.
3. Voice-based risk reminder: allows users to choose whether to accept voice notifications. Upon acceptance, an AI-powered customer service representative provides risk-related information through voice prompts.
4. Payment Cooling-off Period: implementing a typically 15-minute waiting period, during which users can't proceed with the payment. Users can undergo identity verification and manual review to bypass the cooling-off period.
5. Disengageable Payment Interception: implementing an initial interception of a payment. After the user confirms and verifies a series of information, the interception is lifted, allowing the payment to proceed.

Experimental design

The experiment will be run on a major online payment platform in China.

We have five experimental groups based on the different intervention:
1. General risk reminder
2. Personalized risk reminder
3. Voice-based risk reminder
4. Payment Cooling-off Period
5. Disengageable Payment Interception

The subjects are divided into two groups. One group consists of individuals with originally low-risk profiles who have not triggered backend intervention mechanisms, serving as the control. The other group comprises individuals at risk whose payment behaviors have triggered intervention mechanisms. Different types of interventions will be randomly assigned within each of these two groups.

Design details

Randomization method
done in office by a computer
Randomization unit
individual
Clusters
About 150,000 individuals
Observations
About 150,000 individuals
Arms
About 30,000 individuals for each treatment arm
Primary outcomes
whether the transaction is successful; whether the user has lodged fraud complaints with the payment platform; the amount involved in each fraud case.

Cited by

No citation data available for this registration yet.

Keywords