Facial Recognition and Loan Approval: Human Judgment with AI Information

Last registered on December 06, 2019
View Trial History

Pre-Trial

Trial Information

General Information

Title
Facial Recognition and Loan Approval: Human Judgment with AI Information
RCT ID
AEARCTR-0005125
Initial registration date
December 05, 2019

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
December 06, 2019, 10:20 AM EST

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

Locations

Country
China
Region
Beijing

Primary Investigator

Name
Zeyang Chen
Affiliation
Peking University
Contact Primary Investigator

Other Primary Investigator(s)

PI Name
Yu-Jane Liu
PI Affiliation
Peking University
Contact Investigator
PI Name
Juanjuan Meng
PI Affiliation
Peking University
Contact Investigator

Additional Trial Information

Status
Completed
Start date
2019-05-19
End date
2019-05-28
Keywords
Finance & Microfinance
Additional Keywords
JEL code(s)
Secondary IDs
Abstract
Based on some historical data of online microloan application, repaying performance and facial information, we designed a lab experiment to investigate whether AI prediction can help humans in predicting repayment behavior. We aim to answer the following questions: (1) Do facial photos contain information related to repayment behavior? How can algorithm help humans improve accuracy of judgment? (2) Do humans trust and make use of algorithm predictions effectively? How humans weight their own predictions and algorithm predictions?
External Link(s)

Registration Citation

Citation
Chen, Zeyang, Yu-Jane Liu and Juanjuan Meng. 2019. "Facial Recognition and Loan Approval: Human Judgment with AI Information." AEA RCT Registry. December 06. https://doi.org/10.1257/rct.5125-1.0
Sponsors & Partners
Experimental Details

Interventions

Intervention(s)
Subjects may see one of the four types of loan application randomly:
(1) background information
(2) background information + facial photo
(3) background information + facial credit score given by algorithm (not seeing facial photo)
(4) background information + facial credit score given by subject oneself (not seeing facial photo)
Intervention Start Date
2019-05-19
Intervention End Date
2019-05-28

Primary Outcomes

Primary Outcomes (end points)
(1) Subjects' evaluation of repayment likelihood of each loan applicant.
(2) Subjects' decision to reject or accept each loan applicant.
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
Within-subject design.
Experimental Design Details
Randomization Method
Randomization done in office by a computer
Randomization Unit
experimental rounds
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
250 subjects
Sample size: planned number of observations
30000 lending decisions (each subject makes 120 decisions)
Sample size (or number of clusters) by treatment arms
7500 lending decisions in each treatment (4 treatment in total)
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Supporting Documents and Materials
IRB

Institutional Review Boards (IRBs)

IRB Name
IRB Approval Date
IRB Approval Number
Analysis Plan

Post-Trial

Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Intervention

Is the intervention completed?
Yes
Intervention Completion Date
May 28, 2019, 12:00 +00:00
Data Collection Complete
Yes
Data Collection Completion Date
May 28, 2019, 12:00 +00:00
Final Sample Size: Number of Clusters (Unit of Randomization)
250 subjects
Was attrition correlated with treatment status?
No
Final Sample Size: Total Number of Observations
30000 lending decisions
Final Sample Size (or Number of Clusters) by Treatment Arms
7500 lending decisions in each treatment (4 treatments in total)
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials