Addressing Public Health Needs by Promoting Prosocial Behaviors Through Social Media

How to cite

Harrell, Stephen. 2021. "Addressing Public Health Needs by Promoting Prosocial Behaviors Through Social Media." AEA RCT Registry. AEARCTR-0008041. https://doi.org/10.1257/rct.8041-1.1

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

Stephen Harrell*

* Primary investigator

Registry ID
AEARCTR-0008041
DOI
https://doi.org/10.1257/rct.8041-1.1
First registered
Last updated
Status
Completed
Trial dates
2018-09-01 – 2019-10-01
Countries
Brazil (), India (), United States of America ()
Registry record
http://www.socialscienceregistry.org/trials/8041

Abstract

Social media has the potential to encourage behaviors that improve the health and well-being of society at an unprecedented scale. We examine the impact of Facebook’s blood donation tool on voluntary blood donation in three large countries. Quasi-experimental evidence in the United States indicates that the tool caused an increase in blood donation among all donors (+4.0% [95% CI: .04% to 8.0%]), and particularly among new blood donors (+18.9% [95% CI: 4.7% to 33.1%]) during the tool’s three-month staggered rollout. Longitudinal evidence from Brazil and India indicates that the share of donors who both received a message from the tool and stated they were influenced by Facebook to donate blood increased from 0% to 14.1% [95% CI: 12.1% to 16.2%] in the first year. These findings demonstrate that social media can support public health by encouraging offline prosocial behavior at scale.

Experimental design

In 2019, Facebook planned to launch the blood donation tool in the United States. The United States presented a unique measurement opportunity in that there was enough consolidation and coordination within the sector to implement a staggered (partly randomized) rollout of the blood donation tool by geographic location. Furthermore, the blood banks in the United States had a standardized method of collecting ongoing donation data already in place. These attributes make it much more feasible to implement a quasi-randomized (i.e., staggered difference in differences) experimental design in the United States. As with many product launches, however, there were some go-to-market parameters that undermined pure randomization and a select number of cities were launched before the randomized design for the purposes of quality control and to facilitate critical aspects of Facebook’s partnership with blood banks.

Design details

Randomization method
Randomization done in office by a computer
Randomization unit
State level randomization
Clusters
50
Observations
333 blood donation facilities
Arms
196 treatment group blood donation facilities and 137 control group blood donation facilities
Primary outcomes
Number of donations from all donors and from new donors

Cited by

No citation data available for this registration yet.

Keywords