The Minimal Effects of Public Health Campaigns on Travel During the COVID-19 Pandemic
How to cite
Zelizer, Adam; Mesquita, Ethan Bueno de; Shadmehr, Mehdi; Shaver, Benjamin. 2023. "The Minimal Effects of Public Health Campaigns on Travel During the COVID-19 Pandemic." AEA RCT Registry. AEARCTR-0012148. https://doi.org/10.1257/rct.12148-1.1
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
Governments and public health experts have used mass media campaigns to limit the spread of COVID-19 by encouraging social distancing. This paper explores the effectiveness of a health campaign via a large-scale field experiment conducted in Cook County, Illinois. In December 2020, approximately 125,000 households were sent mail from the University of Chicago Medicine encouraging social distancing during the holiday season. Within targeted geographies, we varied both the dosage of treatments as well as publicity --- whether recipients were informed that their neighbors also received the messages --- to highlight the strategic considerations underlying social distancing. We find minimal effects of the campaign on travel during the holidays and on other social distancing behavior. We consider our results in light of a similar intervention reported in Breza et al (2021) and conclude that travel was unlikely to be influenced by messages from public health experts prior to the 2020 holiday season.
Intervention
In December 2020, approximately 125,000 households were sent mail from the University of Chicago Medicine encouraging social distancing during the holiday season. Within targeted geographies, we varied both the dosage of treatments as well as publicity --- whether recipients were informed that their neighbors also received the messages --- to highlight the strategic considerations underlying social distancing.
Experimental design
Mailers were sent to residents of Cook County, Illinois. We obtained contact information for the 4.4 million adults residing in Cook County from a private data vendor. Nearly all residents had valid addresses, while slightly less than half had phone numbers. To avoid sending multiple mailers to the same household, we randomly selected one adult per household to receive the mailer and be called for the survey.
The unit of treatment assignment was the Census Block Group (CBG). Cook County has approximately 4,000 CBGs, which range in size from one household to nearly 3,800. To improve statistical power and optimize our budget, we dropped particularly small and large CBGs from the study. Our population consists of the 3,600 CBGs with between 98 and 799 households. This sample covers nearly 3.6 million people. 100 CBGs were randomly assigned to each of the four treatment arms. We clustered treatment assignment at this level because one of our outcome measures, cell phone tracking data, is aggregated at the CBG-level.
Design details
Cited by
No citation data available for this registration yet.