How COVID-19 Information affects Behavior and Policy Preferences
How to cite
Darden, Michael; Rothwell, Jonathan. 2022. "How COVID-19 Information affects Behavior and Policy Preferences." AEA RCT Registry. AEARCTR-0009318. https://doi.org/10.1257/rct.9318-1.0
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
The purpose of the research is twofold. First, our goal is to better understand how information affects attitudes and behavior. Our study will contribute to this literature in the context of COVID-19, with an emphasis on behavior that has clear economic implications (e.g., eating at restaurants, traveling, etc.). Second, our work will shed light on the role of beliefs in explaining hesitancy to get vaccinated against COVID-19. Lack of vaccine uptake remains a serious problem around the world, and our experiment will help test which messages might be effective or counter-productive in encouraging people to take the vaccine.
We will randomize survey respondents into 12 different information arms (plus one control arm in which no information is provided). Each information arm includes some factual information on COVID-19 vaccines and COVID-19 transmission. We will then follow up with several behavioral and attitudinal questions, including plans for in-person dining and in-person gathering. We will test how the effect of information on behavioral and attitudinal beliefs varies by a.) prior beliefs and b.) political affiliation. Prior beliefs will be measured both with numerical assessments of the COVID-19 hospitalization rate as well as categorical assessments on the relative severity of the Omicron variant.
We will also ask respondents to voluntarily reveal whether they use Twitter and, if so, whether they will share their Twitter handle (ID). For the selected sub-sample who agree to provide their Twitter identity, we will merge Twitter handles with internal Twitter information that is made freely available to scholars. These data have been used by other researchers to study, for example, the spread of misinformation. The idea is to interact characteristics of behavior on Twitter, including usage, likes, retweets, and followers, with treatment in our experiment.
There is also a longitudinal component to our study. After two-months, we will re-survey our original sample to test if the effects of information last only within the experiment itself or for as long as two-months after the experiment, when participants are re-interviewed.
Experimental design
See analysis plan.
Design details
Cited by
No citation data available for this registration yet.