RAPID; Information and Implications for Protection Motivation and Action During the COVID-19 Outbreak

Social, Behavioral and Economic Sciences - The current spread of, and alarm about, the COVID-19 virus provides a unique and ephemeral opportunity to obtain meaningful time-series survey data on public beliefs, attitudes, behaviors, and the receipt of information of various kinds about the disease and its effects on taking protective action. The National Institute for Risk and Resilience (NIRR) utilizes its on-going Twitter data collection associated with coronavirus (collected since January 2020), and undertakes a series of monthly nation-wide surveys on public views to test the broader publics? receipt of, trust in, and use of information about the virus posted on social media. The surveys will include questions about protective action behavior, trust in key actors, perceptions of risk associated with the outbreak, and perceptions of information accuracy/inaccuracy. The complementary survey and social media data streams will allow tracking the spread and penetration of information over time and as the disease spreads in order to match various narratives as they emerge on social media along with beliefs measured in the contemporaneous survey data. The time sensitive data will permit testing of hypotheses about the dynamic relationships between the spread of information in social media, broader public beliefs and behaviors, and effects on protective behaviors that may influence the spread of contagious diseases.

The goal of this study is to measure and track the influence of information about the COVID-19 pandemic on Twitter among members of the broader US public. The study integrates two complementary streams of data to systematically examine the impact of information bubbles and various forms of information on protection motivation and actions in response to the COVID-19 outbreak in the US. First, since January 2020 ,the research team has collected all messages on Twitter that relate to COVID-19, by establishing a connection with the Twitter streaming API. The team obtains all posts and metadata that include any of the following key words: coronavirus, COVID-19, SARS-CoV-2, #coronavirus, #2019_nCov, and #COVID-19. From January 27 to Feb 24, the team collected more than 31 million different messages about the virus. The Twitter posts provide a continuous flow of data about the evolution of information networks and the promulgation and spread of information, but they do not provide information on the extent to which these factors are affecting protective motivations in the broader public and shaping the perceptions that drive them (such as trust in perceived risk). Second, the team collects online rolling nationwide surveys of the broader public?s understanding of COVID-19, with special attention to beliefs about the information that appears on Twitter, over the span of the next year. There are 10 nationwide surveys in all, one each month (time-series cross-sections), with collections timed to obtain 250 responses each week to increase the ability to quickly identify changes in beliefs, perceptions and associated protective behaviors. The surveys are designed to allow pairing the changing pattern of information of various sorts on social media with the receipt and belief of that information among the broader public. The experiments draw from the rise and spread of different kinds of information on Twitter.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.