Advancing the reliability of social media discourse as a measure of public sentiment about COVID-19

The proposed research is to develop a method for making valid, representative estimates of public opinion from sufficiently large and diverse non-probability samples such as social media data. We aim to demonstrate the theoretical, methodological, and practical contributions to the field of public opinion research by demonstrating the reliability and utility of this method in the context of measuring public responses to COVID-19. The measurement of public opinion as it pertains to COVID-19 is critical to strengthening understandings of the pandemic’s broad social and economic implications. The proposed research will use statistical and machine learning methods to demonstrate how user-generated content collected from social media platforms can be utilized in a principled way so as to produce and extrapolate low-bias, representative measures of public opinion related to COVID-19 and other contemporary societal issues.

Faculty Supervisor:

Clifton van der Linden

Student:

Partner:

Vox Pop Labs Inc

Discipline:

Sociology

Sector:

New and Digital Media; Public Service, Policy, and Governance; Information and Communications Technology

University:

McMaster University

Program: