Combining Survey and Social Media Data: Respondents' Opinions on COVID-19 Measures and Their Willingness to Provide Their Social Media Account Information.

Abstract

Research on combining social survey responses and social media posts has shown that the willingness to share social media accounts in surveys depends on the mode of the survey and certain socio-demographics of the respondents. We add new insights to this research by demonstrating that the willingness to share their Facebook and Twitter accounts also depends on the respondents' opinions on specific topics. Furthermore, we extend previous research by actually accessing their social media accounts and checking whether survey responses and tweets are coherent. Our analyses indicate that survey respondents who are willing to share their social media accounts hold more positive attitudes toward COVID-19 measures. The same pattern holds true when comparing their sentiments to a larger Twitter collection. Our results highlight another source of sampling bias when combining survey and social media data: a bias due to specific views, which might be related to social desirability.