Online Contact Patterns and Network Resources: Integrating Social Media Records into a Large-Scale Sampling Survey

Tuesday, 17 July 2018: 08:45
Oral Presentation
Yang-chih FU, Academia Sinica, Taiwan
Ming-Yi CHANG, Academia Sinica, Taiwan
The role of social media interactions in social network analysis has been an important yet puzzling issue. As social media become a regular and convenient platform that facilitates an increasing proportion of interactions, for example, to what extent do online contacts help explain why some people get access to more social sources through their personal networks? The question is difficult to answer without integrating actual contact records on social media into data analysis. In this study we show how integrating Facebook contact records and self-reported data from a large sampling survey could contribute to understanding an individual’s social resources in general, and social capital in particular. Data on network resources and social capital come from the 2017 module on Social Networks and Resources of the Taiwan Social Change Survey (TSCS). As a member of the International Social Survey Programme (ISSP), the TSCS team collected rich information about social resources and social capital, along with other network measures, from a representative probability sample of the adult population in Taiwan (N=2,000). Among those respondents who used Facebook regularly, about 80% also authorized the use of their contact records on Facebook. With detailed Facebook contact records from about 800 respondents, we plan to extract and construct various features at individual, tie, and contact levels, and examine how different online contact patterns help differentiate the respondents’ network resources and social capital. The information about resources and social capital will be derived from both position generator and resource generator in the survey data. The study is expected to reveal key connections between recorded data on social media and self-reported survey results, both from identical individuals in a nationally representative sample.