An Empirical Study on Filter Bubbles in the YouTube Comments Network: Using Social Network Analysis

Introduction

Among the various video platforms, YouTube's growth rate is increasing noticeably. The number of YouTube users worldwide, which reached 1.5 billion in June 2017, rose to 1.8 billion in May 2018, showing a growth of about 34 percent (Statista, 2018). YouTube is the most widely used app in all ages in Korea (Lee, 2019), a platform that competes with Facebook in various countries for the first or second most frequently used app (Miller et al., 2016). YouTube threatens existing media and transcends the position of press so it can no longer be called a simple video-sharing site. Not only that, it is developing its influence as an information retrieval and educational tool (embrain, 2018). However, some point out that the influence of YouTube's recommendation algorithm could lead to a filter bubble that allows users to access videos within limited subjects or fields. Because of the YouTube algorithm, which recommends only the content that users often see and are interested in as related videos, various opinions and expressions cannot be accepted. Thus, it plays a role that lead to polarization among users (Tufekci, 2018).

Based on the existing research on the filter bubble phenomenon, our research aims to find the bias of information acquisition and communication of users through the filter bubble in YouTube, and to suggest an alternative and marketing strategies from corporate perspectives.