The Pervasive Presence of Chinese Government Content on Douyin Trending Videos

31 Pages Posted: 8 Mar 2021

See all articles by Yingdan Lu

Yingdan Lu

Stanford University

Jennifer Pan

Stanford University

Date Written: February 28, 2021

Abstract

The proliferation of social media has expanded the strategies for government propaganda, but quantitative analyses of the content of digital propaganda continue to rely predominantly on textual data. In this paper, we use a multi-modal approach that combines analysis of video, text, and meta-data to explore the characteristics of Chinese government activities on Douyin, China's leading social video-sharing platform. We apply this multi-modal approach on a novel dataset of 50,813 videos we collected from the Douyin Trending page. We find that videos from the Douyin accounts of Chinese state media, government, and Communist Party entities (what we call state-affiliated accounts) represent roughly half of all videos featured on the Douyin Trending page. Videos from state-affiliated accounts focus on political information and news while other Trending videos are dominated by entertainment content. Videos from state-affiliated accounts also exhibit features, including short duration, brightness, and high entropy, found in prior research to increase attention and engagement. However, videos from state-affiliated accounts tend to exhibit lower average levels of audience engagement than Trending videos from other types of accounts. The methods and substantive findings of this paper contributes to an emerging literature in communication on the computational analysis of video as data.

Keywords: china, propaganda, video, Douyin, computer vision

Suggested Citation

Lu, Yingdan and Pan, Jennifer, The Pervasive Presence of Chinese Government Content on Douyin Trending Videos (February 28, 2021). Available at SSRN: https://ssrn.com/abstract=3794898 or http://dx.doi.org/10.2139/ssrn.3794898

Yingdan Lu (Contact Author)

Stanford University ( email )

Stanford, CA 94305
United States

Jennifer Pan

Stanford University ( email )

Stanford, CA 94305
United States

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
143
Abstract Views
481
rank
255,082
PlumX Metrics