Coverage of Social Unrest on Telegram

• In Event: From Crypto to AI: Mobilization On and Against New Platforms and Technologies (Virtual Session)

Abstract

This paper connects online political activism to offline contentious actions in autocratic contexts through a multi-country analysis of Telegram – an online communication platform deemed the ‘protesters app’. Telegram was developed by the creators of VKontakte (VK) - Pavel and Nikolai Durov - in response to repression they faced from Russian authorities for refusing to remove opposition group’s pages from VK. This history of Telegram, which is steeped in anti-regime dissent and multiple instances where the creators have denied requests from the government to share its data, seems to have made it popular amongst dissenters, keen on escaping state surveillance. Recent anecdotal evidence suggests that rising national anxiety and unrest events led to increases in downloads of the application and anti-regime protesters in countries as different as Iran, Hong Kong, Belarus and Russia use the application to organize offline action in the face of severe government censorship and repression.

But there is little research on which protest events are covered on Telegram out of the many possible events to observe. In fact, research on attention to conflict events on social media is lacking even for more mainstream platforms like Twitter and Facebook. More broadly, studies at the intersection of social media use and contentious politics have focused on a single country or a single event, providing less generalizable insights on attributes which predict event coverage in digital spaces. Extensive research of traditional media shows that high levels of conflict intensity, large number of participants, the sensationalist nature of the event and the geographic nearness from the media company’s office correlate with coverage of protest events (Myers & Caniglia, 2016; Ortiz et al., 2005; Rafail et al., 2019; Snyder & Kelly, 1977). However, given the decentralized nature of social media platforms where anyone can post about events, without constraints placed by other competing stories or monetary incentives, it is expected that the results from traditional media studies may not replicate on these online platforms.

Towards this end and with the goal of understanding the use of encrypted messaging applications in confined contexts, I examine the attention to protest events on Telegram, such as those found in crisis monitoring datasets like ACLED. I use protest, government, and country attributes to explain the coverage of on-ground events on this platform.

I hypothesize that the affordances of encryption and privacy as well as the social media-like features that allow individuals to connect and disseminate information, impact the types of events which are covered on Telegram. With the assumption that citizens are motivated to use Telegram to escape possible detection for discussing sensitive topics, I first hypothesize that it is more likely to be used to share information about protests that target central regime figures and those that involve higher levels of state-sponsored repression than protests that target diffuse actors or those which involve non-violent citizen-state interactions. Second, given the anecdotal claims that Telegram is a “protesters app” and is used for organizing anti-regime action, I hypothesize that this platform is more likely to cover conflict events which are anti-government than events which are pro-government. Lastly, I hypothesize that regime attributes that facilitate government information collection, such as surveillance capacity, should increase coverage of protest events on Telegram.

To conduct this analysis, I collect posts from “channels” dedicated to discussing conflict events in each of the countries under study and then conduct a panel analysis. I use the protest events enumerated in ACLED as the population of all events. And to identify relevant channels I create an initial list of keywords using the event summaries provided in the ACLED data. For the pilot analysis of Belarus, I employed the fighting word statistic (Monroe et al., 2008) and fed the ACLED summaries of protest events and random Wikipedia articles as two groups to identify words/phrases that are distinctly used to describe protest events in Belarus. I then search the Telegram analytics database and curate the list of channels which mention these keywords. I subsequently test if the volume of discussion relates to the features of protests and the context in which they occur, as described in my hypothesis above. My main independent variables comprise indicators of (1) if the event targets regime leaders or not (2) if it involves state sponsored repression (3) if the event is anti or pro regime and (4) the level of surveillance in the country. In the process of this analysis, I also develop multilingual text analysis methods to connect relevant discussions on Telegram to recorded protest events in event datasets and provide a framework for analyzing content from this important but understudied platform.

Author

• Ishita Gopal, The Pennsylvania State University