Seminar Summary – Social Media and Support for Dominant Political Incumbents: Evidence from Uganda

By Sayuri Kataoka

Initial optimism regarding the potential of social media to act as “liberation technologies” has been tempered by the efficacy with which repressive regimes have exerted their control over such platforms. The efforts to limit citizen access to social media are widespread – whether through a ban on access to particular platforms, wholesale Internet blackouts or elevating the financial costs of access. Furthermore, social media – combined with state-orchestrated censorship – can distract, misinform and polarize citizens in a way that benefits autocrats. To limit the potentially destabilizing effects of social media, the incumbents in competitive authoritarian regimes often limit citizens’ access.

In theory, social media as a network operates when users follow like-minded people, contributing to echo chambers and polarization. Additionally, social media as a news source is harder for autocrats to control than traditional media. Regime supporters update the most when they are exposed to less skewed information, while social media as a device for coordinating collective action means that increased use leads to increased protest and dissent.

Pia Raffler, Assistant Professor of Government at Harvard University, and her colleagues Jeremy Bowles, Post-doctoral Fellow at Stanford University, and John Marshall, Associate Professor of Political Science at Columbia University, investigated how variation in access to social media affected political attitudes during the 2021 election period in Uganda versus the comparatively “normal” post-election period. On November 15, as a part of the Fall 2023 Human Capital Initiative Research Seminar Series, Raffler presented key findings from the study, revealing that National Resistance Movement (NRM) partisans, or the subscribers to the dominant party whose social media use was subsidized for three months, came to view the NRM more negatively.

For context, social media in Uganda is widely used for finding political information, as 71 percent say political news is the main reason for using social media with 52 percent of Ugandans enjoying mobile Internet access in late 2020. Of these Internet users, 79 percent use Facebook, 78 percent use WhatsApp and 17 percent use X (formerly known as Twitter). In 2018, a social media tax was introduced, making cell phone data expensive. Following the tax, 55 percent of Internet users reported that the tax limited their social media usage, and 68 percent reported high data costs limited social media usage.

To determine whether the views on the NRM were positive or negative, the researchers analyzed the effect of social media exposure on attitudes toward the regime. They randomized the payment of social media tax and mobile data bundles among a panel of approximately 1,400 voters in 11 districts and found that regime supporters became more critical after more social media use.

The researchers also investigated how voters responded to a social media “ban” during the election season. Leveraging a difference-in-differences design during a 30-day social media ban with a two-wave panel survey of approximately 1,300 voters at the climax of the election campaign, they found that the respondents who were able to maintain social media access through virtual private networks (VPNs) came to view the dominant NRM party relatively more positively, compared to those who did not have access to VPNs at baseline. Contrarily, a field experiment implemented half a year later finds that NRM partisans whose social media use was subsidized for three months came to view the NRM more negatively. The researchers posit that this is due to the fact that social media content is relatively opposition-friendly, compared to traditional media outlets, which are more easily controlled by the dominant regime.

The election-time finding, on the other hand appears to reflect both a relative increase in pro-NRM social media content during the social media ban and a backlash effect among those who were cut off from social media. Their findings suggest that social media can bolster the support for opposition parties in competitive authoritarian contexts. At the same time, they highlight the trade-off dominant incumbents face between controlling social media content and minimizing the backlash – especially in politically salient moments.

*

Sayuri Kataoka is a Fall 2023 Communications Fellow with the Boston University Global Development Policy Center. 

Never miss an update: Subscribe to the Human Capital Initiative newsletter.