Bridging the Digital Divide: A Randomized Control Trial on Home Internet Connections in South African Townships

This pre-analysis plan describes the data collection and analysis strategy for an endline survey to a randomized control trial in a township in Durban, KwaZulu-Natal, South Africa. The endline is a follow-up to the previously registered study "#YourVoiceYourVote: Experiment on Social Media GOTV Campaign in South Africa" (#YourVoiceYourVote: Experim…), which focused on a post-election turnout survey during the 2024 national elections in South Africa. The study is a field experiment in a township in KwaZulu-Natal. We randomly assigned households in the township to two interventions. First, in partnership with the South African NGO Project Isizwe (projectisizwe.org/), we provide households with a home WiFi router and unlimited internet connection for the duration of the study. Second, we encouraged the participants in study households sampled during baseline voters via text messages to access a GOTV campaign on social media during the 2024 national elections. We partnered with a South African social media influencer with more than 350,000 followers on Facebook to design this campaign. The two interventions are cross-randomized, forming a full factorial design. We measure our key outcome in an endline survey in early 2025. We seek to look at the effects of the two interventions on several sets of outcomes, including political participation, political attitudes, news consumption and misinformation, engagement in national vs local politics, access to public services, as well as employment and education.