🔎 What Was Studied

Social media use has surged among citizens and politicians in Brazil, and platforms like Twitter became central arenas for debate and propaganda during the highly polarized 2014 legislative and presidential elections. This study asks whether the decision to 'follow' a profile on Twitter can be used to estimate politicians' ideological positions and whether that approach can expose variation within a fragmented legislative body such as the Brazilian Chamber of Deputies.

🧭 How Ideology Was Estimated on Twitter

• Constructed a social network from politicians' follow ties on Twitter and applied a Bayesian spatial model developed by Barberá (2015).
• The model treats follow decisions as informative signals about ideological proximity and locates actors on a latent ideological dimension.

📊 Data and Application

• Uses Twitter follow data from political actors in Brazil to build the network used for estimation.
• Applies Barberá's Bayesian spatial approach to produce ideal points for both elected representatives and public actors who engage in political debate but are not professional politicians.

✨ Key Findings

• The Twitter-based ideal points successfully capture differences between parties and between individual political actors.
• Estimated positions from Twitter follow networks align closely with patterns observed in roll-call vote-based ideal points.
• The method reveals ideological variance within a highly fragmented Chamber of Deputies and allows placement of actors who do not cast roll-call votes.

🧾 Why It Matters

This approach offers a viable, complementary way to measure political ideology when roll-call data are unavailable or incomplete and extends measurement to non-legislative participants in public debate. The successful application of Barberá's model to Brazil demonstrates the method's usefulness in polarized and fragmented political environments.