Example-Presentation: Example Barbera (2015)
- Your presentations
- Go through questions and provide answers
- Use paper’s visualizations if there are any nice ones
- The paper, Google scholar
- My (wrong) impressions back in 2013… [Think out of the box!]
- Background: What is the background of that study? Research area?
- Classic task of measuring politicians’ and voters’ policy positions
- Usually to position on a single latent dimension “ideology”
- Was mostly done for legislators (voting data) & individuals (survey data)
- Research question: What is the overall research question?
- Can we use Twitter to estimate the policy positions of citizens?
- Underlying: What is the ideology of individual i?
- Data: What kind of data is used?
- Twitter data (back then a new platform of political communication)
- Users (individual & politicians) embedded in a common social network
- Hypotheses/assumptions: Are there hypotheses, if yes which ones?
- Main assumption: Twitter users prefer following politicians that are close ideologically
- Hypothesis: We can reliably estimate voter’s ideology using their twitter network (who they follow)
- Analytical & empirical approach
- Estimate individuals’ ideology using Twitter
- Cross-validate Twitter estimates with external information to see whether method correctly classifies/scales Twitter users on the left or right side of the ideological dimension
- Legislators/parties: Compare twitter ideology to voting behavior in parliaments
- Mass ideology: Compare patterns between surveys and Twitter in terms of ideological patterns
- Individuals: Compare campaign spending to Twitter ideology
- Findings: What are the findings?
- Method seems to work
- Twitter conversation is dominated by those with strong ideology
- Innovation: What is so innovative (CSS!)? What struck you about this study? What did you find interesting?
- New form of data (Twitter)
- Classic question (identifying ideology) answered with new data
- Weaknesses/Problems: What are potential weaknesses of the study?
- Representativeness of Twitter users, assumptions, only applicable to users who follow at least one account etc.
- Other thoughts: Any additional thoughts on the study?