#Polar Scores: Tweets Used to Measure Political Polarization

#Polar scores provide a novel approach to measuring political polarization using Twitter content. The dataset contains 55,244 tweets used to develop the algorithm, with fields including tweet metadata, user information, and party labels. The method and open-source Python code were developed by Aron Culotta and colleagues for the Journal of Information Technology & Politics.

Use Cases

• Measuring political polarization of users based on Twitter content and party labels.
• Analyzing partisanship of hashtags using text and hashtag frequency data.
• Developing chamber-agnostic polarization metrics applicable beyond the U.S. Congress.
• Training or benchmarking text-based algorithms for estimating political ideology.

Strengths

• Dataset contains 55,244 tweets with associated metadata.
• Includes explicit party labels and user demographic information such as sex, state, and chamber.
• Accompanied by open-source Python code for calculating the #Polar scores.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.
• The file contains one row per hashtag, meaning tweets may appear multiple times, which could complicate analysis.

Provenance

Source

Twitter content, as described in the paper by Hemphill, Culotta, and Heston.

Collection Method

Leverages party labels, word frequency, and hashtags from tweets to calculate polarization scores.