Exploring New
Media Environments
with Data Science

Michael W. Kearney, PhD Candidate
Center for Research Methods & Data Analysis
Department of Communication Studies
University of Kansas

My research program

 

 

Where I've been

Publications

Partisan media

Partisan networks

 

 

Where I am

Selective exposure on Twitter

Hypotheses

  1. User networks on Twitter will cluster according to partisanship
    • Partisan network homogeneity
  2. Network polarization will increase with proximity to election
    • Change in partisan network homogeneity

Method

  1. Identified every follower from 12 source, or origin, accounts representing 3 groups
  2. Randomly sampled 10,000 followers from each group
  3. Filtered out inactive/bot-like users for final sample of 3,000  
    • Democrat partisans (n = 1,000)
    • Republican partisans (n = 1,000)
    • Entertainment non-partisans (n = 1,000)

 

flowchart

 

 

 

 

flowchart

 

Results

Partisan clusters (H1)

 

H1 results

Simulated null model

Simulated alternate model

Real data

Source accounts

Clustered elites

Change in polarization (H2)

 

H2 results

Network polarization

Network polarization (weighted)

 

Where I'm going

Underlying research question

Dissertation, Twitter, and rtweet

Other projects

 

That's it \o/

Thanks!