Classification of Users in Social Media: Employing Mixed Methods to Categorize Participants in Twitter Discussions

• In Event: AERA Virtual Poster Session 4
  In Poster Session: Benefits of Technology in Facilitating Mixed and Multi-Method Research

Abstract

Social media is an important source of health-related information. However, it is also subject to influencers who are trying to affect communication flows, possibly hampering constructive, open communication, based on ideological differences or for-profit considerations. Using a mixed-methods approach, we will analyze Twitter data on the general topic of breastfeeding, as an exemplary case study, in trying to assess the extend with which communication flows are possibly affected by influencers. We will categorize types of participants to assess how underlying attributes and characteristics might affect shared information. Our findings will constitute a proof-of-concept that a mixed-methods approach, including machine learning, can be instrumental in categorizing types of users and describing how they can affect communication flows on Twitter.

Authors

• Martin Rehm, University of Regensburg

• Hamid Karimi, Michigan State University

• Sara Moukarzel

• Anita Caduff, University of California - San Diego

• Miguel Del Fresno, Universidad Nacional de Educación a Distancia

• Rafael Pérez-Escamilla, Yale University