Bumper Stickers on the Twitter Highway: Analyzing the Speed and Substance of Profile Changes

Bumper Stickers on the Twitter Highway: Analyzing the Speed and Substance of Profile Changes

Abstract

We describe a novel longitudinal study of the frequency and significance of social media users’ profile changes. Drawing upon two formative theories from communication and psychology: self-construal and signaling theory, we examine the likelihood that users will change their profiles and what constitutes a significant profile change. Our findings indicate that users are more likely to change their Profile Summaries and Display Names than their Locations and Screen Names (i.e. handles). Further, we used topic modeling to partition users based on their profiles to identify themes and explored how profile changes differ among these thematic groups (e.g., Trump supporters). Last, we identified the most significant word changes by users in their profiles. Our findings provide valuable baseline data for further study of Twitter profiles, including the spread of social contagion through these profiles.

Publication
In International AAAI Conference on Web and Social Media (ICWSM), ACM.
Date
Avatar
Ryan Wesslen
Machine learning engineer