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

Abstract

We describe a novel observational study of the frequency and significance of social media users' profile changes. Drawing upon literature from impression management, specifically two formative theories: self-construal and signaling theory, our research examines the likelihood that users will change their profiles, what constitutes a significant profile change, and how profile changes could correspond to crisis events. Using Twitter's public API, we created a longitudinal dataset of profile snapshots for 2.3MM active Twitter profiles across a two-week period. Our findings indicate users are more likely to change their Profile Summaries and Display Names than their Locations and Screen Names (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 and explored the possible effects of crisis events, in this case the Las Vegas shooting of October 2017. Using a representative sample of Twitter users, our study is the first to longitudinally examine Twitter profile changes. Our findings provide valuable baseline data for further study of Twitter profiles, including social contagion spread through these profiles and the impact of external events on users' identities.

Misleading Repurposing on Twitter

“…twitter.com/justinbieber). Additionally, prior work has shown that this attribute is fairly stable for most users [38]. We found that only 13.3% (59 million) of users in the archive dataset change changed their scree name, confirming this finding.…”

“…Previous work has analyzed how users change their profile attributes, primarily to uncover when and under which circumstances these changes are made [25,30]. These studies find that screen name and location are the most stable profile attributes [38] and that users change their attributes for many reasons, including maintaining multiple accounts, changing user identifiability (anonymizing or deanonymizing oneself) and username squatting [17]. Regarding screen names in particular, Mariconti et al found that adversaries hijack the screen names of popular users who recently changed their screen name in order to gain visibility, often with malicious intent.…”

Misleading Repurposing on Twitter

“…Misleading repurposing employs attribute changes to signal a change of identity or purpose. Previous work analyzed how users change their profile attributes, primarily to uncover under which circumstances these changes are made (Shima, Yoshida, and Umemura 2017;Neha et al 2019;Wesslen et al 2018). Jain et al (2016) found that users change their attributes to maintain multiple accounts, change user identifiability, and for username squatting.…”

Cross-Country Analysis of User Profiles for Graph-Based Location Estimation

“…We assume that the amount of self-declared information is related to the number of online relationships. Wesslen et al [23] reported that users add words to or remove words from their description. For example, Trump supporters add the word "Trump" to their descriptions, so we assume that users with a long description have more context.…”