Detecting psychological change through mobilizing interactions and changes in extremist linguistic style

Social media interactions are popularly implicated in psychological changes like radicalization. However, there are currently no viable methods to assess whether social media interactions actually lead to such changes. The purpose of the current research was to develop a methodological paradigm that can assess such longitudinal change in individuals’ social media posts. Using this method, we analyzed the longitudinal timelines of 110 Twitter users (40,053 tweets) who had expressed support for Daesh (also known as Islamic State, or ISIS) and we compared them to a baseline sample of twitter timelines (215,008 tweets by 109 users) to investigate the factors associated with within-person increases in conformity to the vernacular and linguistic style of tweets that supported violent extremism. We found that conformity to both extremist group vernacular and linguistic style increased over time, and with mobilizing online interactions. Thus, we show how to detect within-person changes over time in social media data and suggest why these changes occur, and in doing so, validate a methodological paradigm that can detect and predict within-person change in psychological group memberships through social media interactions.