Conflict and polarization thrive on social media. To reduce conflict and bring communities back to healthy discussions, researchers are developing algorithms that identify people who can act as bridges between contrasting communities. But what are the key characteristics of people who can thwart conflict most effectively?
A possible question to this answer lies in the concept of Integrative Complexity (IC). IC is a psychometric that measures the degree to which a person can recognize of multiple perspectives for the issue at hand and, having identified such perspectives, can integrate them in a new perspective, thus identifying paths for conflict resolution.
Based on psychology literature, we built a machine learning model for automated scoring of any piece of text in terms of IC. We conduct for the first time an analysis of IC at scale by applying our tool to over 400k textual snippets from Reddit. Results obtained on a support-based forum focused on mental health match theoretical expectations, thus providing initial evidence of external validity to our tool and setting the stage for its usage in the context of large-scale social media analytics.