Mining Online Discourse Related to Transgender Exclusive Policies in Interscholastic Sport: an Exploratory Natural Language Processing Study.

Introduction: In recent years, many US states have introduced legislation to restrict the access of transgender people to the interscholastic sport. Methods: We compiled a corpus of Twitter posts (n = 17,182) between October 2021 and March 2022 related to trans-exclusionary policies in sport. We performed a cluster analysis on these tweets to explore salient topics in the data. Machine learning algorithms were used to identify automated bots in each topic as well as to measure sentiment in each tweet. Results: Four major clusters of tweets were present in the corpus. Two were primarily driven by human-authored and focused on fairness, both supportive and exclusive of trans athletes. The others were largely driven by bot accounts and included news sharing and vitriolic media sharing. Much of the conversation was driven by automated bot accounts. Conclusions: Primary arguments surrounding trans athletes focus on fairness. However, at least on Twitter, the issue has been given outsized salience through the use of automated bot accounts. This increased issue of salience has influenced a spate of exclusionary legislation, despite the known benefits of participating in sports and physical activity. Policy Implications: Social media has the potential to influence policy-making and legislators but is increasingly polarized. With respect to trans sport, much of the conversation was driven by bot accounts, which do not reflect the true opinions of individual users. Policymakers should be cautious in the use of social media as a tool for generalizing public opinion on contentious issues. [ABSTRACT FROM AUTHOR]