Winning! Adjudication and Dialogue in Social Media

Abstract

This article introduces the concept of adjudication to define the act of granting or denying ownership of an outcome to individuals or groups in social media. We extend existing models of political dialogue to explain differences between winners and losers when elections are adjudicated. We use Twitter data on three elections in Argentina, Brazil, and the United States. We show an increase in event salience upon adjudication, followed by more extensive dialogue among winners and disengagement among losers. Further, we show differences in the network structure of dialogue, with dialogue in winning communities being less hierarchical and dialogue in losing communities being more hierarchical. Our model provides information about the quality of the adjudicator, information drift prior to adjudication, and the level of user engagement. Finally, we describe further extensions to areas of toxic speech and sports. We identify the causal effects of adjudication using a regression discontinuity design

Publication
Working paper, Under Review

Abstract

This article introduces the concept of adjudication to define the act of granting or denying ownership of an outcome to individuals or groups in social media. We extend existing models of political dialogue to explain differences between winners and losers when elections are adjudicated. We use Twitter data on three elections in Argentina, Brazil, and the United States. We show an increase in event salience upon adjudication, followed by more extensive dialogue among winners and disengagement among losers. Further, we show differences in the network structure of dialogue, with dialogue in winning communities being less hierarchical and dialogue in losing communities being more hierarchical. Our model provides information about the quality of the adjudicator, information drift prior to adjudication, and the level of user engagement. Finally, we describe further extensions to areas of toxic speech and sports. We identify the causal effects of adjudication using a regression discontinuity design