Key points are not available for this paper at this time.
We use an experiment across the Arab Twittersphere and a nationally representative survey experiment in Lebanon to evaluate what types of counter-speech interventions are most effective in reducing sectarian hate speech online. We explore whether and to what extent messages priming common national identity or common religious identity, with and without elite endorsements, decrease the use of hostile anti-outgroup language. We find that elite-endorsed messages that prime common religious identity are the most consistently effective in reducing the spread of sectarian hate speech. Our results provide suggestive evidence that religious elites may play an important role as social referents—alerting individuals to social norms of acceptable behavior. By randomly assigning counter-speech treatments to actual producers of online hate speech and experimentally evaluating the effectiveness of these messages on a representative sample of citizens that might be incidentally exposed to such language, this work offers insights for researchers and policymakers on avenues for combating harmful rhetoric on and offline.
Siegel et al. (Mon,) studied this question.