No existing model of political rhetoric fully captures the complex interplay between the mainstream-populism divide and appealing to emotions like fear and anger. We present a new conceptualization and procedure that defines populism in relation to governmentalism, operationalizes both through communication frames, and allows for the analysis of emotions. We separate governmentalist-populist contestation from contestation between government and opposition, solving a longstanding theoretical and empirical problem. Analyzing one million tweets by politicians and their audiences, we fine-tune and employ supervised machine learning (transformer models) to classify populist and governmentalist communication. We find that populist tweets appeal more to anger and more to fear than governmentalist tweets. While we deploy our approach for tweets about Coronavirus in the UK, the procedure is transferable to other contexts and communication platforms.
