Measuring morality in text is a complex and context-specific task that many researchers andcomputational social scientists have attempted to tackle. Understanding morality in the context of political violence adds an additional layer of nuance worth further exploration. My research investigates the moral language utilized in online spaces as it relates to acts of political violence, as well as the methods used to conduct those investigations. Using text-as-data and natural language processing methods, this dissertation categorizes and represents the features of language most prevalent in responses to YouTube videos depicting either perceived left- or right-leaning perpetrators of these acts. Moreover, this work outlines the opportunities and challenges associated with the common methods used for measurement, including dictionaries, large language models and word embeddings. Each chapter of this dissertation is dedicated to a specific method, and describe the results and takeaways that can be gleaned from using them. Finally, this work introduces an embedding based translation method for mapping intra-language terms, ideas and values between divergent groups.