The idea of using text as alternative data in economic and social science research is slowly becoming part of the mainstream. In management research, this means that textual sources like company reports, press releases and transcripts of earnings calls are used in addition to standard performance indicators. In this talk I will present some of our work at the Mannheim Center for Data Science where we explore the impact of linguistic uncertainty indicators in financial documents on the perceived risk of investing in a company and their impact on investment behaviour and market volatility. In particular, we developed a method for creating sector-specific refinements of existing uncertainty dictionaries that better capture specific characteristics of the respective section. Further, we created a neural network-based model for predicting market volatility from textual and standard financial indicators. Finally, we establish a link between uncertainty indicators in text and the investment behaviour of subjects in the context of a user study.