This thesis examines the influence of sentiment in earnings calls on financial market reactions. The objective is to analyse the extent to which the tone used in presentations and Q&A segments affect abnormal returns, and how differences in investors’ information environments shape these effects. The study is based on transcripts from companies in the S&P Composite 1500, whose tone is recorded using finance-specific lexicons and evaluated within an event-study framework. The results suggest that sentiment in the Q&A section in particular triggers significant immediate market reactions, with analysts’ questions having the greatest influence. Conversely, the sentiment of the presentation assumes significance during the drift period and frequently results in subsequent corrections. In addition, investors show increased sensitivity to verbal signals when it comes to companies with a higher degree of uncertainty. The thesis contributes to the literature by demonstrating that context-specific sentiment measurement provides valuable insights for the evaluation of qualitative company information. Furthermore, it highlights the limitations of generic text lexicons.
