Beyond deception, bias and variance empirical perception on winning argument sentiment reception

Argument mining has seen significant advancements in recent years, particularly in how arguments are extracted and analyzed. However, there has been limited research on the behaviour of persuasive argument components, such as contexts, premises, and claims. This paper aims to explore the variability of sentiment within these components in relation to their persuasiveness. Additionally, it examines how model bias effects this sentiment variance and its implications for determining the persuasiveness of arguments. Furthermore, the study investigates the impact of model bias on multiple hypothesis concerning the use of sentiment from various argument components to gauge overall persuasiveness.