Cascate di informazione: giocatori razionali ed irrazionali e comportamento di gregge

Informational cascades occur when individuals follow the decisions of others regardless of their own information signals. This phenomenon often leads to a convergence of decisions, irrespective of the underlying information or preferences, with potential consequences ranging from conformity bias to collective errors. While informational cascades can stifle diversity of opinions and result in suboptimal outcomes, recent research suggests that interventions can positively influence their dynamics. This thesis investigates the dynamics of group decision-making in the presence of informational cascades, focusing on the interplay between perfectly rational Bayesian decision-makers and individuals exhibiting non-conformist behaviour. Through simulations and quantitative analyses, we explore whether the inclusion of non-conformist individuals, who rely on private information or oppose the majority, can enhance the overall accuracy of group decisions. Our findings reveal that, even in a regime where a small proportion of participants base their choices on private information, the group tends to make more accurate decisions, albeit at the expense of individual performance. This trend is not observed when participants choose to oppose the majority. These results underscore the importance of decision independence and diversity of opinions in improving group decision-making accuracy. While this study sheds light on the dynamics of information cascades and their potential implications for group performance, further research is needed to explore their interactions with factors such as information asymmetry, group polarization, and social influence. Understanding these complexities can provide valuable insights into strategies for enhancing decision-making accuracy in collective settings.