The study of Bayesian reasoning through the use of interaction analysis methods.

  • Manuele Reani

Student thesis: Phd

Abstract

Humans find reasoning about uncertainty difficult and often struggle to understand concepts such as conditional probability, which underpin the interpretation of important procedures such as medical tests. Bayesian reasoning is the process of updating one's belief about an uncertain situation in light of new evidence. This way of reasoning may look trivial and intuitive. However, a wealth of research shows that many people struggle with it, perhaps because they are prone to certain biases. In the media, and also in tools specifically developed for analysts (e.g., decision support systems and software for intelligence analysis) graphical representations are commonly used to display uncertainty. Visualisation methods have proved effective in communicating risk at many levels, but their role in facilitating Bayesian reasoning is complex, with contradictory evidence as to their utility, and a lack of understanding as to how they are best applied. Whilst theoretical work has tried to address this issue, there remains a lack of consensus about how people approach Bayesian reasoning tasks, especially when these are presented in graphical formats. The increasing necessity of communicating statistical data to the general public for making informed decisions highlights the importance of studying risk reasoning.
Date of Award1 Aug 2019
Original languageEnglish
Awarding Institution
  • The University of Manchester
SupervisorCaroline Jay (Supervisor) & Niels Peek (Supervisor)

Cite this

'