This is a session of Topic 4: Statistics education at the post secondary (tertiary) level
(Tuesday 13th, 14:00-16:00)
Integrating Bayesian methods with traditional statistics education
- Laura Martignon (Germany)
AbstractBayesian inference is as old as probability theory but it fell into disfavour in the nineteenth and early twentieth centuries due to a quest for objectivity on one hand, and a flurry of experiments in cognitive psychology on the other, apparently demonstrating that the human mind is not hard wired for Bayesian reasoning. Its recent resurgence is due to new insights on cognitive aspects of human Bayesian reasoning as well as to the capability of Bayesian inference models to integrate various types of information providing a unified framework for risk assessment and decision making.
In many real life instances, not just at the individual but at the institutional level, Bayesian strategies are now being applied consistently and successfully. Thus integrating Bayesian reasoning in the standard teaching of statistical methods has become a necessity. Our session will be devoted to the steps for tackling this challenge at the tertiary level. Three aspects will be recurrent in this session:
- Bayesian Statistics as an integrating component of the standard statistics curriculum: how large and how deep should such a component be?
- Biases in Bayesian reasoning and how to make them disappear (e.g., base rate neglect)
- Bayesian Networks and software for learning them from data (e.g., Netica)
|Paper||Title||Presenter(s) / Author(s)|
|4I1||Psychology students’ understanding of elementary Bayesian inference||Carmen Díaz (Spain)|
|4I2||Comparing the Bayesian and likelihood approaches to inference: a graphical approach||Bill Bolstad (New Zealand)|
|4I3||The very beginning of a class on inference: classical vs Bayesian||Lisbeth Cordani (Brazil)|
|4I4||Teaching young grownups how to use Bayesian networks||Stefan Krauss (Germany)|
Georg Bruckmaier (Germany)
Laura Martignon (Germany)