Deciphering Suboptimal Updating: Structure, Sequencing, and Complexity
Speaker: Prof Marina Agranov,
Affiliation: Caltech
Location: Online via Zoom
Zoom link: https://uqz.zoom.us/j/84008636651
Abstract:
We study the underlying reasons for the failure of individuals to adhere to Bayes' rule and break down this departure into three elements: (i) information structure, (ii) timing of information release, and (iii) complexity of the task at hand. In a series of controlled experiments, we alter all three elements and quantify their magnitude. We address task complexity by introducing a novel notion driven by the nonlinear calculations required for Bayesian updating. We experimentally explore our notion of complexity and find empirical support for it. Examining the possibility of counteracting one bias with another, we find that the optimal way of releasing information depends on the underlying complexity of the environment.
About the presenter:
If you would like to meet with Prof Agranov, please contact Dr Zachary Breig
About Centre for Unified Behavioural and Economic Sciences (CUBES) e-seminars
An online seminar series on Experimental and Behavioural Economics organized by the Centre for Unified Behavioural and Economic Sciences (CUBES) of the School of Economics at The University of Queensland.
Our seminars take place fortnightly via Zoom on Wednesdays at 10 am or 5 pm (AEST), depending on whether the guest speaker is streaming from US/Australia or Europe respectively.
Seminars consist of a 60-minute presentation followed by 15 minutes of Q&A.
Local time for seminars
You can check the corresponding times for your own time zone using the following links for each session: 10am, 5pm.
How to register
Clicking the button below gives you the option to register for: (i) all seminars (ii) seminars that take place at 10am or 5pm, or (iii) individual seminars.
Email invitations with a ZOOM link to the event will be sent 48 hours before each seminar.
If you wish to attend an upcoming seminar within the 48-hour window, please drop an email to Lionel Page.