Optimal Sequential Decision with Limited Attention


Speaker:Konrad Mierendorff, University College London
Date: Friday 23 March 2018
Time: 3.30pm
Location: Kolade Teaching Room

Further details

We consider a dynamic model of information acquisition. Before taking an action, a decision maker may direct her limited attention to collecting different types of evidence that support alternative actions. The optimal policy combines three strategies: (i) immediate action, (ii) a contradictory strategy seeking to challenge the prior belief, and (iii) a confirmatory strategy seeking to confirm the prior. The model produces a rich dynamic stochastic choice pattern as well as predictions in settings such as jury deliberation and political media choice. Keywords: Wald sequential decision problem, choice of information, contradictory and confirmatory learning strategies, limited attention