Konstantinos Georgalos, da Lancaster University, vai apresentar o seu trabalho de investigação. 

Heuristics and Biases in Risky Choice: Testing the
Cognitive Toolbox Model

Rich evidence from economic and psychology experiments has confirmed that human
beings often deviate from the Bayesian rational decision-making model. As a response,
the theoretical literature, in an effort to explain the observed violations of rationality, has
suggested either the use of generalisations of the standard Expected Utility model based
on optimisation behaviour, such as Prospect Theory, or the use of fast and frugal heuristics
(rules of thumb). While the empirical literature has provided exhaustive tests of one-to-one
comparisons between Prospect Theory and a particular heuristic (e.g. the Priority Heuristic),
there is a lack of empirical evidence on how well a cognitive toolbox model of risky
choice would perform against Prospect Theory. In this paper we use data on risky choice
from various experimental studies, in various domains, and using Bayesian Hierarchical
estimation techniques, we rigorously test the performance of the toolbox model. We identify
the conditions within which the model performs well, as well as, the size (number of
tools) and the tools (heuristics) that subjects are using when they choose.