Trial and Error: use and misuse of probability in criminology (Leila Schneps, Paris)

24.09.2019 15:00

In the course of a criminal trial, prosecution and defense present multiple pieces of evidence of different types, from notoriously unreliable eyewitness testimony to established facts that may or may not be directly relevant to purely scientific elements such as analysis of chemical or biological traces from the crime scene. The jury is confronted with the task of evaluating the weight of the evidence taken as a whole, which is a supremely difficult task given that the pieces of evidence are usually interdependent. Numerous studies have shown that in fact, normal human intuition is generally far off in making this kind of estimation, and what is worse, everybody seems to err in the same direction. Using several real-life cases and audience participation, we will illustrate this problem, and discuss how to approach it with a probabilities Bayesian model.

Lieu

Bâtiment: Battelle

Villa Battelle, Séminaire "Groupes de Lie et espaces des modules"

Organisé par

Faculté des sciences
Section de mathématiques

Intervenant-e-s

Leila Schneps, Paris

entrée libre

Classement

Catégorie: Séminaire