In this episode, we dive into the intriguing world of Simpson's paradox, a statistical phenomenon that reveals how trends in an overall population can reverse when broken into subgroups. We explore how this paradox, more common than previously thought, can lead to misleading conclusions, affecting everything from research findings to public policy. Using real-world examples, we’ll discuss the dangers of relying solely on group-level data and why it's crucial to dig deeper into subgroups. We’ll also highlight the solutions proposed by researchers, including a new software tool, "Simpsons," designed to help detect and manage this paradox. Tune in to discover how understanding Simpson's paradox can improve data interpretation and decision-making across fields like psychology, medicine, and beyond.

More info: https://blog.univ-reunion.fr/alessioguarino/2019/06/05/le-paradoxe-de-simpson-statistique-analyse-des-donnees/

Ver todos los episodios