webinaire sur l'inférence causale

lien vers l'inscription en ligne

9h45: présentation de la cellule FouilleData

10h: séminaire de David Makowski

Titre: “Comparison of methods for causal analysis and application to assess the impact of drought on maize yields”

Abstract: “In the absence of factorial experiments, it is tempting to rely on observational data to estimate the effect of a given factor (e.g., a climatic event, an agricultural practice) on an outcome of interest (e.g., crop yield, soil carbon content, water pollution). However, the risk of confounding and biased conclusions is often high, particularly in the environmental and agricultural sciences, where several interacting factors often have a simultaneous impact on a given outcome. Here, I present several methods to deal with this issue and reduce the risk of bias in causal analysis. The different methods are illustrated and compared in a case study quantifying the effect of drought on maize yields using agricultural surveys.”