S2, 6 ECTS
Alexandre Grellet et Marc Gurgand
Course given in English
The aim of this course is to train students in statistical methods for estimating causal relationships, based on experiments or quasi-experiments. These methods involve either implementing controlled experimental protocols; or mobilizing statistical data to exploit an experiment given by « nature » or social, economic, institutional events, etc. which, under certain hypotheses, produce differentiated exposures to treatment in different populations, and which make a causal interpretation plausible.
The course will describe the objects of interest (counterfactuals, causal effects), the assumptions under which they can be identified in a given context, the corresponding estimation methods and the statistical properties of the estimators. It constantly articulates statistical tools and their empirical use, notably by systematically reproducing published articles, and by having researchers present papers applying each of the main methods. It also mobilizes the Equipex IDEE to present concrete examples of experimental designs and the practical implementation of controlled experiments.
This course is given on the Jourdan Campus, is compulsory for the Economics department’s L3, and is an elective course of the Cogmaster. Cogmaster students will have tutorials in the DEC’s premises; L3 students will prepare an empirical project with a support of a devoted TA.
Provisional program of the course:
- Rubin causal model
- Randomized control trials (RCT)
- RCT with imperfect compliance
- The many designs of an experiment (assuré par l’Equipex IDEE)
- Power calculations, surveys, attrition and further threats (assuré par l’Equipex IDEE)
- RCT : researcher presentation
- Difference in difference (DID)
- DID : researcher presentation
- Regression discontinuity design (RDD)
- RDD : researcher presentation
- Synthetic control, propensity score matching (PSM)
- An introduction to design based and population based inference ; and to multiple testing issues
- An introduction to heterogenous effects models
Ce cours est obligatoire dans le cadre de la L3 d’économie
L’inscription à ce cours est obligatoire en écrivant à economie@ens.psl.eu