因果ダイアグラム:結論の前に仮定をかこう

構造的因果モデルの勉強を少しずつはじめている関係で,edXにあがっているCausal Diagrams: Draw Your Assumptions Before Your Conclusionsという講義を受けてみました.講師はMiguel Hernánで他にもRobins,Pearl,VanderWeeleへのインタビューやショートレクチャーもあるように,なんとも豪華な顔ぶれ.9週でデザインされたコースのようだが,まとまった時間があれば数日で終わる内容だった.テクニカルな話はほとんどないので,真面目に勉強するならHernán and Robins (2018) Causal Inference を読んでねという流れ(AcknowledgmentsにTadayoshi Fushiki先生のお名前を発見).パールの肉声を初めて聞いたけどもう81歳なんですね.ちなみにシラバスは以下の通り.

Course Description

This course introduces causal diagrams as tools for researchers who study the effects of treatments, exposures, and policies. The course focuses on translating expert knowledge into a causal diagram, drawing causal diagrams under different assumptions, and using causal diagrams to identify common biases and guide data analysis. The first part of the course introduces the theory of causal diagrams and describe its applications to causal inference. The second part of the course presents a series of case studies that highlight the practical applications of causal diagrams to real-world questions from the health and social sciences.

Course Outline

Lesson 1: Causal Diagrams
Released on September 26, 2017

Lesson 2: Confounding
Released on October 3, 2017

Lesson 3: Selection Bias
Released on October 10, 2017

Lesson 4: Measurement Bias/ Putting it All Together
Released on October 17, 2017

Lesson 5: Time-varying Treatments
Released on October 24, 2017

Cases:
Released on October 31, 2017

The Birth Weight Paradox with Dr. Allen Wilcox
Measurement Bias in Memory Loss with Dr. Maria Glymour
Confounding in Mediation Analysis with Dr. Tyler VanderWeele
Genes as Instrumental Variables with Dr. Sonja Swanson