CausalML Weekly

This episode explores the foundational concepts of linear regression as a tool for predictive inference and association analysis. It details the Best Linear Prediction (BLP) problem and its finite-sample counterpart, Ordinary Least Squares (OLS), emphasizing their statistical properties, including analysis of variance and the challenges of overfitting when the number of parameters is not small rel...

This episode explores a powerful method for identifying  causal effects  in non-experimental settings. The authors, affiliated with various universities, explain the  basic RDD framework , where treatment assignment is determined by a  running variable  crossing a  cutoff value . The text highlights how modern  machine learning (ML) methods  can enhance RDD analysis, particularly when dealing with...

This episode introduces and explains the Difference-in-Differences (DiD) framework , a widely used method in social sciences for  estimating causal effects  in situations with treatment and control groups over multiple time periods. It elaborates on the  core assumption of "parallel trends"  and discusses how  Debiased Machine Learning (DML) methods  can be used to incorporate high-dimen...

This episode focuses on  methods for estimating and validating individualized treatment effects , particularly using  machine learning (ML) techniques . It explores various "meta-learning" strategies like the  S-Learner, T-Learner, Doubly Robust (DR)-Learner, and Residual (R)-Learner , comparing their strengths and weaknesses in different data scenarios. The text also discusses  covariat...

This episode  focuses on  Conditional Average Treatment Effects (CATEs) , which are crucial for understanding how treatments affect different subgroups. It contrasts CATEs with simpler average treatment effects, highlighting the complexity and importance of  personalized policy decisions . The text details  least squares methods for learning CATEs , including  Best Linear Approximations (BLAs)  an...

This episode explores advanced econometric methods for causal inference using  Double/Debiased Machine Learning (DML) . It focuses on applying DML to  instrumental variable (IV) models , including partially linear IV models and interactive IV regression models (IRM) for estimating  Local Average Treatment Effects (LATE) . A significant portion addresses  robust DML inference under weak identificat...

This episode examines  methods for causal inference  when unobserved variables, known as  confounders , complicate identifying true causal relationships. It begins by discussing  sensitivity analysis  to assess how robust causal inferences are to such unobserved confounders. The text then introduces  instrumental variables (IVs)  as a technique to identify causal effects in the presence of these h...

This episode focuses on  causal inference and the selection of control variables  within the framework of Directed Acyclic Graphs (DAGs). It explains  various strategies for constructing valid adjustment sets  to identify average causal effects, such as conditioning on parents or common causes of treatment and outcome variables. The text  differentiates between "good" and "bad"...

This episode focuses on  feature engineering , a technique that transforms complex data like text and images into numerical representations called  embeddings  for use in predictive and causal applications. It begins by explaining  principal component analysis  and  autoencoders  as methods for generating these embeddings. The text then specifically addresses  text embeddings , detailing early met...

This episode focuses on  Double/Debiased Machine Learning (DML) methods  for statistical inference on predictive and causal effects in complex regression models. It introduces  Neyman orthogonality  and  cross-fitting  as key ingredients to mitigate bias in high-dimensional settings, providing theoretical foundations and practical algorithms for  Partially Linear Regression Models (PLM)  and  Inte...

This episode explores modern nonlinear regression methods crucial for predictive inference in causal analysis . It  focuses on tree-based techniques  like regression trees, random forests, and boosted trees,  as well as neural networks and deep learning . The text  discusses the theoretical guarantees of these methods, particularly concerning their approximation quality and convergence rates under...

This episode explores causal inference through the lens of directed acyclic graphs (DAGs) and nonlinear structural equation models (SEMs) . It highlights how these models provide a  formal, nonparametric framework  for understanding causal relationships, moving beyond simpler linear assumptions. The text  introduces concepts like counterfactuals and conditional ignorability , explaining how they a...

This episode introduces linear structural equation models (SEMs) and causal diagrams, also known as Directed Acyclic Graphs (DAGs) . The text explains how these models can be used for  causal inference, particularly in economics , using examples like gasoline demand and wage gap analysis. It highlights the importance of  conditional exogeneity and the potential pitfalls of "collider bias&quot...

This episode focuses on methods for  identifying average causal effects  in observational studies. It explores the concept of  conditional ignorability , explaining how  adjusting for observed covariates  can help mitigate selection bias, making non-randomized data comparable to randomized control trials. The text further discusses the  propensity score  as a key tool, detailing its use in  reweig...

This episode focuses on  high-dimensional linear regression models , specifically discussing  causal effects and inference methods . The core of the text explains the  Double Lasso procedure , a technique utilizing Lasso regression twice to estimate predictive effects and construct confidence intervals, emphasizing its reliance on  Neyman orthogonality  for low bias. The authors illustrate its app...

This episode focuses on predictive inference using linear regression methods in high-dimensional settings where the number of predictors (p) often exceeds the number of observations (n). The text primarily explores Lasso regression, explaining its mechanism for variable selection and reducing overfitting by penalizing coefficient magnitudes. It also compares Lasso to other penalized regression tec...

This episode provides a comprehensive overview of causal inference using Randomized Controlled Trials (RCTs), often considered the gold standard in establishing cause-and-effect relationships. The text begins by explaining the potential outcomes framework and the concept of Average Treatment Effects (ATEs), contrasting them with Average Predictive Effects (APEs) and highlighting how random assignm...

This podcast, generated by NotebookLM, summarizes the Causal ML book by Victor Chernozhukov, Christian Hansen, Nathan Kallus, Martin Spindler, and Vasilis Syrgkanis. Disclosure The CausalML Book: Chernozhukov, V. & Hansen, C. & Kallus, N. & Spindler, M., & Syrgkanis, V. (2024): Applied Causal Inference Powered by ML and AI. CausalML-book.org; arXiv:2403.02467. Audio summary is gene...