PEARL VS RUBIN (GELMAN)

Similar documents
UCLA Department of Statistics Papers

Controlling for latent confounding by confirmatory factor analysis (CFA) Blinded Blinded

ANALYTIC COMPARISON. Pearl and Rubin CAUSAL FRAMEWORKS

An Introduction to Causal Analysis on Observational Data using Propensity Scores

Since the seminal paper by Rosenbaum and Rubin (1983b) on propensity. Propensity Score Analysis. Concepts and Issues. Chapter 1. Wei Pan Haiyan Bai

CompSci Understanding Data: Theory and Applications

Targeted Maximum Likelihood Estimation in Safety Analysis

On the Use of the Bross Formula for Prioritizing Covariates in the High-Dimensional Propensity Score Algorithm

OUTLINE CAUSAL INFERENCE: LOGICAL FOUNDATION AND NEW RESULTS. Judea Pearl University of California Los Angeles (

Estimating the Marginal Odds Ratio in Observational Studies

DATA-ADAPTIVE VARIABLE SELECTION FOR

Automatic Causal Discovery

Statistical Models for Causal Analysis

An Introduction to Causal Mediation Analysis. Xu Qin University of Chicago Presented at the Central Iowa R User Group Meetup Aug 10, 2016

Causality II: How does causal inference fit into public health and what it is the role of statistics?

Confounding Equivalence in Causal Inference

Propensity Score Methods for Causal Inference

What Causality Is (stats for mathematicians)

Introduction to Causal Calculus

Gov 2002: 4. Observational Studies and Confounding

Authors and Affiliations: Nianbo Dong University of Missouri 14 Hill Hall, Columbia, MO Phone: (573)

Instrumental variables as bias amplifiers with general outcome and confounding

Causal Directed Acyclic Graphs

Selection on Observables: Propensity Score Matching.

Reasoning Under Uncertainty: Bayesian networks intro

Learning in Bayesian Networks

Quantitative Economics for the Evaluation of the European Policy

On a Class of Bias-Amplifying Variables that Endanger Effect Estimates

CAUSAL INFERENCE IN THE EMPIRICAL SCIENCES. Judea Pearl University of California Los Angeles (

Summary and discussion of The central role of the propensity score in observational studies for causal effects

Online Appendix to Yes, But What s the Mechanism? (Don t Expect an Easy Answer) John G. Bullock, Donald P. Green, and Shang E. Ha

Research Design: Causal inference and counterfactuals

Variable selection and machine learning methods in causal inference

Propensity Score Matching

1. what conditional independencies are implied by the graph. 2. whether these independecies correspond to the probability distribution

Challenges (& Some Solutions) and Making Connections

When Should We Use Linear Fixed Effects Regression Models for Causal Inference with Panel Data?

arxiv: v1 [math.st] 28 Feb 2017

Omitted Variables, Countervailing Effects, and The Possibility of Overadjustment

Advanced Quantitative Research Methodology, Lecture Notes: Research Designs for Causal Inference 1

Discussion of Papers on the Extensions of Propensity Score

Causal Analysis in Social Research

CMPT Machine Learning. Bayesian Learning Lecture Scribe for Week 4 Jan 30th & Feb 4th

Ignoring the matching variables in cohort studies - when is it valid, and why?

From Causality, Second edition, Contents

The Impact of Measurement Error on Propensity Score Analysis: An Empirical Investigation of Fallible Covariates

Causal Inference with Big Data Sets

The decision theoretic approach to causal inference OR Rethinking the paradigms of causal modelling

Causal Inference Lecture Notes: Causal Inference with Repeated Measures in Observational Studies

CAUSALITY. Models, Reasoning, and Inference 1 CAMBRIDGE UNIVERSITY PRESS. Judea Pearl. University of California, Los Angeles

Causal Inference Basics

OF CAUSAL INFERENCE THE MATHEMATICS IN STATISTICS. Department of Computer Science. Judea Pearl UCLA

p L yi z n m x N n xi

University of California, Berkeley

Causal Mechanisms Short Course Part II:

Journal of Biostatistics and Epidemiology

CAUSAL INFERENCE IN STATISTICS. A Gentle Introduction. Judea Pearl University of California Los Angeles (

Help! Statistics! Mediation Analysis

A proof of Bell s inequality in quantum mechanics using causal interactions

Weighting Methods. Harvard University STAT186/GOV2002 CAUSAL INFERENCE. Fall Kosuke Imai

Empirical Bayes Moderation of Asymptotically Linear Parameters

Bayesian Networks BY: MOHAMAD ALSABBAGH

A Distinction between Causal Effects in Structural and Rubin Causal Models

Bootstrapping Sensitivity Analysis

Matching via Majorization for Consistency of Product Quality

Front-Door Adjustment

OUTLINE THE MATHEMATICS OF CAUSAL INFERENCE IN STATISTICS. Judea Pearl University of California Los Angeles (

Causal modelling in Medical Research

Graphical models and causality: Directed acyclic graphs (DAGs) and conditional (in)dependence

19 Effect Heterogeneity and Bias in Main-Effects-Only Regression Models

arxiv: v2 [cs.ai] 26 Sep 2018

Combining multiple observational data sources to estimate causal eects

DEALING WITH MULTIVARIATE OUTCOMES IN STUDIES FOR CAUSAL EFFECTS

Recall from last time: Conditional probabilities. Lecture 2: Belief (Bayesian) networks. Bayes ball. Example (continued) Example: Inference problem

When Should We Use Linear Fixed Effects Regression Models for Causal Inference with Longitudinal Data?

Causal Discovery. Richard Scheines. Peter Spirtes, Clark Glymour, and many others. Dept. of Philosophy & CALD Carnegie Mellon

Lecture Discussion. Confounding, Non-Collapsibility, Precision, and Power Statistics Statistical Methods II. Presented February 27, 2018

A Decision Theoretic Approach to Causality

Vector-Based Kernel Weighting: A Simple Estimator for Improving Precision and Bias of Average Treatment Effects in Multiple Treatment Settings

Genetic Matching for Estimating Causal Effects:

Section 10: Inverse propensity score weighting (IPSW)

Covariate Balancing Propensity Score for General Treatment Regimes

A Note on Adapting Propensity Score Matching and Selection Models to Choice Based Samples

A noninformative Bayesian approach to domain estimation

The 2004 Florida Optical Voting Machine Controversy: A Causal Analysis Using Matching

Microeconometrics. C. Hsiao (2014), Analysis of Panel Data, 3rd edition. Cambridge, University Press.

The propensity score with continuous treatments

Confounding Equivalence in Causal Inference

Randomized trials for policy

Recall from last time. Lecture 3: Conditional independence and graph structure. Example: A Bayesian (belief) network.

9/12/17. Types of learning. Modeling data. Supervised learning: Classification. Supervised learning: Regression. Unsupervised learning: Clustering

Graphical Representation of Causal Effects. November 10, 2016

CPSC 340: Machine Learning and Data Mining. Regularization Fall 2017

causal inference at hulu

Notes on causal effects

Some challenges and results for causal and statistical inference with social network data

review session gov 2000 gov 2000 () review session 1 / 38

Validating Causal Models

Comments on Best Quasi- Experimental Practice

Bayesian Networks: Construction, Inference, Learning and Causal Interpretation. Volker Tresp Summer 2016

Transcription:

PEARL VS RUBIN (GELMAN) AN EPIC battle between the Rubin Causal Model school (Gelman et al) AND the Structural Causal Model school (Pearl et al) a cursory overview Dokyun Lee

WHO ARE THEY? Judea Pearl Professor @ UCLA Computer Science V S Don B. Rubin Professor @ Harvard Statistics Probabilistic approach to AI Contributed to the development of Bayesian networks (belief propagation, graphical models - subsumes many stat models such as Kalman filtering, Markov models, Ising models etc) One of the first to mathematize causal modeling in the empirical sciences. Developing a method of causal and counterfactual inference based on structural models Rubin Causal Model Pioneer of Observational studies Causal inference in experiments and observational studies Inference in sample surveys with nonresponse and in missing data problems Advised by Cochran

ACTUALLY INVOLVED Judea Pearl Professor @ UCLA Computer Science V S Andrew Gelman Professor @ Columbia Statistics Probabilistic approach to AI Contributed to the development of Bayesian networks (belief propagation, graphical models - subsumes many stat models such as Kalman filtering, Markov models, Ising models etc) One of the first to mathematize causal modeling in the empirical sciences. Developing a method of causal and counterfactual inference based on structural models Advised by Don Rubin Prominent Bayesian Statistician with a famous blog Some may already know him from STAT 542 textbook. Applies Bayesian analysis to Political Science

SOME POSTS TO GIVE YOU AN IDEA Andrew Gelman @ Columbia University author of many books including Bayesian Data Analysis used in STAT 542 Larry Wasserman @ CMU author of the purple book All of Statistics + other book Also involved (some indirectly, some only by their papers): Philip Dawid, Jeff Wooldridge, Dehejia and Wahba, Imbens, Michael Sobel + lot more, of course Paul Rosenbaum is mentioned many times. Total 6 long blog entrees, 91 comments by many leader of the field, many research letters and notes

SOME MORE BACKGROUND Causality 2009 Second edition by Judea Pearl: The method of propensity score is based on a simple, yet ingenious, idea of purely statistical character [...] The condition was articulated in the cryptic language of potential outcome, stating that the set [X] must render [Z] Strongly ignorable, i.e., {Y_0,Y_1} ind [Z] [X]. As stated several times in this book, the opacity of ignorability is the Achilles heel of the potentialoutcome approach - no mortal can apply this condition to judge whether it holds even in simple problems, with all causal relationships correctly specified, let alone in partially specified problems that involve dozens of variables.

THE BEGINNING 1: 2008, Letter to the editor of Statistics in Medicine, Ian Shrier presented a question to Don Rubin. Is it possible that, asymptotically, the use of Propensity Scores (PS) methods may actually increase, not decrease, overall bias, compared with the crude, unadjusted estimate of a causal effect? 2: Shrier, Sjolander, and Pearl sent three separate letters to Statistics in Medicine in which M-bias was explained and exemplified 3: 2009 Rubin in response: To avoid conditioning on some observed covariates in the hope of obtaining an unbiased estimator because of phantom but complementary imbalances on unobserved covariates, is neither Bayesian nor scientifically sound but rather it is distinctly frequentist and nonscientific ad hocery. 4: 2009, Judea Pearl Myth, Confusion, and Science in Causal Analysis 2009 Statistics in Medicine. Of course, Yes; the M-graph model presented by Shrier provides a simple such example [...] Rubin pleaded to be puzzled and confused by the terminology, by the example, and by graphs in general

WHAT IS THIS M-BIAS? Rests on Berkson Paradox Two independent causes of a common effect become dependent when we observe the effect; information refuting one cause should make the other more likely. e.g. outcome: late to Stat 921 one reason: woke up super late second reason: had to save the world again woke up late saved world P(save world = yes late = yes )!= P(save world = yes late = yes,woke up late = no ) Thus save world is not independent of woke up late given late

BAYES BALL ALGORITHM

SUPER SIMPLIFIED VERSION OF THEIR PHILOSOPHY Judea Pearl Professor @ UCLA Computer Science V S Don B. Rubin Professor @ Harvard Statistics Causality doesn t come without manipulation Granger Causality is not causality Set up a causal model with graphical models and determine the relationship between the variables. In words of Gelman: The research programme under which all causal inference problems can be framed in terms of graphs, colliders, the do operator, and the like What we ve been learning all along in the course. In words of Gelman: The research programme under which all causal inference problems can be framed in terms of potential outcomes

ONE ISSUE RAISED AMONG MANY (BACK TO M-BIAS) Rubin: Condition on all pre-treatment variables Pearl: Do not condition on all information lest some confounders introduce more bias. graphical models: helpful or not.

SOME PAPERS ON THIS ISSUE Brookhard, M. Alan, Sebastian Schneeweiss, Kenneth J. Rothman, Robert J. Glynn, Jerry Avorn, & Til Stu rmer. 2006. Variable selection for propensity score models. American Journal of Epidemiology 163 (June): 1149-1156. Kevin A. Clarke, Brenton Kenkel, and Miguel R. Rueda. Misspecification and the propensity score: when to leave out relevant pre-treatment variables. preprint, 2010. Soko Setoguchi, Sebastian Schneeweiss, M. Alan Brookhart, Robert J. Glynn, and E. Francis Cook. Evaluating uses of data mining techniques in propensity score estimation: a simulation study. Pharmacoepidemiology and drug safety, 17(6):546 555, June 2008. etc

SOME CLAIMS AMONG COUNTLESS MANY Judea Pearl Andrew Gelman Rubin model is a particular case of Pearl model. Rubin model is not explicit when it comes to ignorability condition Judea Pearl may have proven equivalence of the Rubin model and the Pearl model but assumptions are wrong or irrelevant for some real world problems.

STRUCTURAL CAUSAL MODEL BOOKS Causality By Judea Pearl (UCLA), 2009, Cambridge University Press. Targeted Learning by Mark Van Der Laan (UCB) and Sherri Rose (Johns Hopkins), 2011, Springer Series in Statistics

REFERENCES 1. http://andrewgelman.com/2009/07/disputes_about/ 2. http://andrewgelman.com/2009/07/philip_dawids_t/ 3. http://andrewgelman.com/2009/07/more_on_pearls/ 4. http://andrewgelman.com/2009/07/more_on_pearlru/ 5. http://andrewgelman.com/2009/07/pe 6. http://www.cs.ucla.edu/~kaoru/r348.pdfarls_and_gelm/ 7. http://www.stat.columbia.edu/~gelman/research/ published/causalreview4.pdf