Statistics 209B / EPI 239 / Education 260A                                         Winter 2021

Applications of Causal Inference Methods

Winter 2021 Remote Asynchronous Instruction

Instructor.  David Rogosa (usually)Sequoia 224,   rag{AT}stanford{DOT}edu   
TA   Ismael Lemhadri     lemhadri {AT} stanford {DOT} edu
Course web page:

Course Welcome and Logistics (first day stuff, posted in December, call it Week0)
      Lecture slides, week 0 (pdf)             Audio companion, week 0   
For recreation of in-classroom experience, linked below are youtube versions of the music I play
before starting lecture   and    after lecture concludes.      Some may wish to reverse that ordering.

Registrar's Information
Statistics 209B (also EPI 239, EDUC 260A)   2 units
Title: Applications of Causal Inference Methods
Description: Application of potential outcomes formulation for causal inference to research settings including: mediation,
compliance adjustments, time-1 time-2 designs, encouragement designs, heterogeneous treatment effects, aggregated data, 
instrumental variables, analysis of covariance regression adjustments, and implementations of matching methods.
see class website  
 suggested prerequisite STAT209A/MSE327 or other introduction to causal inference methods.
Course Overview
Note: This course was structured before the new world order of March 2020 as one two-hour class meeting per week.
The way things worked out this quarter, reality dictates Thursday for complete posting for the week's lecture and materials.

Brief Course Outline
Unit 1. Extensions of RCT
  Analyzing Encouragement Designs 
  Assessing Mediation in experimental studies 
  Identifying Moderation in experimental studies (heterogeneous treatment effects)
  The wisdom of Compliance Adjustments (for binary and measured compliance); 
  Analysis of Regression Discontinuity Designs (systematic assignment based on a covariate) 
Unit 2. Legacy Methods for Causal Inference from Observational Data (good and bad)
  Regression adjustments (analysis of covariance) in observational studies
  Instrumental Variables methods for observational studies
  Failures of traditional Path Analysis (and Structural Equation Models)
  Interpreting associations: Spurious Correlation and Simpson's Paradox
  Multilevel data and Contextual Effects
  Reciprocal Causal Effects and non-recursive models
Unit 3. Historical and Modern Methods for Matching in Observational Data 
  Case-Control, case-referent matching sudies
  Subclassification matching
  Propensity Score matching examples
Unit 4. Time-1, Time-2 data in experimental and observational studies.    
  Experimental Designs     Cross-over designs
  Experimental Designs     Comparing groups on time-1, time-2 measurements: repeated measures anova vs lmer OR the t-test
  Observational studies    Economist's differences in differences (or diffs in diffs with matching) for observational studies.  
  Observational studies    Lord's paradox; pre-post group comparisons.
  Observational studies    Exogenous Variables and Correlates of Change (use of lagged dependent variables)
  Additional Special topics  Interrupted Time-series designs;  Current implementations of value-added analysis 

Lectures,  Course Files,  and Readings
this page is where course content resides

Grading, Exams, and Credit Units
Stat209B/EPI239/Ed260A is listed as Letter or Credit/No Credit grading for 2-units
For Winter 2021 grading for the 2-units will be based on a 'take home'(i.e. do at home) Problem Set.
   Also as you will see, for each week's content a number of Review Questions with Solutions are posted.

Course Problem Set 2021    full posting 3/10/21

Statistical computing
Class presentation will be in, and students are encouraged to use, R (occasionally, some references to SAS and Mathematica).
Current version of R is R version 4.0.3 released 2020-10-10
    For references and software: The R Project for Statistical Computing   Closest download mirrors in the past, UCLA and Berkeley, seem no longer avaliable, pick your fave anywhere in the world.
The CRAN Task Views provide useful overviews/navigation for the almost infinite set of R-packages. Task Views that cover contents of this course include CRAN Task View: Statistics for the Social Sciences; CRAN Task View: Econometrics  and  CRAN Task View: Psychometric Models and Methods .

Relevent Texts (optional).    
Causal Inference in Statistics, Social and Biomedical Sciences: An Introduction, Guido Imbens and Don Rubin, 1st Edition (Cambridge University Press)   Stanford access
Design of Observational Studies, Paul Rosenbaum, 1st Edition (Springer). Available online:   Stanford access

To see full course materials from legacy Stat209 (2005-2019)
Statistical Methods for Group Comparisons and Causal Inference   go here