Applied Causal Analysis (with R)
Preface
1
Introduction: About this seminar/book
1.1
About me
1.2
Your turn
1.3
Seminar: Script & Material
1.4
Motivation: The causal inference ‘revolution’
Notes
1.5
Objectives
1.6
Overview of some readings
1.7
Teaser: A seminar as treatment?
Notes
1.8
Datasets for term paper
2
Introduction: Open-source source R
2.1
Why R?
Notes
2.2
Reproducibility & Replicability
2.3
Where/how to study R?
2.4
Installation and setup of R
3
Introduction: Fundamental statistical concepts
3.1
Descriptive vs. causal questions
3.1.1
Descriptive questions
3.1.2
Causal questions
3.2
Measurement: Fundamentals
Notes
3.3
Measurement: Contingency tables
Notes
3.4
Measurement: Distribution(s) of measurements
Notes
3.5
Measurement: Scenarios, planned and realized measurements
Notes
3.6
Measurement: Exercise
Notes
3.7
Data: Basics
3.8
Data: Table format
3.9
Data: (Empirical) Univariate distributions
3.10
Data: (Empirical) Joint distributions
3.11
Data: One more joint distribution
3.12
Data: Probability Distributions
3.13
Data: Probability Distributions & Inference
3.14
Data: Exercise
3.15
Models: What is a model?
3.16
Models: Example: Mean as a model
3.17
Models: Example: Linear model (Equation)
3.18
Models: Example: Linear model (Visualization)
Notes
3.19
Models: Estimation
3.20
Models: Prediction
3.21
Models: Estimand, estimator and estimation (skip)
3.22
Models: Associational vs. causal inference
3.23
Models: Assumptions
3.24
Models: Exercise
4
Causal Analysis: Concepts & Definitions
4.1
The term “causality”
4.2
Deterministic vs. probabilistic causation
4.3
Causal chains & causal mechanism (1)
4.4
Causal chains & causal mechanism (2)
4.5
Causal chains & causal mechanism (3)
4.6
Potential outcomes framework (1)
4.7
Potential outcomes framework (2)
4.8
Potential outcomes (multiple)
4.9
Potential outcomes (multiple treatment values) (skip!)
4.10
Definition of Treatment/Causal Effect
4.11
Precision of causal statements
4.12
ITE: Estimation
4.13
ITE: Exercise
4.14
ITE: Summary
4.15
ATE: Average Treatment Effect
4.16
Why moving from ITE to ATE? (Wikipedia)
4.17
ATE: Naive Estimate
4.18
ATE: Decomposition
4.19
ATE: Identification Problem
4.20
Assumptions
4.21
Assumptions: Independence Assumption (IA)
4.22
Assumptions: Independence Assumption (IA)
4.23
Independence assumption & random assignment
4.24
Assumptions: SUTVA
4.25
Assumptions: SUTVA
4.26
Assumptions: Exercise
4.27
ATT: Average Effect of the Treatment on the Treated and the Control
4.28
Other types of treatment effects
4.29
Exercise: Treatments/outcome as trajectories
4.29.1
Lessons
4.30
Summary
4.31
Causes: Which variables are causes? (Discussion)
4.32
Causes: No causation without manipulation
4.33
Causes: Manipulable causes (Discussion)
4.34
Identification Analysis & Strategy
4.35
Design-based Approach
5
Randomized Experiments
5.1
Basics
5.2
Randomization & independence
5.3
Lab: Sampling & randomizing
5.4
Lab: How randomization induces independence
5.4.1
Long-run randomization & balance (not finished)
5.5
Estimation
5.6
Examples & further reading
5.7
Lab: Analyzing experimental data
5.7.1
Study
5.7.2
Data
5.7.3
Summary Statistics
5.7.4
Analysis
5.8
Exercise
6
Ideal, Natural and Field experiments
6.1
Ideal experiments: Basics
6.2
Ideal experiments: Exercise
6.3
Ideal experiments: Possible “ideal” design
6.4
Ideal experiments: Benefits
6.5
Ideal experiments: Examples
6.6
Natural experiments: Basics
6.7
Natural experiments: Examples
6.8
Natural experiments: Challenges & Criticism
6.9
Field experiments: Basics
6.10
Field experiments: Examples
7
Selection on Observables
7.1
Basics
7.2
Covariates & Missing pot. outcome perspective
7.3
Covariates & Bias
7.4
Covariates: Confounding/overcontrol bias
7.5
Covariates: Endogenous selection bias
7.6
Pervasive problem: Example post-treatment bias
7.7
Estimation
7.8
Examples & Further reading
7.9
Lab
7.9.1
Study
7.9.2
Data
7.9.3
Summary Statistics
7.9.4
Descriptive exploration
7.9.5
Naive estimate of the ATE
7.9.6
Controlling/Conditioning
7.9.7
Controlling: Conceptually
7.10
Exercise
8
Matching
8.1
Basics
8.2
Why Matching?
8.3
Overlap/common support
8.4
Steps in Matching
8.5
Exercise: Logic of & varieties of matching
8.6
What should I match on?
8.7
Propensity Score Matching
8.8
Examples & Further Readings
8.9
Packages & Functions
8.10
Lab
8.10.1
Study
8.10.2
Lab: Data
8.10.3
R Code
8.10.4
Propensity score matching
8.10.5
Genetic Matching: Single variable
8.10.6
Matching on more variables and polynomials..
8.11
Exercise
9
Data over time
9.1
Random quotes for a starter
9.2
Basics 1
9.3
Basics 2
9.4
Wide format data
9.5
Long format data
9.6
Lab: From wide to long and back
10
Difference-in-differences
10.1
Basics
10.2
Estimation & Data
10.3
Examples & Further reading
10.4
Packages & Functions
10.5
Lab
10.5.1
Study
10.5.2
Visualization of data
10.5.3
Data
10.5.4
R Code
10.5.5
Figure 1 and 2
10.5.6
Table 3
10.5.7
Table 4
10.6
Exercise
11
Panel data
11.1
Basics 1
11.2
Estimation: First-difference (FD) estimator
11.3
Estimation: Fixed-effects (FE) estimator
11.4
Estimation & Basics 1
11.5
Estimation & Basics 2
11.6
Exercise: Understanding first-differences and de-meaning
11.7
First-differenced and de-meaned data
11.8
Exercise: Basics
11.9
Exercise: Problem of
11.10
Examples & Further reading
11.11
Packages & Functions: plm
11.12
Packages & Functions: wfe
11.13
Packages & Functions: PanelMatch
11.14
Lab1
11.14.1
Pooled model
11.14.2
First-differences (FD) Model
11.14.3
Fixed effect (FE) Model (This is outdated since provision of PanelMatch package)
11.14.4
Digging into control and treatment group specification
11.14.5
Quick overview of PanelMatch
11.15
Lab2
11.15.1
Study
11.15.2
Data
11.15.3
R-Code
11.16
Exercise
12
IV: Instrumental variables
12.1
Prototype Example
12.2
Compliers and non-compliers
12.3
ITT & CACE/LATE
12.4
Assumptions
12.5
Estimation
12.6
Examples & Further reading
12.7
Packages & Functions
12.8
Lab: Study
12.9
Lab: Data
12.10
Lab: R Code
12.10.1
Clustered Standard Errors
12.10.2
Load data
12.10.3
Summary stats & graphs
12.10.4
Table 1, Column 1
12.10.5
Table 2, Column 1 and 2
12.10.6
Table 5: IV
12.11
Homework
13
RDD: Regression Discontinuity Design
13.1
Basics
13.2
Basics
13.3
Basics
13.4
Estimation: Continuity-Based Approach
13.5
Estimation: Randomization-Based Approach
13.6
Validating assumptions
13.7
Examples & Further reading
13.8
Packages & Functions
13.9
Lab: Study
13.10
Lab: Data
13.11
Lab: R Code
13.11.1
Run the following functions
13.11.2
Load data & install packages
13.11.3
Explore data and subsetting
13.11.4
Summary statistics + graphs
13.11.5
Continuity-based analysis
13.12
Exercise
References
Published with bookdown
Applied Causal Analysis (with R)
3.8
Data: Table format
Name
trust2006
threat2006
education2006
Emily
4
0
8
Sheridan
5
1
1
Sierra
0
0
0
Andy
5
0
9
William
7
0
4
Tameema
5
0
1
Alexandra
2
0
4
Jessie
7
0
5
..
..
..
..
Q: How many rows should this table have if N = 6633?