Causal Inference in R: Decipher complex relationships with advanced R techniques for data-driven decision-making

Causal Inference in R
Contributors
About the author
About the reviewer
Preface
   Who this book is for
   What this book covers
   To get the most out of this book
   Download the example code files
   Conventions used
   Get in touch
   Share Your Thoughts
   Download a free PDF copy of this book
Part 1:Foundations of Causal Inference
1
Introducing Causal Inference
   Defining causal inference
   Historical perspective on causal inference
   Why do we need causality?
   Is it an association or really causation?
   Deep dive causality in real-life settings
   Exploring the technical aspects of causality
      Simpson’s paradox
      Defining variables
   Summary
   References
2
Unraveling Confounding and Associations
   A deep dive into associations
   Causality and a fundamental issue
      Individual treatment effect
      Average treatment effect
   The distinction between confounding and associations
   Discussing the concept of bias in causality
   Assumptions in causal inference
   Strategies to address confounding
      Regression adjustment
      Propensity score methods
   Summary
   References
3
Initiating R with a Basic Causal Inference Example
   Technical requirements
   What is R? Why use R for causal inference?
   Getting started with R
      Setting up the R environment
      Navigating the RStudio interface
   Basic R programming concepts
      Data types in R
      Advanced data structures
      Packages in R
   Preparing for causal inference in R
      Preparing and loading data
      Exploratory data analysis (EDA)
   Simple causal inference techniques
      Comparing means (t-tests)
      Regression analysis
      Propensity score matching
   Case study – a basic causal analysis in R
      Data preparation and inspection
      Understanding the data
      Performing causal analysis
   Summary
   References
Part 2: Practical Applications and Core Methods
4
Constructing Causality Models with Graphs
   Technical requirements
   Basics of graph theory
      Types of graphs – directed versus undirected
      Other graph typologies
      Why we need DAGs in causal science
   Graph representations of variables
      Mathematical interpretation
      Representing graphs in R
      Bayesian networks
      Conditional independence
   Exploring Graphical Causal Models
      Comparison with Bayesian networks
      Assumptions in GCMs
   Case study example of a graph model in R
      Problem to solve using graphs
      Implementing in R
      Interpreting results
   Summary
   References
5
Navigating Causal Inference through Directed Acyclic Graphs
   Technical requirements
   Understanding the flow in Graphs
      Chains and forks
      Colliders
   Adjusting for confounding in graphs
      D-separation
      Do-operator
      The back door adjustment
      The front door adjustment
   Practical R example – back door versus front door
      Synthetic data
      Back door adjustment in R
      Front door adjustment in R
   Summary
6
Employing Propensity Score Techniques
   Technical requirements
   Introduction to propensity scores
      A deep dive into these scores
      Balancing confounding variables
      Check for confounding using propensity scores
      Challenges and caveats
   Stratification and subsampling
      Theory
      Application of propensity scores in R
   Understanding Propensity Score Matching
      Considerations and limitations
      Practical application of PSM in R
      Balancing methods
      Sensitivity analysis
      Visualizing the results
   Weighting in PSM using R
   Summary
   References
7
Employing Regression Approaches for Causal Inference
   Technical requirements
   Role of regression in causality
   Choosing the appropriate regression model
      Understanding the nature of the outcome variable
      Consideration of confounding and interaction effects
      Model complexity, parsimony, and assumptions
   Linear regression for causal inference
      The theory
      Application of regression modeling in R
      Single versus multivariate regression
      Treatment orthogonalization
      Example of the FWL theorem
   Model diagnostics and assumptions
   Non-linear regression for causal inference
      Other types of non-linear models
      Application of a non-linear regression problem in R
   Important considerations in regression modeling
      Which covariates to consider in the model?
      Dummy variables? What are they?
      Orthogonalization effect in R
   Summary
   References
8
Executing A/B Testing and Controlled Experiments
   Technical requirements
   Designing and conducting A/B tests
      Concepts
      Planning your A/B test
      Implementation details
   Controlled experiments and causal inference
      Enhancing causal inference
      Beyond A/B testing – multi-armed bandit tests and factorial designs
      Ethical considerations
   Common pitfalls and challenges
      Strategies for dealing with incomplete data
      Mitigating spill-over effects
      Adaptive experimentation – when and how to adjust your experiment
   Implementing A/B test analysis in R
      Step 1 – Generating synthetic data
      Step 2 – Exploratory data analysis (EDA)
      Step 3 – Statistical testing
      Step 4 – Multivariate analysis
      Step 5 – Interpreting results
      Step 6 – Checking assumptions of the t-test
      Step 7 – Effect-size calculation
      Step 8 – Power analysis
      Step 9 – Post-hoc analyses
      Step 10 – Visualizing interaction effects
   Summary
9
Implementing Doubly Robust Estimation
   Technical requirements
   What is doubly robust estimation?
      An overview of DR estimation
      Technique behind DR
      Comparison with other estimation methods
   Implementing doubly robust estimation in R
      Preparing data for DR analysis
      Implementing basic DR estimators
      Calculating weight
      Crafting the DR estimator
   Discussing doubly robust methods
      Estimating variance
      Advanced DR techniques (using the tmle and SuperLearner packages)
      Balancing flexibility and reliability with DR estimation
   Summary
   References
Part 3: Advanced Topics and Cutting-Edge Methods
10
Analyzing Instrumental Variables
   Technical requirements
   Introduction to instrumental variables
      The concept of instrumental variables
      The importance of instrumental variables in causal inference
   Criteria for instrumental variables
      Relevance of the instrumental variable
      Exogeneity of the instrumental variable
      Exclusion restriction
   Strategies for identifying valid instrumental variables
      Relevance condition
      Exogeneity condition
   Demonstrating instrumental variable analysis in R
      Using gmm for generalized method of moments
      Diagnostics and tests in instrumental variable analysis
      Interpretation of results
   Challenges and limitations of instrumental variable analysis
      Weak instrumental variables
      Measurement errors in instrumental variables
      Interpretation of instrumental variable estimates
   Summary
   References
11
Investigating Mediation Analysis
   Technical requirements
   What is mediation analysis?
      Definition and overview
      The importance of mediation analysis
   Identifying mediation effects
      Criteria for mediation
      Testing for mediation
   Mediation analysis in R
      Setting up the R environment
      Preparing data for mediation analysis
      Conducting mediation analysis
      Interpretation and further steps
      Advanced mediation models
   Summary
   References
12
Exploring Sensitivity Analysis
   Technical requirements
   Introduction to sensitivity analysis
      Why do we need sensitivity analysis?
      Historical context
   Sensitivity analysis for causal inference
      How do we use sensitivity analysis?
      Types of sensitivity analysis
      Key concepts and measures
   Implementing sensitivity analysis in R
      Using R for sensitivity analysis
      Visualizing our findings
      Case study
   Practical guidelines for conducting sensitivity analysis
      Choosing parameters for sensitivity analysis
      Limitations and challenges
   Advanced topics in sensitivity analysis
      Venturing beyond binary treatment
      ML approaches
      Future directions
   Summary
   References
13
Scrutinizing Heterogeneity in Causal Inference
   Technical requirements
   What is heterogeneity?
      Definition of heterogeneity in causality
      Case studies and discussion
      Examples (more of them)
   Understanding the types of heterogeneity
      Pre-treatment heterogeneity
      Post-treatment heterogeneity
      Contextual heterogeneity
   Heterogeneous causal effects deep dive
      Interaction terms in regression models
      Subgroup analysis
      ML techniques
   Estimation methods for identifying HCEs
      Regression Discontinuity Designs
      Instrumental variables
      Propensity Score Matching
   Case study – Heterogeneity in R
      Generating synthetic data
      Exploratory data analysis
      Matching for causal inference
      Estimating the ATE
   Tailoring interventions to different groups
      Conceptual framework
      Case study 1 – Educational interventions and their varied effects on different student demographics
      Case study 2 – Public health campaigns and their differential impacts on various population segments
   Summary
   References
14
Harnessing Causal Forests and Machine Learning Methods
   Technical requirements
   Introduction to causal forests for causal inference
   Historical development and key researchers
   Theoretical foundations of causal forests
      Conditions necessary for causal forest applications
      Advantages and limitations
   Understanding the math behind causal forests
      Deep-dive into causal forests
      Scenario – classroom cohort
   Using R to understand causal forests
      Installing and loading necessary packages
      Simulating data
      Training a causal forest
      Estimating treatment effects
      Validating the model
      Extracting leaf indices
   Machine learning approaches to heterogeneous causal inference
   Impact of social media using causal forests in R
      Setting up the environment
      Data preparation and preprocessing
      Building and tuning causal forest models
      Interpreting results and model validation
   Summary
   References
15
Implementing Causal Discovery in R
   Technical requirements
   Introduction to causal discovery
      Definition and importance
      Historical background
      Theoretical foundations
   Methods for causal discovery
      Constraint-based methods
      Score-based methods
      Hybrid methods
      Functional Causal Models
      Which causal discovery method should we use?
   Implementing causal discovery with Bayesian networks in R
      Using R packages
      Scenario for our problem
      Creating the dataset
      Implementing PC algorithm
      Using bnlearn for Bayesian networks
      More causal discovery methods
      Estimating causal effects
   A multi-algorithm comparative approach to causal discovery in R
      Setting up and generating data
      Constraint-based methods in R
      Score-based methods in R
      Hybrid methods in R
      Visualizing causal relationships
      Interpretation from code
      Future steps
   Summary
   References
Index
   Why subscribe?
Other Books You May Enjoy
   Packt is searching for authors like you
   Share Your Thoughts
   Download a free PDF copy of this book