Causal Inference in R


Malcolm Barrett

Lucy D’Agostino McGowan

Travis Gerke


November 28, 2023


Welcome to Causal Inference in R. Answering causal questions is critical for scientific and business purposes, but techniques like randomized clinical trials and A/B testing are not always practical or successful. The tools in this book will allow readers better make causal inferences with observational data with the R programming language. By its end, we hope to help you:

  1. Ask better causal questions.
  2. Understand the assumptions needed for causal inference
  3. Identify the target population for which you want to make inferences
  4. Fit causal models and check their problems
  5. Conduct sensitivity analyses where the techniques we use might be imperfect

This book is for both academic researchers and data scientists. Although the questions may differ between these settings, many techniques are the same: causal inference is as helpful for asking questions about cancer as it is about clicks. We use a mix of examples from medicine, economics, and tech to demonstrate that you need a clear causal question and a willingness to be transparent about your assumptions.

You’ll learn a lot in this book, but ironically, you won’t learn much about conducting randomized trials, one of the best tools for causal inferences. Randomized trials, and their cousins, A/B tests (standard in the tech world), are compelling because they alleviate many of the assumptions we need to make for valid inferences. They are also sufficiently complex in design to merit their own learning resources. Instead, we’ll focus on observational data where we don’t usually benefit from randomization. If you’re interested in randomization techniques, don’t put away this resource just yet: many causal inference techniques designed for observational data improve randomized analyses, too.

We’re making a few assumptions about you as a reader:

  1. You’re familiar with the tidyverse ecosystem of R packages and their general philosophy. For instance, we use a lot of dplyr and ggplot2 in this book, but we won’t explain their basic grammar. To learn more about starting with the tidyverse, we recommend R for Data Science.
  2. You’re familiar with basic statistical modeling in R. For instance, we’ll fit many models with lm() and glm(), but we won’t discuss how they work. If you want to learn more about R’s powerful modeling functions, we recommend reading “A Review of R Modeling Fundamentals” in Tidy Modeling with R.
  3. We also assume you have familiarity with other R basics, such as writing functions. R for Data Science is also a good resource for these topics. (For a deeper dive into the R programming language, we recommend Advanced R, although we don’t assume you have mastered its material for this book).

We’ll also use tools from the tidymodels ecosystem, a set of R packages for modeling related to the tidyverse. We don’t assume you have used them before. tidymodels also focuses on predictive modeling, so many of its tools aren’t appropriate for this book. Nevertheless, if you are interested in this topic, we recommend Tidy Modeling with R.

There are also several other excellent books on causal inference. This book is different in its focus on R, but it’s still helpful to see this area from other perspectives. A few books you might like:

The first book is focused on epidemiology. The latter two are focused on econometrics. We also recommend The Book of Why Pearl and Mackenzie (2018) for more on causal diagrams.


Modern R Features

We use two modern R features in R 4.1.0 and above in this book. The first is the native pipe, |>. This R feature is similar to the tidyverse’s %>%, with which you may be more familiar. In typical cases, the two work interchangeably. One notable difference is that |> uses the _ symbol to direct the pipe’s results, e.g., .df |> lm(y ~ x, data = _). See this Tidyverse Blog post for more on this topic.

Another modern R feature we use is the native lambda, a way of writing short functions that looks like \(.x) do_something(.x). It is similar to purrr’s ~ lambda notation. It’s also helpful to realize the native lambda is identical to function(.x) do_something(.x), where \ is shorthand for function. See R for Data Science’s chapter on iteration for more on this topic.


The plots in this book use a consistent theme that we don’t include in every code chunk, meaning if you run the code for a visualization, you might get a slightly different-looking result. We set the following defaults related to ggplot2:

  # set default colors in ggplot2 to colorblind-friendly
  # Okabe-Ito and Viridis palettes
  ggplot2.discrete.colour = ggokabeito::palette_okabe_ito(),
  ggplot2.discrete.fill = ggokabeito::palette_okabe_ito(),
  ggplot2.continuous.colour = "viridis",
  ggplot2.continuous.fill = "viridis",
  # set theme font and size
  book.base_family = "sans",
  book.base_size = 14


# set default theme
    base_size = getOption("book.base_size"),
    base_family = getOption("book.base_family")
  ) %+replace%
      panel.grid.minor = element_blank(),
      legend.position = "bottom"

We also mask a few functions from ggdag that we like to customize:

theme_dag <- function() {
  ggdag::theme_dag(base_family = getOption("book.base_family"))

geom_dag_label_repel <- function(..., seed = 10) {
    aes(x, y, label = label),
    box.padding = 3.5,
    inherit.aes = FALSE,
    max.overlaps = Inf,
    family = getOption("book.base_family"),
    seed = seed,
    label.size = NA,
    label.padding = 0.1,
    size = getOption("book.base_size") / 3,