causalplot

Tired of spending endless time customizing a causal process graph in Word or trying to decipher the inner workings of the TikZ package in LaTeX? The goal of causalplot is to create publication-ready causal process plots using a simple convenience function that demands almost nothing of the user and lets you focus your attention on what actually matters—the theorizing. Users chose between templates and add their own variable and mechanism labels, and the package produces a causal plot. The package also allows for additional customization in terms of text color, box color, fonts, and captions.

There are a number of other R packages that allow users to create causal plots, including ggdag, caugi, and DiagrammeR. These packages are more powerful and useful for complex causal mapping, yet lack the space required for clear mechanism-based causal diagrams.

Installation

You can install causalplot from CRAN with:

install.packages("causalplot")

Or install the development version from GitHub with:

# install.packages("devtools")
devtools::install_github("sebastianvanbaalen/causalplot")

What’s new?

The latest version of causalplot includes several new features and fixes:

• New triple-pathway templates 131 (one step) and 1331 (two steps).

• Reduction of white space around the plots.

• Support for causal plots with two independent variables (templates 211, 221, and 2221).

• Support for the standard 2x2 descriptive or explanatory typology.

• An embedded Shiny app that allows users to make plots without writing any code. Launch by typing causalplot::run_app() or visit https://sebvanbaalen.shinyapps.io/causalplot.

Bug-reports and suggestions?

Are you missing a template or is something not running as it should? Make a new issue in the Github repository or email Sebastian at sebastian.van-baalen@pcr.uu.se.

Example

Making a causal diagram using causalplot is easy. All you need is an idea, your variable names, and labels for your causal mechanisms or causal process steps. Use the type argument to chose between different templates in the plot gallery. Below is the code for making the simplest plot consisting only of three boxes: one for the independent variable, one for the dependent variable, and one for the causal mechanism.

causal_plot(
  type = "111",
  labels = c(
    "Lack of natural resources",
    "Increased dependence on civilians",
    "More responsive rebel governance"
  ),
  wrap_width = 16
)

Gallery

causalplot includes several different templates and requires no manual definition of nodes and arrows, simplifying the plotting experience. Below is an overview of all available templates, as well as examples of how to customize the plots. Feel free to reach out if you are missing a particular set-up.

One-step causal mechanism

A simple one-step causal mechanism plot with an independent variable, dependent variable, and one causal mechanism box.

causal_plot(
  type = "111", # Chose the plot template
  text_size = 5, # Set the text size
  box_ratio = 1 # Increase the box height
)

One-step dual-pathway causal mechanism

A causal mechanism plot with an independent variable, dependent variable, and two parallel causal mechanism boxes.

causal_plot(
  type = "121",
  labels = c(
    "Independent variable",
    "Causal mechanism 1",
    "Causal mechanism 2",
    "Dependent variable"
  ),
  text_size = 4,
  wrap_width = 17
)

One-step triple-pathway causal mechanism

A causal mechanism plot with an independent variable, dependent variable, and three parallel causal mechanism boxes.

causal_plot(
  type = "131",
  labels = c(
    "Independent variable",
    "Causal mechanism 1",
    "Causal mechanism 2",
    "Causal mechanism 3",
    "Dependent variable"
  ),
  text_size = 4,
  wrap_width = 17
)

Two-step causal mechanism

A two-step causal mechanism plot with an independent variable, dependent variable, and two causal process boxes.

causal_plot(
  type = "1111", 
  text_size = 4,
  wrap_width = 17, # Set the number of characters before line skip
  fill_mechanisms = "white" # Set the color of the process boxes
)

Three-step causal mechanism

A three-step causal mechanism plot with an independent variable, dependent variable, and three causal process boxes.

causal_plot(
  type = "11111",
  text_size = 4
)

Dual-pathway causal mechanism

A two-step causal mechanism plot with an independent variable, dependent variable, and two separate causal pathways.

causal_plot(
  type = "1221", 
  labels = c(
    "Independent variable",
    "Path 1: mechanism step 1",
    "Path 2: mechanism step 1",
    "Path 1: mechanism step 2",
    "Path 2: mechanism step 2",
    "Dependent variable"
  ),
  wrap_width = 17,
  text_size = 4,
  fill_mechanisms = "black",
  fill_variables = "black",
  text_color = "white",
  font = "Times" # Set the font
)

Triple-pathway causal mechanism

A two-step causal mechanism plot with an independent variable, dependent variable, and three separate causal pathways.

causal_plot(
  type = "1331",
  labels = c(
    "Independent variable",
    "Path 1: mechanism step 1",
    "Path 2: mechanism step 1",
    "Path 3: mechanism step 1",
    "Path 1: mechanism step 2",
    "Path 2: mechanism step 2",
    "Path 3: mechanism step 2",
    "Dependent variable"
  ),
  wrap_width = 17,
  text_size = 4
)

Dual-pathway causal mechanism with joint first step

causal_plot(
  type = "1121", 
  labels = c(
    "Independent variable",
    "Shared mechanism step 1",
    "Path 1: mechanism step 2",
    "Path 2: mechanism step 2",
    "Dependent variable"
  ),
  wrap_width = 17,
  text_size = 4,
  fill_variables = "grey50", # Set the color for the variable boxes
  fill_mechanisms = "grey80" # Set the color for the process boxes
)

Dual-pathway causal mechanism with joint second step

causal_plot(
  type = "1211", 
  labels = c(
    "Independent variable",
    "Path 1: mechanism step 1",
    "Path 2: mechanism step 1",
    "Shared mechanism step 2",
    "Dependent variable"
  ),
  wrap_width = 17,
  text_size = 4,
  fill_variables = "#59C7EB",
  fill_mechanisms = "#CCEEF9"
)

The classic “bathtub” causal mechanism

The classic “bathtub” visualization of a causal relationship as discussed in Hedström and Ylikoski (2010), with a direct macro-level arrow and a two-step causal mechanism.

causal_plot(
  type = "bathtub",
  text_size = 4,
  fill_variables = "#31A354",
  fill_mechanisms = "#BAE4B3"
)

One-step causal mechanism with moderator

A one-step causal mechanism plot with an independent variable, dependent variable, one causal mechanism box, and a moderator variable.

causal_plot(
  type = "111_moderator",
  labels = c(
    "Independent variable",
    "Causal mechanism",
    "Dependent variable",
    "Moderator variable"
  ),
  text_size = 4
)

One-step causal mechanism with confounder

A one-step causal mechanism plot with an independent variable, dependent variable, one causal mechanism box, and a confounding variable.

causal_plot(
  type = "111_confounder",
  labels = c(
    "Independent variable",
    "Causal mechanism",
    "Dependent variable",
    "Confounding variable"
  ),
  text_size = 4
)

Two independent variables with shared mechanism

A causal diagram with two independent variables that converge on a shared causal mechanism before producing a single outcome. A “+” sign indicates that both IVs are necessary.

causal_plot(
  type = "211",
  labels = c(
    "Independent variable 1",
    "Independent variable 2",
    "Shared causal mechanism",
    "Dependent variable"
  ),
  text_size = 4,
  wrap_width = 17
)

Two independent variables with separate mechanisms

A causal diagram with two independent variables, each operating through its own mechanism, converging on a single outcome. A “+” sign indicates that both mechanisms are necessary.

causal_plot(
  type = "221",
  labels = c(
    "Independent variable 1",
    "Independent variable 2",
    "Causal mechanism 1",
    "Causal mechanism 2",
    "Dependent variable"
  ),
  text_size = 4,
  wrap_width = 17
)

Two independent variables with two-step separate mechanisms

A causal diagram with two independent variables, each going through two sequential mechanism steps, before converging on a single outcome.

causal_plot(
  type = "2221",
  labels = c(
    "Independent variable 1",
    "Independent variable 2",
    "Path 1: mechanism step 1",
    "Path 2: mechanism step 1",
    "Path 1: mechanism step 2",
    "Path 2: mechanism step 2",
    "Dependent variable"
  ),
  text_size = 4,
  wrap_width = 17
)

Typology

In addition to causal diagrams, causalplot also includes a typology() function for creating 2x2 descriptive or explanatory typologies. The function produces a single box divided into four quadrants by dashed lines, with double-headed axis arrows, axis labels, endpoint values, and type labels. The example below replicates the typology in Martin et al. (2021).

typology(
  type_labels = c(
    "Predation",
    "Benevolent warlordism",
    "State capture",
    "Rebels-into-statesmen"
  ),
  x_axis_label = "Strength of local-level ties",
  y_axis_label = "Strength of national-level ties",
  x_axis_values = c("Weak", "Strong"),
  y_axis_values = c("Weak", "Strong"),
  text_size = 4,
  axis_text_size = 4
)

Save your plot

causalplot uses ggplot2 to generate the plots, so they can be saved using the ggsave command.

ggplot2::ggsave("man/figures/causal_plot.png", height = 5, width = 12)

Citation

Did you like the package? No need to cite it, but please add a footnote somewhere with the following text:

“Generated using the causalplot R package created by Sebastian van Baalen.”

References

Hedström, Peter, and Petri Ylikoski. 2010. “Causal Mechanisms in the Social Sciences.” Annual Review of Sociology 36: 49–67. https://doi.org/10.1146/annurev.soc.012809.102632.

Martin, Philip A., Giulia Piccolino, and Jeremy S. Speight. 2021. “Ex-Rebel Authority After Civil War: Theory and Evidence from Côte d’ivoire.” Comparative Politics 53 (2): 209–32. https://www.jstor.org/stable/26976051.