D-relationship between variables — Assess d-separation between variables • ggdag

D-separation is a key concept in causal structural models. Variables are d-separated if there are no open paths between them. The node_d*() functions label variables as d-connected or d-separated. The ggdag_d*() functions plot the results. The *_dconnected(), *_dseparated(), and *_drelationship() functions essentially produce the same output and are just different ways of thinking about the relationship. See dagitty::dseparated() for details.

node_dconnected(
  .tdy_dag,
  from = NULL,
  to = NULL,
  controlling_for = NULL,
  as_factor = TRUE,
  ...
)

node_dseparated(
  .tdy_dag,
  from = NULL,
  to = NULL,
  controlling_for = NULL,
  as_factor = TRUE
)

node_drelationship(
  .tdy_dag,
  from = NULL,
  to = NULL,
  controlling_for = NULL,
  as_factor = TRUE
)

ggdag_drelationship(
  .tdy_dag,
  from = NULL,
  to = NULL,
  controlling_for = NULL,
  ...,
  edge_type = "link_arc",
  node_size = 16,
  text_size = 3.88,
  label_size = text_size,
  text_col = "white",
  label_col = text_col,
  node = TRUE,
  stylized = FALSE,
  text = TRUE,
  use_labels = NULL,
  collider_lines = TRUE
)

ggdag_dseparated(
  .tdy_dag,
  from = NULL,
  to = NULL,
  controlling_for = NULL,
  ...,
  edge_type = "link_arc",
  node_size = 16,
  text_size = 3.88,
  label_size = text_size,
  text_col = "white",
  label_col = text_col,
  node = TRUE,
  stylized = FALSE,
  text = TRUE,
  use_labels = NULL,
  collider_lines = TRUE
)

ggdag_dconnected(
  .tdy_dag,
  from = NULL,
  to = NULL,
  controlling_for = NULL,
  ...,
  edge_type = "link_arc",
  node_size = 16,
  text_size = 3.88,
  label_size = text_size,
  text_col = "white",
  label_col = text_col,
  node = TRUE,
  stylized = FALSE,
  text = TRUE,
  use_labels = NULL,
  collider_lines = TRUE
)

Arguments

.tdy_dag: input graph, an object of class tidy_dagitty or dagitty
from: a character vector, the starting variable (must by in DAG). If NULL, checks DAG for exposure variable.
to: a character vector, the ending variable (must by in DAG). If NULL, checks DAG for outcome variable.
controlling_for: a character vector, variables in the DAG to control for.
as_factor: logical. Should the d_relationship variable be a factor?
...: additional arguments passed to tidy_dagitty()
edge_type: a character vector, the edge geom to use. One of: "link_arc", which accounts for directed and bidirected edges, "link", "arc", or "diagonal"
node_size: size of DAG node
text_size: size of DAG text
label_size: size of label text
text_col: color of DAG text
label_col: color of label text
node: logical. Should nodes be included in the DAG?
stylized: logical. Should DAG nodes be stylized? If so, use geom_dag_nodes and if not use geom_dag_point
text: logical. Should text be included in the DAG?
use_labels: a string. Variable to use for geom_dag_repel_label(). Default is NULL.
collider_lines: logical. Should the plot show paths activated by adjusting for a collider?

Value

a tidy_dagitty with a d_relationship column for variable D relationship or a ggplot

Examples

library(ggplot2)
dag <- dagify(m ~ x + y)
dag %>% ggdag_drelationship("x", "y")

dag %>% ggdag_drelationship("x", "y", controlling_for = "m")


dag %>%
  node_dseparated("x", "y") %>%
  ggplot(aes(x = x, y = y, xend = xend, yend = yend, shape = adjusted, col = d_relationship)) +
  geom_dag_edges() +
  geom_dag_collider_edges() +
  geom_dag_node() +
  geom_dag_text(col = "white") +
  theme_dag() +
  scale_adjusted()


dag %>%
  node_dconnected("x", "y", controlling_for = "m") %>%
  ggplot(aes(x = x, y = y, xend = xend, yend = yend, shape = adjusted, col = d_relationship)) +
  geom_dag_edges() +
  geom_dag_collider_edges() +
  geom_dag_node() +
  geom_dag_text(col = "white") +
  theme_dag() +
  scale_adjusted()


dagify(m ~ x + y, m_jr ~ m) %>%
  tidy_dagitty(layout = "nicely") %>%
  node_dconnected("x", "y", controlling_for = "m_jr") %>%
  ggplot(aes(x = x, y = y, xend = xend, yend = yend, shape = adjusted, col = d_relationship)) +
  geom_dag_edges() +
  geom_dag_collider_edges() +
  geom_dag_node() +
  geom_dag_text(col = "white") +
  theme_dag() +
  scale_adjusted()