rdplot: Flexible RD Plot

In KatoPachi/discreteRD: What the Package Does (One Line, Title Case)

Description

Flexible RD Plot

Usage

rdplot(object, ...)

## S3 method for class 'list_global_lm'
rdplot(
  object,
  usemod,
  treat_label = c("treated", "control"),
  ate_label_size = 5,
  ate_label_digits = 3,
  ate_label_format,
  outcome_label,
  ylim,
  vjust = 0,
  hjust = 0,
  xlab = "Running variable",
  ylab = "Average",
  patchwork = TRUE,
  ncol = NULL,
  nrow = NULL,
  ...
)

## S3 method for class 'list_local_lm'
rdplot(
  object,
  usemod,
  treat_label = c("treated", "control"),
  ate_label_size = 5,
  ate_label_digits = 3,
  ate_label_format,
  outcome_label,
  ylim,
  vjust = 0,
  hjust = 0,
  xlab = "Running variable",
  ylab = "Average",
  patchwork = TRUE,
  ncol = NULL,
  nrow = NULL,
  force = TRUE,
  ...
)

Arguments

object	object
...	arguments of simplegg()
usemod	numeric vector. which models do you want to plot?
treat_label	string vector of treatment and control label
ate_label_size	In-plot text size
ate_label_digits	decimal places of in-plot ATE result.
ate_label_format	string of label format of in-plot ATE result. The format produces labels according to a unique grammar. If you want to embed a numerical value related to the estimation result, you need to enclose it in . {estimate}: embed coefficient value (Local ATE) {std.error}: embed standard error of coefficient {statistic}: embed z-score of coefficient {p.value}: embed p-value of coefficient {star}: show *** if p-value <= 0.01; ** if p-value <= 0.05; * if p-value <= 0.1
outcome_label	Outcome label in plot title
ylim	numeric vector of limits of y-axis
vjust	numeric. Adjust in-plot text vertically
hjust	numeric. Adjust in-plot text horizontally
xlab	label of x-axis
ylab	label of y-axis
patchwork	logical. add a list of plots into one composition by patchwork::wrap_plots
ncol	the dimensions of the grid to create. argument of patchwork::wrap_plots
nrow	the dimensions of the grid to create. argument of patchwork::wrap_plots
force	logical. Whether to ignore error about estimation

Examples

running <- sample(1:100, size = 1000, replace = TRUE)
cov1 <- rnorm(1000, sd = 2); cov2 <- rnorm(1000, mean = -1)
y0 <- running + cov1 + cov2 + rnorm(1000, sd = 10)
y1 <- 2 + 1.5 * running + cov1 + cov2 + rnorm(1000, sd = 10)
y <- ifelse(running <= 50, y1, y0)
bin <- ifelse(y > mean(y), 1, 0)
w <- sample(c(1, 0.5), size = 1000, replace = TRUE)
raw <- data.frame(y, bin, running, cov1, cov2, w)

set_optDiscRD(
  y + bin ~ running,
  discRD.cutoff = 50,
  discRD.assign = "smaller"
)

est <- global_lm(data = raw, weights = w)

rdplot(
  est,
  usemod = c(1, 2),
  treat_label = c("Treated", "Control"),
  outcome_label = switch(x,
    "1" = "Simulated outcome",
    "2" = "Simulated outcome > 0"
  ),
  ylab = "Weighted Average",
  ate_label_size = 4
)

nonpara <- local_lm(
  submod = 1,
  data = raw,
  kernel = "uniform",
  bw = 5,
  order = 1:3
)

rdplot(
  nonpara,
  usemod = 2,
  outcome_label = switch(x,
    "1" = "Order = 1",
    "2" = "Order = 2",
    "3" = "Order = 3"
  )
)

KatoPachi/discreteRD documentation built on Feb. 24, 2022, 12:32 a.m.