crossplot_print_stats: A convience function to print the plot stats and a regression...

In ChandlerLutz/crossplotr: Easily create cross-sectional plots with ggplot2

Description

A convience function to print the plot stats and a regression line

Usage

crossplot_print_stats(p, text.pos, stats = c("slope", "r.squared"),
  reg.color = "black", reg.linetype = "solid", log.reg = FALSE,
  weighted = FALSE, sprintf.format = "%.2f", xlabel = NULL,
  ylabel = NULL, reg.label.se = TRUE, mean.label.se = FALSE, ...)

Arguments

p	a crossplot ggplot2 plot. Can be generated from crossplot
text.pos	the text position to be passed to cowplot::draw_label() as a numeric vector of length 2 with x and y.
stats	character vector with the stats that will be printed (in order). Possible statistics include "slope", "intercept", "r.squared", "adj.r.squared", "mean.x", "mean.y" "var.x", "var.y", "sd.x", "sd.y". Defualt is c("slope", "r.squared")
reg.color	a string with the ggplot2 color for the regression line
reg.linetype	a string with the ggplot2 linetype for the regression line
log.reg	logical. If set to TRUE, the regression will be estimated in logs. Defualt is FALSE
weighted	logical. If set to TRUE, the regression and all statistics will be weighted using the size variable form the ggplot2 plot. Default is FALSE
sprintf.format	character string with the sprintf format.
xlabel	the label for the x-axis variable
ylabel	the label for the y-axis variable
reg.label.se	logical. If set to TRUE, the regression standard errors will be printed. Default is TRUE
mean.label.se	logical If set to TRUE, the standard errors for the means will be printed. Default is FALSE
...	Other arguments to be passed to cowplot::draw_label. See ?cowplot::draw_label for options and more details.

Value

a ggplot2 plot with the regression line and label added

Examples

data(mtcars)
p <- crossplot(mtcars, x.var = "hp", y.var = "mpg", size.var = "wt",
               shapes.var = "cyl")
##Print weighted regression line and weighted stats
crossplot_print_stats(p, text.pos = c(200, 25), weighted = TRUE)
##Print unweighted regression line and unweighted stats
crossplot_print_stats(p, text.pos = c(200, 25), weighted = FALSE)

##passing other arguments to cowplot::draw_label.
##See ?cowplot::draw_label for more details
crossplot_print_stats(p, text.pos = c(200, 25), weighted = FALSE,
                      fontface = "bold")

## -- Controlling for qsec weighted by wt -- ##

##First see the results using a normal regression
mod <- lm(mpg ~ hp + qsec, mtcars, weights = wt)
library(sandwich); library(lmtest)
coeftest(mod, vcov = sandwich)
##Now in a plot
p2 <- crossplot(mtcars, x.var = "hp", y.var = "mpg", size.var = "wt",
                shapes.var = "cyl", control.vars = "qsec")
crossplot_print_stats(p2, text.pos = c(0, 7), weighted = TRUE)

ChandlerLutz/crossplotr documentation built on May 6, 2019, 9:56 a.m.