sm_statCorr {smplot2}

Linear regression slope and statistical values (R or R2 and p values) from a paired correlation test.

Description

This combines two different functions: 1) 'geom_smooth()' from ggplot, and 2) 'stat_cor()' from ggpubr. 'geom_smooth()' is used to fit the best-fit model, whereas 'stat_cor()' is used to print correlation results at an optimized location.

Updates from smplot2 include more flexibility, less input arguments and its pairing with 'sm_hvgrid()' / 'sm_corr_theme()'.

Usage

sm_statCorr(
  ...,
  fit.params = list(),
  corr_method = "pearson",
  alternative = "two.sided",
  separate_by = ",",
  label_x = NULL,
  label_y = NULL,
  text_size = 4,
  show_text = TRUE,
  borders = TRUE,
  legends = FALSE,
  r2 = FALSE,
  R2
)

Arguments

...

Arguments for the properties of regression line, such as 'linetype', 'color', etc. For more information, type ?geom_smooth

fit.params

Paramters for the fitted line, such as color, linetype and alpha.

corr_method

Method of the correlation test. Options include: 'pearson', 'kendall', or 'spearman'.

alternative

Specifies the alternative hypothesis (H1). 'two.sided' is the standard way. 'greater' is a positive association, whereas 'less' is a negative association.

separate_by

This marks how the p- and r- values should be separated. The default option is: ',' For more information, check out stat_cor() from the ggpubr package.

label_x

Location of the statistical value prints along the figure's x-axis. It asks for a number within the x-axis limit.

label_y

Location of the statistical value prints along the figure's y-axis. It requires a number within the y-axis limit.

text_size

Size (numerical value) of the texts from correlation.

show_text

If the statistical result needs to be displayed, the input should be TRUE (default). If the statistical result is not needed, the input should be FALSE.

borders

If the border needs to be displayed, the input should be TRUE. If the border is not needed, the input should be FALSE.

legends

If the legend needs to be displayed, the input should be TRUE. If the legend is not needed, the input should be FALSE.

r2

FALSE or TRUE. TRUE if user wants to compute R2. FALSE if R needs to be computed.

R2

Same as r2.

Value

Plots a best-fitted linear regression on a correlation plot with results from correlation statistical tests.

Examples

library(smplot2)
library(ggplot2)
ggplot(data = mtcars, mapping = aes(x = drat, y = mpg)) +
geom_point(shape = 21, fill = '#0f993d', color = 'white', size = 3) +
 sm_statCorr() # computes R

 ggplot(data = mtcars, mapping = aes(x = drat, y = mpg)) +
geom_point(shape = 21, fill = '#0f993d', color = 'white', size = 3) +
 sm_statCorr(R2 = TRUE) # computes R2