fisher_corr: Generate a correlation matrix and confidence intervals

In amcdavid/AMmisc: Miscellaneous Functions, Templates and Tools

Description

Treatment of NAs is incomplete at this time (in particular, the degrees of freedom will not be correct).

Usage

fisher_corr(x, y, conf.level = 0.95, ..., method = "fisher")

## S3 method for class 'AMcorr'
print(
  x,
  cor_fmt = "%.2f",
  sep = " ",
  include_ci = FALSE,
  include_pstar = FALSE,
  pstar_levels = c(`*` = 0.05, `**` = 0.01),
  ...
)

Arguments

x	AMcorr
y	numeric matrix-like object. If missing y=x.
conf.level	numeric confidence level
...	ignored
method	inference method, currently only 'fisher' is supported
cor_fmt	an sprintf format string for correlation
sep	separators between correlation and confidence intervals, and/or pvalues
include_ci	include upper/lower CI?
include_pstar	include pvalue asterix (set according to pstar_levels)
pstar_levels	named numeric vector determining pvalue formatting

Value

structure containing correlation matrix, lr, ur lower and upper CI, the pvalues from the associated tests pz and the degrees of freedom dof.

character matrix

Methods (by generic)

• print: Pretty print a correlation matrix and optionally its confidence intervals and p values

Examples

data(iris)
fc = fisher_corr(iris[, c('Sepal.Length', 'Sepal.Width')],
iris[, c('Petal.Length', 'Petal.Width')])
print(fc, include_pstar = TRUE)
print(fc, include_pstar = TRUE, include_ci = TRUE, cor_fmt = '%.1f')

amcdavid/AMmisc documentation built on June 1, 2020, 11:04 a.m.