wilcoxon.cut.test: Compare two halves of a population by Wilcoxon test.

In Antoine-Masse/KefiR: KefiR Package

Compare two halves of a population by Wilcoxon test.

Description

Test the global efficiency of an index (group) on a mesure. This function is use in B. Jamet – (2022) to show the efficiency of different banks informations product by ONG on reals financial information (see examples).

Usage

wilcoxon.cut.test(
  x,
  group,
  data = NULL,
  prop = 0,
  verbose = FALSE,
  bootstrap = 0,
  seed = NULL
)

Arguments

x	vector of names of the columns to test in data, or dataframe of the vectors to test, or a name of data column
group	a vector of values, or the name or the number of the column variable in data to use to cut x : a vector with 2 values or a quantitative vector, with same length than x.
data	data.frame with the data.
prop	proportion of values excluded around the median of group. Make sense only if group is numerical, not factorial.
verbose	if TRUE, always print the wilcox.test(), if verbose=F print the result only if p-value < 0.05
bootstrap	0 if no bootstrap, number of bootstrap, only pertient on qualitative groups, not on a 2-values group vector.
seed	NULL, an integer random seed for reproductibility of bootstrap.

Details

wilcoxon.cut.test() take the group vector and cut it in two halves around the median. This group vector may be numerical or factorial with 2 levels. A proportion prop= of data around the median may be possibly excluded. The two remaining samples of data in x, associated for lower an upper parts of the group are passed to the test of Wilcoxon to compare their median. This allow us to validate the efficiency of the index group on the measured x tendency.

Value

a dataframe with wilcoxon p-values, or with confidence intervall of p-values, if bootstraped. Note that a p value is always positive. Here, the minus sign in front of a negative p-value means that the index (in group) is in the opposite order with x.

Author(s)

Julien Bousquet (2021)

Examples

# Call with names of columns of dataframe
#wilcoxon.cut.test(x=c('FEF1','FEF'), group='FEF2', data=X) 
# Call with numbers of columns 
#wilcoxon.cut.test(x=2:4, group='FEF2', data=X, prop=0.2, boot=50) 
# Call data with vectors and data.frame :
#wilcoxon.cut.test(X[,2:3], group=X$FEF2, prop=0.1)

Antoine-Masse/KefiR documentation built on Feb. 22, 2024, 5:54 a.m.