Display Spearman Correlation Coefficients and their Significances

Description

The function computes Spearman rank correlation coefficients and places them in the upper triangle of a printed matrix displayed on the current device, the probabilities that the coefficients are not due to chance (Ho: Coefficient = 0) are printed in the lower triangle. The diagonal is filled with NAs to visually split the two triangles.

Usage

1
gx.spearman(xx, ifclr = FALSE, ifwarn = TRUE)

Arguments

xx

a matrix of numeric data.

ifclr

if ifclr = TRUE the data are Centred Log-Ratio transformed prior to the computation of the Pearson Coefficients. The default is no transformation.

ifwarn

by default ifwarn = TRUE which generates a reminder/warning that when carrying out a centred log-ratio transformation all the data must be in the same measurement units. The message can be suppressed by setting ifwarn = FALSE.

Note

Any less than detection limit values represented by negative values, or zeros or other numeric codes representing blanks in the data, must be removed prior to executing this function, see ltdl.fix.df.

Any data vectors (rows) containing NAs are removed prior to computation.

This function is not recommended for use with closed compositional data sets, i.e. geochemical analyses, unless correlations are sought between a non-compositional variable and individual compositional variables. If it is used with compositional data, it is highly recommended that ifclr be set to TRUE to remove the effects of closure and display the ‘true’ inter-element variability. However, different groups of elements, subsets, of a data set will yield different inter-element correlations for the same pair of elements due to the nature of the clr transform. When carrying out a centred log-ratio transformation it is essential that the data are all in the same measurement units, and by default a reminder/warning is display if the data are centred log-ratio transformed, see ifwarn above.

For working with compositional data sets functions gx.vm and gx.sm are recommended. For visual displays see gx.pairs4parts and gx.plot2parts.

Author(s)

Robert G. Garrett

See Also

ltdl.fix.df, remove.na, clr

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
## Make test data available
data(sind.mat2open)

## Compute Spearman correlation coefficients
gx.spearman(sind.mat2open)

## Note, unlike gx.pearson there is no example with a log
## transformation.  The log transformation is monotonic
## and does not change the ranks

## Compute Spearmann correlation coefficients following
## a centred log-ratio transformation
gx.spearman(sind.mat2open, ifclr = TRUE)

Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.