Canonical Correspondence Analysis

42

7.13

Correlation of Env. Vars

This tab gives the weighted and unweighted Pearson correlations between the

environmental variables. When two environmental variables are highly correlated,

only one of them is likely to be of value as a predictor of species composition. If they

are perfectly correlated (R = 1),  then they both cannot be used in the analysis as it

will result in a singular matrix (i.e. it has no inverse matrix). The best analysis will

tend to be produced following the selection of environmental variables that are not

highly correlated.

When running a Redundancy analysis, the Correlation between Environmental

Variables is shown in a single grid as shown. These are the Simple Pearson

Correlations.

When running a Correspondence Analysis two buttons will be shown, the simple,

unweighted values as above, selected by pressing the  Simple  button, and the

Weighted Pearson Correlations weighted by the row sums of the species matrix,

selected with the  Weighted  button.

These results can be printed or exported from the File menu and copied from the Edit

Copyright 2004  PISCES Conservation Ltd