The Descriptive Statistics in The Unscrambler® X provides some simple and effective plotting tools for gaining an overview of datasets. One plot of importance for understating how variables are related to each other is the Cross Correlation plot. This plot can be accessed from Plot menu after performing descriptive statistics for a data set. The matrix is symmetrical (the correlation between A and B is the same as between B and A) and its diagonal elements contains only values of 1, since the correlation between a variable and itself is 1. All other values are between -1 and +1. A large positive value indicates that the corresponding two variables have a tendency to increase simultaneously. A large negative value indicates an inverse relationship of the variables. A correlation close to 0 indicates that the two variables vary independently from each other.
It is recommended to use a matrix plot consisting of “bars” (used as default) or a “map” for studying cross-correlations. Examples are provided below:
Care must be exercised when interpreting the colour scale of the such plots; not all data sets have correlations varying from -1 to +1. The highest value will always be +1 (diagonal), but the lowest may, in some cases, never go below zero! This may occur when, in for example, process measurements where several measurements that capture similar information are studied. This may include texture or light absorbance in a narrow range. Look at the values on the color scale before reaching any conclusions! A table showing the numerical values of cross-correlation between all variables included in the descriptive statistics analysis is also available.