In the Variable Clustering report, the Color Map on Correlations section displays a color map of the correlations between variables. The variables are arranged in the order in which they are listed in the Cluster Members table. This arrangement ensures that members of the same cluster are adjacent in the correlation plot. See Example of Color Map on Correlations.
Tip: Hover over a square on the color map to see the variables involved in that square and their correlation.
Variables in the same cluster tend to have higher absolute correlations (deeper red or blue colors) than variables in different clusters. Therefore, the squares formed by the cells of the correlation map that correspond to the variables for a given component often stand out along the diagonal.
Correlations are computed using the row-wise method. This method excludes any observation with missing data on any of the variables from the correlation calculation. For more information about the row-wise estimation method, see “Statistical Details for Variance Estimation Methods”.