Correlogram

Use Correlogram to visualize and compare the strength and direction of linear relationships (correlations) between pairs of variables.

In This Topic

Continuous variables
By variables
Gradient type

For information about data considerations, examples, and interpretation, go to Overview for Correlogram.

Continuous variables

Enter the columns that contain the data. You must enter at least two columns of numeric data. Each column must have the same number of rows.

By variables

Enter one or more grouping variables in By variables to create a separate correlogram for each level of the grouping variables. The columns that you enter must be the same length as the columns in Continuous variables. The y-scales for each variable are the same across the multiple correlograms.

Show all combinations

When you enter multiple By variables, Minitab enables the Show all combinations checkbox. Select this option to create a separate correlogram for each combination of groups created by the By variables. If you do not select this option, Minitab creates a plot for each group of each By variable.

For example, the first By variable has 2 groups, Male and Female, and the second By variable has 2 groups, Employed and Unemployed. If you select Show all combinations, Minitab creates 4 separate plots for the combinations of Male/Employed, Male/Unemployed, Female/Employed, and Female/Unemployed. If you do not select Show all combinations, Minitab creates 4 separate plots for Male, Female, Employed, and Unemployed.

Gradient type

Select the color scale for the rectangles.

Automatic

When you select Automatic, the gradient type depends on the calculation of the Pearson correlation coefficients.

If the coefficients are positive and negative, the Gradient Type is Diverging symmetric around 0.
If the coefficients are all positive, the Gradient Type is Sequential from low to high.
If the coefficients are all negative, the Gradient Type is Sequential from high to low.

Diverging symmetric around 0: Rectangles with high values are red, and rectangles with low values are blue. The middle of the scale is centered at 0.
Sequential from low to high: Rectangles with high values are dark blue, and rectangles with low values are light blue and light gray. You can use this option to highlight variables with a strong positive relationship.
Sequential from high to low: Rectangles with low values are dark blue, and rectangles with high values are light blue and light gray. You can use this option to highlight variables with a strong negative relationship.

Display correlation values

Select to display the value of the Pearson correlation coefficient within each rectangle of the graph.