In the Fit Least Squares report, the Compare Slopes option appears when there is at least one nominal term, at least one continuous term, and at least one interaction effect for the fixed effects. This option produces a report that enables you to compare the slopes in an analysis of covariance (ANCOVA) model. If the model contains more than one interaction effect, you must select at least one interaction effect for which to compare slopes. The report compares the slopes of each level of the specified interaction effect to the overall slope. By default, the report compares each slope with the overall average slope and all pairwise combinations of slopes. You can also compare the slopes using a variety of other multiple comparisons options.
The Slope Comparisons red triangle menu contains the following options:
Comparisons with Overall Average
Shows or hides a comparison of each slope to the overall average slope using analysis of means (ANOM). The overall average slope is a weighted average of the slopes, where the weights are inversely proportional to the variances of the slope estimates. These variances are the squared values of the Std Error column in the Differences from Overall Average Slope table. See Comparisons with Overall Average.
Comparisons with Control
Shows or hides a comparison of each slope to the slope for a specified control group. When you select Comparisons with Control, a list appears that enables you to specify a level as the control group. See Comparisons with Control.
All Pairwise Comparisons
Contains the following comparison options:
Tukey HSD
Shows or hides a report of all pairwise combinations of slopes using an approach that ensures a joint significance level for all of the comparisons. See All Pairwise Comparisons.
Student’s t
Shows or hides a report of all pairwise combinations of slopes using an approach that specifies a significance level that applies to each of the comparisons. See All Pairwise Comparisons.
Equivalence Tests
Shows or hides a report of all pairwise combinations of slopes using a two one-sided tests (TOST) method. See Equivalence Tests.
Remove
Removes the Slope Comparisons report.