The effect leverage plot for X is essentially a scatterplot of the X-residuals against the Y-residuals (Figure 2.58). To help interpretation and comparison with other plots that you might construct, JMP adds the mean of Y to the Y-residuals and the mean of X to the X-residuals. The translated Y-residuals are called the Y Leverage Residuals and the translated X-residuals are called X Leverage values. The points on the Effect Leverage plots are these X Leverage and Y Leverage Residual pairs.
Figure 2.56 shows how residuals are depicted in the leverage plot. The distance from a point to the line of fit is the residual for a model that includes the effect. The distance from the point to the horizontal line is what the residual error would be without the effect in the model. In other words, the mean line in the leverage plot represents the model where the hypothesized value of the parameter (effect) is constrained to zero.
Figure 2.56 Illustration of a Generic Leverage Plot
Figure 2.57 Comparison of Significance Shown in Leverage Plots
The term leverage is used because these plots help you visualize the influence of points on the test for including the effect in the model. A point that is horizontally distant from the center of the plot exerts more influence on the effect test than does a point that is close to the center. Recall that the test for an effect involves comparing the sum of squared residuals to the sum of squared residuals of the model with that effect removed. At the extremes, the differences of the residuals before and after being constrained by the hypothesis tend to be comparatively larger. Therefore, these residuals tend to have larger contributions to the sums of squares for that effect’s hypothesis test.
1.
Select Help > Sample Data Library and open Big Class.jmp.
2.
Select Analyze > Fit Model.
3.
Select weight and click Y.
4.
Select height, age, and sex, and click Add.
5.
Click Run.
The Whole Model Actual by Predicted Plot and the effect Leverage Plot for height are shown in Figure 2.58. The Whole Model plot, on the left, tests for all effects. You can infer that the model is significant because the confidence curves cross the horizontal line at the mean of the response, weight. The Leverage Plot for height, on the right, also shows that height is significant, even with age and sex in the model. Neither plot suggests concerns relative to influential points or multicollinearity.
Figure 2.58 Whole Model and Effect Leverage Plots

Help created on 10/11/2018