ORS supports built-in diagnostic regression graphs. These graphs are designed to help the
researcher uncover the overall usefulness of the computed model. To reach
diagnostic regression graphics, follow these steps:
1.
After
solving the regression model, select the Resid/Graph tab.
2.
From
the menu, choose: Charts \ Diagnostics \ Regression. The following menu appears:
The diagnostic graph list
includes:
1.
Constance Variance -- no additional input
required.
2.
Error Bars -- no additional input
required.
3.
Predictive Ability -- no additional input
required.
4.
Residual Analysis -- no additional input
required.
5.
Normal Probability -- no additional input
required.
6.
Linearity Test -- requires the identification of an X variable
in the solution.
7.
Outlier Analysis -- requires the identification of an X variable
in the solution.
8.
Omitted Variable -- requires the identification of an X variable
NOT in the solution.
9.
Confidence Intervals -- no additional input
required.
10.
Component+Residual -- choose an X variable.
11.
Scatter Plot -- this is a 3d rotatational plot. Choose two X in solution variables. After choosing two checkboxes, click inside the window to
indicate done. Click OK to see the
plot.