Running the Analysis

To run GLM Univariate with wild residual bootstrapping, you first need to create residuals.

1. From the menus choose:

   Analyze > General Linear Model > Univariate...

   

2. In the GLM Univariate dialog, select diff as the dependent variable.
3. Select Gender [gender], Employment Category [jobcat], and Minority Classification [minority] as fixed factors.
4. Select Months since Hire [jobtime] and Previous Experience (months) [prevexp] as covariates.
5. Click Model.

   

6. In the Model dialog, select Custom and select Main effects from the Build Terms dropdown.
7. Select gender through prevexp (when variables are listed in file order) as model terms.
8. Click Continue.
9. Click Save in the GLM Univariate dialog.

   

10. In the Save dialog, select Unstandardized in the Residuals group.
11. Click Continue.
12. Click Bootstrap in the GLM Univariate dialog.

    

    The bootstrap settings persist across dialogs that support bootstrapping. Saving new variables to the dataset is not supported while bootstrapping is in effect, so you need to make certain it is turned off.

13. If needed deselect Perform bootstrapping in the Bootstrap dialog.
14. Click OK in the GLM Univariate dialog. The dataset now contains a new variable, RES_1, which contains the unstandardized residuals from this model.
15. Recall the GLM Univariate dialog and click Save.
16. Deselect Unstandardized in the Save dialog, then click Continue and click Options in the GLM Univariate dialog.

    

17. In the Options dialog, select Parameter estimates in the Display group.
18. Click Continue.
19. Click Bootstrap in the GLM Univariate dialog.

    

20. Select Perform bootstrapping.
21. To replicate the results in this example exactly, select Set seed for Mersenne Twister and type 9191972 as the seed.
22. There are no options for performing wild bootstrapping through the dialogs, so click Continue, then click Paste in the GLM Univariate dialog.

Next