Troubleshooting
Problem
I would like to run a hierarchical linear Regression, i.e., a regression where I enter sets of predictors into the model in blocks, or stages. I want to test whether the addition of each block has improved the model's fit to the data over the previous blocks. Can this be done with the SPSS REGRESSION procedure?
Resolving The Problem
Yes, this analysis is very feasible in SPSS REGRESSION.
If you are using the menus and dialog boxes in SPSS, you can run a hierarchical regression by entering the predictors in a set of blocks with Method = Enter, as follows:
Enter the predictor(s) for the first block into the 'Independent(s)' box in the main Linear Regression dialog box. Leave Method set at 'Enter'. Then click the 'Next' button at the top of the 'Independent(s)' box. This clears that box.
Enter the variable(s) for block 2 in your model. If there is a third block, click the 'Next' button again to clear the second block variables from the box and enter the 3rd block variable(s).
Continue this sequence for as many blocks as needed. You don't need to click 'Next' after entering the variable(s) for the last block.
Click the 'Statistics' button in the main Linear Regression dialog box. In the Statistics dialog, check R squared change. This will request a test of the significance of the change in R squared at each successive block in the model.
Note that you don't need to enter the block 1 variables in block 2 or later blocks. The predictor list across blocks is cumulative.
If you're running Regression through a syntax window, the blocks are built by a series of /METHOD = ENTER subcommands, as in:
REGRESSION
/MISSING LISTWISE
/STATISTICS COEFF OUTS R ANOVA CHANGE
/CRITERIA=PIN(.05) POUT(.10)
/DEPENDENT y
/METHOD=ENTER x1 x2 x3
/METHOD=ENTER x4 x5
/METHOD=ENTER x6.
The keyword 'CHANGE' in the /STATISTICS subcommand is the request for a test of the change in R squared.
Related Information
Historical Number
20301
Was this topic helpful?
Document Information
Modified date:
16 April 2020
UID
swg21476749