In my data I want to be able to determine which set of IVs (there are 8 variables total, 4 in each set) does a better job of predicting a set a DVs (there are 4). I know I could use multivariate regression to use all DVs in the same analyses, but I don't know how to group the IVs into 2 sets. On the other hand I could look at incremental validity using linear regression and putting the different IV sets into different blocks, but in SPSS (maybe that is my problem...) you can only have one DV.
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
2
If your DVs are correlated then yes it makes sense to do multivariate multiple regression. See Regression with multiple dependent variables?
-
$\begingroup$ With multivariate multiple regression is there a way to group IVs to test which set of IVs is a better predictor of the set of DVs? $\endgroup$– tebCommented Jan 15, 2014 at 6:02
-
$\begingroup$ Selecting the "best" subset of IVs is known as variable selection. A good set of IVs should be uncorrelated to each other but correlate well with the DVs. $\endgroup$– pontikosCommented Jan 19, 2014 at 22:39