One of our statisticians responded:
"Type I sums of squares in GLM will probably give her what she wants. This is like running the model with one predictor, then two, then three, and giving the additional reduction in error sum of squares to each term in turn, but all tests use the final model error sum of squares (or sums of squares and cross-products in this multivariate case) instead of a different error term and degrees of freedom each time."
------------------------------
Rick Marcantonio
Quality Assurance
IBM
------------------------------
Original Message:
Sent: Fri February 24, 2023 03:49 PM
From: Anita Panov
Subject: How to run hierarchical multivariate multiple regression?
Hello. I'm working on my thesis and have a complex situation with multiple IVs and DVs, specially the latest cannot be grouped to one composite DV due to, i.e., different units of measure and generally it makes sense not to combine them. I have an idea how to run the multivariate regression in SPSS (analyze - general linear model - multivariate...), but I do not know how to do it as hierarchical multivariate regression. Could please someone help me and explain how to put the IVs and control variables in specific order? Also, does anyone has a good guide how to interpret the results then?
I appreciate any help and thank so much for support!
Cheers,
Anita
------------------------------
Anita Panov
------------------------------