I am facing an interpretation issue in my project.
RQ: I want to compare 2 predictor variables (A and B) in terms of their explanatory power towards a criterion variable C. Which one explains more? Which one is the better predictor?
My approach:
- I calculated a full model with A, B as predictors and C as outcome variable. Then I compare effect sizes f2 of A and B.
- I calculated seperate models for A and C and B and C. There, I analyze path coefficients and R2 as well as R2 adjusted.
- Is this (R2 and effect sizes) enough to get an answer for my RQ?
- Am I allowed to compare both R2 values of two seperate models under ceteris paribus conditions?