Assuming I compute a variable (G) by combining 6 variables (A,B,C,D,E,F). Can i run a multiple regression model with G as the dependent variable and A,B,C,D,E,F as independent variables to determine the significant effects each of the independent variables have on G?
Hello Yahya,
If you're asking, can you run six regressions, each with a single IV, the answer is yes. Even simpler, just running bivariate correlations of the six IVs with the DV (G) will give you the same results.
The real question is, what is it you want to know about "...effects [of] each of the independent variables..."? If you want to know how strongly each one relates to the DV, the method above will suffice.
However, if you want to know what each IV can contribute with respect to explanatory power above and beyond the other IVs, then run a set of 6 more regressions, each time omitting just one of the IVs (e.g., run #1: use A,B,C,D,E and omit F. run #2: use A,B,C,D,F and omit E). The difference in R-squared between these models and the R-squared with all six IVs (which you've apparently run already) will let you know the influence of each IV beyond what the others contribute (for your sample).
Good luck with your work.
You can follow stepwise regression. It will divide the model into sub-models based on the number of independent variables.
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running