Hi all,
I want to to study the effect of X on Y with a moderator W. W can take 3 levels (it's not an ordinal variable, it's rather 3 different situations).
Model 1 in Process allows to use a dichotomous moderator. I know that I can change W and transform it into 2 dichotomous variables, W1 and W2
If W=1, W1=1, W2=0
If W=2, W1=1, W2=1
If W=3, W1=0, W2=0. (or should it be 0 and 1? I guess this makes no difference)
But then, should I use Model 2? Is there a specific way to analyze the data?
Thanks in advance,
mk
Hello, I recommend you to use the free program JAMOVI which has the option of factoring the analysis by categories of the variable as you refer.
You can do a basic regression where the first step includes two dummy variables and then the next step adds the two interaction terms. This is all that PROCESS does, but it automates it for you.
Thanks to David and Christian!
I'll check both possibilities and I'll keep you updated.
Best,
mk
Cristian Ramos-Vera , I tried Jamovi (excellent software that I didn't know).
I installed the Medmod module but the moderator can only be a continuous variable. Is there something that I'm missing?
Thanks
There is a package in library (on JAMOVI) that has the program to install that refers to GLM models for the respective analysis.
Sorry if my question is stupid, but in this case, the GLM model is tested and it doesn't consider my variable like a moderator? (using the correct algorithm)
I'm still unclear why you are experimenting with all these specialized programs when all you need is ordinary least squares with two dummy variables and their associated interaction terms.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS