I am using categorical variables as a moderator variable, and my dependent, mediator and IV variables are latent in nature. Which regression test I can apply. As logistic regression is for binary variables. but in my case combination of latent and multigroup (age, gender, income, ethnicity).
Thank you all
Dear Riffat Shahani
If the variables are latent by natural as (i.e, intention), and you are interested to study the effect of a set of moderators and you are willing to test the mediation effect, then I would strongly recommend you to use the structural equation modelling known as SEM, which is a technique that is a combination of confirmatory factor analysis known as CFA, and path analysis which is basically a linear regression.
In addition, SEM is not utilized in the SPSS, instead, there is an extension for the SPSS, so-called AMOS in which you can run SEM.
Best Regards,
Belal Edries
Belal thankyou for your answer,if i use SEM,then i will creat dummies for mediator and moderators?
As your dependent variable is on continous/interval scale hence linear regression/SEM would be appropriate. As far as mediation and moderation analysis is concerned, you can use PROCESS Macro for SPSS or you can also test moderation and mediation models altogether through Smart PLS.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression