Suppose, I have two predictor/independent variables A and B that (i.e. both A and B) affect a dependent variable D. I have also a moderator variable C that influences the relationship between A and D, and B and D. There are some indicators under D. Should I separately (i.e. twice) collect the data or the values of the indicators under moderator D since I am considering the influence of the moderator on two relationships (A and D, and B and D)?
Parag,
there is no need to collect data twice. The same data can be used for both moderation effects.
Assuming that all four variables are latent and are measured through multiple indicators, your goal would be to estimate latent moderation (or interaction). This is a more advanced issue in SEM, but there are several solutions.
If you are using Mplus, you can use the LMS procedure, which is considered among the best, and is very easy to use.
If you are using a different software, I would recommend the following paper, which discusses different modeling options:
Article Three Approaches to Estimate Latent Interaction Effects: Int...
Dear Mr. Sen
You can follow this Video
1. For SmartPLS https://www.youtube.com/watch?v=qBmhQN8WiMw
2. For SPSS https://www.youtube.com/watch?v=sj2XWfZaxek
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling