My model Consist of one independent, one moderator and three dependent variables. Please need guidance to use SEM or Hierarchical linear regression, and what are the main differences between these two techniques. Thank you
SEM; structural equation modeling, is a multivariate statistical analysis technique that is used to analyze structural relationships. It is a combination of factor analysis and multiple regression analysis. A multivariate analysis handles more than one dependent variables against one or more independent variable, even with a covariate or a moderator. A multiple or hierarchical linear regression on the hand, considers more than one independent variable which is not the case in your research. So, I think the answer is obvious. Go for SEM.
SEM is much more flexible. For your case, SEM can easily accommodate three DVs, which is more challenging with regression.
Another way of looking at this is that linear regression is a special (very restricted) case of an SEM. In other words, and broadly speaking, SEM can do anything regression can do, but not vice versa.
While both regression and SEM can easily handle moderation, I agree with Heiko Breitsohl that it is better to use SEM when you have multiple dependent variables.
Thank you very much Maurice Ekpenyong Heiko Breitsohl David L Morgan for your kind responses. I am now pretty much confident to go with SEM. Anyhow I have one more confusion that is; whether through applying SEM one can get R square change or not? Because we can observed R square change step by step entering variables in hierarchical regression analysis. Thanks
- Badges
- Science method