I have two questions regarding indirect effect in a path model: Firstly, in a model where x is exogenous variable, y is mediator, and z is outcome, I understand that the indirect effect Bxz is calculated by Bxy*Byz. However, if y is a continuous variable and z is a binary variable, then Bxy and Byz are based on different metrics. In this case, do I have to tell mplus that they are derived from two different scales? Or mplus will do it for me automatically? Secondly, assuming that x is a 4-category nominal variable, then 3 dummies will be created to represent x. In this case, I would like to know if indirect effect can still be calculated since standardization does not seem available in this particular model. Thank you very much.
You need to read the paper on our website to see how Mplus handles this correctly:
Muthén, B. & Asparouhov, T. (2015). Causal effects in mediation modeling: An introduction with applications to latent variables. Structural Equation Modeling: A Multidisciplinary Journal, 22(1), 12-23. DOI:10.1080/10705511.2014.935843