Message/Author |
|
|
Is it possible to model random effects as a function of covariates in Mplus? For example assume the Level 1 model is as follows: i s | y1@0 y2@1 y3@2 y4@3 y5@4; If I want to model the intercept random effect regressed on a covariate x, what is the proper syntax? i on x; !this would not be correct since it regresses the intercept on the covariate. This question is motivated by Donald Hedeker's work on ecological momentary assessment data using location scale models to study the variability. Thank you. |
|
|
i is not an intercept but an intercept random effect, that is, a latent variable. So say i on x; |
|
|
Thank you for your quick reply. In the UG example 6.10, the following syntax is used: i s on X1 X2; This is described as regressing the growth factor on X1 and X2. How is this the same as regressing the intercept (mean) random effect on X1 and X2? In the above question, I used the term intercept when I should have used the term mean since i and s are latent variables. Thank you. |
|
|
"Growth factor" is the term we use for the random effects. The intercept random effect is i. So i is a latent variable. It has a mean and a variance and since it is a variable it can be part of a regression. Let me know if that is not what you are asking. |
|
Back to top |