Typically, you would examine time-invariant covariates as predictors of the growth (intercept and/or slope) factors. Time-varying covariates could be included with effects directly onto the measured variables (indicators) at each time point. In Mplus, you use ON statements for regressions, for example

int slope ON x y z;

where int and slope are the growth factors and x, y, and z are external variables (covariates/predictors of the growth factors).

Dr. Geiser

Thank you so much for your response.

I created a syntax including the code you mentioned.

Please see attached. I hope I did it right.

By the way, I noticed that this code is similar to how

TVC and TIC is included in a normal second order growth model.

My original syntax had a part that made all the TIC load on

the variable of focus (in my case AE) at their corresponding time point. (as below)

----------------

! AE on AS (TVC1)

AE1 ON AS1;

AE2 ON AS2;

AE3 ON AS3;

AE4 ON AS4;

AE5 ON AS5;

AE6 ON AS6;

AE9 ON AS9;

! AE on TC (TVC2)

AE1 ON TC1;

AE2 ON TC2;

AE3 ON TC3;

AE4 ON TC4;

AE5 ON TC5;

AE6 ON TC6;

AE9 ON TC9;

------------------

I assume that if I wanted to know how the TVCs affected the

value at each time point, including the the above syntax wouldn't cause problems. Is this idea correct?

* I looked at the syntax and found some errors, so I have uploaded a revised file. (test2)

Yes, it is possible to include total indirect effects (TIE) and total direct effects (TDE) as well as total effect (TE) for a latent variable model (LVM) or a latent basis model (LBM) in SEM.

Total indirect and total direct effects can be calculated using the concept of "mediation" and "moderation" respectively. Mediation refers to the extent to which one variable (the mediator) explains the relationship between another variable (the predictor) and a third variable (the outcome). Moderation refers to the extent to which a variable (the moderator) modifies the relationship between another variable (the predictor) and a third variable (the outcome).

The calculation of these effects can be done using specific software, such as AMOS, Mplus, R or Python, that allows the estimation of SEM models. These software's allow to estimate these effects by adding specific parameters, such as indirect and direct effects, to the model, and then running the model to obtain the estimates of these effects.

It is important to note that these calculations assume that the model is correctly specified, and that the assumptions of SEM are met. Therefore, it is recommended to consult with a statistician or SEM expert before attempting to estimate TIE, TDE and TE in LBM.