a question on 'fixed.x = TRUE' in sem,lavaan,...

[asas asas]

Dear lavaan  users and experts,

When I checked the major functions,like 

 'sem' ,'lavParTable',and 'lavaanify', there is an parameter 'fixed.x' whose default value is TRUE.Since my background is

engineering,I am not very clear about the application details of lisrel model.Is the 'fixed.x' often  set to be TRUE in most

application cases?

This will mean the $\eta$ (endogenous latent variables) and $ksi$(exogenous latent variables) in the 

structural model will be pre-determined by the measurement model, then a simultaneous equation model will be fitted.

If 'fxed.x' ==FALSE, then the $\eta$,$ksi$ will be determined jointly with measurement model and structural model.

Why is the 'fixed.x' set to be TRUE as default?Does this mean in most psychology research, the model is considered as this?

I need some explanation on when to set 'fixed.x' = FALSE.

Thanks for your help,

Best regards to you,

Yilun

[Edward Rigdon]

The "x" in "fixed.x" refers to a covariate--an exogenous observed variable. This is distinct from observed variables that are dependent on common factors (or anything else). The lavaan materials are written from an "all-y" perspective, where all indicators of common factors are thought of as "y" variables. From a lavaan perspective, only observed variables which are exogenous covariates are considered to be "x" variables.

--
You received this message because you are subscribed to the Google Groups "lavaan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lavaan+un...@googlegroups.com.
To post to this group, send email to lav...@googlegroups.com.
Visit this group at https://groups.google.com/group/lavaan.
For more options, visit https://groups.google.com/d/optout.

[asas asas]

Thanks for the kind reply.

I want to confirm that,what the exogenous covariates mean.Does it mean the exogenous variables, that are not indicators of any latent variables, but show up in the structural model? So these   exogenous covariates have no measurement errors.Is this correct?

Best,

Yilun

在 2019年1月21日星期一 UTC-5下午2:20:47，Edward Rigdon写道：

[Edward Rigdon]

     Yes, that is precisely correct. In fact, it is important to avoid making mention in syntax of variances or covariances of variables intended to be exogenous covariates, because that may cause lavaan NOT to treat those variables as covariates.

     And yes, these covariates do not have error terms. If you want to specify error terms, then these variables must themselves be dependent on something.

     Covariates may be variables like "treatment category identifier," assigned by researchers rather than being sampled. The point of "fixed.x=T" is that these variables do not have random sampling error, unlike the other variables in the model.