lavaan warning: deleted cases with FIML

[A G]

Dear Community

I have 2 questions regarding the sample that was taken into account with my growth function, which I calculated the following way.

growth( mf_full, data = dtall, sample.nobs = nobs,  estimator = "ml", missing = "fiml") 

As the sample (3478) contains missing, I went for FIML. But how does full information maximum likelihood exclude cases? Of my sample, there are 2040 deleted cases by Lavaan. However, the model is robust and fits well... 

warning

2040 cases were deleted due to missing values in 

 exogenous variable(s), while fixed.x = TRUE.

1) What does that mean? According to which criteria does full information maximum likelihood exclude cases?

2) Is there a way to include the whole sample?

3) Is there a function to investigate which participants were excluded? 

Thank you a lot

[Terrence Jorgensen]

1) What does that mean? According to which criteria does full information maximum likelihood exclude cases?

Exogenous observed variables do not need to be modeled, in the sense that their sample means and (co)variances can be taken as given rather than estimated.  This is the default (fixed.x=TRUE), but you can set fixed.x=FALSE to treat the exogenous variables as endogenous and estimate their summary statistics as well.  

2) Is there a way to include the whole sample?

You can set missing = "fiml.x" to use the FIML approach also for exogenous predictors.

 

3) Is there a function to investigate which participants were excluded? 

lavInspect(fit, "empty.idx")

See the ?lavInspect help page.

Terrence D. Jorgensen

Assistant Professor, Methods and Statistics

Research Institute for Child Development and Education, the University of Amsterdam

http://www.uva.nl/profile/t.d.jorgensen