option lclass ( ) not allowed error

[Karen Hancock]

Novice here!

This may be a simple question, but I couldn't find an answer using either Google or this forum, so I am probably doing something very basic that is wrong.

I am trying to run latent class analyses using gsem and no matter how I specify my model I get the same error message. An example model shown below

gsem (meaning <- inwork owner indsex ), logit lclass(C 3)
option lclass() not allowed
r(198);

All the variables are 0,1 binary variables and I am using about 3,000 observations.
I copied the exact syntax for this code from these instructions from Stata: https://www.stata.com/features/overv...lass-analysis/

What am I not seeing here please?

Thank you for any pointers

Karen

[Clyde Schechter]

I don't see anything wrong with that syntax, and it works on my setup. Perhaps you are using a version of Stata from before the introduction of latent category variables?

[Hemanshu Kumar]

I tried

sysuse auto, clear

foreach var of varlist length mpg price {
    su `var', meanonly
    gen `var'_bin = `var' < r(mean)
}

gsem (foreign <- length_bin mpg_bin price_bin ), logit lclass(C 3)

and it works in both Stata 16.1 and the current Stata 19.5. So perhaps you are using a fairly old version of Stata?

[Karen Hancock]

To test if it's a problem with my data I downloaded Stata's test dataset "use https://www.stata-press.com/data/r19/gsem_lca1"
I then pasted in their code with the same results.
gsem (accident play insurance stock <- ), logit lclass(C 2)
option lclass() not allowed

I am using Stata 14.2/SE, so maybe it's not allowed in my version of Stata?

[Clyde Schechter]

Support for latent category variables was introduced in version 15.