LCM with membership on weighted sample ERROR

[Laetitia Tuffery]

I am contacting you because I have a problem with an LCM with membership on weighted sample.

The estimate is as follows:

Lclogit X Y1 Y2 Y3 Y4, id(hhid) group(cset) membership(Z1 Z2 Z3 Z4) nclass(2) lclogitml, iter(80) pweight(pw)

Stata runs for a long time and ends up with incomplete results and many error messages.

TwoStep: did not reach goal/
Could not calculate numerical derivatives/
Missing values encountered/
Something went wrong in comprob3

Moreover, when running the same program twice in a row, the results are not identical.

Nevertheless, this estimation works perfectly without the “pweights” option (i.e., on the non-weigthed sample).

How can I solve this issue? Thanks very much for your advice.

[Rich Goldstein]

first, if weighted data were allowed, the "pweight(pw)" is not an option and would come before the comma

second, however, weights are apparently not allowed with this command

third, however (again), weights, including pweights, are allowed with -sem- and -gsem-; so,

h sem

and then click on examples and scroll down to examples 50g to 52g for latent class and latent profile models

[Laetitia Tuffery]

Thank you for your answer, we tried with SEM and GSEM commands but it did not work as we could not include the “ID” variable which is necessary for discrete choice estimation.

Do you have any other suggestions?