Fitting a latent class model using SEM path diagram

[Roman Mostazir]

Hi,
I cam across an example of a latent class model presented in the Stata conference 2021 using -gsem-. I understand what is being done but can't find a way to fit the model using Stata's -sem-path diagram (model building and estimation). Any suggestion on how to fit such model using the path diagram is much appreciated. The codes for the model and the data example given below:

#delimit ;
gsem (1.C: depress <- r@0 depbase@c1 risk@c2)
  (2.C: depress <- r depbase@c1 risk@c2)  
  (1.C: comp <- _cons@-15, logit)  
  (2.C: comp <- _cons@15, logit)
  (C <- age educ motivate econ assert single nonwhite),
  lclass(C 2)
 ;
 #delimit cr

The sample data is below:

* Example generated by -dataex-. For more info, type help dataex
clear
input float(depress depbase risk age educ motivate econ assert single nonwhite r c comp)
-1.28 2.73 1.67 28.75 13   4 3.33  4.5 1 0 0 1 .
 -.27 2.18 1.55 28.27 13   5    4 3.25 1 0 0 1 .
 -.19 2.64  1.6 26.33 14 6.5 1.67  2.5 1 0 0 1 .
 -.82 2.18  1.5 39.09 16 6.5    5    5 1 1 0 1 .
  .36 2.64 2.01 37.92 11 5.5    5 2.75 1 1 0 1 .
-1.18 2.45 1.89 28.81 12   4 4.33    2 1 0 0 1 .
  .45 2.91 2.15  38.7 12 5.5    4 1.75 1 0 0 1 .
    0 2.82 1.88 39.27 12   4 2.33 1.75 0 0 0 1 .
-1.45 2.45  1.7 29.92 17 4.5 3.67    3 1 0 0 1 .
 -.73 2.55 1.61 39.78 16 5.5 2.33 2.75 0 0 0 1 .
 -.81 2.45 1.62 33.97 13 4.5 3.67 3.75 1 0 0 1 .
 -.64 2.73 1.97 42.13 16 6.5    5 3.75 1 0 0 1 .
-1.37 2.82 1.84 32.27 14 4.5    4 4.25 0 0 0 1 .
-1.18 2.18  1.7  31.7 14 4.5    5    3 1 0 0 1 .
-1.09 2.27 1.52  32.8 13 4.5 4.33  4.5 0 1 0 1 .
-1.37 2.64 1.73 34.21 15 5.5 2.67  2.5 1 1 0 1 .
  .72 2.64 1.66 22.04 12   5 2.33 2.75 1 1 0 1 .
 -.27 1.91  1.5 25.04 13 5.5 4.33  2.5 0 0 0 1 .
 -.54  2.9 2.04 37.85 14   6 4.33 3.25 1 1 1 1 1
-1.27 2.82    2 35.33 13 5.5 4.67  3.5 0 1 1 0 0
  .27 2.64 1.82 43.13 12 4.5 4.67 4.25 1 0 1 0 0
-1.27 2.45 1.85  26.3 13 5.5 4.67 2.75 1 0 1 0 0
  .09 2.73 1.74 42.05 12 6.5 3.67 4.25 1 0 1 1 1
  .37 2.36 1.56 18.79 12   6    3    3 1 0 1 0 0
 -.36 2.27 1.49 39.99 13 6.5 3.33  3.5 1 0 1 1 1
-1.37 2.55 1.67 34.89 13 3.5    4 4.25 1 0 1 0 0
  .54 2.73  1.9 35.57 12   5 4.67    4 1 0 1 1 1
 1.45    2 1.44 18.47 11   5 3.67 2.75 1 0 1 0 0
   -1 2.55 1.71 38.67 12 5.5 3.33    3 0 0 1 0 0
 -.82 2.82 1.93 37.99 13   5 3.33  2.5 1 0 1 1 1
-1.16 2.45  1.8 71.38 10 5.5 4.67 3.25 0 0 1 1 1
 -.18 2.45 1.85 47.61 10   5 4.67 2.75 0 0 1 1 1
  .82 2.45 1.62 39.88 13   5 3.67 3.75 1 0 1 1 1
  .46 2.18 1.47 50.96  9   5    3 2.75 1 0 1 1 1
 -.72 2.27 1.61 31.05 10 4.5 3.67 2.75 1 0 1 1 1
 1.27 2.55 1.82 20.58 12 6.5 4.33 3.25 0 0 1 0 0
 -.46  2.1 1.58 24.98 14   5    5 3.75 1 1 1 0 0
-1.64 2.82  1.9 28.15 16 5.5 3.67 3.25 1 0 1 0 0
  .18 2.09 1.48 26.96 12   5 4.33 3.75 1 0 1 0 0
 -.91 2.27 1.45 33.42 17 5.5    3  3.5 1 0 1 1 1
  .46 2.45 1.39 27.91 16 4.5 2.67 4.75 0 0 1 0 0
 -.73 2.55 1.63  31.6 16 4.5 3.67 4.25 1 0 1 1 1
  -.5  2.2 1.57 57.77 13   5 4.67    4 1 0 1 1 1
-1.82    3  2.1 51.13 16   6 4.33 3.25 1 0 1 1 1
 -.55 2.55 1.82 39.87 12   5    4 2.75 0 0 1 0 0
 -.55 2.73 1.71 48.98 17   4    4    5 1 0 1 0 0
   -1 2.73 1.55  21.7 13   5    2    4 0 0 1 0 0
  .18 2.55 1.86 40.59 16   4 4.67 3.25 1 1 1 1 1
-1.09 2.45 1.51 27.93 14 4.5    3    4 1 0 1 0 0
-1.37 2.55 1.84 48.29 13   6 4.33    3 1 0 1 1 1
 -.09 2.09 1.67 44.82 17   6    4  1.5 1 0 1 1 1
end

Thanks.

[Roman Mostazir]

More specifically, what I am asking is how to define a 2-class solution for a latent variable using SEM path diagram? I am referring to the ( ,lclass(C 2) ) option part in the codes above

 
 , lclass(C 2)

[Roman Mostazir]

More specifically, what I am asking is how to define a 2-class solution for a latent variable using SEM path diagram? I am referring to the ( ,lclass(C 2) ) option part in the codes above

Sorted. We can't. P.36 -sem- manual says, "we cannot draw path diagrams in the Builder to fit latent class models. Instead, we fit
latent class models using the command language".