Dear all,
Greetings!
I have three concerns
1. Does Stata version 16.1 can handle two step approach for latent class model with covariates ?
2. Does the latent classes beta coefficient & P- value implicate any thing on the model selection? Please refer my output below
3. Shall I have to worry about multilevel latent class model ,If I account clusters & strata by using complex survey design in the model
svy:gsem ( KIstata KJstata KKstata KLstata KMstata KNstata KOstata KPstata KQstata KRstata KSstata KTstata KUstata KVstata < , logit) (
> C <- ), lclass(C 3) iterate(50)
(running gsem on estimation sample)
Survey: Generalized structural equation model
Number of strata = 11 Number of obs = 2,341
Number of PSUs = 96 Population size = 2,341.0002
Design df = 85
Linearized
Coef. Std. Err. t P>t [95% Conf. Interval]
1.C (base outcome)
2.C
_cons -.0089854 .1029819 -0.09 0.931 -.213741 .1957702
3.C
_cons -.4321705 .1185585 -3.65 0.000 -.6678966 -.1964445
Greetings!
I have three concerns
1. Does Stata version 16.1 can handle two step approach for latent class model with covariates ?
2. Does the latent classes beta coefficient & P- value implicate any thing on the model selection? Please refer my output below
3. Shall I have to worry about multilevel latent class model ,If I account clusters & strata by using complex survey design in the model
svy:gsem ( KIstata KJstata KKstata KLstata KMstata KNstata KOstata KPstata KQstata KRstata KSstata KTstata KUstata KVstata < , logit) (
> C <- ), lclass(C 3) iterate(50)
(running gsem on estimation sample)
Survey: Generalized structural equation model
Number of strata = 11 Number of obs = 2,341
Number of PSUs = 96 Population size = 2,341.0002
Design df = 85
Linearized
Coef. Std. Err. t P>t [95% Conf. Interval]
1.C (base outcome)
2.C
_cons -.0089854 .1029819 -0.09 0.931 -.213741 .1957702
3.C
_cons -.4321705 .1185585 -3.65 0.000 -.6678966 -.1964445
