LCA with covariates and distal outcome

http://www.statmodel.com/discussion/messages/13/26233.html?1589496566

Mplus Discussion >> BCH with covariate interpretation

Sara Suzuki posted on Sunday, September 30, 2018 - 4:31 am

I tried both regressing the distal on covariates (syntax A) and not regressing the distal on covariates (syntax B) and got very different Wald test results depending on what I did.

Could you help me understand which model is better?

Is it correct that syntax A, where I regress the distal on covariates as well, gives a more "pure" Wald test, showing whether the means of the distal differ between classes after accounting for selection into classes based on covariates and effect of covariates on the distals?

Syntax A:
DATA: FILE=manualBCH.dat;

USEVARIABLES ARE
covariate1
covariate2
distal1
distal2
distal3
w1-w3;

TRAINING = w1-w3(BCH);

CLASS = C(3);

ANALYSIS: TYPE = MIXTURE;

MODEL:

%OVERALL%
C ON covariate1 covariate2;
distal1 distal2 distal3 ON covariate1 covariate2;

%C#1%
distal1 distal2 distal3 ON covariate1 covariate2;

%C#2%
distal1 distal2 distal3 ON covariate1 covariate2;

%C#3%
distal1 distal2 distal3 ON covariate1 covariate2;

Syntax B:
DATA: FILE=manualBCH.dat;

USEVARIABLES ARE
covariate1
covariate2
distal1
distal2
distal3
w1-w3;

TRAINING = w1-w3(BCH);

CLASS = C(3);

ANALYSIS: TYPE = MIXTURE;

MODEL:

%OVERALL%
C ON covariate1 covariate2;

 Bengt O. Muthen posted on Sunday, September 30, 2018 - 2:54 pm
You are right about syntax A - it is better when there are direct effects.

 Scot Seitz posted on Saturday, December 14, 2019 - 12:25 pm
I recently ran the third step of the BCH method. My auxiliary model has multiple covariates (both dichotomous and continuous) and 3 continuous outcomes. I am trying to examine the role of the covariates in predicting the latent classes. In the output in the "Categorical Latent Variables" section, I see that each class is regressed on each covariate (comparing each class to the last class, which seems to be the reference class). I have a few questions:

1) How can I change the reference class so that I can examine the role of each covariate between different classes?
2) Are the estimates in the "Categorical Latent Variables" section regression coefficients?
3) In the "Categorical Latent Variables" section, can I obtain standardized estimates? I included "output: stdyx" in my syntax and standardized results are not reported for this section (although standardized results are reported for other sections).
4) Do you suggest any other way to examine the role of the covariates in predicting the latent classes in the third step of the BCH method?

 Tihomir Asparouhov posted on Monday, December 16, 2019 - 2:46 pm
1. Reorder the weights like this
training=w1 w2 w4 w3(bch);

2. It is multinomial logistic regression
https://en.wikipedia.org/wiki/Multinomial_logistic_regression

3. I would recommend using
define: standardize x;
so the "model results" will be standardized.

4. You can use the 3-step (R3step option) and the 1-step estimation. If those estimates disagree you might have direct effects from the covariate to the latent class variable indicators. See Table 4
http://www.statmodel.com/download/3stepOct28.pdf

 shonnslc posted on Sunday, May 10, 2020 - 1:17 pm
Hi,

I am trying the manual BCH method to examine if profiles differ in a distal outcome. I know how to do that in an automatic way, but not sure how to do that in a manual way:

usevar=
y BCHW1 BCHW2 ;
CLASSES = c(2);
training=BCHW1 BCHW2 (bch);

Analysis:
Type = mixture;
starts = 0;

model:
%overall%

%c#1%
[y];
%c#2%
[y];

This will only give me the mean of the distance outcome in each class, but how can I get the differences between them. Thank you!

 Bengt O. Muthen posted on Monday, May 11, 2020 - 1:55 pm
Use Model Constraint.

%c#1%
[y] (y1mean);
%c#2%
[y] (y2mean);

Model constraint:

new(diff);
diff= y1mean - y2mean;

 shonnslc posted on Thursday, May 14, 2020 - 8:45 am
Thank you, Muthen. But isn't this approach (using the model constraint command) based on the Wald test instead of the chi-square test used in the automatic approach? Does that matter? Thank you!

 Bengt O. Muthen posted on Thursday, May 14, 2020 - 3:49 pm
No, the Wald test is used in Model Tests. Model Constraint uses a regular z-test for each New parameter.