Mplus Discussion >> Regression mixture model

Topics
Last Day
Last 3 Days
Last Week
Tree View

Edit Profile


Regression mixture model

Mplus Discussion > Latent Variable Mixture Modeling >

Message/Author

Miriam Forbes posted on Monday, March 06, 2017 - 1:27 pm

Hi Drs Muthén,

I'm trying to use a mixture model to detect classes that differ on the strength of their relationship between two continuous variables (X and Y). I can't get my head around how to specify the model so that the groups differ only on the slope of X->Y, and not based on the intercept.

Could you please point me in the right direction to write model syntax that separates classes based only on the slope of x->y, and allows the intercept (and variable means and variances) to vary within class?

Thanks for your help!

Miri

Bengt O. Muthen posted on Monday, March 06, 2017 - 6:05 pm

Model:

%c#1%
y on x;
y;
[y] (1);
%c#2%
y on x;
y
[y] (1);

Miriam Forbes posted on Tuesday, March 07, 2017 - 9:31 am

Beautiful! Thanks, Dr Muthén.