title: this is an example of a linear growth
model for a categorical outcome.
this is the same example as in section 9.2 of
Mplus Web Note #4, page 18, using 4-category
outcomes
montecarlo:
names = u11-u14;
generate = u11-u14(3);
categorical = u11 - u14;
nobs = 500;
nreps = 1;
save = ex6.5.dat;
analysis:
parameterization = theta;
model population:
i s | u11@0 u12@1 u13@2 u14@3;
[u11$1-u14$1*-.7];
[u11$2-u14$2*0];
[u11$3-u14$3*.7];
u11*.5 u12*.6 u13*.9 u14*1.4;
[i@0 s*-.5];
i*.5; s*.1; i with s*0;
model:
i s | u11@0 u12@1 u13@2 u14@3;
[u11$1-u14$1*-.7] (1);
[u11$2-u14$2*0] (2);
[u11$3-u14$3*.7] (3);
u11@.5 u12*.6 u13*.9 u14*1.4;
! in the theta parameterization used here,
! the residual variance needs to be fixed for
! one time point for identification purposes.
! here, it is fixed at the true value for the first
! time point (.5), although in the real-data
! analysis the default is fixed at 1 which makes
! the estimates come out in a different metric
[i@0 s*-.5];
i*.5; s*.1; i with s*0;
output:
tech9;