title: this is an example of a quadratic growth
model for a continuous outcome
montecarlo:
names = y11-y14;
nobs = 500;
nreps = 1;
save = ex6.9.dat;
model population:
i s q | y11@0 y12@1 y13@2 y14@3;
[y11-y14@0];
y11-y14*1;
[i*.5 s*1 q*.5];
i*1; s*.4; q*.2; i with s*.1;
! i and s with q defaults to zero (free)
! this model gives many replications with
! non-pos def cov matrices, with low
! power (see last column) for the s and y14
! variances, suggesting that more time points
! may be needed for this quadratic model
model:
i s q | y11@0 y12@1 y13@2 y14@3;
[y11-y14@0];
y11-y14*1;
[i*.5 s*1 q*.5];
i*1; s*.4; q*.2; i with s*.1;
output:
tech9;