Documentation Example 6 for PROC NLMIXED
/****************************************************************/
/* S A S S A M P L E L I B R A R Y */
/* */
/* NAME: nlmex6 */
/* TITLE: Documentation Example 6 for PROC NLMIXED */
/* Simulated nested linear random effects model */
/* PRODUCT: STAT */
/* SYSTEM: ALL */
/* KEYS: */
/* PROCS: NLMIXED */
/* DATA: Little et al(2006) */
/* */
/* SUPPORT: Raghavendra Rao Kurada (Raghu) */
/* REF: */
/* MISC: */
/****************************************************************/
%let na = 100;
%let nb = 5;
%let nr = 2;
data nested;
do A = 1 to &na;
err1 = 3*rannor(339205);
do B = 1 to &nb;
err2 = 2*rannor(0);
do rep = 1 to &nr;
err3 = 1*rannor(0);
resp = 10 + err1 + err2 + err3;
output;
end;
end;
end;
run;
proc nlmixed data = nested;
bounds vara >=0, varb_a >=0;
mean = intercept + aeffect + beffect;
model resp ~ normal (mean, s2);
random aeffect ~ normal(0,vara) subject = A;
random beffect ~ normal(0,varb_a) subject = B(A);
run;