楼主的问题我已用SAS帮你实现，程序如下：
%Macro randomsurvey(n=,school=,classroom=,child=,bi=,bij=,bijk=,dataset=);
options nonotes nosource;
%do loop=1 %to &n.;
data  survey&loop.;
length id $ 200;
    retain _seed_ &loop.;
    do school=1 to &school.;
u_i=&bi*rannor(_seed_);
do classroom=1 to &classroom.;
u_ij=&bij*rannor(_seed_);
do child=1 to &child.;
u_ijk=&bijk*rannor(_seed_);
cova=0+(1-0)*ranuni(_seed_);

meduc1=0;
meduc2=0;
meduc3=0;
if 0<cova<0.5 then do; meduc1=1;mother_educ=0;end;
else if cova<0.8 then do; meduc2=1;mother_educ=1;end;
else do; meduc3=1; mother_educ=2;end;
score=7+0*meduc1+0.5*meduc2+1*meduc3+u_i+u_ij+u_ijk;

replicate=&loop.;
id=cats(_seed_,school,classroom,child);
class=input(cats(_seed_,school,classroom),best12.);
            output;
end;
end;
end;
/* keep id classroom school child score mother_educ replicate;*/
run;
%end;
data  &dataset.;
    set  survey: ;
run;
proc datasets lib=work nolist;delete survey:;run;quit;
%Mend;

牛逼！！！！！！