下面是我的程序：

data rpw;
pb=0.1;
do pa=pb to 1 by 0.1;
 do j=1 to 3;
  na=1;  /*initial number of A*/
  nb=1;  /*initial number of B*/
  retain seed1 123 seed2 345 seed3 678;
  do i=1 to 100;
     call ranuni(seed1,allo_p);
     if allo_p<na/(na+nb) then do;
        t='A';
        call ranuni(seed2,succ_pa);
     if succ_pa<pa then na=na+1;
     else nb=nb+1;
        end;
     else do;
        t='B';
        call ranuni(seed3,succ_pb);
     if succ_pb<pb then nb=nb+1;
     else na=na+1;
     end;
  end;
  allo_pa=na/(na+nb);
  output;
 end;
end;
run;
proc print;
var pa allo_pa;
run;

接下来是运行结果：
                                    Obs     pa    allo_pa

                                       1    0.1    0.57843
                                       2    0.1    0.49020
                                       3    0.1    0.51961
                                       4    0.2    0.52941
                                       5    0.2    0.53922
                                       6    0.2    0.50980
                                       7    0.3    0.55882
                                       8    0.3    0.53922
                                       9    0.3    0.55882
                                      10    0.4    0.61765
                                      11    0.4    0.54902
                                      12    0.4    0.54902
                                      13    0.5    0.63725
                                      14    0.5    0.68627
                                      15    0.5    0.62745
                                      16    0.6    0.74510
                                      17    0.6    0.70588
                                      18    0.6    0.68627
                                      19    0.7    0.81373
                                      20    0.7    0.72549
                                      21    0.7    0.78431
                                      22    0.8    0.79412
                                      23    0.8    0.78431
                                      24    0.8    0.81373
                                      25    0.9    0.92157
                                      26    0.9    0.88235
                                      27    0.9    0.92157
                                      28    1.0    0.99020
                                      29    1.0    0.97059
                                      30    1.0    0.99020

我想做的是对每个pa,求allo_pa的方差，然后以pa为横轴，allo_pa的方差为纵轴作图，能不能在对pa的循环里加入means过程啊，我那样做的输出来的不是想要的结果。。。头疼死了，用的SAS9.0也不支持自定义函数。。。求大侠指点。。