ＬＺ给的答案好像不对吧。ｓｅｅｄ应该是取０位或５位或６位的奇数。生成10000个均值为M,标准为S的正态随机数的程序：

data norm;
keep value;
do i=1 to 10000;
  value=Ｍ+Ｓ*normal(2367);
  output;
end;
run;

[此贴子已经被作者于2007-11-12 21:35:23编辑过]

NORMAL Function

generates a pseudo-random normal deviate

NORMAL( seed)

where seed is a numeric matrix or literal. The seed argument can be any integer value up to 2³¹-1.

The NORMAL function is a scalar function that returns a pseudo-random number having a normal distribution with a mean of 0 and a standard deviation of 1. The NORMAL function returns a matrix with the same dimensions as the argument. The first argument on the first call is used for the seed (or if that is 0, the system clock is used for the seed). This function is synonymous with the DATA step function RANNOR. The Box-Muller transformation of the UNIFORM function deviates is used to generate the numbers.

感谢！我用的是固定的常数值，我查了一下书，说是可以直接取0或者5位、6位、7位奇数。

还有个问题

假如我现在有x1 x2 x3、、、x201，201个离散数据，我想生成1000000个随机数就在前面的201个数据中取，但必须使前面201个数据取到的机会向等。

能够直接写程序生成吗？还是得先生成均匀随机数，然后再匹配呢？

哎现在才知道自己又多菜~~~！！