另外想知道，SAS怎么能把变量的分布图画出来。。
谢谢！！

Many new statistics have been added to PROC UNIVARIATE, and new distribution fitting facilities enable you to fit a wide range of parametric distributions as well as nonparametric kernel density estimators.
http://support.sas.com/rnd/app/da/new/daunivariate.html

univariate 是不是只能验证正态分布？
谢谢

   proc univariate data=dta1;
      histogram Length / beta   ctext  = blue;
   run;

Primary Options for Parametric Fitted Distributions
BETA(beta-options) Fits beta distribution with threshold parameter , scale parameter , and shape parameters  and  
EXPONENTIAL(exponential-options) Fits exponential distribution with threshold parameter  and scale parameter  
GAMMA(gamma-options) Fits gamma distribution with threshold parameter , scale parameter , and shape parameter  
LOGNORMAL(lognormal-options) Fits lognormal distribution with threshold parameter , scale parameter , and shape parameter  
NORMAL(normal-options) Fits normal distribution with mean  and standard deviation  
WEIBULL(Weibull-options) Fits Weibull distribution with threshold parameter , scale parameter , and shape parameter c

Really?
I will try it , thanks!