post your dataset

sas的文件不能上传，转成了excel，O(∩_∩)O谢谢！

5# zhaojing830712

I think the problem may be from the denominator degrees of freedom for the tests of fixed effects. The degrees of freedom were mostly set as 0 in this setting.
When a random statement was specified, it is the default for SAS to compute denominator degrees of freedom by the containment method, which provides correct degrees of freedom only for ‘iid’ designs. Apparently, yours is not this case.
I would suggest not use ‘DDFM = con’ as the default does but choose other settings, such as 'DDFM = KR' in the model statement.

JingJu

多谢指教，如果添加'DDFM = KR' 选项，能运行出结果，
但是日志提示：NOTE: 122 observations are not included because of missing values.
NOTE: Convergence criteria met but final hessian is not positive definite.
NOTE: Asymptotic variance matrix of covariance parameter estimates has been found to be singular and a generalized inverse was used. Covariance parameters
      with zero variance do not contribute to degrees of freedom computed by DDFM=KENWARDROGER.
NOTE: 上述消息是针对以下 BY 组的:
      sex=1
NOTE: 117 observations are not included because of missing values.
NOTE: Convergence criteria met but final hessian is not positive definite.
NOTE: Asymptotic variance matrix of covariance parameter estimates has been found to be singular and a generalized inverse was used. Covariance parameters
      with zero variance do not contribute to degrees of freedom computed by DDFM=KENWARDROGER.
NOTE: 上述消息是针对以下 BY 组的:
      sex=2

10# zhaojing830712

This message is pretty common in mixed procedure, particularly when using both random and repeated statement in the model (in more than 90% cases).
Non-positive definite may be caused by zero variance whereas the model may still be able to correctly estimate the variance and thus many people are quite comfortable with this message.

哦，这样啊，真是万分感谢！！