SAS mixed model 结果显著，但是多重比较结果组间却无差异不显著，求教原因，不甚感激

boe 发表于 2013-7-20 12:22
正常，参阅马克思主义哲学整体与局部的关系

但是统计上是无法讲通的，这个不能这样给人家解释了

I guess the result is not from SAS. I cannot understand AIC = -2LL in SAS definition. I only see the residual in COVariance, does it mean you not using RANDOM effect in your model? Anyway, I am not fully understanding your model here. But i will guess a large block effect (random, if there is) will inflate the SE in least square means.
JingJu