y****1
发帖数: 400 | 1 I recently began to use this procedure. I understand that the model=
statement specifies the fixed effects, and the random statement specifies
the random effects. But what does it mean if a variable shows in both
places? How can it be both fixed and random effects?
Thank you so much. I have been bothered by this for a few days. |
P****D
发帖数: 11146 | 2 Because your understanding, that "model=statement specifies the fixed
effects, and the random statement specifies the random effects" is wrong.
model= statement needs to include BOTH fixed and random effects. You then
put the random effects in the RANDOM statement as the second step.
【在 y****1 的大作中提到】
: I recently began to use this procedure. I understand that the model=
: statement specifies the fixed effects, and the random statement specifies
: the random effects. But what does it mean if a variable shows in both
: places? How can it be both fixed and random effects?
: Thank you so much. I have been bothered by this for a few days.
|
s*r
发帖数: 2757 | 3 趁没什么人看见，赶快自己删了吧
【在 P****D 的大作中提到】
: Because your understanding, that "model=statement specifies the fixed
: effects, and the random statement specifies the random effects" is wrong.
: model= statement needs to include BOTH fixed and random effects. You then
: put the random effects in the RANDOM statement as the second step.
|
s*r
发帖数: 2757 | 4 似乎逻辑上不可以，但你放到sas里面他也能算出来
【在 y****1 的大作中提到】
: I recently began to use this procedure. I understand that the model=
: statement specifies the fixed effects, and the random statement specifies
: the random effects. But what does it mean if a variable shows in both
: places? How can it be both fixed and random effects?
: Thank you so much. I have been bothered by this for a few days.
|
y****1
发帖数: 400 | 5 咦，是吗？因为SAS里面给的example有那种比如说有3个变量，然后2个出现在model=里
面，剩下一个在random里面，就是model=里面不一定要包含所有的变量。我就觉得那有
什么区别呢？
多谢多谢哈，嘻嘻。
【在 P****D 的大作中提到】
: Because your understanding, that "model=statement specifies the fixed
: effects, and the random statement specifies the random effects" is wrong.
: model= statement needs to include BOTH fixed and random effects. You then
: put the random effects in the RANDOM statement as the second step.
|
y****1
发帖数: 400 | 6 不是说我吧？
很心虚的样子。。。
【在 s*r 的大作中提到】
: 趁没什么人看见，赶快自己删了吧
|
y****1
发帖数: 400 | 7 这句话不是很懂额~
【在 s*r 的大作中提到】
: 似乎逻辑上不可以，但你放到sas里面他也能算出来
|
P****D
发帖数: 11146 | 8 啊？我还真没试过model=里不放random effect的。一会儿我去试试。
【在 s*r 的大作中提到】
: 趁没什么人看见，赶快自己删了吧
|
P****D
发帖数: 11146 | 9 刚刚试了，真的可以哎！
【在 s*r 的大作中提到】
: 似乎逻辑上不可以，但你放到sas里面他也能算出来
|
s*r
发帖数: 2757 | 10 proc glm 和 proc mixed有不同的要求
【在 P****D 的大作中提到】
: 刚刚试了，真的可以哎！
|
y****1
发帖数: 400 | 11 那放不放random effects在model=里面，结果有区别么？
比如如果我只有一个variable，是random的，那：
1. model=var ... ;
random var ... ;
和
2. model= ... ;
random var ... ;
有区别吗？
【在 P****D 的大作中提到】
: 刚刚试了，真的可以哎！
|
