lsmestimate in proc mixed

[Coooooo_Lee]

Hi there,

I just got below code, does anyone can help explain a little bit on the matrix following lsmestimate statement.

What does the matrix stands for?

From statistical perspective, how to explain the matrix?  Thanks!

you could simply set time         has 4 values like 1 ,2, 3, 4
                                 trt             has 2 values as 'A' 'B'

                                 subgroup has 3 values as 'High' 'Low' 'Mid'

	       proc mixed data = test method = reml order = internal cl covtest;

	           class id time trt (ref = 'A')  subgroup  ;

	           model CHANGE = BASE*time trt*time*subgroup / solution ddfm=KR singular=1e-10;

	           repeated time / type=un subject=id;
	           lsmeans     trt*time*subgroup / pdiff om cl alpha=0.05;
	           lsmestimate trt*time*subgroup 
	         0  0  0  0  0  0
		 0  0  0  0  0  0
	         0  0  0  0  0  0
	        -2  1  1  2 -1 -1,
	         0  0  0  0  0  0
		 0  0  0  0  0  0
	         0  0  0  0  0  0
	         1 -2  1 -1  2 -1,
	         0  0  0  0  0  0
		 0  0  0  0  0  0
	         0  0  0  0  0  0
	         1  1 -2 -1 -1  2 /divisor=3 joint elsm cl e alpha=0.05;	
	       run; 

[Rick_SAS]

The documentation for the LSMESTIMATE statement explains the meaning of the values. 

I suggest you start by understanding the ideas for a single categorical variable, then try to understand a two-way interaction.

Your example shows a three-way interaction. There are three separate estimates specified. Each estimate involves 24 (=4*2*3) coefficients. The three estimates are separated by commas.

In addition to the doc, I recommend "CONTRAST and ESTIMATE Statements Made Easy: The LSMESTIMATE Statement"
by Kiernan, Tobias, Gibbs, and Tao (2001)