Computation of R2 using Proc Nested

[GSD]

Hi All,

I am using the following SAS to code to compute the variance components of four categorical variables. I need to compute the R2 of the model. Can I directly use the Sum Squares Error and Sum Squares Total given in the output or should I be using a different technique?

proc sort data=ANOVATEST;

by Y I_Code C_name S_C;

proc nested data=Anovatest;

class Y I_Code C_name S_C;

var S_RA;

run;

quit;

In the above, S_RA is a continuous variable and Y represents year, I_Code C_name S_C are categorical variables representing effects.

Variance Source	DF	Sum of Squares
Total	23436	8120.03
Y	11	11.704
I_Code	785	626.329
C_Name	18735	6542.12
S_C	3532	893.299
Error	373	46.5771

Thanks in advance for your time.

[SteveDenham · Accepted solution]

Without testing, R^2 = (TSS - ESS)/TSS.  Explained variation over total variation.

Steve Denham