really?
I did not realize such an intermediate sas version existing...But anyway, since just a linear model, a transfomration will do the work as well:
The REG Procedure
Model: model1
Test test1 Results for Dependent Variable Weight
Mean
Source DF Square F Value Pr > F
Numerator 1 1265.91603 9.55 0.0070
Denominator 16 132.50623
\
The REG Procedure
Model: model2
Test test2 Results for Dependent Variable Weight
Mean
Source DF Square F Value Pr > F
Numerator 1 1265.91603 9.55 0.0070
Denominator 16 132.50623
- data t;
- set sashelp.class;
- t_height = height/1.5;
- run;
- proc reg data =t;
- var weight age height t_height;
- model1: model weight =age height;
- test1_: test age+1.5*height =0;
- run;
- /* since your SAS works in this type*/
- model2: model weight =age t_height;
- test2_: test age+t_height =0;
- run;
- quit;
复制代码