factorial split-plot with a strip factor

[rajabi83]

Greetings and Regards My design is factorial split-plot with a strip factor I need to a correct model Data a; Input r a b c d tkw; Cards; proc glm; class r a b c d; model tkw=r a r*a b c b*c r(b c) a(b c) r*a(b c) d a*d d(b c) a*d(b c)/ss3; test h=r a e=r*a; test h=b c b*c e=r(b c); test h=a(b c) e=r*a(b c); means a/duncan alpha=0.05 e=r*a; means b/duncan alpha=0.05 e=r(b c); means c/duncan alpha=0.05 e=r(b c); means d/duncan alpha=0.05; means (b*c)/duncan alpha= 0.05 e= r(b c); run;

[rajabi83]

Greetings and Regards
My design is factorial split-plot with a strip factor
I need to a correct model

Data a;
Input r a b c d tkw;
Cards;



proc glm;
class r a b c d;
model tkw=r a r*a b c b*c r(b c) a(b c) r*a(b c) d a*d d(b c) a*d(b c)/ss3;
test h=r a e=r*a;
test h=b c b*c e=r(b c);
test h=a(b c) e=r*a(b c);
means a/duncan alpha=0.05 e=r*a;
means b/duncan alpha=0.05 e=r(b c);
means c/duncan alpha=0.05 e=r(b c);
means d/duncan alpha=0.05;
means (b*c)/duncan alpha= 0.05 e= r(b c);
run;

[StatsMan · Accepted solution]

This thread should help.  It is a model very similar to yours.  Moving your model to PROC MIXED is a more modern way of tacking any spit-plot design.