topic Randomized block design in Statistical Procedures

Randomized block design

deleted_user — Fri, 21 Aug 2009 13:05:12 GMT

Dear all,

Pl let me the procedure to do the randomized block design in two replicates.

Thanks,
SSN

Re: Randomized block design

Paige — Fri, 21 Aug 2009 13:54:25 GMT

Do you want to create such a design or analyze such a design? Can you give us more details?

Re: Randomized block design

deleted_user — Sat, 22 Aug 2009 14:00:27 GMT

Hi,

I want to analyze the design.

My data is like

Treatment has 8 categories, Blocks has 4 categories and two Replications.

Pl let me know my code is correct or not.

proc glm data=temp;
class Rep treat block;
model resp = treat block Rep ;
run;
quit;

Many Thnaks,
SSN

Re: Randomized block design

Paige — Mon, 24 Aug 2009 15:01:53 GMT

Seems to me that replicates are nested within treatment and block. If that's not correct, then please correct me.

If that is the case, then you want to change the model statement to indicate that fact.

model resp=treat block treat*block;

You most likely also want to test treatment against the treat*block term, not against the residual error (replicate error) which is the default. The treat*block term is the experimental error.

test h=treat e=treat*block;

Or you could do the whole thing in PROC MIXED.

Re: Randomized block design

Paige — Mon, 24 Aug 2009 15:02:03 GMT