treatment means with proc mixed

[djakub]

Hello, 

I was handed a past research project that never got finished and keep getting differences in treatment means with the code I am using. It's been a while since I have used SAS and done statistics. My differences seem to be in the model statement. I cannot remember what the "P=" is supposed to mean but my differences seem to be resulting from that as my data and model is exactly the same except that one statement has P=0.61 and the other has P=0.9361. This is probably a dumb question but I am having a hard time remembering. Could anyone clarify this for me?

proc mixed;

class trt rep pvt year;

model iw=trt year; *trt*year (P=0.61);

random pvt;

lsmeans trt year /pdiff;

run;.

proc mixed;

class trt rep pvt year;

model ibw=trt year; *trt*year (P=0.9361);

random pvt;

lsmeans trt year /pdiff;

run;

[PaigeMiller]

It would be more meaningful if you showed us a screen capture of the two outputs from SAS. You have shown us p-values, but we don't know what test the p-values are for.

In general, when the p-value is large than some threshold (often 0.05, but other thresholds can be used) then you do not have a statistical difference between the treatment means; or in other words, you accept the hypothesis of no difference.