DF in Repeated Measures ANOVA

sms16

New Member
#1
Hi all! :wave: In advance: thank you to all those who take the time to even read this post, as well as answer it:

Additional to my sunfish study (I talked about in previous posts) I am also conducting a restoration project dealing with treatment and control sites over multiple years.

I ran a PROC MIXED repeated measures ANOVA in SAS (see code below) with sites as random effects and years as fixed:

data REST;
proc mixed covtest cl;
class Site Treatment Year;
model response=Treatment Year Treatment*Year;
repeated Year/subject=Site r rcorr type=cs;
lsmeans Treatment*Year / pdiff adjust = tukey;
run;

My Question is: How does this function calculate df? My Treatment df is what I expect it to be for both the numerator and denominator, but my Year df is not what I expected it to be in the denominator (but they ARE correct for the numerator). I was told the df calculations in the PROC MIXED function can be tricky and not to worry if I didn't understand the numbers, but I'd really like to if anyone has any insight! :yup:

P.S. here is an example of the df I got

Site: 5--plots
Treatment: 2--control or treatment
Year: 7--from 2003 through 2010

Effect Num DF Den DF
Treatment 1 3
Year 6 18
Interaction 6 18