Thanks very much for the prompt reply! I have now managed to get the Fit Statistics. cDNA is copies of genes in a sample. The copies, initially would be a count, but I express them as copies/100mL groundwater. The values have a huge range (they can be anywhere between around 10 copies to billions of copies). "Date" is actually meant to represent different seasons and thus I included it as a fixed effect. The same wells are sampled each date, but for testing to see if seasons have an effect on gene copies. "Well" is also a treatment since the wells have varying degrees of contamination. I see what you mean about using /subject=well now. I had included it because the well is also the unit I am measuring. Now that I think of it, would adding corresponding block values to the wells and using "/subject=block" be a better idea? class well date block;
model cDNA = well date date*well / dist=negbin link=log;
random intercept / subject =block;
lsmeans date*well / tdiff adjust=tukey lines;
output out=second predicted=pred residual=resid residual (noblup)=mresid student=studentresid student(noblup)=smresid;
run; And actually, since it seems like this data is continuous, is there any way I can adjust the lognormal distribution method? (I've attached the output from the lognormal distribution analysis and the coding is below). I'm sorry to be a pest, I'm a beginner with SAS and just running stats in general. Wrapping my head around it is proving tricky. Thanks again, A title2 'cDNA';
proc glimmix data=first;
class well date block;
model cDNA = well date date*well / dist=lognormal ddfm=kr;
random _residual_ / subject = block;
lsmeans date*well / tdiff adjust=tukey lines;
output out=second predicted=pred residual=resid residual (noblup)=mresid student=studentresid student(noblup)=smresid;
run;
... View more