topic Re: How to reverse the direction of difference using pdiff option in proc mixed in Statistical Procedures

How to reverse the direction of difference using pdiff option in proc mixed

Arthur2 — Tue, 06 Oct 2020 08:01:59 GMT

Hi,

I'm using the following code.

ods graphics on;
ods output lsmeans=estimate
diffs=diff;

proc mixed data = mydata (where=(treatn in (1,2,3))) plots=all;
by estremn estrem atypen atype paramcd param avisitn avisit ;
class treatn;
model aval = treatn base/residual outp=resid;
lsmeans treatn / pdiff=control('3');
run;
ods output close;
ods graphics off;

The aim is to get the LS means, SE and the difference of LS means and SEs from plcebo(treatn=3) to each of the treatments (treatn=1,2) as 'placebo - treatment. I believe currently the code is generating the diffrence other way round as (treatment - placebo).
Need to know how to reverse the direction of difference to get the difference.

Thanks

Re: How to reverse the direction of difference using pdiff option in proc mixed

PGStats — Tue, 06 Oct 2020 22:17:31 GMT

From the top of my head:

1) Change the sign of variable aval (this will also change the sign of residuals),

2) Change the sign of LS means and confidence interval limits (if any) in lsmeans and diffs datasets.

Make sure you label those differences appropriately in your result tables, since you are going against standard practice.

Re: How to reverse the direction of difference using pdiff option in proc mixed

SteveDenham — Tue, 13 Oct 2020 12:57:49 GMT

Use an LSMESTIMATE statement that utilizes the coefficients in the direction of interest. For example,

lsmestimate treatn 'Placebo - trt 1' -1 0 1,
                   'Placebo - trt 2' 0 -1 1;

SteveDenham