Hi,
I'm trying to run a proc mixed covtest, and keep recieving errors.
First, I tried running:
proc mixed covtest data=rsa1;
where ans_ctas=1;
class couple_id women session;
model UFBs_total=
male male*x male*y male*x_x male*x_y male*y_y male*lUFBs_totalmc
female female*x female*y female*x_x female*x_y female*y_y female*lUFBs_totalmc
/s ddfm=satterth noint;
random
/*male male*x male*y male*x_x male*x_y male*y_y male*lUFBs_totalmc*/
/*female female*x female*y female*x_x female*x_y female*y_y female*lUFBs_totalmc*/
male male*x male*y male*lUFBs_totalmc
female female*x female*y female*lUFBs_totalmc
/subject = couple_id type=vc gcorr ;
repeated women session/subject = couple_id type = un@ar(1);
estimate 'male a1' male*x 1 male*y 1;
estimate 'male a2' male*x_x 1 male*x_y 1 male*y_y 1;
estimate 'male a3' male*x 1 male*y -1;
estimate 'male a4' male*x_x 1 male*x_y -1 male*y_y 1;
estimate 'female a1' female*x 1 female*y 1;
estimate 'female a2' female*x_x 1 female*x_y 1 female*y_y 1;
estimate 'female a3' female*x 1 female*y -1;
estimate 'female a4' female*x_x 1 female*x_y -1 female*y_y 1;
estimate 'male vs. female a1' male*x 1 male*y 1 female*x -1 female*y -1;
estimate 'male vs. female a2' male*x_x 1 male*x_y 1 male*y_y 1 female*x_x 1 female*x_y 1 female*y_y -1; ;
estimate 'male vs. female a3' male*x 1 male*y -1 female*x 1 female*y 1;
estimate 'male vs. female a4' male*x_x 1 male*x_y -1 male*y_y 1 female*x_x -1 female*x_y 1 female*y_y -1; ;
run;
and I recieved - WARNING: Stopped because of too many likelihood evaluations.
than I added to the first line:
proc glimmix covtest data=rsa1 maxiter=1000 maxfunc=5000;
where ans_ctas=1;
class couple_id women session;
model UFBs_total=
male male*x male*y male*x_x male*x_y male*y_y male*lUFBs_totalmc
female female*x female*y female*x_x female*x_y female*y_y female*lUFBs_totalmc
/s ddfm=satterth noint;
random
/*male male*x male*y male*x_x male*x_y male*y_y male*lUFBs_totalmc*/
/*female female*x female*y female*x_x female*x_y female*y_y female*lUFBs_totalmc*/
male male*x male*y male*lUFBs_totalmc
female female*x female*y female*lUFBs_totalmc
/subject = couple_id type=vc gcorr ;
repeated women session/subject = couple_id type = un@ar(1);
estimate 'male a1' male*x 1 male*y 1;
estimate 'male a2' male*x_x 1 male*x_y 1 male*y_y 1;
estimate 'male a3' male*x 1 male*y -1;
estimate 'male a4' male*x_x 1 male*x_y -1 male*y_y 1;
estimate 'female a1' female*x 1 female*y 1;
estimate 'female a2' female*x_x 1 female*x_y 1 female*y_y 1;
estimate 'female a3' female*x 1 female*y -1;
estimate 'female a4' female*x_x 1 female*x_y -1 female*y_y 1;
estimate 'male vs. female a1' male*x 1 male*y 1 female*x -1 female*y -1;
estimate 'male vs. female a2' male*x_x 1 male*x_y 1 male*y_y 1 female*x_x 1 female*x_y 1 female*y_y -1; ;
estimate 'male vs. female a3' male*x 1 male*y -1 female*x 1 female*y 1;
estimate 'male vs. female a4' male*x_x 1 male*x_y -1 male*y_y 1 female*x_x -1 female*x_y 1 female*y_y -1; ;
run;
similar to this question - https://communities.sas.com/t5/SAS-Procedures/mixed-model-warning-message/td-p/99242
and now the error I get (whether I choose mixed or glimmix) is - WARNING: Did not converge.
What am I doing wrong?
Thanks!
