03-31-2017 04:06 AM
I ran the following program and kept getting this warning message WARNING: The initial estimates did not yield a valid objective function, and no chi2 values for the COVTEST.
proc glimmix method = laplace;
class mom dad;
model ontsurv (event = '1')= jsh /S dist = bin link = logit ddfm =bw;
random int/ subject = mom;
random int/ subject = dad(mom);
NLOPTIONS TECH = NRRIDG;
covtest 'var (mom)= 0' 0 .;
covtest 'var (dad(mom)) = 0' . 0;
I then changed the parameter values from 0 to 0.01 and it worked – i.e.
covtest 'var (mom)= 0' 0.0000001 .;
covtest 'var (dad(mom)) = 0' . 0.0000001;
Does anyone know why this happened, and why changing the parameter values help?
Also, I want to make sure that I understand this correctly – I am testing two random slopes (MOM and DAD) which are nested into each other, two sets of slopes?
Thanks in advance,