I‘m learning the meglm for GLMMs. I noticed that we can use the "cov" option to define the variance-covariance matrix of random effects. However, I’m not clear of the matrix. For example, I suppose one predictor, group (1 for treatment and 0 for control) by studies (studyid), for binary disease outcome, and esbalish a random coefficient. and intercept GLMM. Then I use "cov(exc)" to define the matrix. I get the following estimates: _b[group], _se[group], _b[cons], _se[cons] , covariance of group and cons "cov(_cons,group)".
We known the cons means the baseline risk in control group (p1), while the group means treatment effect (ln OR). So, does the cov means the variance-covariance of group and cons, or more detailed, p1 and lnOR？
And if I want to known the covariance of p1 and p2 (risk of events in treatment group, and p2=p1+lnOR), how should I do?
Thanks.
Array