Hi all,

I am having difficulties in obtaining marginal effects of a GLM estimation that I am running. When I run the following, I obtain all coefficients of all covariates specified:

Code:
svy, subpop(if group ==1): glm exportintensity bimodal990##(c.foreign c.govt c.age c.age2 c.size_num c.capacity i.email_d i.tech_d) i.countryyear i.industry, fam(bin) link(logit)
However, as far as I am aware, in GLM since these are not marginal effects, so I run the following:
Code:
margins, dydx(*)
, or
Code:
margins, dydx(*) atmeans
I get marginal effects for all but i.countryyear, and i.bimodal990 and its interactions with the other covariates. For these, Stata tells me they are "not estimable." This is a bit strange because I was obtaining all GLM coefficients perfectly. I am not sure if I am using the wrong command for marginal effects, since there's a bunch of interactions that I am estimating. Is there something I am overlooking here?

If you think the commands are correct and still I am getting problems, I'd be happy to share a part of my data. Thank you!