I have been trying to build a population average model with binary outcomes and some factors. The estimation sample has about 17000-31000 individuals with negative outcomes each wave and about 2000-7000 with positive outcomes each wave. The model building code is:
m2_binary_temp <- gee::gee(binary ~ var1 + var2 + var3 + var4 + var5 + var6 +
as.numeric(var7) + var8 + var9,
data = dd,
maxiter = 25,
family = binomial(link = "logit"),
id = person_id,
corstr = "exchangeable",
scale.fix = TRUE,
scale.value = 1)
summary(m2_gad_temp)
When I ran with only the exposure variable, the model ran just fine, but with all variables in, the kernel crashed. I tried configuring the maximum iterations but didn't help.
Could it be the data causing the issue, or is it more likely my modeling commands?