NA for confidence interval

RStudio IDE
#1

Hello

I have some trouble getting a right confidence interval for my data.
I have a first binary variable : hypo2 and 2 others binary variables p10 and s10.
When I try this piece of code, I have a NA value for superior value of confidence interval.

glm_res2 <- glm(hypo2~p10+s10, data = dataN2, family = binomial)
summary(glm_res2)
confint(glm_res2)

image

Do you have some ideas to explain why it happens and what I can do to overpass this issue ?

Thank you

#2

It's just a guess, but confidence intervals for glm are based on profiled likelihoods, and I guess due to the large standard errors, it's difficult to retrieve a valid value here. You may try Wald-based confidence intervals, e.g. using the parameters package: parameters::ci(model, method = "wald").