How to do a binomial (not binary) logistic regression

binomial_logistic_regression.R

library(tibble)

ilogit <- plogis # inverse logit 

Df <- tibble(corpus = c('bnc_subset', 'subset'),
             count = c(5242, 10092),
             total = c(27449, 56355))


M <- glm(cbind(count, total-count) ~ 0 + corpus, 
         data = Df,
         family = binomial)


# Confidence intervals on the proportions for "subset" and "bnc"
ilogit(confint(M))