Is there an issue with the likelihood computation of the binomial model fit with glm when using cbind or am I missing something. See below for reproducible example:
